Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

DSG is a 3-deoxy progestogenic oral contraceptive steroid. The steroid itself lacks potent progestogenic activity (due to the absence of an oxygen function at the 3-position of the steroid ring), thus requiring bioactivation to 3-KDSG to exert its desired biological effect (Viinika et al., 1976). Limited in vitro data identified the major metabolite of DSG as 3-KDSG with 3-OHDSG and 3-OHDSG as the presumed intermediates, and more polar metabolites also being formed (Viinika, 1979; Madden et al., 1990). The pharmacokinetics of DSG have been extensively studied and support the theory that 3-KDSG is the active metabolite (Back et al., 1987; McClamrock and Adashi, 1993).

The enzymes responsible for bioactivation of DSG have never been formally identified although there is some evidence for the involvement of CYP isoforms (Madden et al., 1990). The CYP family of enzymes is responsible for the oxidation of structurally diverse lipophilic chemicals and plays an important role in the hydroxylation of endogenous steroids resulting in the biosynthesis of all major classes of steroid hormones (Waterman, 1986). Individual CYPs have the ability to hydroxylate endogenous and synthetic steroids with a high degree of stereospecificity and regioselectivity (Waxman et al., 1991). CYP3A4 has been implicated as the major catalyst of steroid 6-hydroxylation being responsible for the 6-hydroxylation of testosterone, androstenedione and progesterone (Waxman et al., 1988), cortisol (Abel and Back; 1993) and the synthetic corticosteroids budesonide (Jönsson et al., 1995) and dexamethasone (Gentile et al., 1996). Hydroxylation can occur at different positions on the steroid ring, which in some cases are CYP3A-mediated reactions (Guengerich, 1988; Waxman et al., 1991; Yamazaki and Shimada, 1997), but can also be mediated by other CYP isoforms such as CYP1A2 (Aoyama et al., 1990). The CYP2C subfamily has been associated with EE₂ 2-hydroxylation (Ball et al., 1990) and recently CYP2C9 has been implicated as the major catalyst in the bioactivation of another steroidal pro-drug, mestranol (Schmider et al., 1997). This compound is a synthetic oestrogen (differing from EE₂ by the presence of a methoxy group at the 3-position), but to become biologically active conversion to EE₂ is required. In vitro studies have suggested that CYP2C9 is responsible for this biotransformation on the basis of inhibition by sulfaphenazole (a selective inhibitor of CYP2C9). Yamazaki and Shimada (1997) have recently investigated the involvement of CYP2C19, CYP2C9 and CYP3A4 in steroid hydroxylation using human liver microsomes and recombinant CYP enzymes. The studies revealed that progesterone 21-hydroxylation was catalysed by CYP2C19 (and CYP2C9 to a lesser extent), as was the conversion of testosterone to androstenedione.

DSG-containing oral contraceptives have been marketed for more than 10 yr now but there is still only limited in vitro data on its metabolism. There is an obvious need to identify the enzymes involved in the formation of the biologically active metabolite 3-KDSG, especially with the expected introduction of oral contraceptives onto the Japanese market (the Japanese population is known for their CYP2C19 polymorphism; Goldstein and Morais, 1994). Thus it was our intention in this study to confirm the pathway of 3-KDSG formation from DSG suggested by Madden et al. (1990) and more importantly, to attempt to identify the enzyme(s) involved in DSG bioactivation in human liver microsomes using the following characterisation techniques 1) the use of specific low molecular weight enzyme inhibitors, 2) correlations with immunochemically determined levels of CYP isoforms and/or probe substrate activities, 3) use of heterologously expressed enzyme systems and 4) inhibitory antibodies raised against specific CYP isoforms. Such techniques have been well documented in the past and usually give a reasonably strong indication of the CYP isoforms involved in the metabolism of compounds (Guengerich and Shimada, 1991).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Chemicals

[16-³H]DSG, [3-²H]DSG (label present at both the - and -positions), unlabeled DSG and its metabolites were kindly provided by Organon (Oss, The Netherlands). Tolbutamide, coumarin, quinidine, chlorzoxazone, methyl p-tolyl sulfide, its sulfoxide, NADP, G6P, glucose 6-phosphate dehydrogenase, diethylenetriaminepentaacetic acid and NADPH were obtained from Sigma-Aldrich Chemical Co. (Poole, England). Furafylline and S-mephenytoin were purchased from Ultrafine Chemicals (Manchester, England). Ketoconazole was kindly donated by Janssen (Beerse, Belgium) and sulfaphenazole was a gift from Ciba-Geigy (Horsham, England). Chlorpropamide was kindly donated by Pfizer Central Research (Sandwich, England). HPLC solvents were of Analar grade and were supplied by Fisons (Loughborough, England). Scintillation fluid (Flo Scint A) was purchased from Canberra Packard (Pangbourne, England). All other chemicals were from BDH (Poole, England).

