Materials and Methods

Chemicals.

Glucose 6-phosphate, glucose 6-phosphate dehydrogenase, NADP⁺, phenacetin, acetamidophenol, coumarin, 7-hydroxycoumarin, and diclofenac were purchased from Sigma (St. Louis, MO). [¹⁴C]-(S)-mephenytoin was obtained from Amersham Life Science (Arlington Heights, IL). Testosterone and 6β-hydroxytestosterone were obtained from Steraloids (Wilton, NH). Paclitaxel, 7-ethoxy-4-trifluoromethylcoumarin, and 7-hydroxy-4-trifluoromethylcoumarin were obtained from Molecular Probes (Eugene, OR). 4′-Hydroxydiclofenac, (±)-bufuralol, and 6α-hydroxypaclitaxel were obtained from Gentest Corp. (Woburn, MA) and (S)-mephenytoin and 1′-hydroxybufuralol were obtained from Ultrafine Chemicals, Ltd. (Manchester, UK). All other chemicals were analytical grade quality and all solvents were high-pressure liquid chromatography (HPLC) grade.

Microsomes.

Microsomes were obtained from Gentest Corp. from metabolically competent human B lymphoblastoid cell lines that stably express human cytochromes P-450 1A1, 1A2, 2A6, 3A4, 2B6, 2C9, 2C19, and 2D6 (catalog nos. M111b, M103c, M104r, M107r, M110a, M118r, M119a, and M117r, respectively) or from baculovirus-insect cell-expressed CYP2C8 (Supersomes, catalog no. P212).

Cytochrome P-450 Assays.

Assays were performed in a final volume of 0.25 ml containing 1.3 mM NADP⁺, 3.3 mM glucose 6-phosphate, 0.4 U/ml glucose 6-phosphate dehydrogenase, 3.3 mM MgCl₂, and 50 mM (for CYP2A6 and CYP2C19) or 100 mM (for CYP1A1, CYP1A2, CYP3A4, CYP2B6, CYP2C8, and CYP2D6) potassium phosphate buffer, pH 7.4. Assays for CYP2C9 were carried out in 100 mM Tris buffer, pH 7.5. Substrate, microsomal protein concentrations, and literature references for the assays were as follows: CYP1A1 and CYP1A2, 100 μM phenacetin and 400 μg/ml protein (Butler et al., 1989); CYP2A6, 400 μM coumarin and 200 μg/ml protein (Greenlee and Poland 1978); CYP3A4, 120 μM testosterone and 200 μg/ml protein (Oldham and Clarke 1977); CYP2B6, 50 μM 7-ethoxy4-trifluoromethylcoumarin and 400 μg/ml protein (DeLuca et al., 1988); CYP2C8, 10 μM paclitaxel and 30 μg/ml protein (Rahman et al., 1994); CYP2C9, 15 μM diclofenac and 17.2 μg/ml protein (Leeman et al., 1993); CYP2C19, 100 μM [¹⁴C]-(S)-mephenytoin (specific activity, 5.7 mCi/mmol) and 1 mg/ml protein (Wrighton et al., 1993); CYP2D6, 10 μM (±)-bufuralol and 100 μg/ml protein (Kronbach et al., 1987). Enzyme concentrations and incubation times were selected to give less than 10% conversion into the measured metabolite to ensure first order reaction rate kinetics. Incubations were carried out for 30 min at 37°C, except for 20 min with the CYP2C8 assay.

The reactions were stopped by the addition of 50 μl (CYP1A1, CYP1A2, CYP2C8, CYP2C19) or 125 μl (CYP3A4) acetonitrile, 50 μl of 20% trichloroacetic acid (CYP2A6, CYP2B6), 50 μl of 94% acetonitrile-6% glacial acetic acid (CYP2C9), or 25 μl of 70% perchloric acid (CYP2D6). After cooling on ice for 5 min, the incubations were centrifuged at 10,000g for 3 min and aliquots were removed for further analysis.

HPLC and Fluorometric Analysis.

