Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

All-trans-retinoic acid (t-RA)¹, the highly receptor-active metabolite of all-trans-retinol (vitamin A₁, t-ROH), can trigger retinoid-mediated signal transduction pathways via binding to nuclear retinoic acid receptors (Petcovich et al., 1987; Means and Gudas, 1995). This t-RA-induced, retinoid receptor-mediated signal transduction is believed to be responsible for many if not all of the observed potent therapeutic and teratogenic effects of retinoids in humans (Soprano and Soprano, 1995).

Oxidative conversions of t-ROH directly control endogenous levels of t-RA and thus play an important role in retinoid receptor-mediated gene expression. Conversion of t-ROH to t-RA in adult hepatic tissues comprises two consecutive reactions (Kim et al., 1992; Napoli, 1996). t-ROH is first oxidized to all-trans-retinal (t-RAL), and the formed t-RAL is then rapidly oxidized to t-RA. It is generally agreed that, under normal conditions, the first reaction is the rate-limiting step and that its rate controls rates of biosynthesis of t-RA from t-ROH.

In hepatic tissues, the major metabolic site of retinoid biotransformation, it is well accepted that conversion of t-ROH to t-RAL can be catalyzed by alcohol dehydrogenase isoenzymes, and that subsequent conversion of t-RAL to t-RA is mainly catalyzed by aldehyde dehydrogenases (Blaner and Olson, 1994; Napoli, 1996; Duester, 1998). In rats, a microsomal retinol dehydrogenase also has been characterized (Napoli and Race, 1990; Napoli et al., 1992), and that enzyme catalyzes the conversion of retinol to retinal effectively. Recently, a human microsomal NAD⁺-dependent dehydrogenase was identified as an efficient retinol dehydrogenase (Gough et al., 1998). Because these enzymes have shown low K_m (4-22 µM) and high V_max/K_m pmol/min/µM, it is widely accepted that these microsomal retinol dehydrogenases may play a major role in biosynthesis of t-RA from retinol in both rodent and human hepatic tissues.

Recent studies have demonstrated that various cytochrome P-450 (CYP)-dependent monooxygenases that are recognized as important xenobiotic-etabolizing enzymes, can also catalyze the second step in the biosynthesis of t-RA from t-ROHthe conversion of t-RAL to t-RA (Roberts et al., 1992; Raner et al., 1996). These findings are important because they suggest an alternative mechanism for endogenous generation of t-RA in addition to the alcohol dehydrogenase/ADLH-dependent pathways. Many CYP isoforms shown to catalyze the reaction are important drug-metabolizing enzymes and it thus seems likely that catalysis of retinoid biotransformation by CYP could result in significant drug-retinoid interactions. However, whether and to what extent CYP isoforms can catalyze conversion of t-ROH to t-RAL (the rate-limiting step) have not been reported to our knowledge.

In an earlier study (Shih and Hill, 1991), it was reported that rat liver microsomes were able to catalyze conversion of retinol to retinal and the reaction was characterized as oxidase-dependent. However, the same reaction was also shown to be NADPH-dependent, sensitive to inhibition by ketoconazole and -naphthoflavone (ANF), and was inducible by 3-methylcholanthrene. Those interesting observations suggested to us that the reaction was primarily, if not exclusively, catalyzed by P-450-dependent monooxygenases rather than an oxidase(s). The importance of oxidation of t-ROH to t-RAL and the possible catalysis by CYP for the reaction appeared to warrant a rigorous and thorough investigation.

In this study, we investigated the potential for CYP-catalyzed oxidative conversion of t-ROH to t-RAL in adult human hepatic tissues. The objectives of this study were: 1) to investigate whether t-ROH can be oxidatively converted into t-RAL via a cytochrome P-450-catalyzed monooxygenation; and 2) if the P-450-dependent reactions proceeded, then which individual CYP isoforms were primarily responsible for the catalysis. To accomplish these two objectives, pooled human liver microsomes (HLMS) were used for establishing the CYP-dependent oxidation of t-ROH to t-RAL. Secondly, selective chemical inhibitors and human cDNA expressed-CYP enzymes (CYP SUPERSOMES) were used for evaluating the significance of individual CYP enzymes in HLMS-catalyzed reactions.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals and CYP Enzymes. t-ROH was purchased from Aldrich Chemical Co. (St. Louis, MO). t-RAL and t-RA were purchased from Sigma Chemical Co. (St. Louis, MO). Furafylline was purchased from Gentest Corp. (Woburn, MA). All other chemicals were purchased either from Aldrich Chemical Co. or from Sigma Chemical Co. Reagents and solvents used were of the highest purity commercially available.

