C-25 hydroxylation of 1α,24(R)-dihydroxyvitamin D3 is catalyzed by 25-hydroxyvitamin D3-24-hydroxylase (CYP24A1): metabolism studies with human keratinocytes and rat recombinant CYP24A1

ABB Archives of Biochemistry and Biophysics 431 (2004) 261–270 www.elsevier.com/locate/yabbi

C-25 hydroxylation of 1a,24(R)-dihydroxyvitamin D3 is catalyzed by 25-hydroxyvitamin D3-24-hydroxylase (CYP24A1): metabolism studies with human keratinocytes and rat recombinant CYP24A1 Norbert Asteckera, Ekaterina A. Bobrovnikovab, John L. Omdahlb, Lynn Gennaroc, Paul Vourosc, Inge Schusterd, Milan R. Uskokovice, Seiichi Ishizukaf, Guochun Wangg, G. Satyanarayana Reddya,* b

a Department of Chemistry, Brown University, Box H, Providence, RI 02912, USA Department of Biochemistry and Molecular Biology, University of New Mexico School of Medicine, Albuquerque, NM 87131-5221, USA c Barnett Institute and Department of Chemistry and Chemical Biology, Northeastern University, Boston, MA 02115, USA d Institute of Pharmaceutical Chemistry, University of Vienna, Vienna, Austria e Bioxell Inc., 340 Kingsland Street, Bldg. 76/13, Nutley, NJ 07110, USA f Teijin Institute for Bio-Medical Research, 191-8512 Tokyo, Japan g Department of Engineering, Brown University, Providence, RI 02912, USA

Received 19 July 2004, and in revised form 18 August 2004 Available online 25 September 2004

Abstract Recently, 25-hydroxyvitamin D3-24-hydroxylase (CYP24A1) has been shown to catalyze not only hydroxylation at C-24 but also hydroxylations at C-23 and C-26 of the secosteroid hormone 1a, 25-dihydroxyvitamin D3 (1a,25(OH)2D3). It remains to be determined whether CYP24A1 has the ability to hydroxylate vitamin D3 compounds at C-25. 1a,24(R)-dihydroxyvitamin D3 (1a,24(R)(OH)2D3) is a non-25-hydroxylated synthetic vitamin D3 analog that is presently being used as an antipsoriatic drug. In the present study, we investigated the metabolism of 1a,24(R)(OH)2D3 in human keratinocytes in order to examine the ability of CYP24A1 to hydroxylate 1a,24(R)(OH)2D3 at C-25. The results indicated that keratinocytes metabolize 1a,24(R)(OH)2D3 into several previously known both 25-hydroxylated and non-25-hydroxylated metabolites along with two new metabolites, namely 1a,23,24(OH)3D3 and 1a,24(OH)2-23-oxo-D3. Production of the metabolites including the 25-hydroxylated ones was detectable only when CYP24A1 activity was induced in keratinocytes 1a,25(OH)2D3. This finding provided indirect evidence to indicate that CYP24A1 catalyzes C-25 hydroxylation of 1a,24(R)(OH)2D3. The final proof for this finding was obtained through our metabolism studies using highly purified recombinant rat CYP24A1 in a reconstituted system. Incubation of this system with 1a,24(R)(OH)2D3 resulted in the production of both 25-hydroxylated and non-25-hydroxylated metabolites. Thus, in our present study, we identified CYP24A1 as the main enzyme responsible for the metabolism of 1a,24(R)(OH)2D3 in human keratinocytes, and provided unequivocal evidence to indicate that the multicatalytic enzyme CYP24A1 has the ability to hydroxylate 1a,24(R)(OH)2D3 at C-25.
2004 Elsevier Inc. All rights reserved. Keywords: 1a,24(R)(OH)2D3; 1a,25(OH)2D3; Keratinocytes; Tacalcitol; Metabolism; CYP24A1

*

Corresponding author. Fax: +401-751-7076. E-mail address: [email protected] (G.S. Reddy).

0003-9861/$ - see front matter
2004 Elsevier Inc. All rights reserved. doi:10.1016/j.abb.2004.08.023

262

N. Astecker et al. / Archives of Biochemistry and Biophysics 431 (2004) 261–270

It is well established that 1a,25-dihydroxyvitamin D3 (1a,25(OH)2D3, calcitriol),1 the hormonally active form of vitamin D3, exerts not only classical functions in calcium homeostasis but also potent effects on cell proliferation and differentiation [1,2]. So far, toxic effects such as hypercalcemia prevented the clinical use of 1a,25(OH)2D3 as an antiproliferative/prodifferentiating agent. As a consequence, numerous vitamin D analogs have been synthesized with the goal of generating specific vitamin D compounds that inhibit proliferation of cells and induce differentiation as effectively as 1a,25(OH)2D3 without causing significant hypercalcemia [3]. In this context, 1a,24(R)-dihydroxyvitamin D3 (1a,24(R)(OH)2D3, tacalcitol) a synthetic vitamin D3 analog, has been developed as a drug for topical use in the treatment of psoriasis [4,5]. The mechanisms by which vitamin D analogs exert specific biological activities that are different from that of 1a,25(OH)2D3 are not completely understood. Besides the differences in their binding properties to VDR (vitamin D receptor) and DBP (vitamin D binding protein), the target cell metabolism differences have been shown to influence significantly the biological activity profile of a given vitamin D analog [6–8]. At present, target cell metabolism of 1a,24(R)(OH)2D3 has not been fully elucidated. In a previous study, we reported that 1a,24(R)(OH)2D3 is metabolized in rat kidney via two pathways [9]. The first pathway is initiated by C-25 hydroxylation and proceeds further via C-24 oxidation pathway similar to that described for the natural hormone 1a,25(OH)2D3 [10]. In the second pathway, 1a,24(R) (OH)2D3 undergoes metabolism via the C-24 oxidation pathway without prior hydroxylation at C-25. In the same study, we also reported that the rat kidney failed to convert 1a-hydroxyvitamin D3 into 1a,25(OH)2D3. This finding indicated that the rat kidney lacked the classical C-25 hydroxylase activity [9]. Thus, the identity of the enzyme that catalyzes C-25 hydroxylation of 1a,24(R)(OH)2D3 in rat kidney was not determined. However, we speculated that the multicatalytic enzyme 25-hydroxyvitamin D324-hydroxylase (CYP24A1) may be playing a role in 1 Abbreviations used: 25OHD3, 25-hydroxyvitamin D3; 24(R), 25(OH)2D3, 24(R),25-dihydroxyvitamin D3; 1a,25(OH)2D3, 1a,25dihydroxyvitamin D3; 1a,25(OH)2-3-epi-D3, 1a,25-dihydroxy-3-epivitamin D3; 1a,24(R),25(OH)3D3, 1a,24(R),25-trihydroxyvitamin D3, 1a,25(OH) 2 -24-oxo-D 3 , 1a,25-dihydroxy-24-oxovitamin D 3 ; 1a,23(S),25(OH)3-24-oxo-D3, 1a,23(S),25-trihydroxy-24-oxovitamin D3, 1a,23(OH)2-24,25,26,27-tetranor D3 or C-23 alcohol, 1a,23-dihydroxy-24,25,26,27-tetranorvitamin D3; 1aOH,23COOH-24,25,26,27tetranor D3 or calcitroic acid, 1a-hydroxy-23-carboxy-24,25,26, 27-tetranorvitamin D3; 1a,23(S),25(OH)3D3, 1a,23(S),25-trihydroxyvitamin D3; 1a,25(OH)2D3-26,23(S)-lactone, 1a,25-dihydroxyvitamin D3-26,23(S)-lactone; 1a,23,24(OH)3D3, 1a,23,24-trihydroxyvitamin D3; 1a,24(OH)2-23-oxo-D3, 1a,24-dihydroxy-23-oxo-vitamin D3; 1a,24(R)(OH)2D3, 1a,24(R)-dihydroxyvitamin D3; 1aOH-24-oxo-D3, 1a-hydroxy-24-oxovitamin D3; 1a,23(OH)2-24-oxo-D3, 1a,23-dihydroxy-24-oxovitamin D3.