The HPLC system consisted of Waters model 510 pumps, a Waters model 717 plus Autosampler, and a Waters model 486 tunable absorbance detector or a model 996 photo diode array detector. This system was controlled by an IBM-compatible computer using Waters (Milford, MA) Milennium Version 2.15 software. Instead of absorbance detectors, a Packard Radiomatic Flo-One flow scintillation analyzer was used to detect peaks in the 2C19 assay and an Optical Technical Devices scanning spectrofluorometer (Elmsford, NY) was used in the 2D6 assay.

Aliquots of supernatant from all except the CYP2A6 and CYP2B6 assays were injected onto a Supelco 4.6 × 250-mm Nucleosil C₁₈ 5-μm HPLC column (Bellefonte, PA) and separated at 45°C with a flow rate of 1 ml/min. The HPLC column was protected by a Waters 3.9 × 20 mm 60A 4-μm Sentry guard column. Metabolite peaks were separated with 10% methanol (mobile phase A), 100% methanol (mobile phase B), or 30% acetonitrile-70% water-0.16 ml/liter 70% perchloric acid (mobile phase C) as follows:

CYP1A1/CYP1A2 assay.

This assay was as follows: initial conditions of 100% A changed to 85% A:15% B over 6 min, then changed to 100% B over 1 min and held for 5 min. The acetamidophenol product (R_t = 10 min) was detected by absorbance at 244 nm.

CYP2C8 assay.

This assay was as follows: initial conditions of 45% A:55% B (flow rate = 0.5 ml/min) changed to 42.5% A:57.5% B (flow rate = 1.0 ml/min) over 5 min, then changed to 35% A:65%B over 20 min. The 6α-hydroxypaclitaxel product (R_t = 15 min) was detected by absorbance at 230 nm.

CYP2C9 assay.

This assay was as follows: initial conditions of 30% B:70% C changed to 100% B over 15 min and held for 1 min. The 4′-hydroxydiclofenac product (R_t = 12 min) was detected by absorbance at 280 nm.

CYP2C19 assay.

This assay was as follows: initial conditions of 75% A:25% B changed to 100% B over 8 min and held for 3 min. The metabolite product (R_t = 8 min) was detected by flow liquid scintillation detection.

CYP2D6 assay.

This assay was as follows: initial conditions of 100% C held for 7 min, then changed to 100% B over 1 min and held for 5 min. The 1′-hydroxybufuralol product (R_t = 5.5 min) was detected by fluorescence with excitation at 252 nm and emission at 302 nm.

CYP3A4 assay.

This assay was as follows: initial conditions of 47% A:53% B changed to 42% A:58% B over 8 min, then changed to 100% B over 0.1 min and held for 5.9 min. The 6β-hydroxytestosterone product (R_t = 8 min) was detected by absorbance at 254 nm.

Aliquots (100 μl) from the CYP2A6 and CYP2B6 assays were diluted with 1.9 ml of 0.1 M Tris buffer, pH 9.0. Fluorescence was measured at 368-nm excitation and 456-nm emission for CYP2A6 and at 410-nm excitation and 510-nm emission for CYP2B6. Standard curves were generated for each assay using 7-hydroxycoumarin for CYP2A6 and 7-hydroxy-4-trifluoromethylcoumarin for CYP2B6.

Data Analysis.

The results were obtained from at least four determinations (from duplicate incubations from at least two separate experiments) for each solvent concentration, using different lots of microsome preparations. Data are expressed as percentage of inhibition compared to control incubations containing no solvent.

Results

The effect of methanol, ethanol, DMSO, and acetonitrile on the catalytic activities of nine human cDNA-expressed cytochrome P-450s are shown in Table 1. Initially, experiments were performed at solvent concentrations of 0.3, 1, and 3%. In most instances where there was >20% inhibition at 0.3% solvent, additional experiments were done with overlapping solvent concentrations down to 0.1%. Acetonitrile was tested at higher concentrations (up to 10%) because preliminary experiments suggested that acetonitrile was a comparatively weak inhibitor of cytochrome P-450 activity.

