- Email: [email protected]
Omeprazole, an inducer of human CYP1A1 and 1A2, is not a ligand for the Ah receptor
Vol. 188, No. 2, 1992 October
BIOCHEMICAL
AND BIOPHYSICAL
30, 1992
RESEARCH COMMUNICATIONS Pages 820-825
OMEPRAZOLE, AN INDUCER OF HUMAN CYPlAi AND lA2, IS NOT A LIGAND FOR THE Ah RECEPTOR Martine Daujat’, Bogumila Peryt*, Pierre L.esca*, Gilles Fourtanie$, Jacques Domergue4, and Patrick Maurel’ 1 INSERM U-128, CNRS, BP 5051, Montpellier 34033,France 2 Laboratoire de Pharmacologic-Toxicologic, INRA, BP 3, Toulouse 3 193 &France 3Service de Chirurgie Digestive, Hopital de Rangueil, Toulouse 31054,France 4 Service de Chirurgie C, Hopital Saint Eloi, Montpellier 34059,France
Received
September
17,
1992
Summary. Omeprazole is a benzimidazole derivative which induces both P450 1Al and lA2 in human liver in vitro and in vivo. Northern blot analysis of polyA RNA prepared from primary cultures of human hepatocytes indicates that both 1Al and lA2 messages are induced by Dnaphthoflavone and omeprazole. Co-treatment of cells with these inducers and with actinomycin D or cycloheximide results in no accumulation of both n-RNA or superinduction of 1AlmRNA, respectively. 9S enriched fraction of cytosol was prepared either from human hepatocytes in culture or from human liver tissue and analyzed by sucrose density gradient sedimentation for its capacity to bind 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), omeprazole or omeprazole sulfone (a metabolite of omeprazole in man). Whereas 2 p.M TCDD displaced almost totally [3H]TCDD from the Ah receptor, both omeprazole and omeprazole sulfone did not, even at 5000-fold molar excess. In addition, when [r4C] omeprazole was incubated with 9S enriched fraction of human liver or hepatocyte cytosol, no interaction could be detected in sucrose density gradient. These experiments suggest that omeprazole is not a ligand for the human liver Ah 0 1992 Academic Press, Inc. receptor
Omeprazole was previously shown to be an aryl hydrocarbon-like inducer of CYPlAl and lA2 in human hepatocytes in primary culture and in vivo (1). Induction of P450 1Al and lA2 is known to be mediated through the cytosolic Ah receptor (2) which binds the inducer, translocates to the nucleus where it interacts with specific 5’DNA regulatory sequences. This interaction results in the activation of the transcription of a battery of genes including glutathione S-transferase, UDP glucuronyl transferase, NADPH menadione oxidoreductase, in addition to CYPlAl and lA2 (3). The Ah receptor has been previously shown to bind specifically planar polycyclic aromatic hydrocarbons. Other inducers of cytochromes P450, for example phenobarbital and dexamethasone, do not compete with these molecules for the Ah receptor (2). The finding that omeprazole, a benzimidazole derivative which does not apparently fit the structural requitements for binding to the receptor, induces P450 1Al and lA2 raises the question of whether or not this compound is a ligand for the Ah receptor. In a previous work (1) we tentatively addressed this question by using two strains of mice, C57BL/6 and DBA/2, known to be responsive and non 0006-291X/92 $4.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
820
Vol.
188, No. 2, 1992
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
responsive to TCDD, respectively (4, 5). However, omeprazole failed to induce both 1Al and lA2 in both strains, making this investigation unconclusive. We therefore decided to investigate directly the putative interaction of omeprazole with the human liver Ah receptor. In this communication we report that while the human liver cytosol does contain a functional Ah receptor which binds TCDD, a prototypical inducer of CYPlA 1, no interaction could be detected between this receptor and omeprazole. MATERIALS
AND METHODS
Human
liver samples and primary cultures of human hepatocytes. Human liver tissue was obtained from a 30 year-old male who became an organ donor after cerebral aneurysm. The use of this tissue for scientific purpose was authorized by the French National Ethics Committee, the liver of this subject being unproper for transplantation. Hepatocyte cultures FT13, FT31 and FT33 were prepared from lobectomies resected for: a primary hepatocellular carcinoma on a 63 year-old man, a hepatic adenoma on a 24 year-old woman, and a hydatic cyst on a 54 year-old woman, respectively. Hepatocytes were isolated and cultured in a serum free medium as previously described (1). Isolation of polyA RNA and northern blot analysis. Cells were maintained in culture for 72 H after plating before any treatment. At this time, cells were either untreated (UT) or treated with 50 jtM B-naphthoflavone (BNF) or omeprazole (OM) or both, for an additional 18 H in the absence or in presence of 5 pg actinomycin D (ActD) or 10 l.tg cycloheximide (CHX). Poly A RNA was isolated as described (1) and 15 pg were analyzed in northern blot with two cDNA probes: 1) PCRlA which hybridizes with both 1Al and lA2 mRNAs: a 632 bp long fragment of gene 1Al exon 2 from nucleotide 2868 to 3500 made by polymerase chain reaction from human genomic DNA