hydroxylating 1a,24(R)(OH)2D3 at C-25. To address this issue, we investigated the metabolism of 1a,24(R) (OH)2D3 in human keratinocytes which are the main target cells for 1a,24(R)(OH)2D3 [11,12]. Furthermore, human keratinocytes are well studied in terms of the expression and possessing the activities of both CYP24A1 [13,14] and sterol C-27 hydroxylase (CYP27A1) [15–17]. Finally, in order to provide unequivocal proof that the multicalalytic enzyme CYP24A1 has the ability to hydroxylate 1a,24(R)(OH)2D3 at C-25, we investigated the metabolism of 1a,24(R)(OH)2D3 using a highly purified recombinant rat CYP24A1. Materials and methods Vitamin D compounds and chemicals 1a,25-Dihydroxyvitamin D3 [1a,25(OH)2D3], 24(R), 25-dihydroxyvitamin D3 [24(R)25(OH)2D3], and 25hydroxyvitamin D3 [25OHD3] were synthesized at Hoffmann-LaRoche. 1a,24(R)-Dihydroxvitamin D3 (1a,24(R)(OH)2D3), 1a,25(OH)2[1b-3H]D3 (15 Ci/mmol) and 1a,24(R)(OH)2[1b-3H]D3 (15 Ci/mmol) were synthesized at the Teijin Institute for Biomedical Research. All the synthetic vitamin D compounds were found to be pure by HPLC analysis. All the known natural metabolites of 1a,25(OH)2D3 and 1a,24(R)(OH)2D3, which include 1a,24(R),25-trihydroxyvitamin D3 [1a,24(R), 25(OH)3D3], 1a,25-dihydroxy-24-oxovitamin D3 [1a, 25(OH)2-24-oxo-D3], 1a,23(S),25-trihydroxy-24-oxovitamin D3 [1a,23(S),25(OH)3-24-oxo-D3], 1a,23-dihydroxy-24,25,26,27-tetranorvitamin D3 [1a,23(OH)2-24, 25,26,27-tetranor D3 or C-23 alcohol], 1a-hydroxy-24oxovitamin D3 (1aOH-24-oxo-D3) and 1a,23-dihydroxy-24-oxovitaminD3 (1a, 23(OH)2-24-oxo-D3) were synthesized biologically in the rat kidney perfusion system as described previously [9,18]. 1a,23(S),25(OH)3D3 and 1a,25(OH)2-3-epi-D3 were synthesized by one of us (M.R.U., Hoffmann-La Roche, Nutley, NJ). Human keratinocytes Normal human neonatal foreskin keratinocytes were purchased from Clonetics (San Diego, CA) and cultivated in serum free keratinocyte growth medium (KGM) (Clonetics, San Diego, CA) containing 0.06 mM calcium. For all experiments, 90% confluent cell cultures were used in the second passage. High-performance liquid chromatography High-performance liquid chromatography (HPLC) analysis was performed with a Waters System Controller (Model 600E) equipped with a photodiode array detector (Model PDA 990 and Model PDA 996) to monitor ultraviolet (UV) absorbing material at 265 nm. In the

N. Astecker et al. / Archives of Biochemistry and Biophysics 431 (2004) 261–270

studies in which radiolabeled substrates were used, 1-min fractions (2 ml each) were collected and radioactivity was measured for each fraction in a scintillation counter (Beta Trac, TM Analytic, Elk Grove Village, IL) after the addition of 5 ml Enviro Safe (ANOROC Scientific, Hackensack, NJ). The samples were analyzed using different HPLC systems as follows: HPLC system-I: Zorbax SIL column (250 mm · 4.6 mm); solvent; hexane/isopropanol (93:7, vol/vol); flow rate, 2 ml/min. HPLC system-II: Zorbax SIL column (250 mm · 4.6 mm); solvent; methylene chloride/isopropanol (96:4, vol/vol); flow rate, 2 ml/min. HPLC system-III: Zorbax SIL column (250 mm · 4.6 mm); solvent, hexane/isopropanol (94:6, vol/vol); flow rate, 2 ml/min. HPLC-system-IV: Zorbax ODS column (250 mm · 4.6 mm); solvent, water/methanol (20:80, vol/vol); flow rate, 1 ml/min. GC/MS analysis GC/MS analysis was performed using a Hewlett– Packard GC-MSD system, composed of a series 6890 GC, a 5973 mass selective detector and 7683 autosampler. Standards and metabolites of 1a,24(R)(OH)2D3 were dried under nitrogen, reconstituted in 15 ll 1:1 acetonitrile/Power-Sil-Prep (Alltech Associates, Deerfield, IL). The samples were incubated at 70 C for 15 min to ensure the complete derivatization. Derivatives were then transferred to autosampler vials and analyzed on a DB-5 ms capillary (30 m · 0.25 mm · 0.25 mm) column (J&W Scientific) using UHP helium as a carrier gas. Synthesis of standard 1a,23,24(OH)3D3 One microgram of 1a,23(OH)2-24-oxo-D3 was dissolved in 100 ll ethanol containing approximately 1 mg of NaBH4. After 30 min at 25 C, the reaction product was dried under nitrogen and dissolved in hexane/isopropanol (93:7, vol/vol). The sample was chromatographed on a Zorbax-SIL column using HPLC-system I. The NaBH4 reduction product was collected and its structure was confirmed by GC/MS as 1a,23,24(OH)3D3. Metabolism of 1a,25(OH)2D3 and of 1a,24(R)(OH)2D3 in human keratinocytes To identify and quantitate various metabolites of both 1a,25(OH)2D3 and 1a,24(R)(OH)2D3, human keratinocytes (cell culture flask: 150 cm2) were incubated with 20 lg 1a,25(OH)2D3 or 1a,24(R)(OH)2D3 (1 lM) for 24 h. Lipids from both cells and media were extracted according to the procedure of Bligh and Dyer [19]. To assess recovery from the samples 1 lg unlabeled 25OHD3 was added as internal standard. Lipid extracts were subjected to HPLC analysis. Isolation and purification of products were achieved using different HPLC systems (HPLC-system I-IV). Metabolites produced from