Oxidation of Retinoids Catalyzed by HLMS or CYP SUPERSOMES. In this study, a concentration of 35 µM t-ROH was used for most experiments for two reasons. Firstly, this concentration is within in the range of physiological concentrations of t-ROH in human hepatic tissues (11-4000 µM) (Tanumihardjo et al., 1990). Secondly, this concentration permitted facile analyses for metabolite generation.

Inhibition of Oxidation of t-ROH and t-RAL Catalyzed by HLMS or CYP SUPERSOMES. Experiments designed to assess inhibition by carbon monoxide (CO) followed the same procedures described above except that the incubations were in an atmosphere of CO (80%) and O₂ (20%). The ratio and flow rate of gases were regulated with a gas regulator and incubations opened to air (N₂/O₂, 80:20) served as controls. For chemical inhibition, inhibitors were added to incubation vessels and were preincubated with substrate plus HLMS or CYP SUPERSOMES for 3 min at 37°C before the reactions were initiated by adding NADPH. For mechanism-based inhibition, inhibitors were preincubated with HLMS or CYP SUPERSOMES and NADPH for 15 min at 37°C, and the reactions were initiated by the addition of substrate. Termination of the reactions and extraction of retinoids followed the same procedures as those described above. As suggested by Masimirembwa et al. (1999), a concentration of 10 µM CYP inhibitor was chosen for a one-concentration screen study. Other concentrations were then tested as dictated by the initial results. Incubations without inhibitors served as controls. For heat inactivation experiments, suspensions of HLMS or CYP SUPERSOMES were heated at 100°C for 3 min before addition to incubation vessels.

HPLC Procedures. The solvent delivery system for HPLC consisted of two model 100 A dual piston Beckman pumps and was interfaced with a Shimadzu SPD-10A UV detector (set at a wavelength of 354 nm) and a Shimadzu C-R5A Chromatopac data processor. The HPLC system was equipped with a Beckman mixing chamber and manual injector. Chromatographic conditions described by Kim et al. (1992) were adopted with slight modifications. Analytical eluents consisted of solvent A (acetonitrile/H₂O/acetic acid, 49.75:49.75:0.5, by volume) and solvent B (acetonitrile/H₂O/acetic acid, 90:10:0.04, by volume), both containing 10 mM ammonium acetate. The HPLC elution conditions were as follows: 80% solvent A plus 20% solvent B with a flow rate of 1.2 ml/min for 25 min. Identification and quantitation of retinoids were conducted with two HPLC columns. The first column was a Zorbax octadecylsilane (ODS) column (4.6 × 250 mm) (MAC-MOD Anal. Inc., Chadds Ford, PA) and a flow rate of 1.2 ml/min was used. The second column was a Beckman Ultrasphere Octyl column (2.0 × 250 mm) and a flow rate of 0.2 ml/min was used. One hundred microliters of a mixture of authentic t-ROH, t-RAL, and t-RA or organic supernatant extract of incubation mixtures was loaded onto the HPLC column. Elution times of t-RA, t-ROH, and t-RAL were approximately 14, 16, and 22 min with the Zorbax column and were approximately 18.5, 20.5, and 26.0 min with the Beckman column. Quantitation of retinoids with both HPLC columns was highly consistent. A 50% dilution with HPLC eluent (20% A plus 80% B) of the supernatant before injection on the HPLC column was helpful to achieve a better separation of retinoids. An additional method described by Roberts et al. (1992) was adopted to further confirm the identity of t-RAL generated from the reaction. Briefly, HPLC-purified t-RAL was vortexed with NaBH₄ (a reducing agent for aldehyde, 0.1 mg in methanol) and t-RAL was reduced to t-ROH. Identity of t-ROH was confirmed by HPLC and UV/visible spectroscopy.