and specific oligonucleotides; 2) pGAPDH: a plasmid harboring a cDNA coding for rat hepatic glyceraldehyde phosphate dehydrogenase. Preparation of cytosol and sucrose density gradient analysis. Cytosol was prepared from human liver tissue or from cultured untreated hepatocytes essentially as described (6), by centrifugation of the homogenates for 1 H at 100 000 x g in HEDGM buffer (25 mM HEPES, pH 7.4, 1.5 mM EDTA, 1 mM dithiothreitol, 20 mM sodium molybdate, and 10% v/v glycerol). Enriched 9s fraction was prepared by sedimentation of cytosol on a 5-20% sucrose gradient for 3 H at 372 000 x g on a VTi vertical rotor. Fractions 18-22 were pooled and concentrated in CF25 centiflo membrane cones (Amicon). This 9s enriched fraction (0.4 ml in HEDGM, 0.5-l mg protein) was then incubated for 1 H with 20-25 nM [3H] TCDD or of 940 nM [14C] omeprazole, in the absence or presence of either 2 yM TCDD, 2.5-25 PM TCDF, 100 PM omeprazole or omeprazole sulfone. After treatment with charcoal-dextran (0.5 mg/ml), the sample was loaded on a lo-30% sucrose gradient and centrifuged again for 2 H at 372 000 x g. Twenty two fractions (282 pl each) were collected and analyzed for radioactive bound material in a scintillation counter. Chemicals. TCDD, [3H] TCDD (37 Ci/mmol) and TCDF (2,3,7,8-tetrachlorodibenzofuran), were purchased from Chemsyn Science Lab. (Lenexa, TX). Omeprazole was from Astra Haeslle AB (Moelndal, Sweden). [14C] omeprazole (53 mCi/mmol) and omeprazole sulfone were obtained from Glaxo Research Group Ltd. (Ware, England). Other chemicals were from Sigma (Saint Louis, MO). RESULTS When 72 H-old primary cultures of human hepatocytes were treated for 18 H in the presence of 50 PM &naphthoflavone or omeprazole, both 1Al and lA2 mRNA were significantly induced (Figure 1 left). Previous experiments demonstrated that the kinetics and concentrationdependance of induction were the same whatever the inducer, l3-naphthoflavone or omeprazole (7). Both mRNAs remained undetectable in untreated cells. Actinomycin D fully blocked the induction of both mRNAs, strongly suggesting a transcriptional activation of CYPlAl and lA2 by B-naphthoflavone and omeprazole. In contrast, cycloheximide produced a superinduction (8, 821
Vol.
188,
No.
2,
1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
- lA2 - IA1
-1A2 -lAl
- GAPDH
-
A UT
C
-
BNF
AC-AC BNF
OM
MC
OM
MC
BNF
OM
&ggre G Northern blot analysis of polyA RNA from human hepatocyte in primary culture treated with various compounds. UT, untreated; BNF, 50pM l3naphtholfavone; OM, 50 WM omeprazole; MC, 50fiM 3-methylcholanthrene; A, Z&g/ml actinomycin D; C, lOlg/ml cycloheximide. Experiments carried out with cultures FT13 and FT33.
9) of 1Al mRNA accumulation with either inducer, while concomitantly, it decreased the accumulation of lA2 mRNA. Interestingly, cycloheximide alone, in the absence of inducer, produced a detectable accumulation of the 1Al mRNA, as previously found in rabbit hepatocyte cultures (10). Finally, co-treatment of cells with &naphthoflavone and omeprazole (or LL naphthoflavone and 3-methylcholanthrene) did not produce a higher accumulation of 1Al or lA2 mRNA than with either compound used alone (Figure 1 right ). In parallel experiments, omeprazole sulfone, one of the most abundant circulating metabolite of omeprazole in man (1 l), did not induce 1Al nor lA2 mRNA nor protein accumulation, under the same conditions (not shown). Expression of GAPDH mRNA was used as a control in these experiments: this gene is constitutively expressed in our culture system, it is not inducible by B-naphthoflavone nor omeprazole and the GAPDH messenger half-life is much larger than 18 H since the treatment with actinomycin D did not affect its accumulation. In sum, these results demonstrate that omeprazole behaves as an aryl hydrocarbon-like inducer of human cytochromes P450, the typical inducer being represented here by &naphthoflavone. This prompted us to evaluate the capacity of omepramle to bind to the Ah receptor. When the 9s enriched fraction of human liver cytosol was incubated with [ 3H] TCDD and analyzed on a sucrose density gradient, a peak of radioactivity reflecting the TCDD-Ah receptor complex was obtained in the 9s region, i. e. fraction 12 in our assay (Figure 2). As expected, the peak almost totally disappeared when the incubation was carried out in the presence of a large excess of unlabeled ligand. However, when either omeprazole or omeprazole sulfone were used as competitors in the same experiments, even at a 5000-fold excess with respect to the radiolabeled ligand, the peak of radioactivity was not affected. Although this was not surprising for the sulfone, which is not an inducer, it was so for omeprazole. In a second series of experiments [ 14C] omeprazole was used as the radioligand in parallel with [3H] TCDD. Again, incubation of the 9s enriched fi-action with [ 3H] TCDD resulted in a sharp peak in the expected position, this peak being displaced by an excess of TCDF (a TCDD derivative which is a strong inducer of 1Al). However, no peak of radioactive material was detected following incubation of 822
Vol.