263

1a,24(R)(OH)2D3 were identified by GC/MS analysis by comparing their mass spectra with those of authentic standards. Metabolism studies using 1a,25(OH)2[1b-3H]D3 (15 Ci/mmol) and 1a,24(R)(OH)2[1b-3H]D3 (15 Ci/ mmol) as substrates are performed as follows: human keratinocytes (six-well plates: well
10 cm2) were incubated with either 0.1 lCi of 1a,25(OH)2[1b-3H]D3 (7 nM) or 1a,24(R)(OH)2[1b-3H]D3 (7 nM) for 1, 3, or 4 h. Incubations were terminated with methanol. To assess recovery from the samples 1 lg of unlabeled 1a,25(OH)2D3 was added as internal standard. Lipids from both cells and media were extracted according to a procedure of Bikle et al. [14] and spiked with authentic standards and analyzed on HPLC-system I. In the experiments in which 1a,24(R)(OH)2[1b-3H]D3 was used, fractions eluting between 0 and 20 min were collected and analyzed on HPLC-system III. For the metabolism studies performed under basal and induced conditions, keratinocytes were treated prior to the incubation with either vehicle (0.5% ethanol) or unlabeled 1a, 25(OH)2D3 (10 nM) for 4 h. The medium was changed and incubated with 1a,25(OH)2[1b-3H]D3 or 1a,24(R) (OH)2[1b-3H]D3 for 1 h. Metabolism of 1a,25(OH)2D3 and 1a,24(R)(OH)2D3 by recombinant CYP24A1 Metabolism studies by CYP24A1 were performed in a 50 mM phosphate buffer solution (pH 7.4) containing 0.2 lM recombinant rat CYP24A1, 4 lM recombinant bovine adrenodoxin (ADX), 0.8 lM recombinant bovine adrenodoxin reductase (ADR) and 1 mM NADPH. Rat CYP24A1, and bovine ADX and ADR were produced and purified as recently described [20]. The substrates 1a,25(OH)2D3 and 1a,24(R)(OH)2D3 were added at a final concentration of 10 lM. Reactions were initiated by addition of NADPH and were carried out at 37 C for various time intervals. Control incubations were conducted in the absence of ADR. Reactions were terminated by freezing samples in liquid nitrogen. Vitamin D metabolites were extracted from the incubation solution with a mixture of methanol/methylene chloride (1:2). To assess recovery from the samples, 1 lg unlabeled 25OHD3 was added prior to the extraction. Lipid extracts were subjected to HPLC analysis using HPLC-system I. Metabolites produced from 1a,24(R)(OH)2D3 were isolated and identified by GC/MS analysis as described above.

Results Metabolites of 1a,24(R)(OH)2D3 and 1a,25(OH)2D3 produced by human keratinocytes Fig. 1 shows HPLC profiles of organic extracts from cultures of primary human keratinocytes incubated

264

N. Astecker et al. / Archives of Biochemistry and Biophysics 431 (2004) 261–270

Fig. 1. HPLC profiles of organic extracts from incubations of human keratinocytes with either 1a,25(OH)2D3 (A1, A2) or 1a,24(R)(OH)2D3 (B1, B2) at 1 lM for 24 h. (A, B) HPLC was performed using a Zorbax-SIL column (4.6 · 250 mm) eluted with hexane/isopropanol (93:7, vol/vol) (HPLC system I). Vitamin D metabolites were detected by monitoring UV absorbance at 265 nm. Metabolites produced from 1a,24(R)(OH)2D3 are indicated by the numbers 1–8 and the letter M. Peak * did not exhibit a characteristic vitamin D spectrum. The fraction eluting between 30 and 38 min (indicated by dotted line) was collected and analyzed on a second HPLC system. (A1, B1) HPLC profile of the collected fraction analyzed on a HPLC system using a Zorbax Sil column (4.6 · 250 mm) eluted with methylene chloride/isopropanol (96:4, vol/vol) (HPLC system II).

with either 1a,25(OH)2D3 (Panels A, A1) or 1a,24(R) (OH)2D3 (Panel B, B1) at a concentration of 1 lM for 24 h. As shown in our previous publication [21], 1a,25(OH)2D3 is metabolized by human keratinocytes to 1a,24(R),25(OH)2D3, 1a,25(OH)2-24-oxo-D3, 1a,23(S),25(OH)3-24-oxo-D3, and C-23 alcohol derived via the C-24 oxidation pathway, along with 1a,25(OH)2-3-epi-D3 and 1a,23(S),25(OH)3D3. Under the same conditions, the synthetic analog 1a,24(R) (OH)2D3 was converted by human keratinocytes into nine metabolites (1–8, M) possessing an intact 5,6-cistriene vitamin D chromophore (UV-spectra: kmax at

265 nm, kmin at 228 nm). Metabolites were isolated using different HPLC systems (described under Materials and methods) and identified by GC/MS analysis. Table 1 gives a summary of the molecular ions and characteristic fragments detected in the mass spectra of the trimethylsilyl derivatives of 1a,24(R)(OH)2D3 and its metabolites. Metabolites 4, 5, 7, and 8 obtained from incubations with 1a,24(R)(OH)2D3 were identical to metabolites obtained from incubations with 1a,25(OH)2D3. They were identified as 1a,25(OH)2-24-oxo-D3 (metabolite 4), 1a,23(OH)2-24,25,26,27-tetranor-D3 (metabolite 5), 1a,23(S),25(OH)3-24-oxo-D3 (metabolite 7) and

