Pharmacology, Vol. 52, pp. 753-761,
`Biochemical
`CopyrIght 0 1996 Elsevier Science
`Inc.
`
`1996.
`
`+ 0.00
`ISSN 0006-2952/96/$15.00
`PII SOOOS-2952(96)00357-7
`
`Identification of CYP3A4 as the
`(RU 486)
`Principal Enzyme Catalyzing Mifepristone
`Oxidation
`in Human Liver Microsomes
`
`R.
`and
`*DEPARTMENT OF BIOPHARMACEUTICAL SCIENCES, SCHWL OF PHARMACY, UNIVERSITY
`OF CALIFORNIA, SAN FRANCISCO, CA 94143; AND ~DEPARTMENT OF DRUG DISPOSITION,
`LILLY RESEARCH LABORATORIES, ELI LILLY & COMPANY, INDIANAPOLIS, IN 46285, U.S.A.
`
`ABSTRACT.
`(CYP)
`the major cytochrome P450
`approaches were used to elucidate
`Various complementary
`(RU 486) d emethylation
`enzyme responsible
`for mifepristone
`and hydroxylation
`in human
`liver microsomes:
`chemical
`and immunoinhibition
`of specific CYPs; correlation
`analyses between
`initial
`rates of mifepristone
`metabolism
`and relative
`immunodetectable
`CYP
`levels and rates of CYP marker substrate metabolism;
`and
`evaluation of metabolism by cDNA-expressed
`CYP3A4. Human
`liver microsomes
`catalyzed
`the demethylation
`of mifepristone with mean (f SD) apparent K,,, and V,,,
`values of 10.6 f 3.8 FM and 4920 f 1340 pmol/min/mg
`protein,
`respectively;
`the corresponding
`values for hydroxylation
`of the compound were 9.9 f 3.5 FM and 610
`k 260 pmol/min/mg protein. Progesterone
`and midazolam
`(CYP3A4
`substrates)
`inhibited metabolite
`formation
`by up to 77%. The CYP3A
`inhibitors
`gestodene,
`triacetyloleandomycin,
`and 17a-ethynylestradiol
`inhibited
`mifepristone demethylation
`and hydroxylation
`by 70-80%;
`antibodies
`to CYP3A4
`inhibited
`these reactions by
`approximately
`82 and 65%, respectively.
`In a bank of human
`liver microsomes
`from 14 donors, rates of mife-
`pristone metabolism
`correlated
`significantly with relative
`immunodetectable
`CYP3A
`levels, rates of midazolam
`l’- and 4-hydroxylation
`and rates of erythromycin N-demethylation, marker CYP3A
`catalytic activities
`(all r2
`> 0.85 and P < 0.001). No significant
`correlations were observed
`for analyses with relative
`immunoreactive
`levels or marker catalytic
`activities
`of CYPlA2,
`CYP2C9, CYP2C19,
`CYP2D6,
`or CYPZEl.
`Recombinant
`CYP3A4
`catalyzed mifepristone
`demethylation
`and hydroxylation with apparent K,,, values 7.4 and 4.1 p,M,
`respectively. Collectively,
`these data clearly support CYP3A4
`as the enzyme primarily
`responsible
`for mifepris-
`tone demethylation
`and hydroxylation
`in human
`liver microsomes. BIOCHEM PHARMACOL 52;5:753-761,
`1996.
`
`KEY WORDS. mifepristone; cytochrome P450; CYP3A4; demethylation; hydroxylation; human liver micro-
`somes
`
`[17@
`properties of mifepristone
`The antiprogestational
`hydroxy- 11 P-(4-dimethylaminophenyl)-17a-(
`l-propynyl)-
`estra-4,9-dien-3-one,
`Fig. 11, the first antiprogestin used
`clinically, were discovered
`somewhat
`serendipitously
`in
`1980 by scientists at Roussel Uclaf characterizing a series of
`antiglucocorticoids
`[l]. Accordingly,
`the molecule antago-
`nizes progesterone with more than 2-fold greater binding
`affinity
`to the human endometrial progesterone
`receptor
`and cortisol with over lo-fold higher affinity to the human
`placental glucocorticoid
`receptor
`[2]. Because of the recog-
`nized promise of an antiprogestin
`in the areas of pregnancy
`termination
`and fertility control,
`the use of mifepristone
`and newer antiprogestins
`for these
`indications has been
`studied most thoroughly. Currently, mifepristone,
`in com-
`bination with a synthetic prostaglandin analog, is safely and
`effectively used as an abortifacient
`in France, Great Britain,
`
`(415) 476-3853; FAX (415) 476.8887.
`$ co~esponding author.
`Received 29 November 1995; accepted 4 April 1996.