Confirmation of t-RAL and t-RA by Liquid Chromatography-Mass Spectroscopy (LC-MS). t-RAL and t-RA generated from incubations of t-ROH and t-RAL plus HLMS were confirmed by LC-MS using a fully integrated Hewlett-Packard Series 1100 LC/mass selective detector system (Hewlett-Packard, Palo Alto, CA). The HPLC system, composed of a binary solvent pump, a vacuum degasser, and a temperature-controlled autosampler and column compartment, was serially interfaced to an Hewlett-Packard 1100 diode array detector (DAD) and mass selective detector quadrupole mass spectrometer. A Beckman Ultrasphere Octyl column (2.2 mm i.d. × 250 mm, 5-µm particle size) was used and operated under the isocratic conditions described above at a flow rate of 0.200 ml/min using a sample injection volume of 10 µl. The column effluent was subjected to positive mode electrospray ionization using nitrogen as a drying gas at a temperature of 350°C and a flow rate of 10.0 liters/min with a nebulizer pressure of 25 psi and a capillary voltage of 4 kV. Signal responses for the analytes were optimized as a function of fragmentor voltage resulting in a setting of 50 V for all analytes. The selection of ion windows, span (±0.3 Da), and dwell (443 ms/ion) for selected ion monitoring data acquisition and processing was accomplished using Hewlett-Packard ChemStation software operating in the high resolution mode. Ions monitored were m/z 285.2 and 301.2, corresponding to the [M+H]⁺ for t-RAL and t-RA, respectively.

Demethylation of Imipramine. Imipramine demethylation was conducted in reaction vessels containing potassium phosphate buffer (0.1 M, pH 7.4) and 0.1 mM imipramine as substrate. HLMS or CYP SUPERSOMES were preincubated with imipramine at 37°C for 3 min. Final volumes of the incubation mixtures were 0.5 ml. Reactions were initiated by the addition of 1 mM NADPH and were continued for 20 min. The reactions were terminated by freezing, and imipramine metabolites were extracted according to the methods described by Sequeira and Strobel (1995). Briefly, 0.25 ml of Na₂CO₃ (2 M, pH 12) was added to the incubation vessels and mixed well. Ethyl ether (2.5 ml) was added to the vessels and vortexed. The incubation mixtures were centrifuged at 3000g for 10 min. The ether layer was decanted and back-extracted into 0.5 ml HCl (0.1 M). Extraction efficiency for the demethylated metabolite was approximately 60 to 68% (Sequeira and Strobel, 1995). After extraction, the ether layer was decanted and dried under N₂. The HPLC mobile phase (500 µl) was used for suspension of the metabolites.

HPLC Analyses of Metabolites of Imipramine Oxidation. With the same solvent delivery system described previously, identification and quantitation of desipramine were conducted with a Supelco LC-PCN column (5 µm, 250 × 4.6 mm) (Supelco, Supelco Park, Bellefonte, PA) following the procedures described by Sequeira and Strobel (1995). Metabolites of imipramine oxidation were separated by HPLC then identified by matching elution times with the standard compounds. The mobile phase was pumped at a flow rate of 1.4 ml/min and consisted of acetonitrile/methanol/K₂HPO₄ (10 mM, pH 7.0) = 40:30:25, and the UV detector was set at 214 nm.

Statistical Analyses. All experimental data were expressed as means ± S.D. for three or four experimental measurements. A Microsoft Excel statistics package (version 5.0, Microsoft Corp., Redmond, WA) was used for all statistical analyses.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Identifications and Quantitations of t-RAL. An HPLC chromatogram of authentic t-RA, t-ROH, and t-RAL is shown in Fig. 1A. Under the chromatographic conditions used, t-ROH was completely separated from t-RAL and t-RA, thus quantitations of each individual retinoid were readily achieved. The mass spectra of authentic t-RAL and t-RA (metabolites of oxidative conversion of t-ROH) are presented in Fig. 1, B and C, respectively. The major ion fragments were detected at 285.2 and 301.2 m/z, resulting from the positive ionization of t-RAL and t-RA ([M+H]⁺), respectively. Detection of [M+H]⁺ at 285.2 and 301.2 m/z were then used for confirmation of the presence of t-RAL and t-RA in incubation mixtures.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   HPLC profiles of authentic t-RA, t-ROH, and t-RAL from a Beckman Ultrasphere Octyl column (2.2 mm i.d. × 250 mm, 5-µm particle size) obtained at 354 nm (A); mass spectra of authentic t-RAL (B); and t-RA (C). For chromatographic and LC-MS conditions, see Materials and Methods.