Table 1 Molecular ions and characteristic ion fragments in the mass spectra of the trimethylsilyl derivatives of 1a,24(R)(OH)2D3 and its metabolites Substrate and metabolites

1a,24(R)(OH)2D3 and identified compounds

M+

[M–90]+ (M+–TMSOH)

[M–131]+ (A-ring cleavage)

A-ring cleavage

C23/C24cleavage

C24/C25cleavage

Substrate 1 2 3 4 5 6 7 8 M

1a,24(R)(OH)2D3 1aOH-24-oxo-D3 1a,24(OH)2-23-oxo-D3a 1a,23(S)(OH)2-24-oxo-D3 1a,25(OH)2-24-oxo-D3 C-23 alcohol 1a,23,24(OH)3D3a 1a,23(S),25(OH)3-24-oxo-D3 1a,24(R),25(OH)3D3 Isomer of 1a,24(R)(OH)2D3a

632 558 646 646 646 576 720 734 720 632

542 468 556 556 556 486 630 644 630 542

501 427 515 515 515 445 589 603 589 501

217 217 217 217 217 217 217 217 217 217

145 — 145 — — — 145 — — 145

— — — — 131 — — 131 131 —

a

New metabolites (identified in the present study).

N. Astecker et al. / Archives of Biochemistry and Biophysics 431 (2004) 261–270

265

Fig. 2. Mass spectra of TMS-derivatives of 1a,23,24(OH)3D3 (metabolite 6) (A) and 1a,24(OH)2-23-oxo-D3 (metabolite 2) (B).

1a,24(R),25(OH)2D3 (metabolite 8). Metabolites 1, 2, 3, 6, and M were unique to the incubations with 1a,24(R)(OH)2D3. Two of these metabolites, metabolites 1 and 3, co-migrated on HPLC with 1aOH-24oxo-D3 and 1a, 23(OH)2-24-oxo-D3. Their identity as 1aOH-24-oxo-D3 and 1a,23(OH)2-24-oxo-D3 was confirmed by comparing their mass spectra with those of authentic standards. Due to the lack of suitable standards, metabolites 2, 6, and M required separate structural identification shown in the next section. Structural identification of 1a,23,24(OH)3D3 (metabolite 6), 1a,24(OH)2-23-oxo-D3 (metabolite 2), and of a stereoisomer of 1a, 24(R)(OH)2D3 (metabolite M) Fig. 2 shows the mass spectra of the trimethylsilyl derivatives of metabolites 6 and 2. The mass spectrum of the derivatized metabolite 6 (Panel A) yielded a molecular ion at m/z 720 (M+). Characteristic fragments were detected at m/z 630, 589, 395, 217 and 145. The fragment at m/z 630 (M+–90) indicated the loss (in the form of trimethylsilanol) of one trimethylsilylated hydroxyl group. Fragments at m/z 589 (M+–131) and m/z 217 derived from a cleavage in the A-ring and denoted the presence of hydroxyl groups at C-1 and C-3. The molecular ion and the fragments at m/z 630, 589, and 217 indicated a structure similar to that of 1a,24(R),25(OH)3D3 (see Table 1). However, the fragment at m/z 131 characteristic of C-25 hydroxylated vitamin D3 metabolites was not detected. This excluded the possibility of a C-25 hydroxyl group in metabolite 6 (see Table 1). The fragments at m/z 395 (M+–145–90– 90), also detected in the mass spectrum of 1a,23(S) (OH)2-24-oxo-D3 (not shown), and the fragment at m/z 145 suggested instead a preferred cleavage site

across C-23/C-24. In order to identify metabolite 6 definitely, we chemically synthesized standard 1a,23,24(OH)3D3. Metabolite 6 comigrated on HPLC with the standard 1a,23,24(OH)3D3 and their mass spectra were identical. Thus, metabolite 6 was unequivocally identified as 1a,23,24(OH)3D3. The mass spectrum of the derivatized metabolite 2 (Panel B) exhibited a molecular ion at m/z 646 and fragments at m/z 556 (M+–90), m/z 515 (M+–131), m/z 217, and at m/z 145. This fragmentation pattern indicated a vitamin D compound with a hydroxyl and a keto group in the side chain (see Table 1). Like in the spectrum of metabolite 5, the major fragment at m/z 145 indicated a hydroxyl group at C-24. NaBH4 reduction of metabolite 2 yielded a compound exhibiting the mass spectrum of 1a,23,24(OH)3D3. This is further indication that the keto group of metabolite 2 is located at C-23. Thus, metabolite 2 was unequivocally identified as 1a,24(OH)2-23-oxo-D3. The mass spectrum of the trimethylsilyl derivative of metabolite M (not shown) exhibited the same molecular ion and the same major ion fragments as observed in the mass spectrum of the parent compound 1a,24(R) (OH)2D3. The different retention behavior of metabolite M on HPLC when compared with 1a,24(R)(OH)2D3 led us to identify metabolite M as a stereoisomer of 1a,24(R)(OH)2D3. Final structural identification of metabolite M was not carried out in this study. Relative amounts of unmetabolized 1a,24(R)(OH)2D3 and 1a,25(OH)2D3 and their metabolites produced by human keratinocytes The relative amounts of unmetabolized 1a,24(R) (OH)2D3 and 1a,25(OH)2D3 and of their metabolites

266

N. Astecker et al. / Archives of Biochemistry and Biophysics 431 (2004) 261–270

Table 2 Side chain oxidized metabolites of 1a,25(OH)2D3 1a,24(R)(OH)2D3 produced by human keratinocytes

Unmetabolized substrate 1aOH-24-oxo-D3 1a,24(R),25(OH)3D3 1a,23(OH)2-24-oxo-D3 1a,24(OH)2-23-oxo-D3 1a,25(OH)2-24-oxo-D3 1a,23,25(OH)3-24-oxo-D3 C-23 alcohol 1a,23(S),25(OH)3D3 1a,23,24(OH)3D3

and

1a,25(OH)2D3 (nmol)

1a,24(R)(OH)2D3 (nmol)

