Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The pleiotropic biological responses to retinoic acids arise after their binding to retinoid receptors (Sporn and Roberts, 1994). Some examples of these responses include: 1) the inhibition of abnormal squamous differentiation, which has therapeutic potential (Mayne and Lippman, 1997); 2) modulation of cell surface adhesion molecules (Dinoto et al., 1996); 3) a key role in embryonic development, including that of the posterior hindbrain and nervous system (Sucov and Evans, 1995; Maden, 1996); 4) morphogenetic signaling (Osumiyamashita, 1996); 5) regulation of the proliferation of human neuroblastoma cells (Melino et al., 1997); 6) reconstitution of immune function (Ross and Hämmerling, 1994); and 7) regulation of vitamin A absorption (Levin and Davis, 1997).

Despite this array of important functions of retinoic acids, the pathways of biosynthesis in humans of the prototypic retinoic acid, all-trans-retinoic acid, and its 9-cis-isomer, have not been completely resolved and are still controversial. Cellular retinoic acids can be derived from several sources, as reviewed by Duester (1996): 1) vitamin A or retinol is oxidized via retinal to retinoic acid; 2) -carotene is cleaved to form retinal (Wang et al., 1991, 1992), and in turn is oxidized to retinoic acid; and 3) retinoic acid is absorbed in the small intestine (Skare and DeLuca, 1983). The oxidative conversion of retinals to the corresponding retinoic acids is generally considered to occur via two pathways (Blaner and Olson, 1994; Duester, 1996)a cytosolic pathway catalyzed by a NAD-dependent aldehyde dehydrogenase, and an NADPH-dependent, endoplasmic reticulum-bound pathway catalyzed by cytochrome P-450s (CYPs)¹.

There is relatively little available data on CYP-catalyzed oxidation of retinals to retinoic acids. Of eight purified rabbit liver and nasal CYPs, reconstituted with NADPH-P-450 reductase and cytochrome b₅ (b₅), and tested for their capacity to metabolize retinals to retinoic acids (Roberts et al., 1992), only forms CYP1A2 and 3A6 catalyzed the oxidation, and b₅ did not facilitate metabolism. Subsequently, rabbit CYP1A1 was demonstrated to be most effective in catalyzing retinoic acid formation from retinals in the order 9-cis- > 13-cis- > all-trans retinal, with K_m values of 18, 29, and 14 µM, respectively (Raner et al., 1996). The corresponding value with purified rat liver CYP1A1 and all-trans-retinal is 11.6 µM (Tomita et al., 1996). We reported that purified rat small intestinal CYP2J4 catalyzed the oxidation of all-trans- and 9-cis-retinals to the corresponding retinoic acids with K_m values of 54 and 49 µM, respectively (Zhang et al., 1998). There is no available data on human CYP-catalyzed metabolism of retinals to retinoic acids.

In contrast to the dearth of information on the CYP-catalyzed metabolism of retinals to retinoic acids in humans, CYP-catalyzed human catabolic metabolism of retinoic acids has been reported. Human breast cancer T47D cells metabolize all-trans-retinoic acid to 4-hydroxy- and 18-hydroxy all-trans-retinoic acid. The authors suggest that novel CYPs may be involved (Han and Choi, 1996). A possible candidate is CYP2C8, which in a reconstituted system does catalyze 4-hydroxylation of all-trans-retinoic acid (Leo et al., 1989). Human liver microsomes catalyze the 4-hydroxylation of all-trans-retinoic acid, and the reaction is competitively inhibited by 9-cis- and 13-cis-retinoic acids with K_i/K_M ratios of 3.5 and 6.3, respectively (Nadin and Murray, 1996). This result implies that the two cis isomers are weaker substrates than all-trans-retinoic acid for the CYP that catalyzes the 4-hydroxylase activity. A novel CYP, CYP26, is a specific 4-hydroxylase of all-trans-retinoic acid and is induced by its substrate (White et al., 1997; Marikar et al., 1998; Sonneveld et al., 1998).

The presence of several retinoid-binding proteins in cells that carry out metabolic conversions of retinoids adds to the complexity of retinoid metabolism. The most extensively characterized of these binding proteins are cellular retinol-binding protein, cellular retinol-binding protein type two [CRBP(II)], cellular retinoic acid-binding protein, and cellular retinoic acid-binding protein type two (Ong, 1994). CRBP(II) binds retinol and retinal, but not retinoic acid (MacDonald and Ong, 1987; Levin et al., 1988; Li et al., 1991), and is found only in small intestinal enterocytes (Crow and Ong, 1985).