BIOCHEMICAL
188, No. 2, 1992
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
10 Fractions
0
0
0 0
10 Fractions
20
0
20
10 Fractions
20
Figure Sucrose density gradient of 9s enriched fractions of cytosol from human liver or hepatocyte in primary culture. A and B: 9s fraction of cytosol from human liver tissue. C: 9S fraction of cytosol from human hepatocytes in primary culture (FT31). 1.20nM [3H] TCDD; 2.2OnM [ 3H] TCDD plus 2uM TCDD; 3.20nM [ 3H] TCDD plus 1OOuM omeprazole; 4. 20nM [ 3H] TCDD plus 1OOuM omeprazole sulfone; 5. 20nM [3H] TCDD; 6 and 7. 20nM [3H] TCDD plus 2.5 and 25uM TCDF, respectively; 8.940nM [ 14C] omeprazole. TCDD: 2,3,7,8-tetracholodibenzo-p-dioxin.
cytosol with 940 nM [14C] omeprazole. In further experiments the concentration of omeprazole was increased to 940 l.tM without success. These experiments strongly suggest that omeprazole is not a ligand for the Ah receptor. DISCUSSION We show in this communication that while omeprazole behaves as a typical inducer of P450 1Al and lA2 in human hepatocytes, it does not apparently bind to the Ah receptor. Induction of CYPlAl and lA2 by this compound in our culture system is characterized by the following criteria: i) omeprazole induces both CYPlAl and lA2 in 100% of cultures tested, ii) except for another benzimidazole derivative of close chemical structure (unpublished results), it is the only non planar non polycyclic aromatic chemical that induces CYPl As; iii) it is effective in the same range of concentration (lo-50 p.M) as l3-naphthoflavone or 3-methylcholanthrene; iv) it induces CYPlAl and lA2 messages with the same kinetics as I&naphthoflavone or 3methylcholanthrene; iv) it exhibits a superinduction when used in association with cycloheximide; v) its effect is not additive with that of &naphthoflavone. The Ah receptor has been characterized for a number of years. Its central role in the process leading to 1Al induction has been established from the use of different animal strains (4, 5) or cell lines (12, 13) some of them lacking the functional receptor and accordingly exhibiting either no response or atypical response to the aryl hydrocarbons. Briefly, after binding the inducer, the ligand-Ah receptor complex translocates to the nucleus where it interacts with two so-called AhREl or 2 or XRE (aryl hydrocarbon or xenobiotic responsive element) of the 5’ flanking region of the gene (14). These elements were shown to behave as enhancers of the gene transcription (15). In addition, deletion analysis of the transcriptional activity of the 5’ flanking 823
Vol.
188, No. 2, 1992
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
region revealed another element, NRE or AhRE3 (negative responsive element), whose effect should be to repress the expression of the gene, presumably through interaction with one or several factors (14). The superinduction of 1Al mRNA observed when cells or animals are treated concomitantly with an aryl hydrocarbon and cycloheximide was interpreted as reflecting the release of a labile factor from the NRE or AhRE3 thus suppressing its inhibitory effect (8,9, 14). In sum, transcriptional activation of CYPlAl might occur either through induction by an aryl hydrocarbon activating the XREs, or through derepression via a process involving the interaction between the NRE and its specific binding factor(s). Recently, on the basis of the structural similarities between sterols and benzo(a)pyrene, the prototypic substrate of P450 lA1, Nebert and co-workers (16) reasoned that endogenous sterols could represent the natural substrates regulating the CYPlAl gene. Mevinolin, an inhibitor of HMG-CoA reductase, the rate-limiting enzyme in cholesterol biosynthesis, was accordingly used to deplete the cells of sterols. Unexpectedly, mevinolin was shown to induce CYPlAl independently of its inhibitory effect on HMG-CoA reductase. Moreover, this molecule was not able to displace TCDD from the Ah receptor. Interaction of the AhRE3 sequence of the CYPlAl gene with nuclear extracts was investigated by the gel mobility shift assay. The results were interpreted as reflecting a derepressive effect of mevinolin or metabolites on the gene transcription. Omeprazole and mevinolin do not exhibit obvious structural similarities. On the other hand, omeprazole sulfone is not an inducer of CYPlAl and lA2 (nor does