5.81 ± 0.01

8.27 ± 0.29 4.15 ± 0.18 2.34 ± 0.17 1.34 ± 0.21 1.04 ± 0.11 <0.5 <0.5 <0.5

6.18 ± 0.17

1.14 ± 0.12 1.18 ± 0.14 <0.5 1.05 ± 0.04

0.83 ± 0.05

were determined from the HPLC profiles shown in Fig. 1. Each metabolite was quantified by comparing its peak area with the corresponding peak area in a standard curve. Results are shown in Table 2 and represent the mean values of triplicate incubations. From this data, it is concluded that 1a,24(R)(OH)2D3 exhibits higher metabolic stability than 1a,25(OH)2D3 in human keratinocytes. Same conclusion was also obtained in our previous study comparing the metabolism of 1a,25(OH)2D3 with that of 1a,24(R)(OH)2D3 in the isolated perfused rat kidney [9]. Moreover, it was noted that 1aOH-24-

oxo-D3 and 1a,24(R),25(OH)3D3 were the two major products obtained from the incubations with 1a,24(R) (OH)2D3. The amount of 1aOH-24-oxo-D3 exceeded that of 1a,24(R),25(OH)3D3. This finding indicates that in the metabolism of 1a,24(R)(OH)2D3, the preferred oxidation site is at C-24, followed closely by C-25. A time course metabolism study using 1a, 25(OH)2[1b-3H]D3 and 1a,24(R)(OH)2[1b-3H]D3 as substrates After the identification of the metabolites obtained from incubations with 1a,24(R)(OH)2D3 at micromolar concentrations, a time course study comparing the metabolism of 1a,25(OH)2D3 with that of 1a,24(R) (OH)2D3 was performed at nM concentrations. For these experiments, the 3H-labeled substrates 1a,25(OH)2 [1b-3H]D3 and 1a,24(R)(OH)2[1b-3H]D3 were used. Fig. 3 shows HPLC profiles illustrating the metabolism of 1a,25(OH)2[1b-3H]D3 and 1a,24(R) (OH)2[1b-3H]D3 in human keratinocytes at three different incubation periods (1, 3, and 4 h). As shown in our recent publication [22], 1a,25(OH)2D3 was converted by human keratinocytes at 1 h incubation period primarily into 1a,25(OH)2-3epi-D3 (Fig. 3A, 1 h). At 3 and 4 h incubation periods, 1a,25(OH)2D3 was metabolized into its various metabo-

Fig. 3. HPLC profiles of various 3H-labeled vitamin D metabolites produced by human keratinocytes incubated with either 1a,25(OH)2[1b-3H]D3 (A) or 1a,24(R)(OH)2[1b-3H]D3 (B) over a time course of 1–4 h. HPLC was performed using a Zorbax-SIL column (4.6 · 250 mm) eluted with hexane/isopropanol (93:7, vol/vol) (HPLC system I). Fractions (1 min) were collected and [3H]-radioactivity in each fraction was measured. Bars indicate the total radioactive counts in each fraction. Counts of substrates are not fully shown. Arrows above the bars indicate elution positions of comigrated standards. In the incubations with 1a,24(R)(OH)2D3 the fraction eluting between 0 and 20 min was collected and analyzed on a different HPLC system (HPLC system III: hexane/isopropanol [96:4, vol/vol]). The respective HPLC profiles are shown in the insets.

N. Astecker et al. / Archives of Biochemistry and Biophysics 431 (2004) 261–270

267

lites along the C-24 oxidation pathway, as follows: 1a,25(OH)2D3 ) 1a,24(R),25(OH)3D3 ) 1a,25(OH)224-oxo-D3 ) 1a,23(S),25(OH)3-24-oxo-D3 ) C-23 alcohol (Fig. 3A, 3, 4 h). Similar to 1a,25(OH)2D3, the synthetic analog 1a,24(R)(OH)2D3 was converted at 1 h incubation period into a less polar metabolite (Fig. 3B, 1 h). This metabolite co-migrated with metabolite M, that has been identified in a previous section as a stereoisomer (possible 3-epimer) of 1a,24(R)(OH)2D3. At 3 and 4 h incubation periods, 1a,24(R)(OH)2D3 was converted into metabolites that co-migrated with the authentic standards of both non-25-hydroxylated (1aOH-24oxo-D3, 1a,23(S)(OH)2-24-oxo-D3, and C-23 alcohol) and the 25-hydroxylated metabolites (1a,25(OH)2-24oxo-D3, 1a,23(S),25(OH)3-24-oxo-D3, and 1a,24(R), 25(OH)3D3). From this time course experiment, it becomes obvious that like in the case of the natural hormone 1a,25(OH)2D3, no significant side chain metabolism of 1a, 24(R)(OH)2D3 occurred at 1 h incubation period, but only at extended incubation periods (>3 h). Metabolism of 1a,25(OH)2D3 and 1a,24(R)(OH)2D3 in human keratinocytes with or without prior exposure to 1a,25(OH)2D3 Metabolism of 1a,25(OH)2D3 via the C-24 oxidation pathway, was not observed in human keratinocytes under basal conditions. As shown in the previous section, when keratinocytes were incubated with 1a,25(OH)2D3 for 1 h only, metabolites derived via the C-24 oxidation pathway were below the detection limit (Fig. 3A; 1 h). However, when keratinocytes were incubated with 1a,25(OH)2D3 for a longer period (>3 h), 1a,25(OH)2 D3 induced CYP24A1 activity and as a result it can be observed that 1a,25(OH)2D3 is converted into 1a,24(R),25(OH)3D3 and various other metabolites via the C-24 oxidation pathway (Fig. 3A; 3 h). This observation was reexamined in the following experiment. Human keratinocytes were treated for 4 h with either vehicle (basal conditions) or 1a,25(OH)2D3 (induced conditions) at 10 nM concentration. The medium was changed and cells were incubated for 1 h with 3H-labeled 1a,25(OH)2[1b-3H]D3. The amount of 3H-labeled 1a,24(R),25(OH)3D3 produced from 1a,25(OH)2 [1b-3H]D3 was assessed. As illustrated in Fig. 4A, little or no 1a,24(R),25(OH)3[1b-3H]D3 was produced under vehicle treated conditions (basal conditions), whereas under 1a,25(OH)2D3 treated conditions (induced conditions) 1a,24(R),25(OH)3[1b-3H]D3 was produced to a substantial amount at 1 h of incubation (P < 0.001). This confirmed that pretreatment of keratinocytes with 1a,25(OH)2D3 induces CYP24A1 activity and as a result greatly increases its own metabolism into 1a,24(R), 25(OH)3D3 via the C-24 oxidation pathway.