Human Liver Samples

Histologically normal livers were obtained from transplant donors. Consent for their removal was obtained from the donor's relatives and Ethics Committee approval was granted for their use in this study. Livers were stored as 10- to 20-g portions at 80°C until use.

Washed microsomes were prepared by the classical differential sedimentation method (Purba et al., 1987). Human liver microsomal protein content was determined by the method of Lowry et al. (1951).

Incubations with Deuterated DSG

Incubations contained [²H]DSG (10 µM), 1 mg · ml¹ microsomal protein, MgCl₂ (3 mM), an NADPH-regenerating system (G6P 5 mM, NADP 500 µM and G6P-dehydrogenase 0.63 U) and 0.067 M phosphate buffer to a final volume of 5 ml. The NADPH-regenerating system was incubated for 2 min before commencement of the incubation to allow for generation of NADPH. The reaction was terminated after 15 min by freezing on solid CO₂. Samples were stored at 20°C until use.

Solid Phase Extraction of Metabolites for LCMS

After allowing the sample to thaw on ice, it was loaded onto a C₁₈ solid phase cartridge and washed with one volume of 100% double distilled water (ddH₂O) followed by one volume of a mixture of ddH₂O:acetonitrile (90:10%; v/v). The sample was then eluted with one volume of a mixture of ddH₂O:acetonitrile (10:90%; v/v). The eluate was evaporated under nitrogen and reconstituted into mobile phase for analysis by liquid chromatography-mass spectrometry (LCMS).

LCMS

[²H]DSG and its metabolites were eluted from a Hypersil ODS column (100 × 2.1 mm i.d.). Elution was isocratic for 7 min using a mobile phase of ddH₂O:acetonitrile (55:45; v/v), increasing to 10:90 v/v over a 5-min linear gradient. A flow rate of 0.25 ml · min¹ was used.

Mobile phase was delivered by a HP1090 pump unit to the interface of a triple quadrupole mass spectrometer API 300 (Perkin-Elmer Sciex Instruments, Holland/USA). Nebulizing gas (N₂) and drying gas (air) were delivered at 50 and 200 l · hr¹, respectively. The interface temperature was 480°C and the orifice and ring voltages 15 and 200 V, respectively. Ion current chromatograms were collected between 250 to 350 amu.

In Vitro Metabolism of DSG by Human Lymphoblastoid Cell Line Microsomes

Incubations were performed using commercially available microsomes prepared from AHH-1 TK ± human lymphoblastoid cell lines expressing individual enzymes (Gentest Corporation, Woburn, MA). As a preliminary screen, the microsomes used were from cell lines expressing CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP3A4 and NADPH-CYP reductase. A control cell line, cHol (which natively expresses low levels of CYP1A1), was also used.

DSG (1 µM; 0.2 µCi; in ethanol, final volume 1%) was incubated at 37°C for a period of 15 to 60 min in the presence of MgCl₂ (10 mM), microsomal protein (0.1 mg) NADPH (1 mM) and 0.067 M phosphate buffer to a final volume of 500 µl. [Incubations with the CYP2C9 microsomes were performed using Tris-HCl buffer (0.1 M; pH 7.4)]. Reactions were initiated by the addition of NADPH.

Reactions were terminated by extraction of DSG and its metabolites into ethylacetate (3 ml). After separation by centrifugation, the organic phase was evaporated to dryness and the samples reconstituted into mobile phase before analysis by radiometric HPLC.

HPLC Analysis

DSG and its metabolites were resolved on a Prodigy 5ODS-2 column (5 µm, 15 cm × 4.6 mm i.d.). Elution consisted of solvent A:solvent B (25:75; v/v) increasing to 40:60 over a 5-min linear gradient, further increasing to 70:30 over a 12-min linear gradient followed by an increase to 100% solvent A over a 5-min linear gradient. Elution was then decreased to 25:75 over a 5-min period followed by 10-min isocratic elution. Solvent A, acetonitrile:methanol (70:30; v/v); Solvent B-ddH₂O (pH 3.0). The flow rate was 1.3 ml · min¹ and the column was maintained at 50°C. Mobile phase was delivered by a Kontron pump system 325 linked to an on-line A250 Flo-One radiomatic detector.