it bind the Ah receptor) in spite of its close structural similarity with omeprazole. Would omeprazole and mevinolin act through the same mechanism? Derepression of CYPlAl gene by omeprazole through an hypothetical interaction with factor(s) binding the NRE can be tested experimentally: 1) co-treatment of cells with R-naphthoflavone and omeprazole would result in a superinduction; 2) co-treatment of cells with omeprazole and cycloheximide would not result in a superinduction of 1Al messenger. This is in contrast with what we observed. However, cycloheximide might exert a number of direct and/or undirect effects on the cell metabolism so that its mode of action described above, in terms of CYPlAl superinduction, is perhaps simplistic and must remain hypothetical. Experiments aimed at determining whether or not omeprazole is able to interact in some way with the NRE of human CYPlAl by gel mobilityshift assay are in progress. Another possibility is that an unidentified metabolite of omeprazole (other than the sulfone) is the actual inducer. We are currently investigating the metabolism of this compound in our hepatocyte culture system. Our results (unpublished) show that in 72 H-old untreated cultures, 8 to 15% of the initial amount of omeprazole (50 jtM) is found as oxidized metabolites in the extracellular medium after 1 H incubation, as determined by HPLC analysis. Since the extend of metabolism can be induced (by treatment of cells with rifampicin or phenobarbital but not by Bnaphthoflavone) or inhibited (by ketoconazole or cyclosporine for example), we are currently addressing the question of whether or not the oxidative metabolism of this molecule modulates its inducing potency in our culture system. It is ironic that while omeprazole was the first drug to be characterized as an aryl hydrocarbon-like inducer in man, it is also one of the first of the CYPlA inducers to apparently not interact with the Ah receptor. Clearly, inducers of CYPl Al gene are not limited to planar aromatic hydrocarbons. The cases of omeprazole and mevinolin suggest that the screening for CYPlAl inducers should be extended to other classesof compounds and that the mechanism(s) of induction of this gene is (are) far more complex than previously suspected. 824
Vol.
188,
No.
2,
1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
ACKNOWLEDGMENTS We thank Drs. Daniel W. Nebert (University of Cincinnati) and Pave1 Anzenbacher (Institute of Biopharmacy, Hradec KralovC) for critical reading of the manuscript. This work was supported in part by a grant from Glaxo Research Group Ltd (Ware, England). The authors wish to thank Dr. J. M. Blanchard (USTL, Montpellier, France) for providing the GAPDH cDNA probe used in this work. REFERENCES 1. Diaz, D., Fabre, I., Daujat, M., Saint Aubert, B., Bories, P., Michel, H., and Maurel, P. (1990) Gastroenterology 99,737-747. 2. Poland, A. P., Glover, E., and Kende, A. S. (1976) J. Biol. Chem. 251,4936-4946. 3. Nebert, D. W., and Gonzalez, F. J. (1987) Ann. Rev. Biochem. 56,945-993. Nebert, D. W., and Gelboin, H. V. (1969) Arch. Biochem. Biophys. 134,76-89. i: Poland, A. P., Glover, E., Robinson, J. R., and Nebert, D. W. (1974) J. Biol. Chem. 249, 5599-5606. Lesca, P., Fernandez, N., and Roy, M. (1987) J. Biol. Chem. 262,4827-4835. ;: Daujat, M., Fabre, I., Diaz, D., Pichard, L., Fabre, G., Fabre, J. M., Saint Aubert, B., and Maurel, P. (1990) Biochem Pharmacol. (Life Sci. Adv.) 9, 3 15-326. Nebert, D. W., and Gelboin, H. V. (1970) J. Biol. Chem. 245, 160-168. i: Israel, D. I., Estolano, M. G., Galeazzi, D. R., Whitlock, J. P. (1985) J. Biol. Chem. 260, 5648-5653. lO.Daujat, M., Clair, P., Astier, C., Fabre, I., Pineau, T., Yerle, M., Gellin, J. and Maurel, P. (1991) Eur. J. B&hem. 200, 501-510. 1 l.Regardh, C. G., Andersson, T., Lagerstrom, P. O., Lundborg, P., and Skanberg, I. (1990) Ther. Drug Monitor. 12, 163-172. 12Hankinson, 0. (1985) Somat. Cell Genet. 9,497-514. 13.Hankinson, O., Andersen, R. D., Birren, B., Sander, F., Negishi, M,, and Nebert, D. W. (1985) J. Biol. Chem. 260, 1790-1795. 14.Jones, P. B., Galeazzi, D. R., Fisher, J. M., and Whitlock, J. P. (1985) Science 227, 14991502. 15.Fujisawa-Sehara, A., Sogawa, K., Yamane M., and Fujii-Kuriyama, Y (1987) Nucl. Acid Res. 15, 4179-4191. 16.Puga, A., Raychaudhuri, B., and Nebert, D. W. (1992) FASEB J. 6,777-785.