Fig. 4. Major 3H-vitamin D metabolites produced by human keratinocytes incubated with either 1a,25(OH)2[1b-3H]D3 (A) or 1a,24(R)(OH)2[1b-3H]D3 (B) under basal and induced conditions. Keratinocytes were treated with either vehicle (basal condition) or 10 nM 1a,25(OH)2D3 (induced condition) for 4 h, then incubated with the 3H-labeled substrates for 1 h. Lipid extracts were analyzed on HPLC as described in Fig. 3. Bars indicate the radioactive counts that co-eluted with the respective authentic standards. Results represent the mean (±SD) of three different incubations. *P < 0.001.

To address the question whether metabolism of the synthetic analog 1a,24(R)(OH)2D3 via C-24 oxidation and especially via C-25 hydroxylation requires prior induction of CYP24A1, a similar experiment was conducted using 3H-labeled 1a,24(R)(OH)2[1b-3H]D3 as substrate. The results of this experiment are illustrated in Fig. 4B. As shown in the previous section, the side chain metabolism of 1a,24(R)(OH)2D3 is primarily initiated by oxidation at C-24 and hydroxylation at C-25. This leads to the production of 1aOH-24-oxoD3 and 1a,24(R),25(OH)3D3. Fig. 4B shows the amounts of these two metabolites produced from 1a,24(R)(OH)2[1b-3H]D3 in human keratinocytes. Under vehicle treated conditions, no significant amounts of either 1aOH-24-oxo-[1b-3H]D3 or 1a,24(R),25(OH)3 [1b-3H]D3 were produced. Under 1a,25(OH)2D3 treated conditions, a great increase in the production of both metabolites was noted (P < 0.001). From these results, it can be concluded that oxidation at C-24 as well as hydroxylation at C-25 of 1a,24(R)(OH)2D3 are only operative in human keratinocytes in which CYP24A1 activity was induced by prior exposure to 1a,25(OH)2D3.

268

N. Astecker et al. / Archives of Biochemistry and Biophysics 431 (2004) 261–270

Metabolism of 1a,25(OH)2D3 and 1a,24(R)(OH)2D3 by recombinant CYP24A1 Recent studies have shown that CYP24A1 is a multicatalytic enzyme catalyzing the entire six-step oxidation from 1a,25(OH)2D3 to calcitroic acid [25,26]. The multicatalytic properties of CYP24A1 and the results obtained in previous sections suggested that CYP24A1 is the enzyme responsible for the metabolism of 1a,24(R) (OH)2D3 into both 25-hydroxylated and non-25-hydroxylated metabolites. In order to prove this assumption, comparative metabolism studies with 1a,25(OH)2D3 and 1a,24(R)(OH)2D3 were conducted using highly purified recombinant rat CYP24A1. Selected HPLC profiles obtained from these studies are shown in Fig. 5. Panel A shows the results obtained from incubations performed with the natural hormone 1a,25(OH)2D3. As previously reported, 1a,25(OH)2D3 is metabolized in this system via the C-24 oxidation pathway leading to the production of 1a,24(R),25(OH)3D3, 1a,25(OH)2-

24-oxo-D3, 1a,23(S),25(OH)3-24-oxo-D3 and 1a, 23(OH)2-24,25,26,27-tetranor-D3 (C-23 alcohol) [21]. Panel B shows the results obtained from incubations with the synthetic analog 1a,24(R)(OH)2D3. It can be seen that 1a,24(R)(OH)2D3 was converted in this system into several products. Metabolites produced from 1a,24(R) (OH)2D3 were isolated and identified by GC/MS analysis as 1aOH-24-oxo-D3, 1a,23(OH)2-24oxo-D3, 1a,25(OH)2-24-oxo-D3, 1a,23,25(OH)3-24-oxoD3, 1a, 23(OH)2-24,25,26,27-tetranor-D3 (C-23 alcohol), and 1a,24(R),25(OH)3D3. From this information, it can be inferred that CYP24A1 catalyzes oxidation at C-24 to the oxo-form, hydroxylation at C-25 as well as subsequent oxidations in the metabolism of 1a,24(R)(OH)2D3 as described in our previous publication [9]. Furthermore, it was noted that 1a,24(R) (OH)2D3, unlike 1a,25(OH)2D3, is still detected at 10 min of incubation indicating higher metabolic stability for 1a,24(R)(OH)2D3 when compared to 1a,25(OH)2D3.

Fig. 5. HPLC profiles of organic extracts from incubations of 1a,25(OH)2D3 (A1–A3) and 1a,24(R)(OH)2D3 (B1–B3) in the reconstituted system containing CYP24A1, adrenodoxin (ADX), adrenodoxin reductase (ADR), and NADPH. Incubations were performed for 10 min (A2, B2) and 30 min (A3, B3). Control incubations were conducted in the absence of ADR (A1, B1). Lipid extracts of the incubation solutions were subjected to HPLC analysis. HPLC was performed using a Zorbax-SIL column (4.6 · 250 mm) eluted with hexane/isopropanol (93:7, vol/vol) (HPLC system I). Numbers indicate elution positions of various standards shown in A1 and B1.