`
`therapeutic uses for
`Sweden, and China. Numerous other
`antiprogestins have been considered, as reviewed in a 1993
`report from the Institute of Medicine
`[3]. Numerous studies
`suggest that antiprogestins may be effective as continually,
`cyclically,
`or post-coitally
`administered
`contraceptive
`agents [4-71. Antiprogestins
`also lend themselves
`to a great
`number of other promising, potential uses unrelated
`to fer-
`tility control. Chief among these is antineoplastic agents for
`certain
`types of breast cancer
`[8, 91, prostate cancer
`[lo],
`[l 1, 121, and uterine
`meningioma
`leiomyoma
`[13]. Impor-
`tantly, mifepristone has been found recently
`to reverse P-
`glycoprotein-mediated
`drug resistance
`oitro [14, 151, a
`characteristic
`that could enhance
`its effectiveness
`as an
`anticancer
`agent alone or in combination
`therapy. Anti-
`progestins may also have a role in the treatment of endo-
`metriosis [13]. Finally, because of its antiglucocorticoid
`ac-
`tivity, mifepristone has been studied as a potential
`treat-
`ment for Gushing’s syndrome
`161.
`These probable
`indications
`for mifepristone would entail
`long-term administration
`of the drug; we thus thought
`it
`
`1
`
`TEVA1022
`
`ELSEVIER
`Gruhmn
`Jung,” Steven A. Wrightonj
`Leslie 2. Benet*#
`Tel
`in
`[
`

`

`754
`
`“37
`
`R.
`
`H3C,
`
`mifepristone 0- hydroxylated monodemethylated didemethylated
`
`FIG. 1. Mifepristone and its three major metabolites in humans, produced via successive demethylations of the 11 P-dimethyl-
`aminophenyl and hydroxylation of the 17cw-propynyl groups.
`
`is primarily
`isoform
`to determine which CUPS
`important
`for its metabolism
`in humans, hypothesizing
`responsible
`that it is CYP3A4. This enzyme and others of the CYP3A
`subfamily are known
`to catalyze steroid oxidations
`in hu-
`mans [17, 181 as well as the metabolism of a great number
`of structurally diverse xenobiotics
`including nifedipine
`[19],
`the immunosuppressants
`cyclosporine
`[20] and tacrolimus
`[21], midazolam and
`triazolam
`[22],
`the antiarrhythmic
`agents
`lidocaine
`[23], amiodarone
`[24, 251 and quinidine
`[26], taxol [27], etoposide
`[28], vinblastine and other vinca
`alkaloids
`[29, 301. Thus, CYP3A4
`involvement
`in the me-
`tabolism of mifepristone could have important
`implications
`for potential drug-drug
`interactions.
`to three major
`In humans, mifepristone
`is metabolized
`metabolites
`through successive demethylations
`of the 1 lJ3-
`dimethylaminophenyl
`group and hydroxylation of the 17a-
`propynyl moiety (Fig. 1). To date, the CYPs involved in the
`formation of these metabolites have only been investigated
`in the rat. Using
`inducers of various CYP
`isoforms,
`the
`involvement of members of the 2B, 2C, and 3A subfamilies
`was suggested [3 11, while immunoinhibition
`experiments
`in
`a subsequent work implicated a major role for CYP2Bl
`[32].
`The related human
`isoform CYP2B6
`comprises only -0.2%
`of expressed
`liver CYPs
`[33], and
`its role
`in xenobiotic
`
`$ Abbreuiadons: CYP, cytochrome P450; 17EE, 17a-erhynylestradiol; and TAO, triacetyloleandomycin.
`
`metabolism may be very limited [34]. Further studies in rat
`hepatoma variants support the ability of all three
`impli-
`cated rat subfamilies
`to catalyze mifepristone oxidations
`[35, 361. CYP-mediated
`steroid metabolism
`in the rat is
`known to involve multiple subfamilies [18], perhaps making
`this species an inappropriate model for the study of syn-
`thetic
`(or endogenous)
`steroid metabolism when extrapo-
`lations to humans are to be made.
`For the present work, a variety of approaches were em-
`ployed to determine
`the CYP primarily responsible for mife-
`pristone metabolism
`in human
`liver microsomes: chemical
`and
`immunoinhibition
`of specific
`isoforms;
`correlation
`analyses between
`initial rates of mifepristone metabolite
`formation and relative
`immunodetectable CYP
`levels and
`rates of CYP
`isoform marker substrate metabolism;
`and
`evaluation of metabolism by cDNA-expressed CYP3A4.
`
`MATERIALS AND METHODS
`Chemicals and Specimens
`
`and hydroxylated
`and its monodemethylated
`Mifepristone
`metabolites were gifts from Roussel Uclaf
`(Romainville,
`France). Didemethylated mifepristone and gestodene were
`supplied by Schering AG
`(Berlin, Germany). Midazolam
`was a gift from Hoffmann-LaRoche
`(Nutley, NJ, U.S.A.).
`7,8-Benzoflavone,
`quinidine,
`sulfinpyrazone,
`17EE, TAO,
`disulfiram, progesterone,
`deoxycorticosterone,
`NADPH,
`and sodium phosphate were purchased
`from
`the Sigma
`
`2
`
`G.