In this article we have investigated the role of human CYPs in the oxidation of all-trans- and 9-cis-retinal to the corresponding retinoic acids using a series of cDNA-expressed CYPs. The results for 9-cis-retinal metabolism were confirmed in human hepatic microsomal systems. The physiological relevance of CYP metabolism of retinals is investigated and discussed.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Microsomal preparations (supersomes) containing cDNA-expressed human CYP1A1, 1A2, 1B1, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1, 3A4, 3A5, or 4A11 and coexpressed NADPH-CYP reductase were obtained from Gentest Co. (Woburn, MA). Some CYP3A4 supersome preparations also had b₅ coexpressed. Human liver microsomes were obtained from XenoTech (Kansas City, KS) and were characterized for CYP1A2, 2A6, 2C9, 2C19, 2D6, 2E1, 3A4/5, 3A4, and 4A9/11 activities by the company. The microsomes were received frozen and were stored at 80°C until use. Human recombinant b₅ was obtained from Panvera (Madison, WI). All-trans- and 9-cis-retinals and retinoic acids were purchased from Sigma Chemical Co. (St. Louis, MO). All other reagents were obtained as described previously (Zhang et al., 1998). CRBP(II) was provided by Dr. David Ong, Vanderbilt University (Nashville, TN). The rat recombinant protein was expressed in Escherichia coli strain JM103 and purified as described previously (Dew et al., 1993).

Metabolism of All-trans- and 9-cis-retinal. Retinal metabolism was assayed using HPLC essentially as described previously (Zhang et al., 1998). Reaction mixtures in a total volume of 500 µl contained 50 mM potassium phosphate buffer, pH 7.4, 1 mM L-ascorbic acid, 30 µg of 1,2-dilauroyl-sn-glycero-3-phosphorylcholine, 0.1 µM cDNA-expressed CYP with coexpressed NAPDH-P-450 reductase, and 20 µM all-trans- or 9-cis-retinal, or as indicated in the figure legends. In some experiments, b₅ was added at a ratio of 5 nmol/nmol CYP or was coexpressed with the CYP. For human liver microsomal metabolism, reaction mixtures contained 0.5 mg of microsomal protein and 100 µM 9-cis-retinal. All mixtures were preincubated at 37°C for 1.0 min before the reaction was initiated with 20 µl of a 25 mM NADPH stock solution. Control experiments were performed in which NADPH was omitted. In studies where CRBP(II) was added to the metabolic reactions, the binding protein was added to the substrate at room temperature for 15 min before the addition of CYP. All reactions and other procedures were carried out in the absence of overhead light. Retinoic acids were quantified using the peak area at 360 nm, and all standard curves were linear over the concentration range of the retinoic acid products.