825
BIOCHEMICAL
AND BIOPHYSICAL
30, 1992
RESEARCH COMMUNICATIONS Pages 820-825
OMEPRAZOLE, AN INDUCER OF HUMAN CYPlAi AND lA2, IS NOT A LIGAND FOR THE Ah RECEPTOR Martine Daujat’, Bogumila Peryt*, Pierre L.esca*, Gilles Fourtanie$, Jacques Domergue4, and Patrick Maurel’ 1 INSERM U-128, CNRS, BP 5051, Montpellier 34033,France 2 Laboratoire de Pharmacologic-Toxicologic, INRA, BP 3, Toulouse 3 193 &France 3Service de Chirurgie Digestive, Hopital de Rangueil, Toulouse 31054,France 4 Service de Chirurgie C, Hopital Saint Eloi, Montpellier 34059,France
Received
September
17,
1992
Summary. Omeprazole is a benzimidazole derivative which induces both P450 1Al and lA2 in human liver in vitro and in vivo. Northern blot analysis of polyA RNA prepared from primary cultures of human hepatocytes indicates that both 1Al and lA2 messages are induced by Dnaphthoflavone and omeprazole. Co-treatment of cells with these inducers and with actinomycin D or cycloheximide results in no accumulation of both n-RNA or superinduction of 1AlmRNA, respectively. 9S enriched fraction of cytosol was prepared either from human hepatocytes in culture or from human liver tissue and analyzed by sucrose density gradient sedimentation for its capacity to bind 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), omeprazole or omeprazole sulfone (a metabolite of omeprazole in man). Whereas 2 p.M TCDD displaced almost totally [3H]TCDD from the Ah receptor, both omeprazole and omeprazole sulfone did not, even at 5000-fold molar excess. In addition, when [r4C] omeprazole was incubated with 9S enriched fraction of human liver or hepatocyte cytosol, no interaction could be detected in sucrose density gradient. These experiments suggest that omeprazole is not a ligand for the human liver Ah 0 1992 Academic Press, Inc. receptor
Omeprazole was previously shown to be an aryl hydrocarbon-like inducer of CYPlAl and lA2 in human hepatocytes in primary culture and in vivo (1). Induction of P450 1Al and lA2 is known to be mediated through the cytosolic Ah receptor (2) which binds the inducer, translocates to the nucleus where it interacts with specific 5’DNA regulatory sequences. This interaction results in the activation of the transcription of a battery of genes including glutathione S-transferase, UDP glucuronyl transferase, NADPH menadione oxidoreductase, in addition to CYPlAl and lA2 (3). The Ah receptor has been previously shown to bind specifically planar polycyclic aromatic hydrocarbons. Other inducers of cytochromes P450, for example phenobarbital and dexamethasone, do not compete with these molecules for the Ah receptor (2). The finding that omeprazole, a benzimidazole derivative which does not apparently fit the structural requitements for binding to the receptor, induces P450 1Al and lA2 raises the question of whether or not this compound is a ligand for the Ah receptor. In a previous work (1) we tentatively addressed this question by using two strains of mice, C57BL/6 and DBA/2, known to be responsive and non 0006-291X/92 $4.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
820
Vol.
188, No. 2, 1992
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
responsive to TCDD, respectively (4, 5). However, omeprazole failed to induce both 1Al and lA2 in both strains, making this investigation unconclusive. We therefore decided to investigate directly the putative interaction of omeprazole with the human liver Ah receptor. In this communication we report that while the human liver cytosol does contain a functional Ah receptor which binds TCDD, a prototypical inducer of CYPlA 1, no interaction could be detected between this receptor and omeprazole. MATERIALS
AND METHODS
Human
liver samples and primary cultures of human hepatocytes. Human liver tissue was obtained from a 30 year-old male who became an organ donor after cerebral aneurysm. The use of this tissue for scientific purpose was authorized by the French National Ethics Committee, the liver of this subject being unproper for transplantation. Hepatocyte cultures FT13, FT31 and FT33 were prepared from lobectomies resected for: a primary hepatocellular carcinoma on a 63 year-old man, a hepatic adenoma on a 24 year-old woman, and a hydatic cyst on a 54 year-old woman, respectively. Hepatocytes were isolated and cultured in a serum free medium as previously described (1). Isolation of polyA RNA and northern blot analysis. Cells were maintained in culture for 72 H after plating before any treatment. At this time, cells were either untreated (UT) or treated with 50 jtM B-naphthoflavone (BNF) or omeprazole (OM) or both, for an additional 18 H in the absence or in presence of 5 pg actinomycin D (ActD) or 10 l.tg cycloheximide (CHX). Poly A RNA was isolated as described (1) and 15 pg were analyzed in northern blot with two cDNA probes: 1) PCRlA which hybridizes with both 1Al and lA2 mRNAs: a 632 bp long fragment of gene 1Al exon 2 from nucleotide 2868 to 3500 made by polymerase chain reaction from human genomic DNA and specific oligonucleotides; 2) pGAPDH: a plasmid harboring a cDNA