Acetonitrile was observed to be only slightly inhibitory to CYP2B6 and not inhibitory to any of the other hepatic enzymes tested at a concentration of 0.3%. However, at 0.3% acetonitrile, 26% inhibition of the extrahepatic enzyme, CYP1A1, was observed. At 1%, a slight inhibition (<20%) of CYP1A2, CYP2A6, CYP2C19, CYP2D6, and CYP3A4 was observed. While at 3% only CYP2C9 was not inhibited; this lack of inhibition extended to 5% acetonitrile. At 10%, acetonitrile inhibited CYP2C9 activity by 62% and the other cytochrome P-450 activities by 92 to 100% (data not shown).

Methanol was not inhibitory to any enzyme at a concentration of 0.3%. At 1%, approximately 25% inhibition of CYP1A1 and CYP2D6 was observed with lower levels of inhibition noted for CYP2B6 and CYP2C9 (19% and 12%, respectively). CYP1A2, CYP2A6, and CYP2C8 were not inhibited by 3% methanol.

DMSO was slightly inhibitory (13%) to CYP2B6 and quite inhibitory (25–38%) to CYP2C19, CYP2D6, and CYP3A4 at a concentration of 0.3%. Further testing revealed that significant inhibition (17–23%) was still present at 0.1% DMSO. At 1 and 3% DMSO, only CYP1A2, CYP2A6, and CYP2C8 were not inhibited.

Ethanol was quite inhibitory (19–28%) to CYP1A1, CYP2B6, and CYP2C19 at a concentration of only 0.1%. At 0.3%, 24% inhibition of CYP2D6 was observed. At 1%, 18% inhibition of CYP2A6 was observed. While at 3%, only CYP2C9 was not inhibited but the inhibition of CYP2C8 was relatively small (16%).

Discussion

These results show that there is no consistent effect of any of the four common organic solvents examined on the activities of nine cDNA-expressed human cytochromes P-450 over a moderate solvent concentration range (0.1–3%). Any given solvent has a variable inhibitory effect on the different cytochromes examined and any given cytochrome P-450 activity is variably inhibited by the different solvents tested.

If one groups the effects by enzyme, CYP1A2, CYP2C8, and CYP2C9 appear to be the most resistant to inhibition by the tested organic solvents. CYP1A1, CYP2B6, CYP2C19, and CYP2D6 appear to be the most sensitive to inhibition by these solvents. The latter enzymes tolerate no solvents at concentrations at 3% and many solvents are substantially inhibitory at 1% or below. CYP2A6 and CYP3A4 have an intermediate response. CYP2A6 is inhibited by greater than 1% ethanol or acetonitrile, whereas CYP3A4 is inhibited by all solvents at 3% and by only DMSO at lower concentrations.

Clearly, purely aqueous incubation conditions are preferred for cytochrome P-450 incubations. If an organic solvent must be used, methanol appears to be the most suitable for this purpose. This conclusion is based on our observation that none of the cytochromes P-450 tested were inhibited more than 10% by 0.3% methanol. Likewise, acetonitrile demonstrates modest inhibitory effects. There was no substantial inhibition (<10%) of most cytochromes P-450 at 0.3% acetonitrile, except for CYP2B6 (13%) and the extrahepatic form, CYP1A1 (26%). Both ethanol and DMSO substantially inhibited (10–30%) some cytochromes as low as 0.1% solvent.

Two similar studies on the effects of organic solvents on a number of cytochrome P-450 activities in human liver microsomes have been recently reported (Chauret et al., 1998; Hickman et al., 1998). Chauret et al. (1998) examined multiple solvent concentrations. Hickman et al. (1998) examined only a single concentration (1%). Although there are differences in enzyme source and some of the assay conditions, there were many similarities in the results of the three studies. For example, CYP1A2 activity was not affected substantially by methanol at concentrations ≤1%. Coumarin 7-hydroxylase (CYP2A6) activity was affected by methanol or DMSO at ≤5%. A similar observation at 1% solvent has also been reported (Draper et al., 1997). In the testosterone 6β-hydroxylase (CYP3A4) assay, there was little effect with methanol and acetonitrile at levels ≤0.5%, and DMSO was strongly inhibitory at low (0.2%) concentrations (this report and Chauret et al., 1998). For CYP2C19 activity, strong inhibition at low concentrations of DMSO was observed, as well as little inhibition by methanol and acetonitrile at concentrations ≤1%. Ethanol as a solvent and the other solvent effects on CYP1A1, CYP2B6, or CYP2C8 activities were not tested by Chauret et al. (1998) or Hickman et al. (1998).