Correlation Study. Correlations between hepatic microsomal metabolic activities toward 9-cis-retinal and those of CYP1A2, 2A6, 2C9, 2C19, 2D6, 2E1, 3A4/3A5, and 4A9/11 were determined with 12 different human liver microsomal preparations. Human hepatic microsomal CYP activities were determined by XenoTech. Correlation coefficients were determined by Pearson Moment correlations using SigmaStat Software (SPSS Inc., Chicago, IL).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Rates of biotransformation by CYP1A1, 1A2, 1B1, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1, 3A4, 3A4+b₅, 3A5, and 4A11 of all-trans- and 9-cis-retinal to the corresponding retinoic acid are shown in Fig. 1. Conversion of all-trans-retinal (20 µM) to all-trans-retinoic acid was only detected with CYP1A1, 1A2, 1B1, and 3A4+b₅, with turnover numbers of 0.53, 0.18, 0.20, and 0.41 nmol/min/nmol P-450, respectively. In the absence of b_5, CYP3A4 did not produce detectable quantities of products. With 9-cis-retinal as substrate, CYP1A1, 1A2, 2C19, and 3A4+b₅ catalyzed the formation of 9-cis-retinoic acid, with turnover numbers of 0.40, 1.58, 0.27, and 0.41 nmol/min/nmol P-450, respectively. CYP1A2 exhibited a much higher turnover number than other CYPs toward 9-cis-retinal. The coexpressed b₅ was essential for CYP3A4 activity with 9-cis- and all-trans-retinal. Representative chromatograms of the reaction products are shown in Fig. 2, for CYP1A1 and 3A4+b₅, using all-trans-retinal as substrate, and for CYP1A2, using 9-cis-retinal as substrate. The chromatograms indicate that all reactions were NADPH-dependent, and that there was a slight isomerization of all-trans-retinal substrate to 13-cis-retinal. By comparison with the retention times of 4-hydroxylated retinals and retinoic acids, it is clear that no 4-hydroxy retinals or retinoic acids were produced by those CYPs that did not produce detectable retinoic acid products. Thus the failure to detect retinoic acids was not due to their further metabolism to 4-hydroxy products. For those CYPs that produced retinoic acid products, CYP1A1, 1A2, 1B1, and 2C19 did not yield any detectable further metabolism to 4-hydroxy products, whereas CYP3A4+b₅ did catalyze further metabolism. With all-trans- and 9-cis-retinal as substrates CYP3A4+b₅ supersomes converted 21 and 24%, respectively, of the retinoic acid metabolites to 4-hydroxy products.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Rates of aldehyde oxidation of all-trans- () and 9-cis-retinals () to the corresponding retinoic acids by various cDNA-expressed human CYPs. Activities were measured at 37°C. Reaction mixtures contained 50 mM phosphate buffer, pH 7.4, 20 µM all-trans- or 9-cis-retinals, 1 mM ascorbic acid, 0.1 µM expressed CYP with coexpressed NADPH-P-450 reductase, and 30 µg of phospholipid. Extraction and HPLC analysis of metabolites were carried out as described in Experimental Procedures.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Metabolites of all-trans- and 9-cis-retinal generated by human CYP1A1 (A, B), CYP3A4+b5 (C, D), and CYP1A2 (E, F). Substrates and their metabolites were resolved by HPLC. Reaction conditions are as described in the legend to Fig. 1 except that substrate concentrations were 30 µM all-trans- (A and B, C and D) or 9-cis-retinal (E and F). Reactions were performed at 37°C for 15 min in the presence (A, C, and E) or absence (B, D, and F) of NADPH (1 mM). ATRAL, all-trans-retinal; 9-c-RAL, 9-cis-retinal; ATRH, all-trans-retinoic acid; 9-c-RH, 9-cis-retinoic acid; and 13-c-RAL, 13-cis-retinal.

The activity of CYP3A4+b₅ supersomes was approximately 4-fold higher than that of CYP3A4 supersomes when measured by testosterone 6--hydroxylation as reported by Gentest Co.; this effect of b₅ in potentiating CYP3A4 activity was also detected with all-trans- and 9-cis-retinal as substrates. The absence of commercially available CYP3A5+b₅ supersomes and the failure of CYP3A5 supersomes to catalyze the oxidation of either all-trans- or 9-cis-retinal, prompted us to examine the effects of added b₅ to CYP3A5 and 3A4 supersomes. Inclusion of b₅ in the reaction mixture increased retinoic acid synthetic activities of CYP3A4 and 3A5 to detectable levels when 50 µM all-trans-retinal was used as substrate, with turnover numbers of 0.32 and 0.13 nmol/min/nmol P-450, respectively. Although CYP3A4 exhibited higher activity than CYP3A5, it was still much less active than the CYP3A4+b₅ preparation, which had a turnover number of 0.91 nmol/min/nmol P-450 at the same substrate concentration.

The kinetics of retinal oxidation were determined for CYP1A1, 1A2, and 3A4, with concentrations of all-trans- and 9-cis-retinal varying between 30 and 200 µM. The data are presented in Table 1. Apparent K_m values for the formation of all-trans-retinoic acid catalyzed by CYP1A1, CYP1A2, and 3A4+b₅ were 55, 356, and 255 µM, respectively; and the apparent V_max values were 2.0, 8.3, and 6.3 nmol/min/nmol P-450, respectively. The apparent K_m values for the formation of 9-cis-retinoic acid catalyzed by CYP1A1, CYP1A2, and CYP3A4+b₅ were 77, 91, and 368 µM, respectively; and the apparent V_max values were 2.7, 9.7, and 7.6 nmol/min/nmol P-450, respectively. The catalytic efficiency of the CYP enzymes, expressed as V_max/K_m, suggested that CYP1A2 is the most efficient for 9-cis-retinoic acid formation.