coding for rat hepatic glyceraldehyde phosphate dehydrogenase. Preparation of cytosol and sucrose density gradient analysis. Cytosol was prepared from human liver tissue or from cultured untreated hepatocytes essentially as described (6), by centrifugation of the homogenates for 1 H at 100 000 x g in HEDGM buffer (25 mM HEPES, pH 7.4, 1.5 mM EDTA, 1 mM dithiothreitol, 20 mM sodium molybdate, and 10% v/v glycerol). Enriched 9s fraction was prepared by sedimentation of cytosol on a 5-20% sucrose gradient for 3 H at 372 000 x g on a VTi vertical rotor. Fractions 18-22 were pooled and concentrated in CF25 centiflo membrane cones (Amicon). This 9s enriched fraction (0.4 ml in HEDGM, 0.5-l mg protein) was then incubated for 1 H with 20-25 nM [3H] TCDD or of 940 nM [14C] omeprazole, in the absence or presence of either 2 yM TCDD, 2.5-25 PM TCDF, 100 PM omeprazole or omeprazole sulfone. After treatment with charcoal-dextran (0.5 mg/ml), the sample was loaded on a lo-30% sucrose gradient and centrifuged again for 2 H at 372 000 x g. Twenty two fractions (282 pl each) were collected and analyzed for radioactive bound material in a scintillation counter. Chemicals. TCDD, [3H] TCDD (37 Ci/mmol) and TCDF (2,3,7,8-tetrachlorodibenzofuran), were purchased from Chemsyn Science Lab. (Lenexa, TX). Omeprazole was from Astra Haeslle AB (Moelndal, Sweden). [14C] omeprazole (53 mCi/mmol) and omeprazole sulfone were obtained from Glaxo Research Group Ltd. (Ware, England). Other chemicals were from Sigma (Saint Louis, MO). RESULTS When 72 H-old primary cultures of human hepatocytes were treated for 18 H in the presence of 50 PM &naphthoflavone or omeprazole, both 1Al and lA2 mRNA were significantly induced (Figure 1 left). Previous experiments demonstrated that the kinetics and concentrationdependance of induction were the same whatever the inducer, l3-naphthoflavone or omeprazole (7). Both mRNAs remained undetectable in untreated cells. Actinomycin D fully blocked the induction of both mRNAs, strongly suggesting a transcriptional activation of CYPlAl and lA2 by B-naphthoflavone and omeprazole. In contrast, cycloheximide produced a superinduction (8, 821
Vol.
188,
No.
2,
1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
- lA2 - IA1
-1A2 -lAl
- GAPDH
-
A UT
C
-
BNF
AC-AC BNF
OM
MC
OM
MC
BNF
OM
&ggre G Northern blot analysis of polyA RNA from human hepatocyte in primary culture treated with various compounds. UT, untreated; BNF, 50pM l3naphtholfavone; OM, 50 WM omeprazole; MC, 50fiM 3-methylcholanthrene; A, Z&g/ml actinomycin D; C, lOlg/ml cycloheximide. Experiments carried out with cultures FT13 and FT33.
9) of 1Al mRNA accumulation with either inducer, while concomitantly, it decreased the accumulation of lA2 mRNA. Interestingly, cycloheximide alone, in the absence of inducer, produced a detectable accumulation of the 1Al mRNA, as previously found in rabbit hepatocyte cultures (10). Finally, co-treatment of cells with &naphthoflavone and omeprazole (or LL naphthoflavone and 3-methylcholanthrene) did not produce a higher accumulation of 1Al or lA2 mRNA than with either compound used alone (Figure 1 right ). In parallel experiments, omeprazole sulfone, one of the most abundant circulating metabolite of omeprazole in man (1 l), did not induce 1Al nor lA2 mRNA nor protein accumulation, under the same conditions (not shown). Expression of GAPDH mRNA was used as a control in these experiments: this gene is constitutively expressed in our culture system, it is not inducible by B-naphthoflavone nor omeprazole and the GAPDH messenger half-life is much larger than 18 H since the treatment with actinomycin D did not affect its accumulation. In sum, these results demonstrate that omeprazole behaves as an aryl hydrocarbon-like inducer of human cytochromes P450, the typical inducer being represented here by &naphthoflavone. This prompted us to evaluate the capacity of omepramle to bind to the Ah receptor. When the 9s enriched fraction of human liver cytosol was incubated with [ 3H] TCDD and analyzed on a sucrose density gradient, a peak of radioactivity reflecting the TCDD-Ah receptor complex was obtained in the 9s region, i. e. fraction 12 in our assay (Figure 2). As expected, the peak almost totally disappeared when the incubation was carried out in the presence of a large excess of unlabeled ligand. However, when either omeprazole or omeprazole sulfone were used as competitors in the same experiments, even at a 5000-fold excess with respect to the radiolabeled ligand, the peak of radioactivity was not affected. Although this was not surprising for the sulfone, which is not an inducer, it was so for omeprazole. In a second series of experiments [ 14C] omeprazole was used as the radioligand in parallel with [3H] TCDD. Again, incubation of the 9s enriched fi-action with [ 3H] TCDD resulted in a sharp peak in the expected position, this peak being displaced by an excess of TCDF (a TCDD derivative which is a strong inducer of 1Al). However, no peak of radioactive material was detected following incubation of 822
Vol.