Inhibitor and Kinetic Studies with CYP2C9 and CYP2C19 Cell Line Microsomes

Studies were performed to determine apparent K_m and V_max values for DSG 3-hydroxylation by CYP2C9 cell line microsomes. Enzyme activity was measured (under linear conditions) for seven DSG concentrations (1-50 µM; 0.2 µC_i) using 75 µg microsomal protein, MgCl₂ (10 mM) and NADPH (1 mM) in a final volume of 0.5 ml for 10 min. K_m and V_max values were determined by nonlinear regression analysis using an iterative program (GRAFIT) which fits the Michaelis-Menten equation to the experimental data.

A study to determine the inhibitory potential of SULF (0-1.0 µM) against the formation of 3-OHDSG by CYP2C9 cell line microsomes was undertaken using DSG (1 µM; 0.2 µC_i) and the same incubation conditions as mentioned above. IC₅₀ values, i.e., concentration producing 50% inhibition of control enzyme activity, were determined using an iterative program (GRAFIT). In addition, the apparent K_i was determined for SULF concentrations of 0 to 1.0 µM and DSG concentrations of 1 to 30 µM using ENZPAK (version 3.0; BIOSOFT).

The apparent K_m and V_max values for TOL 4-hydroxylation by CYP2C9 cell line microsomes were determined. TOL (50-500 µM) was incubated for 30 min using the same incubation conditions as DSG. The reaction was terminated by the addition of 6 N HCl (25 µl). Chlorpropamide (10 µl of 12.5 µg · ml¹) was used as the internal standard. Unmetabolized TOL, 4-hydroxy-TOL and chlorpropamide were extracted into methyl-tert butyl ether (3 ml). After evaporation of the organic solvent, the residue was reconstituted into 150 µl of mobile phase and analysed by HPLC (Back et al., 1988). K_m and V_max values were calculated. In addition, IC₅₀ values were determined for the inhibition of TOL (100 µM) metabolism by SULF (0-1.0 µM).

DSG (1 µM; 0.2 µC_i) metabolism by microsomes from lymphoblasts containing cDNA-expressed CYP2C19 and the inhibitory potential of S-MEPH (a substrate for CYP2C19; 0-200 µM) was also investigated. A 30-min incubation was used.

Human Liver Microsomal Incubations

DSG (1 µM; 0.1 µCi) was incubated in the presence of microsomal protein (1 mg · ml¹), MgCl₂ (10 mM), NADPH (1 mM) and 0.067 M phosphate buffer in a final volume of 1 ml at 37°C. Initial studies were performed to determine the extent of metabolite formation at different times (0-30 min).

In subsequent studies, an incubation time of 5 min was selected for optimum 3-OHDSG formation and an incubation time of 30 min was selected for optimal 3-KDSG formation or accumulation (fig. 1).

Correlation studies. The relative CYP3A4 content and DEX 6-hydroxylase activity (a probe for CYP3A4) of a panel of human livers has been previously determined in our laboratory (Gentile et al., 1996). Studies were performed to determine either the formation of 3-OHDSG or 3-KDSG in the same panel of human livers. These activities were then compared with CYP3A4 blot/activity using the Spearman rank (Rs) correlation test.

Inhibitor studies. The formation of 3-KDSG was assessed when DSG was coincubated in the presence of the following selective CYP inhibitors/substrates: ketoconazole (0.1 and 1.0 µM; CYP3A4 inhibitor), sulfaphenazole (1 and 10 µM; CYP2C9 inhibitor), furafylline (1 and 10 µM; CYP1A2 mechanism based inhibitor, preincubated for 15 min in the absence of DSG), chlorzoxazone (50 and 100 µM; CYP2E1 substrate), quinidine (1 and 10 µM; CYP2D6 inhibitor) and coumarin (1-10 µM; CYP2A6 substrate).