To estimate the relative importance of the CYP enzymes in retinoic acid formation in vivo, we have examined potential correlations between the metabolite formation rates for the respective CYP isoform-selective substrates and the 9-cis-retinoic acid formation rates in human liver microsomal preparations. Due to the presence of microsomal retinol dehydrogenase, we were unable to determine all-trans-retinoic acid formation activities with human liver microsomes. As shown in Fig. 3, there is a very good correlation (r = 0.96, P < .001) between the 9-cis-retinoic acid formation and 7-ethoxyresorufin O-dealkylation activity in 12 different human liver microsomal preparations. In contrast, there was no correlation between 9-cis-retinoic acid formation and testosterone 6-hydroxylation, a marker reaction for CYP3A4, or for the various other human CYP marker activities (Table 2).

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Correlation between the rates of human liver microsomal 7-ethoxyresorufin O-deethylation and 9-cis-retinoic acid formation. Rates for the 9-cis-retinoic acid formation were determined as described in the legend to Fig. 1, except that 100 µM substrate and 0.5 mg of microsomal protein were used. CYP1A2-catalyzed 7-ethoxyresorufin O-deethylase activity data were provided by XenoTech for the 12 human liver microsomal preparations.

The role of CYP1A2 in human liver microsomal metabolism of 9-cis-retinal to 9-cis-retinoic acid was further confirmed by an inhibition study with -naphthoflavone, a specific inhibitor of CYP1A. As shown in Fig. 4, addition of -naphthoflavone to microsomal reactions lead to a dose-dependent inhibition of 9-cis-retinoic acid formation. CRBP(II), at a molar ratio of 1:1 with 9-cis-retinal, caused a 32% inhibition in the cDNA-expressed CYP1A2-mediated oxidation of this substrate.

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Inhibition of 9-cis-retinoic acid formation in a pool of human liver microsomal preparations by -naphthoflavone (ANF). Retinal metabolism was carried out as described in the legend to Fig. 3. Activities are shown as a percentage of control reactions in the absence of inhibitor.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Although it is generally accepted that oxidation of aldehydes to carboxylic acids in biological systems is catalyzed by aldehyde dehydrogenase, it has been recognized for several years that CYPs are also capable of catalyzing this reaction. The first report of such a reaction appeared in 1991 when 11-oxo-⁸-tetrahydrocannabinol was oxidized to ⁸-tetrahydrocannabinol-11-oic acid by a CYP (Watanabe et al., 1991), subsequently identified as CYP2C29 (Matsunaga et al., 1994). In 1992 we reported that an unidentified rat hepatic CYP (CYP2E1 was eliminated as a possibility) catalyzed the conversion of 2,2,2-trifluoroacetaldehyde to trifluoroacetic acid (Kaminsky et al., 1992).

Studies with rabbit (Roberts et al., 1992; Raner et al., 1996) and rat (Tomita et al., 1996) CYPs demonstrated that oxidation of retinoid aldehydes to retinoic acids could also be catalyzed by CYPs. These studies lead to the incorporation of a CYP-mediated pathway into the overall scheme of retinoic acid biosynthesis (Duester, 1996). The current studies were undertaken to determine whether human CYPs could catalyze the oxidation of retinals to retinoic acids and to gain some insight into the physiological relevance of such putative reactions. Of the 14 human CYPs investigated, five yielded detectable rates of all-trans-retinal to all-trans-retinoic acid metabolism-CYP1A1, 1A2, 1B1, and CYP3A4 and 3A5, the latter two only in the presence of b₅. The ratio of the CYP1A1/1A2 turnover numbers for all-trans-retinal of approximately 3 is comparable to the corresponding ratio for the rabbit enzymes (Raner et al., 1996). The K_m for the human CYP1A1-catalyzed oxidation of all-trans-retinal to all-trans-retinoic acid (55 µM) is slightly higher than the value of 14 µM reported for rabbit CYP1A1 (Raner et al., 1996) and the value of 11.6 µM reported for rat CYP1A1 (Tomita et al., 1996). Our observation that CYP3A4 and 3A5 catalyze the oxidation of all-trans-retinal to the corresponding retinoic acid is consistent with the reported observation that rabbit CYP3A6 catalyzes the same reaction (Roberts et al., 1992). However, in the case of the human enzymes, b₅ is essential for detectable activity, in contrast to the rabbit CYP3A6, which does not require b₅ (Roberts et al., 1992). It is well known that b₅ stimulates CYP3A4 and 3A5 activity with some substrates but not with others (Guengerich, 1999).