Characterizations of HLMS-Catalyzed Oxidation of t-ROH. The HLMS-catalyzed oxidation of t-ROH to t-RAL was characterized and the results are presented in Table 1. The reaction was catalyzed after additions of either NADPH or NADH, but NADPH appeared to be a much more effective cofactor than NADH. Omission of these cofactors resulted in undetectable reactions. The reaction was also abolished with heat inactivation and was strongly inhibited by carbon monoxide (80:20 CO/O₂ versus 80:20 N₂/O₂). Activity after addition of an NADPH-regenerating system was higher but not statistically different from that observed with direct additions of NADPH alone.

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Effect of concentration of protein on velocity of HLMS-catalyzed oxidation of t-ROH to t-RAL. t-ROH (35 µM) was incubated with HLMS (0.0-1.0 mg of protein) plus NADPH (1 mM) in potassium phosphate buffer (0.1 M, pH 7.4) at 37°C for 20 min in the dark. Results are means ± S.D. of three to four measurements. For details, see Materials and Methods.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Eadie-Hofstee plots of NADPH-dependent (A) and NAD-dependent (B) conversion of t-ROH to t-RAL in human hepatic microsomes. Km and Vmax for NADPH-dependent reaction were 19.4 µM and 52 pmol/min/mg protein (r2 = 0.77). Km and Vmax for NAD-dependent reaction were 41.5 µM and 280 pmol/min/mg protein (r2 = 0.96).

Chemical Inhibition of HLMS-Catalyzed Oxidation of t-ROH. Table 2 presents the effects of selective CYP inhibitors on the HLMS-catalyzed oxidation of t-ROH to t-RAL. At final concentrations of 10 µM, selective inhibitors for CYP2A6, CYP2B6, or CYP2E1 did not produce significant effects on the reaction. On the other hand, selective inhibitors for CYP family 1 (CYP1A1, CYP1A2, and CYP1B1), the CYP2C and CYP3A subfamilies, and CYP2D6 each produced significant inhibitory effects on reaction rates. Among them, furafylline (highly selective for CYP1A2) and troleandomycin (TAO; highly selective for CYP3A isoforms) each produced profound inhibitory effects on the HLMS-catalyzed oxidation of t-ROH to t-RAL (70-75% inhibition).

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Effects of varying concentrations of chemical inhibitors on HLMS-catalyzed oxidation of t-ROH to t-RAL. ANF (), furafylline (), TAO (), tranylcypromine (), sulfaphenazole (), quercetin (), or quinidine () were incubated with t-ROH (35 µM), HLMS (0.25 mg protein), and NADPH (1 mM) in potassium phosphate buffer (0.1 M, pH 7.4) at 37°C for 20 min in the dark. Results are means ± S.D. of three to four measurements. Statistical differences between control and control plus inhibitor were evaluated with t tests (P < .05). For details, see Materials and Methods.

CYP SUPERSOMES-Catalyzed Conversions of t-ROH to t-RAL and t-RAL to t-RA. To further ensure that the CYP SUPERSOMES used in the study were enzymatically active and to standardize the activities, CYP SUPERSOMES were incubated with imipramine under the same reaction conditions as those used for retinoid biotransformation. Activities of CYP SUPERSOMES for demethylating imipramine to desipramine are presented in Fig. 5. All CYP SUPERSOMES investigated in this study exhibited readily measurable imipramine demethylase activities.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Demethylation of imipramine catalyzed by CYP SUPERSOMES. Imipramine (0.1 mM) was incubated with CYP SUPERSOMES (0.025 nmol) plus NADPH (1 mM) in potassium phosphate buffer (0.1 M, pH 7.4) at 37°C for 20 min. Results are means ± S.D. of three to four measurements. For details, see Materials and Methods.