N. Astecker et al. / Archives of Biochemistry and Biophysics 431 (2004) 261–270

Discussion The mitochondrial cytochrome P450 enzyme 25hydroxyvitamin D3-24-hydroxylase (CYP24A1) is well known for catalyzing hydroxylation at C-24 in the metabolism of 25OHD3 and 1a,25(OH)2D3 [23,24]. Recent studies have shown that recombinant CYP24A1 catalyzes not only a single hydroxylation but multiple oxidative reactions at C-24 as well as at C-23 and C26 in the side chain of 1a,25(OH)2D3 [20,24–28]. This finding suggested that CYP24A1 is responsible for all oxidative reactions involved in the C-24 and C-23 oxidation pathways of 1a,25(OH)2D3 metabolism [10,29,30]. In the present study, we critically evaluated the role of CYP24A1 in the metabolism of 1a,24(R)(OH)2D3 in human keratinocytes and especially in its ability to catalyze C-25 hydroxylation of 1a,24(R)(OH)2D3. We found that 1a,24(R)(OH)2D3 is metabolized in human keratinocytes into both 25-hydroxylated and non-25-hydroxylated metabolites in the same way as in rat kidney reported in our previous study [9]. In addition to all previously identified metabolites of 1a,24(R)(OH)2D3 produced in rat kidney, we also identified in this study two new metabolites namely 1a,23,24(OH)3D3 and 1a,24(OH)223-oxo-D3. The production of these two new metabolites initiated by C-23 oxidation appears to be unique to keratinocytes as we did not detect their production either in our previous study using the isolated perfused rat kidney system, or in our present study using the highly purified rat recombinant CYP24A1 reconstituted system. From recent studies [27,28] investigating the metabolism of 1a,25(OH)2D3 using the Escherichia coli expression system for both rat CYP24A1 and human CYP24A1, it was discovered that subtle differences exist between the type of metabolites produced by rat CYP24A1 vs. human CYP24A1. It was shown that metabolites produced through the C-23 oxidation pathway predominate in the case of human CYP24A1 when compared to rat CYP24A1. This finding explains why the rat CYP24A1 did not produce the two new C-23 oxidized metabolites, identified in this study. Furthermore, we also identified along with the C-24 and C-23 oxidized metabolites the production of a stereoisomer of 1a,24(R)(OH)2D3 in human keratinocytes. The exact structure identification of this metabolite was not achieved in this study because of limited amount material available for NMR studies. In keratinocytes, like in other target cells, CYP24A1 is not detected under basal conditions. CYP24A1 expression and activity is induced by the natural hormone 1a,25(OH)2D3 as well as by biologically active vitamin D analogs like 1a,24(R)(OH)2D3 [14] (Fig. 3). The induction of the expression and activity of CYP24A1 in keratinocytes resulted in not only C-24 oxidation but also surprisingly in C-25 hydroxylation of 1a,24(R)(OH)2D3. This finding suggested that

269

CYP24A1 catalyzes both C-24 oxidation as well as C25 hydroxylation of 1a,24(R)(OH)2D3. Sterol C-27 hydroxylase (CYP27A1), a cytochrome P450 enzyme that has been shown to catalyze C-25 hydroxylation of vitamin D3 compounds, is constitutively expressed in human keratinocytes but its activity is low and not induced by 1a,25(OH)2D3 [16]. Under our experimental conditions, CYP27A1 does not substantially contribute to the C-25 hydroxylation of 1a,24(R)(OH)2D3 (Fig. 3B, 1 h). However, our data indicate that C-25 hydroxylation of 1a,24(R)(OH)2D3 by CYP24A1 is operative in keratinocytes which are the main target cells for 1a,24(R)(OH)2D3. The ability of CYP24A1 to perform C-25 hydroxylation of 1a,24(R)(OH)2D3 was unequivocally demonstrated using highly purified recombinant CYP24A1 reconstituted system. These results show for the first time that CYP24A1 is capable of catalyzing C-25 hydroxylation of certain vitamin D compounds. C-25 hydroxylation is a crucial step in the activation of natural vitamin D3 in which vitamin D3 undergoes C25 hydroxylation in liver before it is converted into the hormonally active form 1a,25(OH)2D3 in kidney [30]. The synthetic analog 1a,24(R)(OH)2D3 also has been shown to undergo C-25 hydroxylation in liver [31] and in target tissues ([9], present study). The identity of the enzyme that is responsible for C-25 hydroxylation of vitamin D3 in liver has not been fully elucidated. Five different hepatic cytochrome P450 enzymes, CYP2C11 [32,33], CYP27A1 [17], CYP2D25 [34], CYP2R1 [35], and CYP3A4 [36] have been reported to catalyze C-25 hydroxylation of vitamin D3 compounds. At present, the expression and the activity of these different cytochrome P40 enzymes except CYP27A1 are unknown in human keratinocytes. In spite of our lack of thorough understanding of the various enzymes involved in hydroxylating vitamin D compounds at C-25; we provide in our present study unequivocal evidence to indicate that C-25 hydroxylation of 1a,24(R)(OH)2D3 in target cells can also be catalyzed by CYP24A1. C-25 hydroxylation is not required for 1a,24(R)(OH)2D3 to exert biological activities. 1a,24(R)(OH)2D3 is biologically active without a hydroxyl group at C-25. The hydroxyl group at C-24 of 1a,24(R)(OH)2D3 acts as a surrogate C-25 hydroxyl function [37]. The biological activity profile of 1a,24(R)(OH)2D3 as well as its binding affinities for VDR and vitamin D binding protein (DBP) are well documented [38–41]. The fact that CYP24A1 catalyzes C-25 hydroxylation of 1a,24(R)(OH)2D3 raises the question whether CYP24A1 can accept other non 25-hydroxylated vitamin D compounds as its substrates. Earlier studies have reported that CYP24A1 does not hydroxylate vitamin D3 or the synthetic analog 1a (OH)D3 at C-25 [9,23,42,43]. From these studies, it became obvious that the vitamin D compounds have to be hydroxylated first at C-25 prior to their subsequent metabolism by CYP24A1. However, in our

270

N. Astecker et al. / Archives of Biochemistry and Biophysics 431 (2004) 261–270

present study, we show that CYP24A1 metabolizes 1a,24(R)(OH)2D3 that lacks the hydroxyl group at C25. This finding indicates that CYP24A1 can still accept certain non 25-hydroxylated vitamin D compounds as its substrates. In summary, we identified in the present study CYP24A1 as the main enzyme responsible for the metabolism of the antipsoriatic drug 1a,24(R)(OH)2D3 in human keratinocytes. 1a,24(R)(OH)2D3 is metabolized via several oxidation attacks in the side chain at C-23, C-24, and C-25. These modifications lead to the production of metabolites that are unique to 1a,24(R)(OH)2D3 along with metabolites that are identical to those produced from the natural hormone 1a,25(OH)2D3. Furthermore, we provide unequivocal evidence to indicate that the multicatalytic enzyme CYP24A1 is capable of hydroxylating 1a,24(R)(OH)2D3 at C-25. Acknowledgments This work was supported by United States Public Health Service Grant DK52488 to G.S.R. and Grant AR45455 to J.L.O. References [1] E. Abe, C. Miyaura, H. Sakagami, M. Takeda, K. Konno, T. Yamazaki, S. Yoshiki, T. Suda, Proc. Natl. Acad. Sci. USA 78 (1981) 4990–4994. [2] E.L. Smith, N.C. Walworth, M.F. Holick, J. Invest. Dermatol. 86 (1986) 709–714. [3] R. Bouillon, W.H. Okamura, A.W. Norman, Endocr. Rev. 16 (1995) 200–257. [4] T. Kato, M. Rokugo, T. Terui, H. Tagami, Br. J. Dermatol. 115 (1986) 431–433. [5] K. Kragballe, Arch. Dermatol. Res. 284 (Suppl. 1) (1992) S30– S36. [6] J.M. Lemire, D.C. Archer, G.S. Reddy, Endocrinology 135 (6) (1994) 2818–2821. [7] M.-L. Siu-Caldera, H. Sekimoto, S. Peleg, C. Nguyen, A.-M. Kissmeyer, L. Binderup, A. Weiskopf, P. Vouros, M.R. Uskokovic, G.S. Reddy, J. Steroid Biochem. Mol. Biol. 71 (1999) 111– 121. [8] A.J. Brown, Curr. Pharm. Des. 6 (7) (2000) 701–716. [9] E.A. Weinstein, D.S. Rao, M.-L. Siu-Caldera, K.-Y. Tserng, M.R. Uskokovic, S. Ishizuka, G.S. Reddy, Biochem. Pharmacol. 58 (12) (1999) 1965–1973. [10] G.S. Reddy, K.-Y. Tserng, Biochemistry 28 (1989) 1763–1769. [11] P.C.M. Van de Kerkhof, Br. J. Dermatol. 132 (1995) 675–682. [12] J.M. Mommers, F.A.C.M. Castelijns, B.A.M.P.A. Seegers, M.M. Van Rossum, C.A.E.M. Van Hooijdonk, P.E.J. Van Erp, P.C.M. Van de Kerkhof, Br. J. Dermatol. 139 (1998) 468–471. [13] M.L. Chen, G. Heinrich, Y.-I. Ohyama, K. Okuda, J.L. Omdahl, T.C. Chen, M.F. Holick, Proc. Soc. Exp. Biol. Med. 207 (1994) 57–61.