The human CYPs exhibit differing specificities for the 9-cis- compared with the all-trans-retinal isomer. Although both CYP1A1 and 1A2 catalyze the oxidation to 9-cis-retinoic acid, the ratio of the turnover numbers for CYP1A1/CYP1A2 is 0.25 as compared with the value of 3 with all-trans-retinal. CYP1B1 yields no detectable product with 9-cis-retinal in contrast to the case with all-trans-retinal and CYP2C19 catalyzes formation of retinoic acid with 9-cis retinal in contrast to the case with all-trans-retinal. CYP3A4+b₅ yields the same turnovers with both all-trans- and 9-cis-retinals. The K_m value of 77 µM for human CYP1A1 metabolism of 9-cis-retinal to the retinoic acid is higher than the value of 18 µM reported for rabbit CYP1A1 (Raner et al., 1996).

Metabolism of the metabolite retinoic acids to the corresponding 4-hydroxy retinoic acids was catalyzed only by CYP3A4+b₅, of all of the human CYPs investigated. Under our reaction conditions, only approximately 20% of the retinoic acids generated were further hydroxylated. For the CYP3A4+b₅-catalyzed metabolism, the reported turnovers are thus probably underestimated and should be approximately 20% higher to reflect the further metabolism of the products.

Our studies with human liver microsomal preparations indicate that with 9-cis-retinal as substrate, CYP1A2 is the major contributor to microsomal oxidation to the 9-cis-retinoic acid. This activity correlated with the CYP1A2 activity in 12 human liver microsomal preparations but not with any of the other activities tested, including those of CYP2C19 and 3A4. The virtually complete inhibition of 9-cis-retinal oxidation in human liver microsomal preparations by -naphthoflavone supports the conclusion of a major role for CYP1A2 in the oxidation of 9-cis-retinal in human liver endoplasmic reticulum. This result probably reflects the situation in vivo in the liver. In extrahepatic organs, however, where CYP1A2 is not expressed, other CYPs probably assume the major role in microsomal metabolism. In the human small intestine where CYP3A4 is the major form of CYP expressed (Zhang et al., 1999), it could be presumed to be the major microsomal catalyst of 9-cis-retinal oxidation.

Although these studies clearly demonstrate that human CYPs have the capability of metabolizing retinals to retinoic acids, both hepatically and extrahepatically, the question of whether such metabolism is physiologically relevant is unresolved. The apparent K_m values for human CYP1A2 oxidation of all-trans- and 9-cis-retinoic acid (356 and 91 µM, respectively) are approximately 25,000- and 14,000-fold higher, respectively, than the values for human liver cytosolic aldehyde dehydrogenase (Klyosov, 1996; Bhat and Samaha, 1999). In situations where the cytosolic enzyme is expressed, it is thus likely that it, rather than the microsomal CYPs, would preferentially catalyze the retinal oxidation. The failure of CRBP(II) binding of 9-cis-retinal to facilitate CYP1A2-mediated metabolism of the retinal to the retinoic acid further mitigates against an in vivo role for CYPs in retinoic acid formation. The observed inhibition of 9-cis-retinal oxidation by CRBP(II) contrasts with the lack of inhibition of retinal dehydrogenase-mediated retinal oxidation by a cellular retinol binding protein (Wang et al., 1996). It was hypothesized that for the retinals to undergo in vivo metabolism catalyzed by relatively weaker binding CYPs, the metabolism would have to be facilitated by the substrate being bound to a form of binding protein. The results of the limited study conducted here would preclude that possibility. Any preferential disposition of protein-bound retinals into the endoplasmic reticulum rather than the cytoplasm in vivo, would, however, favor a role for CYPs in retinal oxidation. More investigation is clearly required to resolve which enzymes catalyze the important oxidation of retinals to retinoic acids in vivo.