View larger version (21K): [in this window] [in a new window]   Fig. 6.   Conversion of t-ROH to t-RAL catalyzed by CYP SUPERSOMES. t-ROH (35 µM) was incubated with CYP SUPERSOMES (0.025 nmol) plus NADPH (1 mM) in potassium phosphate buffer (0.1 M, pH 7.4) at 37°C for 20 min in the dark. Results are means ± S.D. of three to four measurements. For details, see Materials and Methods.

View larger version (14K): [in this window] [in a new window]   Fig. 7.   Conversion of t-RAL to t-RA catalyzed by CYP SUPERSOMES. t-RAL (35 µM) was incubated with CYP SUPERSOMES (0.025 nmol) plus NADPH (1 mM) in potassium phosphate buffer (0.1 M, pH 7.4) at 37°C for 20 min in the dark. Results are means ± S.D. of three to four measurements. For details, see Materials and Methods.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

This study has shown for the first time that P-450 cytochromes in human adult hepatic microsomes will catalyze oxidative conversion of t-ROH to t-RAL. The identity of t-RAL generated from incubations of t-ROH plus HLMS was confirmed with HPLC, LC-MS, and reduction of t-RAL generated to t-ROH by reducing agent (NaBH₄). The catalysis was enzymatic as evidenced from observations that the HLMS-dependent reaction was terminated by heat inactivation and also exhibited substrate saturation. Evidence also strongly indicated that this reaction was catalyzed primarily by CYP enzymes because the reaction was NADPH-dependent and was strongly inhibited by CO as well as by various other CYP inhibitors. It is known that CYP can catalyze conversions of a variety of primary or secondary alcohols to aldehydes or ketones through formation of gem-diols with subsequent dehydration to a keto group (Parkinson, 1996). Because t-ROH is a primary alcohol, CYP-catalyzed conversion of t-ROH to t-RAL may also proceed via the same mechanism.

Recent studies have indicated that hepatic concentrations of total retinoid alcohols could be as high as 2860 nmol/g liver (approximately 4 mM) (Tanumihardjo et al., 1990; Barua et al., 1997). As shown in this study, the apparent K_m and V_max values for the NADPH-dependent, HLMS-catalyzed conversion of t-ROH to t-RAL were approximately 19.4 µM and 52 pmol/min/mg protein, respectively. Interestingly, the NAD-dependent, HLMS-catalyzed conversion of t-ROH to t-RAL exhibited a higher K_m (41.5 µM) and a higher V_max (280 pmol/min/mg protein). Because hepatic concentrations of retinol are far higher than either K_m, V_max is indicative of the catalytic efficiency. Therefore, NAD-dependent catalysis appeared to be a major pathway in human hepatic microsomal synthesis of t-RAL from t-ROH. This observation is consistent with the current concept that microsomal retinol dehydrogenases are the primary catalysts for converting t-ROH to t-RA in the liver (Napoli and Race, 1990; Kim et al., 1992; Napoli et al., 1992; Blaner and Olson, 1994; Napoli, 1996; Gough et al., 1998).

However, CYP-dependent oxidation of t-ROH to t-RAL may be important for hepatic and/or extrahepatic biosynthesis of t-RA. The newly identified CYP-dependent oxidation of t-ROH and t-RAL could represent an important alternative mechanism for biosynthesis of both t-RAL and t-RA in vivo. Recent studies in vivo appear to support the important roles of CYP enzymes in hepatic retinoid metabolism. For example, knockout of aryl hydrocarbon receptor, which regulates genes encoding CYP1A1, CYP1A2, and CYP1B1, resulted in significant changes of hepatic levels of retinoids in mice (Andreola et al., 1997; Gonzalez and Fernandez-Salguero, 1998). In addition, under conditions in vivo, CYP enzymes are likely to be saturated by hepatic t-ROH. Because many CYP enzymes also catalyze metabolisms of various human drugs, potential interactions between retinoids and drugs might have pharmaceutical and toxicological significance.

Most P-450 enzymes involved in xenobiotic biotransformation, such as CYP3A4, members of the CYP2C subfamily, CYP2D6, and 1A2, are primarily localized in liver endoplasmic reticulum (microsomes) (Parkinson, 1996). Some P-450 enzymes, such as CYP1A1 and CYP1B1, are mainly expressed in extrahepatic tissues. For example, CYP1A1 is readily detected in lung, intestine, skin, lymphocytes, and placenta (Parkinson, 1996) whereas CYP1B1 is preferentially expressed in hormone-synthesizing organs such as adrenal gland, ovaries, and testes (Juchau et al., 1998).