Inhibitory antibodies. Anti-CYP reductase. An inhibitory antibody raised against human NADPH CYP reductase in goat (donated by Professor M. McManus) was used to assess the involvement of this enzyme in 3-KDSG formation. The antibody (0-200 µl) was preincubated with microsomal protein (0.5 mg) at 4°C for 30 min. Incubations were performed with DSG (1.0 µM; 0.1 µCi) as previously described.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Figure 1a shows a typical radiomatic chromatogram illustrating the microsomal metabolism of DSG (1 µM; 0.1 µCi) by human liver microsomes after a 5-min incubation. The predominant metabolite (based on retention time by cochromatography with authentic standards) was 3-OHDSG (Rt = 19.9 min) with much less amounts of 3-OHDSG (Rt = 16.5 min) and 3-KDSG (Rt = 17.6 min) being observed. Unmetabolised DSG had a Rt = 26 min. Figure 1b illustrates microsomal metabolism of DSG (1 µM; 0.1 µCi) after a 30-min incubation. The major metabolite observed was 3-KDSG. The more polar metabolites eluted between 2 to 14 min (presumed further hydroxylated metabolites of 3-KDSG).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Radiomatic chromatograms illustrating DSG metabolism (1 µM; 0.1 µCi) by human liver microsomes after a, a 5-min incubation and b, a 30-min incubation.

Mass spectrometry data from the deuterated DSG experiments are summarized in table 1. The pseudomolecular ion ([M + 1]⁺) of the monohydroxylated isomers were not observed; however, fragment ions at m/z 310.3 (H₂O) and 292.3 (2 × H₂O) were in accordance with the spectra obtained for the synthesized [²H]3-OHDSG standard and were 1 amu more than nondeuterated standards. A pseudomolecular ion at m/z 325 was observed that would correspond to a keto derivative of DSG that had lost both deuterium labels from the 3-position. The spectrum was in accordance with that obtained for synthesized 3-KDSG. [²H]DSG gave a pseudomolecular ion at m/z 313. Because the hydroxylated metabolites still contain the [²H] label these results prove that DSG is initially hydroxylated at the 3- and 3-positions and then undergoes dehydrogenation to form 3-KDSG (table 1).

Both CYP2C9 and 2C19 microsomes were able to hydroxylate DSG with the major metabolite being 3-OHDSG (fig. 2). Lesser amounts of 3-OHDSG was observed as was a small amount of 3-KDSG. The microsomes that failed to metabolise DSG were cHol, CYP1A2, 3A4, 2C8 and NADPH-CYP reductase. Phenacetin O-deethylase, dexamethasone 6-hydroxylase and cytochrome c reduction at 550 nm was used as a positive controls for CYP1A2, CYP3A4 and NADPH-CYP reductase respectively (results not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Radiomatic chromatograms illustrating DSG metabolism (1 µM; 0.2 µCi) after a 30-min incubation with a, CYP2C9 cell line microsomes and b, CYP2C19 cell line microsomes.

Kinetic and inhibition data for DSG 3-hydroxylation and TOL 4-hydroxylation by CYP2C9 cell line microsomes is shown in table 2. The average K_m values were found to be 6.5 and 88.7 µM for DSG and TOL, respectively, and the corresponding V_max value were 1269 and 247 pmol · mg¹ · min¹, respectively (data are calculated from two individual experiments). SULF inhibited DSG 3-hydroxylation and TOL 4-hydroxylation producing IC₅₀ values of 0.67 and 0.21 µM, respectively. In addition, the apparent K_i for SULF against DSG 3-hydroxylation was 0.91 µM.

View this table: [in this window] [in a new window]   TABLE 2 Kinetic and inhibition parameters (SULPH) for DSG 3-hydroxylation and TOL 4-hydroxylation by CYP2C9 cell line microsomes

The formation of 3-OHDSG by CYP2C19 cell line microsomes was moderately inhibited by S-MEPH producing an IC₅₀ value of 189 µM (data not shown).

A good correlation was observed between DEX 6-hydroxylation and CYP3A4 content [correlation coefficient (r) = 0.74; P < .005, data not shown]. Conversely, 3-KDSG formation produced a very good negative correlation with the CYP3A4 blot r = 0.85; P = .001 and DEX 6-hydroxylation r = 0.74; P = .0096. There was no significant correlation between 3-OHDSG formation and CYP3A4 content or activity (r = 0.45 and 0.16, respectively).