BIOCHEMICAL
188, No. 2, 1992
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
10 Fractions
0
0
0 0
10 Fractions
20
0
20
10 Fractions
20
Figure Sucrose density gradient of 9s enriched fractions of cytosol from human liver or hepatocyte in primary culture. A and B: 9s fraction of cytosol from human liver tissue. C: 9S fraction of cytosol from human hepatocytes in primary culture (FT31). 1.20nM [3H] TCDD; 2.2OnM [ 3H] TCDD plus 2uM TCDD; 3.20nM [ 3H] TCDD plus 1OOuM omeprazole; 4. 20nM [ 3H] TCDD plus 1OOuM omeprazole sulfone; 5. 20nM [3H] TCDD; 6 and 7. 20nM [3H] TCDD plus 2.5 and 25uM TCDF, respectively; 8.940nM [ 14C] omeprazole. TCDD: 2,3,7,8-tetracholodibenzo-p-dioxin.
cytosol with 940 nM [14C] omeprazole. In further experiments the concentration of omeprazole was increased to 940 l.tM without success. These experiments strongly suggest that omeprazole is not a ligand for the Ah receptor. DISCUSSION We show in this communication that while omeprazole behaves as a typical inducer of P450 1Al and lA2 in human hepatocytes, it does not apparently bind to the Ah receptor. Induction of CYPlAl and lA2 by this compound in our culture system is characterized by the following criteria: i) omeprazole induces both CYPlAl and lA2 in 100% of cultures tested, ii) except for another benzimidazole derivative of close chemical structure (unpublished results), it is the only non planar non polycyclic aromatic chemical that induces CYPl As; iii) it is effective in the same range of concentration (lo-50 p.M) as l3-naphthoflavone or 3-methylcholanthrene; iv) it induces CYPlAl and lA2 messages with the same kinetics as I&naphthoflavone or 3methylcholanthrene; iv) it exhibits a superinduction when used in association with cycloheximide; v) its effect is not additive with that of &naphthoflavone. The Ah receptor has been characterized for a number of years. Its central role in the process leading to 1Al induction has been established from the use of different animal strains (4, 5) or cell lines (12, 13) some of them lacking the functional receptor and accordingly exhibiting either no response or atypical response to the aryl hydrocarbons. Briefly, after binding the inducer, the ligand-Ah receptor complex translocates to the nucleus where it interacts with two so-called AhREl or 2 or XRE (aryl hydrocarbon or xenobiotic responsive element) of the 5’ flanking region of the gene (14). These elements were shown to behave as enhancers of the gene transcription (15). In addition, deletion analysis of the transcriptional activity of the 5’ flanking 823
Vol.
188, No. 2, 1992
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
region revealed another element, NRE or AhRE3 (negative responsive element), whose effect should be to repress the expression of the gene, presumably through interaction with one or several factors (14). The superinduction of 1Al mRNA observed when cells or animals are treated concomitantly with an aryl hydrocarbon and cycloheximide was interpreted as reflecting the release of a labile factor from the NRE or AhRE3 thus suppressing its inhibitory effect (8,9, 14). In sum, transcriptional activation of CYPlAl might occur either through induction by an aryl hydrocarbon activating the XREs, or through derepression via a process involving the interaction between the NRE and its specific binding factor(s). Recently, on the basis of the structural similarities between sterols and benzo(a)pyrene, the prototypic substrate of P450 lA1, Nebert and co-workers (16) reasoned that endogenous sterols could represent the natural substrates regulating the CYPlAl gene. Mevinolin, an inhibitor of HMG-CoA reductase, the rate-limiting enzyme in cholesterol biosynthesis, was accordingly used to deplete the cells of sterols. Unexpectedly, mevinolin was shown to induce CYPlAl independently of its inhibitory effect on HMG-CoA reductase. Moreover, this molecule was not able to displace TCDD from the Ah receptor. Interaction of the AhRE3 sequence of the CYPlAl gene with nuclear extracts was investigated by the gel mobility shift assay. The results were interpreted as reflecting a derepressive effect of mevinolin or metabolites on the gene transcription. Omeprazole and mevinolin do not exhibit obvious structural similarities. On the other hand, omeprazole sulfone is not an inducer of CYPlAl and lA2 (nor does it