[14] D.D. Bikle, M.K. Nemanic, E. Gee, P. Elias, J. Clin. Invest. 78 (1986) 557–566. [15] B. Lehmann, O. Tiebel, M. Meurer, Arch. Dermatol. Res. 291 (1999) 507–510. [16] M. Schuessler, N. Astecker, G. Herzig, G. Vorisek, I. Schuster, Steroids 66 (2001) 399–408. [17] Y.-D. Guo, S. Strugnell, D.W. Back, G. Jones, Proc. Natl. Acad. Sci. USA 90 (1993) 8668–8672. [18] G.S. Reddy, G. Jones, S.W. Kooh, D. Fraser, Am. J. Physiol. 243 (1982) E265–E271. [19] E.G. Bligh, W.J. Dyer, Can. J. Biochem. Physiol. 37 (1959) 911– 917. [20] A. Annalora, E.A. Bobrovnikova-Marjon, R. Serda, L. Lansing, M. Chiu, A. Pastuszyn, S. Iyer, C. Marcus, J.L. Omdahl, Arch. Biochem. Biophys. 425 (2004) 133–146. [21] G.S. Reddy, K.R. Muralidharan, W.H. Okamura, K.Y. Tserng, J.A. McLane, Steroids 66 (2001) 441–450. [22] N. Astecker, G.S. Reddy, G. Herzig, G. Vorisek, I. Schuster, Mol. Cell. Endocrinol. 170 (2000) 91–101. [23] Y. Ohyama, K. Okuda, J. Biol. Chem. 266 (1991) 8690–8695. [24] M. Akiyoshi-Shibata, T. Sakaki, Y. Ohyama, M. Noshiro, K. Okuda, Y. Yabusaki, Eur. J. Biochem. 224 (1994) 335–343. [25] M.J. Beckmann, P. Tadikonda, E. Werner, J. Prahl, S. Yamada, H.F. DeLuca, Biochemistry 35 (1996) 8465–8472. [26] T. Sakaki, N. Sawada, Y. Nonaka, Y. Ohyama, K. Inouye, Eur. J. Biochem. 262 (1999) 43–48. [27] T. Sakaki, N. Sawada, K. Komai, S. Shiozawa, S. Yamada, K. Yamamoto, Y. Ohyama, K. Inouye, Eur. J. Biochem. 267 (2000) 6158–6165. [28] N. Sawada, T. Kusudo, T. Sakaki, S. Hatakeyama, M. Hanada, D. Abe, M. Kamao, T. Okano, M. Ohta, K. Inouye, Biochemistry 43 (15) (2004) 4530–4537. [29] S. Ishizuka, A.W. Norman, J. Biol. Chem. 262 (1987) 7165–7170. [30] G. Jones, S.A. Strugnell, H.F. DeLuca, Physiol. Rev. 78 (1998) 1193–1231. [31] H. Kawashima, K. Hoshina, N. Saitoh, Y. Hashimoto, S. Ishimoto, T. Noguchi, H. Orimo, FEBS Lett. 104 (1979) 367–370. [32] S. Andersson, H. Jornvall, J. Biol. Chem. 261 (1986) 16932– 16936. [33] S. Hayashi, M. Noshiro, K. Okuda, J. Biochem. (Tokyo) 99 (6) (1986) 1753–1763. [34] F. Hosseinpour, K. Wikvall, J. Biol. Chem. 275 (2000) 34650. [35] J.B. Cheng, D.L. Motola, D.J. Mangelsdorf, D.W. Russell, J. Biol. Chem. 278 (2003) 38084–38093. [36] R.P. Gupta, B.W. Hollis, S.B. Patel, K.S. Patrick, N.H. Bell, J. Bone Miner. Res. 19 (4) (2004) 680–688. [37] P.D. Siebert, N. Ohnuma, A.W. Norman, Biochem. Biophys. Res. Commun. 91 (1979) 827–834. [38] T. Matsunaga, M. Yamamoto, H. Mimura, T. Ohta, M. Kiyoki, T. Ohba, T. Nauchi, J. Hosoi, T. Kuroki, J. Dermatol. 17 (1990) 135–142. [39] K. Matsumoto, K. Hashimoto, M. Kiyoki, M. Yamamoto, K. Yoshikawa, J. Dermatol. 17 (1990) 97–103. [40] T. Kobayashi, H. Okamura, Y. Azuma, M. Kiyoki, K. Matsumoto, K. Hashimoto, K. Yoshikawa, J. Dermatol. 17 (1990) 707– 709. [41] R. Horst, S. Prapong, T. Reinhardt, N. Koszewski, J. Knutson, C. Bishop, Biochem. Pharmacol. 60 (2000) 701–708. [42] J.L. Omdahl, L.A. Hunsaker, A.P. Evan, P. Torrez, J. Biol. Chem. 255 (1980) 7460–7466. [43] G.S. Reddy, G. Jones, S.W. Kooh, D. Fraser, H.F. DeLuca, Am. J. Physiol. 254 (1983) E359–E364.