There are examples of differences between our data and that previously reported (Chauret et al., 1998; Hickman et al., 1998). For example,Hickman et al. (1998) observed DMSO to inhibit caffeine 3-demethylase activity, whereas we and Chauret et al. (1998) observed DMSO to be less inhibitory to phenacetin O-deethylase activity. Chauret et al. (1998) and Hickman et al. (1998) observed inhibition of tolbutamide hydroxylation (CYP2C8/9) by methanol and DMSO. In contrast, we observed paclitaxel 6α-hydroxylase (CYP2C8) activity and diclofenac 4′-hydroxylase (CYP2C9) activity were only slightly inhibited (<15%) by up to 3% methanol or by 1% DMSO. Another point of difference was that Chauret et al. (1998) and Hickman et al. (1998) recorded little inhibition of dextromethorphan O-demethylase activity at methanol and DMSO concentrations ≤1%. By contrast, our findings showed significant inhibition of CYP2D6 catalyzed bufuralol 1′-hydroxylase by 0.3% DMSO (38%) and 1% methanol (26%). Hickman et al. (1998) did not observe dextromethorphan N-demethylase activity (CYP3A4) to be inhibited by DMSO, whereas we and Chauret et al. (1998) observed testosterone 6β-hydroxylase activity to be substantially inhibited.

The results from our study with single recombinant human cytochromes P-450 generally agree with the results from previously published studies with human liver microsomes (Chauret et al., 1998; Hickman et al., 1998) only when the enzyme/substrate pairs are the same (i.e., 1A2/phenacetin, 2A6/coumarin, 2C19/(S)-mephenytoin, and 3A4/testosterone). All major points of difference involve a change in substrate, suggesting that the inhibitory effects of solvents are substrate-dependent for a given cytochrome P-450. Further experimentation is required to define the extent of this substrate selectivity.

We wish to emphasize several practical implications from results reported here and those previously reported (Chauret et al., 1998;Hickman et al., 1998).

If one is evaluating the relative rates of metabolism of a substrate to different products, one should select a solvent type and concentration that are not inhibitory to the participating enzymes. It may be desirable to examine two solvent concentrations to verify that the selected conditions do not affect the results.

If one is evaluating potentially selective probe substrates or inhibitors for the CYP1A subfamily, solvent choice may be problematic. Our results and those of Chauret et al. (1998) suggest the use of methanol or DMSO (0.3% or less) is preferred to avoid inhibiting CYP1A1. Hickman et al. (1998) observed a 1% concentration of these solvents to be quite inhibitory to CYP1A2-catalyzed caffeineN-demethylase activity.

DMSO is the solvent of choice for solubilization of chemical libraries for high throughput screening applications for pharmacological activity. Three major drug-metabolizing enzymes (CYP2C19, CYP2D6, and CYP3A4) are profoundly inhibited by this solvent in our study. Therefore, if cytochrome P-450-mediated metabolic data are to be obtained, the chemical library should be prepared in an alternative solvent.

It seems possible that the selectivity of inhibition of organic solvents could be exploited to improve the selectivity of the enzyme-selective probe assays. For example, if a specific biotransformation is catalyzed by two enzymes, and the minor contributing enzyme, but not the major contributing enzyme, is inhibited by a specific concentration of solvent, then the addition of a selected solvent at the correct concentration can improve the selectivity of the biotransformation for the major contributing enzyme.

Commercially available radiolabeled (S)-mephenytoin is supplied as an ethanol solution. Since this substrate is commonly used to assay not only CYP2C19, but also CYP2B6 (Heyn et al., 1996), it is critical that ethanol be removed before use.

Finally, it appears noteworthy that ethanol is a significant inhibitor of CYP1A1, CYP2B6, and CYP2C19 at the lowest solvent concentration tested (0.1%). Similar blood human alcohol concentrations are not uncommon and may have implications to in vivo xenobiotic metabolism and drug pharmacokinetics.