bind the Ah receptor) in spite of its close structural similarity with omeprazole. Would omeprazole and mevinolin act through the same mechanism? Derepression of CYPlAl gene by omeprazole through an hypothetical interaction with factor(s) binding the NRE can be tested experimentally: 1) co-treatment of cells with R-naphthoflavone and omeprazole would result in a superinduction; 2) co-treatment of cells with omeprazole and cycloheximide would not result in a superinduction of 1Al messenger. This is in contrast with what we observed. However, cycloheximide might exert a number of direct and/or undirect effects on the cell metabolism so that its mode of action described above, in terms of CYPlAl superinduction, is perhaps simplistic and must remain hypothetical. Experiments aimed at determining whether or not omeprazole is able to interact in some way with the NRE of human CYPlAl by gel mobilityshift assay are in progress. Another possibility is that an unidentified metabolite of omeprazole (other than the sulfone) is the actual inducer. We are currently investigating the metabolism of this compound in our hepatocyte culture system. Our results (unpublished) show that in 72 H-old untreated cultures, 8 to 15% of the initial amount of omeprazole (50 jtM) is found as oxidized metabolites in the extracellular medium after 1 H incubation, as determined by HPLC analysis. Since the extend of metabolism can be induced (by treatment of cells with rifampicin or phenobarbital but not by Bnaphthoflavone) or inhibited (by ketoconazole or cyclosporine for example), we are currently addressing the question of whether or not the oxidative metabolism of this molecule modulates its inducing potency in our culture system. It is ironic that while omeprazole was the first drug to be characterized as an aryl hydrocarbon-like inducer in man, it is also one of the first of the CYPlA inducers to apparently not interact with the Ah receptor. Clearly, inducers of CYPl Al gene are not limited to planar aromatic hydrocarbons. The cases of omeprazole and mevinolin suggest that the screening for CYPlAl inducers should be extended to other classesof compounds and that the mechanism(s) of induction of this gene is (are) far more complex than previously suspected. 824
Vol.
188,
No.
2,
1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
ACKNOWLEDGMENTS We thank Drs. Daniel W. Nebert (University of Cincinnati) and Pave1 Anzenbacher (Institute of Biopharmacy, Hradec KralovC) for critical reading of the manuscript. This work was supported in part by a grant from Glaxo Research Group Ltd (Ware, England). The authors wish to thank Dr. J. M. Blanchard (USTL, Montpellier, France) for providing the GAPDH cDNA probe used in this work. REFERENCES 1. Diaz, D., Fabre, I., Daujat, M., Saint Aubert, B., Bories, P., Michel, H., and Maurel, P. (1990) Gastroenterology 99,737-747. 2. Poland, A. P., Glover, E., and Kende, A. S. (1976) J. Biol. Chem. 251,4936-4946. 3. Nebert, D. W., and Gonzalez, F. J. (1987) Ann. Rev. Biochem. 56,945-993. Nebert, D. W., and Gelboin, H. V. (1969) Arch. Biochem. Biophys. 134,76-89. i: Poland, A. P., Glover, E., Robinson, J. R., and Nebert, D. W. (1974) J. Biol. Chem. 249, 5599-5606. Lesca, P., Fernandez, N., and Roy, M. (1987) J. Biol. Chem. 262,4827-4835. ;: Daujat, M., Fabre, I., Diaz, D., Pichard, L., Fabre, G., Fabre, J. M., Saint Aubert, B., and Maurel, P. (1990) Biochem Pharmacol. (Life Sci. Adv.) 9, 3 15-326. Nebert, D. W., and Gelboin, H. V. (1970) J. Biol. Chem. 245, 160-168. i: Israel, D. I., Estolano, M. G., Galeazzi, D. R., Whitlock, J. P. (1985) J. Biol. Chem. 260, 5648-5653. lO.Daujat, M., Clair, P., Astier, C., Fabre, I., Pineau, T., Yerle, M., Gellin, J. and Maurel, P. (1991) Eur. J. B&hem. 200, 501-510. 1 l.Regardh, C. G., Andersson, T., Lagerstrom, P. O., Lundborg, P., and Skanberg, I. (1990) Ther. Drug Monitor. 12, 163-172. 12Hankinson, 0. (1985) Somat. Cell Genet. 9,497-514. 13.Hankinson, O., Andersen, R. D., Birren, B., Sander, F., Negishi, M,, and Nebert, D. W. (1985) J. Biol. Chem. 260, 1790-1795. 14.Jones, P. B., Galeazzi, D. R., Fisher, J. M., and Whitlock, J. P. (1985) Science 227, 14991502. 15.Fujisawa-Sehara, A., Sogawa, K., Yamane M., and Fujii-Kuriyama, Y (1987) Nucl. Acid Res. 15, 4179-4191. 16.Puga, A., Raychaudhuri, B., and Nebert, D. W. (1992) FASEB J. 6,777-785.
825