Catalytic roles for individual CYP enzymes were explored by evaluating the effects of selective inhibitors on the HLMS-dependent oxidation of t-ROH to t-RAL. On the basis of chemical inhibition in hepatic microsomes, members of CYP family 1 and CYP3A subfamily appeared to be quite active whereas members of CYP2C subfamily appeared to be much less active. CYP2A6, CYP2B6, and CYP2E1 exhibited virtually no detectable activity. However, the sole application of chemical inhibition became problematic for further identifying particular CYP enzymes primarily responsible for catalyzing the reaction. This is because many chemical inhibitors available are selective but not specific. Indeed, most of them can inhibit more than one CYP enzyme (Parkinson, 1996). To help address this problem, cDNA-expressed human CYP enzymes (CYP SUPERSOMES) were also used for assessing rates of the reaction catalyzed by individual CYP enzymes. Results from CYP SUPERSOMES complemented the data obtained from chemical inhibition, and thus assisted in identifying the CYP enzymes most responsible for catalyzing the reaction in human adult liver microsomes. To assure that the CYP SUPERSOMES used were enzymatically active and to standardize activities, demethylation of imipramine catalyzed by CYP SUPERSOMES was conducted under the same reaction conditions as those used for studies of retinoid oxidation. Imipramine is a common substrate for members of CYP family 1, for CYP2C and CYP3A subfamilies, and for CYP2D6 (Parkinson, 1996) and a body of information pertaining to these activities is now available. Imipramine demethylase activity, therefore, can serve as an indicator for comparative catalytic activities in CYP SUPERSOMES. Ratios of imipramine demethylation to t-ROH oxidation thus provided important information pertaining to the capacity of individual CYP isoform to catalyze t-RA biosynthesis.

CYP1A2 SUPERSOMES exhibited a relatively low K_m and high V_max for conversion of t-ROH to t-RAL. In human adult liver, CYP1A2 accounts for approximately 18% of total CYP enzymes (Parkinson, 1996). Both high catalytic activity and a relatively high hepatic concentration strongly suggest that CYP1A2 acts as one of the major CYP enzymes to catalyze the oxidation of t-ROH to t-RAL in hepatic microsomes. This is consistent with the observation that furafylline, a highly selective inhibitor of CYP1A2 (Racha et al., 1998), produced strong inhibition (70-75%) of the HLMS-catalyzed reaction. ANF exhibited less potency in blocking CYP1A2-catalyzed conversion of t-ROH to t-RAL than furafylline. At the same concentration (10 µM), ANF produced only approximately 35% inhibition. The relatively low inhibitory potency of ANF was confirmed by comparing the effects of ANF and furafylline on CYP1A2 SUPERSOMES-catalyzed reactions.

CYP3A4 and CYP3A5 SUPERSOMES also showed relatively high activities (roughly 150-200 pmol/min/nmol) for catalyzing conversion of t-ROH to t-RAL. When hepatic levels of the CYP3A subfamily (approximately 40% of total hepatic CYP enzymes) are taken into account, it appears that the CYP3A subfamily would also play an important role for catalyzing the reaction in adult human liver. The fact that TAO effectively inhibited the HLMS-catalyzed reaction appeared to support this assessment. In this regard, separate experiments showed that TAO also inhibited the CYP1A2 SUPERSOMES-catalyzed reaction by 20 to 25% at 10-µM concentrations, indicating that part of the inhibitory effect produced by TAO on the HLMS-catalyzed reaction was due to inhibition of CYP1A2 activity.

Although CYP2D6 SUPERSOMES were also active for catalyzing the reaction, its relatively low concentration (approximately 2% of total hepatic CYP enzymes) suggested that CYP2D6 was not a major factor for conversion of t-ROH to t-RAL in human adult liver. This conclusion also was supported by the observation that quinidine produced only a moderate inhibitory effect (approximately 35% inhibition) at a relatively high (10 µM) concentration on the HLMS-catalyzed conversion of t-ROH to t-RAL.