Figure 3 illustrates the effects of CYP inhibitors at low "selective" inhibitory concentrations (or at concentrations exceeding the K_m where a CYP substrate was used), on 3-KDSG formation by human liver microsomes. The most marked inhibition observed was with sulfaphenazole (32% at 10 µM). Ketoconazole showed a dramatic increase in formation/accumulation of 3-KDSG at both 0.1 (168% of control activity) and 1.0 µM (278% of control activity). No overall change in the formation of 3-KDSG was observed when DSG was incubated in the presence of all six inhibitors at once (data not shown).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   The effect of selective CYP inhibitors/substrates on 3-KDSG formation in human liver microsomes (n = 3 except # where n = 5).

The effects of SULF and S-MEPH on the formation of 3-OHDSG by human liver microsomes are illustrated in figure 4. SULF inhibited formation of 3-OHDSG by 28% (at 1 µM) and 48% (at 10 µM). S-MEPH (100 µM) produced less inhibition, only decreasing 3-OHDSG by approximately 15%. An additive inhibitory effect was observed in the presence of both inhibitors, 39% inhibition (1 µM SULF and 100 µM S-MEPH) and 65% (at 10 µM SULF and 100 µM S-MEPH).

View larger version (30K): [in this window] [in a new window]   Fig. 4.   The effect of sulphaphenazole and S-mephenytoin (alone and in combination) on 3-OHDSG formation by human liver microsomes (n = 3).

A CYP reductase inhibitory antibody greatly inhibited 3-KDSG formation in hepatic microsomes from two human livers (fig. 5). Figure 6 illustrates the effects of an inhibitory CYP3A4 antibody on 3-KDSG formation and dexamethasone 6-hydroxylation (probe substrate for CYP3A4). At an antibody:microsomal protein ratio of 2:1, there was no inhibition of 3-KDSG formation whereas DEX 6-hydroxylation was inhibited approximately 40%. At a 4:1 ratio, DEX 6-hydroxylase activity was completely abolished and approximately 30% of 3-KDSG formation was inhibited. Control serum did not influence enzyme activity screened. The effects of an inhibitory CYP2E1 antibody, using CHLOR 6-hydroxylation as the probe substrate, are also illustrated in figure 6. At both a 2:1 and 4:1 antibody:microsomal protein ratio, CHLOR 6-hydroxylation was inhibited approximately 60%. No inhibition of 3-KDSG formation was observed.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   The effect of an inhibitory CYP reductase antibody on 3-KDSG formation in human liver microsomes. (Each point is the average of two human livers.)

View larger version (51K): [in this window] [in a new window]   Fig. 6.   The effect of an inhibitory CYP3A4 and CYP2E1 antibody on 3-KDSG formation and DEX 6-hydroxylation or CHLOR 6-hydroxylation (n = 3).

Figure 7 illustrates the effect of inhibitory CYP2C9 antisera on 3-OHDSG formation after a 5-min incubation, 3-KDSG formation after a 30-min incubation and TOL 4-hydroxylation after a 30-min incubation. One µl of the antisera inhibited approximately 50% of 3-OHDSG formation. Volumes of more than this completely inhibited metabolism. After a 30-min incubation, 1 µl of the antisera inhibited 3-KDSG formation by approximately 20%. Five µl of the antibody were required to completely block metabolism after a 30-min incubation. Tolbutamide 4-hydroxylation was completely inhibited with 3 µl of the antisera; N.B., control sera had no effect on the activities screened, data not shown.

View larger version (14K): [in this window] [in a new window]   Fig. 7.   The effect of an inhibitory CYP2C9 antiserum on 3-OHDSG formation (after 5 min), 3-KDSG (after 30 min) and 4-OHTOL (after 30 min) by human liver microsomes. Determinations were performed in duplicate in one human liver.

View larger version (13K): [in this window] [in a new window]   Fig. 8.   The proposed metabolic pathway of desogestrel bioactivation in vitro.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

This study has shown that the oral contraceptive steroid DSG is extensively metabolized by human liver microsomes to the major metabolite 3-ketodesogestrel. Using deuterated DSG (labeled at the 3- and 3-positions), it was demonstrated that 3-KDSG was formed via hydroxylated metabolites thus confirming the pathway previously suggested (Madden et al., 1990). Analysis of 3-OH- and 3-OHDSG by LCMS indicated the presence of a ²H as well as a hydroxy group (because the fragment ions observed were 1 amu more than those of unlabeled 3-OH and 3-OHDSG). It appears that these intermediates undergo further rapid metabolism to form 3-KDSG such that even at a very short incubation time (5 min), where 3-OHDSG is the predominant metabolite, 3-KDSG formation is still observed. This rapid biotransformation has also been demonstrated in in vivo studies (McClamrock and Adashi, 1993). Volunteers administered identical doses of oral DSG (150 µg) or 3-KDSG (150 µg) both in combination with ethinyloestradiol (30 µg) had comparable 3-KDSG plasma concentrations.

Because the biotransformation of DSG to 3-KDSG is very rapid both in vitro and in vivo, we have examined both 3-OHDSG and overall 3-KDSG formation by human liver microsomes and/or cell line microsomes and aimed to establish which enzyme(s) are responsible for the initial hydroxylation of DSG in vitro using a number of well documented characterization techniques.

Heterologously expressed enzymes were initially used to establish specific isoform involvement. Results from such studies should be interpreted with some caution because although a cDNA expressed enzyme may have the capability of metabolising a given compound, the levels of that enzyme in vivo may be very low and the contribution by that particular enzyme may be negligible. Microsomes from human lymphoblastoid cell lines genetically engineered to express CYPs 1A2, 2C8, 3A4 and NADPH-CYP reductase did not metabolize DSG. However, microsomes from cell lines expressing CYP2C9 and CYP2C19 metabolized DSG primarily to 3-OHDSG with small amounts of 3-OHDSG and 3-KDSG formed (possibly due to the presence of some native hydroxysteroid dehydrogenase).

The kinetics of 3-hydroxylation by CYP2C9 cell line microsomes were examined and apparent K_m (6.5 µM) and V_max (1269 pmol · mg¹ · min¹) values determined. SULF potently inhibited 3-hydroxylation (IC₅₀ = 0.67 µM; K_i = 0.91 µM) which was in accordance with the inhibition of tolbutamide 4-hydroxylase (IC₅₀ = 0.21 µM) in the same microsomes. Clearly these results demonstrate that CYP2C9 plays an important role in DSG 3-hydroxylation.

Correlation studies are often used to further substantiate enzyme involvement in a metabolic biotransformation. The enzyme activity under investigation can be correlated with 1) metabolism of marker substrates and/or 2) with immunochemically determined levels of individual enzymes in a panel of livers. A range of human livers were immunoblotted for relative CYP3A4 content and the metabolism of a probe substrate for CYP3A4 (dexamethasone 6-hydroxylation) determined in the same panel of livers. There was a significant negative correlation between 3-KDSG and CYP3A4 content/activity, suggesting that the further metabolism of 3-KDSG is mediated by CYP3A4. No correlation was observed between 3-OHDSG formation and CYP3A4 content/activity.

A number of well documented low molecular weight inhibitors (and some selective substrates) were screened to assess their inhibitory potential on 3-KDSG formation by human liver microsomes. Of these, only sulfaphenazole caused significant inhibition. Sulfaphenazole is used as a selective inhibitor of human CYP2C9 activity at low micromolar concentrations (Back et al., 1988; Baldwin et al., 1995; Ono et al., 1996; Bourrie et al., 1996). However, a point that should be borne in mind is that other major CYP activities are not affected by sulfaphenazole at concentrations up to 100 µM (Baldwin et al., 1995). The maximum inhibition of 3-KDSG formation we observed was 32% at 10 µM sulfaphenazole. This apparent lack of potency indicates that CYP2C9 is not the only CYP isoform responsible for 3-KDSG formation in human liver microsomes (based on the cell line microsomes, CYP2C19 may also be involved).

Ketoconazole is a selective inhibitor of CYP3A4-mediated reactions at low (micromolar) concentrations (Maurice et al., 1992; Baldwin et al., 1995). Inhibition by ketoconazole was not observed; in fact there was a marked increase in the formation/accumulation of 3-KDSG when DSG was coincubated with ketoconazole. It would appear that the polar metabolites are probably further metabolites of 3-KDSG and this pathway is being inhibited by ketoconazole leading to the apparent increase in 3-KDSG. These results suggest that the further metabolism of 3-KDSG by human liver microsomes is likely to be mediated by CYP3A4 and would explain the significant negative correlation between 3-KDSG formation and CYP3A4 content/activity. When all the inhibitors/substrates were combined in an incubation, there was no evidence of inhibition. The reasoning behind such an experiment was to show that DSG was not an indiscriminate substrate of CYP enzymes, i.e., if one CYP isoform was inhibited, DSG metabolism may continue by an alternative enzyme? Thus it would appear from the inhibitor studies that the CYP1A2, 2A6, 2D6, 2E1 and 3A4 isoforms were not involved in 3-KDSG formation, but there is a role for the CYP2C subfamily.