CYP2C8, CYP2C9, CYP2C18, and CYP2C19 SUPERSOMES each exhibited minimal activities (<100 pmol/min/nmol) for catalyzing conversion of t-ROH to t-RAL, which sharply contrasted with their outstandingly high imipramine demethylase activities (1000-30,000 pmol/min/nmol). This observation suggested that t-ROH was not a good substrate for CYP2C subfamily isoforms in terms of oxidation to t-RAL and that members of CYP2C subfamily were not likely to be quantitatively important in catalyzing the reaction in human adult liver. The moderate inhibitory effects produced by inhibitors of CYP2C in HLMS-catalyzed reactions were probably the sum of inhibition of activities of all four CYP2C isoforms as well as nonspecific inhibition of activities of other CYP enzymes.

As t-RAL can be rapidly oxidized to t-RA, the lower generation of t-RAL via oxidation of t-ROH catalyzed by CYP2D6 or members of the CYP2C subfamily might also be the result of their potential activities for catalyzing conversion of t-RAL to t-RA. As shown in Fig. 7, however, CYP2D6, CYP2C8, CYP2C9, CYP2C18, and CYP2C19 all showed relatively low activities for catalysis of oxidation of t-RAL to t-RA. For these CYP enzymes, therefore, the lower generations of t-RAL shown in Fig. 6 appeared due to their lower catalytic capacities for converting t-ROH to t-RAL.

CYP1A1 and CYP1B1 SUPERSOMES exhibited very high catalytic activities for conversion of t-ROH to t-RAL as well as low K_m values. Because both CYP1A1 and CYP1B1 are expressed at only very low constitutive levels in human adult liver, they are not likely to act as significant participants in hepatic biosynthesis of t-RA from t-ROH. However, in various extrahepatic tissues (such as in brain and fetal tissues), expressions of CYP1A1 and CYP1B1 are readily detected, particularly after exposures to inducing agents (Strobel et al., 1997; Juchau et al., 1998). Therefore, CYP1A1 and CYP1B1 may perform important roles in tissue-specific biosynthesis of t-RA from t-ROH. CYP1A1 and CYP1B1 are highly inducible by environmental chemicals (Parkinson, 1996) and such induction might remarkably increase the rates of biosynthesis of t-RA catalyzed by CYP1A1 and CYP1B1 in human fetal tissues, thereby influencing normal development. On the other hand, as substrates for CYP1A1 and CYP1B1, retinoids might also inhibit xenobiotic oxidation catalyzed by these two enzymes. Indeed, inhibition by retinol of CYP1A1-catalyzed xenobiotic oxidation reactions has been reported recently (Yamazaki and Shimada, 1999).

Conversion of t-RAL to t-RA catalyzed by CYP SUPERSOMES also was investigated to corroborate investigations with HLMS. CYP1A1 and CYP1A2 SUPERSOMES were highly active for catalyzing conversion of t-RAL to t-RA as has also been reported previously (Roberts et al., 1992; Raner et al., 1996). With catalysis by CYP1A1 and CYP1A2 SUPERSOMES, rates of oxidation of t-RAL to t-RA were roughly 2- to 3-fold greater than that of oxidation of t-ROH to t-RAL, suggesting that the latter was the rate-limiting step for biosynthesis of t-RA from t-ROH when catalyzed by these enzymes. Interestingly, for the CYP1B1, CYP3A4, CYP3A5, and CYP2D6 SUPERSOMES-catalyzed reactions, conversion of t-ROH to t-RAL appeared not to be the rate-limiting step because rates of oxidation of t-ROH to t-RAL were roughly the same as those of oxidation of t-RAL to t-RA. Similar to their catalysis of the conversion of t-ROH to t-RAL, CYP SUPERSOMES of members of CYP2C subfamily showed minimal activities in catalyzing conversion of t-RAL to t-RA.

In summary, the conversion of t-ROH to t-RAL catalyzed by P-450 cytochromes in human adult hepatic microsomal fractions indicated that CYP1A2 and CYP3A4 appear to be the primary enzymes responsible for catalyzing t-RA biosynthesis in human adult hepatic microsomes. Because of their excellent catalytic activities, low K_m values, and high inducibility by environmental chemicals and xenobiotics, CYP1A1 and CYP1B1 could play important roles in human extrahepatic and fetal tissue-specific biosynthesis of t-RA from t-ROH.