Because the principal metabolite formed by the CYP2C9 and CYP2C19 cell line microsomes was 3-OHDSG, it was decided to more closely examine the initial hydroxylation step in DSG metabolism. Studies were performed in human liver microsomes using a 5-min incubation time for predominant formation of 3-OHDSG to establish whether inhibitors of CYP2C9 and CYP2C19 could potently inhibit the initial hydroxylation of DSG. SULF was screened as a CYP2C9 inhibitor (1 and 10 µM) and S-MEPH was screened for inhibitory effects as a substrate for CYP2C19 (100 µM) in human liver microsomes. The inhibition of 3-OHDSG formation by SULF was slightly more potent compared to inhibition of overall 3-KDSG formation; however, a maximum of 48% inhibition at 10 µM was observed. As mentioned previously, concentrations of less than this are usually sufficient to potently inhibit CYP2C9-mediated activities. A total of 100 µM S-MEPH produced a weak inhibition (15%) of 3-OHDSG formation, N.B., the K_m value of S-MEPH has been determined as approximately 52-67 µM; Wienkers et al., 1996; however, because S-MEPH inhibition of CYP2C19-mediated 3-OHDSG formation produced an IC₅₀ value of 189 µM, this may explain the apparent lack of potency of the CYP2C19 substrate in human liver microsomes, i.e., a higher concentration should have been used to inhibit 3-OHDSG formation by 50%. When combined with SULF (1 and 10 µM) there was an additive effect, i.e., 39 and 65%, respectively.

The use of inhibitory antibodies is a very useful tool to define the role and relative contribution of individual enzymes in a particular biotransformation. An inhibitory NADPH CYP reductase antibody was used to establish whether CYP alone, or other non-CYP enzymes were involved. The reductase enzyme is responsible for supply of electrons from NADPH to CYP in the catalytic cycle. Thus if it could definitively be said that this enzyme was involved it would give a strong indication of CYP involvement. Studies using this antibody demonstrated that the CYP reductase enzyme is an obligatory requirement in 3-KDSG formation and would thus rule out the involvement of other drug-metabolizing enzymes in DSG bioactivation in vitro.

Results from studies with the CYP3A4 and CYP2E1 antibodies have ruled out the involvement of these isoforms in 3-KDSG formation by human liver microsomes. We have demonstrated that these antibodies are inhibitory against their respective isoforms by inhibiting probe substrate activities [dexamethasone 6-hydroxylase as a probe for CYP3A4 (Gentile et al., 1996)] and chlorzoxazone 6-hydroxylase as a probe for CYP2E1 (Peter et al., 1990). However, the CYP2C9 antiserum (previously shown to also inhibit CYP2C19, Yamazaki and Shimada, 1997), completely inhibited DSG metabolism in human liver after a 5-min incubation. TOL 4-hydroxylase was also completely inhibited with similar amounts of antiserum. When the metabolism of DSG was assessed after 30 min, the total metabolism was greater (as expected) but the effects of the antiserum were slightly diminished, i.e., more antiserum was required to completely inhibit metabolism. These results indicate the other CYP2C isoform may become involved when a longer incubation period is used and this isoform requires an increased amount of antiserum to inhibit its activity. This may be a reflection of differential effects of the antiserum on CYP2C9 and 2C19 activities.

Several approaches have been adopted to identify the enzyme(s) involved. When the pharmacological activity of a drug relies on metabolism to an active metabolite, it is essential that an individual receiving the compound can perform the required biotransformation. Clearly, the presence of polymorphically expressed enzymes within a population has implications for the use of a pro-drug if one of these enzymes is involved in the formation of the biologically active metabolite to a major extent. In the case of DSG, the studies confirm that the in vitro bioactivation of DSG to 3-KDSG is via 3- and 3-hydroxylated intermediates (fig. 8) and the data strongly suggest an important role for CYP2C9 in the initial 3-hydroxylation of DSG with a possible role for CYP2C19. However, despite the inference of involvement of polymorphic CYP2C19 and the clear role of CYP2C9 (for which several variants have been identified, Bhasker et al., 1997), the in vivo pharmacokinetic studies to date have shown that bioactivation of DSG to 3-KDSG is a rapid and efficient process (Hasenack et al., 1986; Back et al., 1987; Bergink et al., 1990).