`Jang et al.
`

`

`Catalysis of Mifepristone Oxidation by CYP3A4
`
`755
`
`Furafylline was
`(St. Louis, MO, U.S.A.).
`Chemical Co.
`obtained from Research Biomedicals
`International
`(Natick,
`MA, U.S.A.). HPLC grade methanol and acetonitrile were
`from Fisher Scientific
`(Pittsburgh, PA, U.S.A.).
`Human
`liver specimens were obtained
`from organ do-
`nors, all of whom had died as a result of head trauma, under
`a protocol approved by the Committee on Human Research
`of the University of California
`at San Francisco. Micro-
`somes were prepared by homogenization
`and differential
`centrifugation,
`following established methods
`[37], of non-
`transplantable
`liver from a 53-year-old male
`(HL-Ol),
`a
`5-year-old male (HL-02) and a 36-year-old
`female (HL-03).
`The microsomes were stored until used at -80”
`in 10 mM
`Tris acetate
`(pH 7.4) containing
`1 mM EDTA and 20%
`(w/v) glycerol. Protein and CYP concentrations were de-
`termined by the Pierce bicinchoninic
`assay (Pierce Chemi-
`cal Co., Rockford,
`IL, U.S.A.)
`and Fe*’ vs Fe’*-CO differ-
`ence spectra [38], respectively.
`from 14 donors,
`The bank of human
`liver microsomes
`used for correlation analyses (designated HL-A
`through N),
`has been previously described and characterized
`for relative
`immunoreactive CYP
`levels and for initial rates of CYP
`isoform marker substrate metabolism
`[39-411. Rabbit anti-
`bodies used in immunoinhibition
`experiments were pro-
`duced as previously described
`[40]. Microsomes
`from a hu-
`man P-lymphoblastoid
`cell line stably transfected
`to coex-
`press CYP3A4
`and NADPH-CYP
`reductase were obtained
`from the Gentest Corp.
`(Wobum, MA, U.S.A.).
`
`Assay for Mifepristone and Metabolites
`
`of mifepris-
`A published HPLC assay for the determination
`tone and
`its three major metabolites
`in serum
`[42] was
`modified
`for measuring
`levels
`in microsomal
`incubations.
`Briefly,
`the mobile phase was methanol:acetonitrile:water
`(35:30:35)
`at a flow rate of 1.4 mL/min through a Beckman
`Ultrasphere C-18 column
`(5 pm x 4.6 mm i.d. x 250 mm)
`with UV monitoring
`(304 nm). The autoinjector,
`pump,
`and detector were Shimadzu models SIL-9A, LC-600,
`and
`SPD-6A,
`respectively. A Hewlett Packard 3392A
`integra-
`tor was used. Quantitation was done with extinction
`coef-
`ficients
`from authentic
`standards.
`
`Incubation Conditions
`
`incubations consisted of 60 kg microsomal pro-
`In general,
`tein (or 200 p,g protein for microsomes containing
`cDNA-
`expressed CYP3A4)
`in 0.1 M Na2HP04 buffer (pH 7.4) at
`37” with substrate
`(mifepristone or its monodemethylated
`metabolite
`in the absence or presence of inhibitors)
`added
`in methanol
`(final concentration
`~2%,
`v/v). Reactions
`were initiated by adding NADPH
`in buffer (to 1 mM, total
`volume
`200 pL) after a 5-min preincubation
`period,
`stopped after 2 min by adding a 2-fold volume of acetoni-
`trile containing deoxycorticosterone
`as internal standard,
`and vortexed. Precipitated proteins were pelleted by cen-
`
`loo-150
`
`FL of the
`
`(5 min at 11,000 g), and
`trifugation
`supernatant was subjected
`to HPLC.
`in-
`For mechanism-based
`inhibitors, catalysis-dependent
`activation was initiated by the addition of NADPH
`(using
`HL-03 microsomes) and carried out for 30 min, followed by
`lo-fold dilution of the microsomes with buffer containing
`mifepristone
`and NADPH.
`Thereafter,
`reactions were
`stopped at 2 min and samples processed as described above.
`In some experiments,
`inhibition of the second demethyl-
`ation was evaluated using the monodemethylated metabo-
`lite (synthetic
`standard) as substrate.
`various amounts of
`In immunoinhibition
`experiments,
`sera from pre-immune and immunized rabbits (to CYP2C9
`and CYP3A4) were incubated with HL-02 microsomes at
`24” for 30 min before the addition of substrate and the assay
`of catalytic activity. The antisera to CYP2C9 was found to
`be maximally
`inhibitory
`(by approximately 75%) of tolbu-
`tamide hydroxylation
`at 75 p,L/mg protein
`(data not
`shown).
`
`Data Analysis
`
`formation, substrate con-
`For characterization of metabolite
`centration was varied up to 200 PM, and kinetic parameters
`were estimated by non-linear
`regression
`analyses
`(with
`Minim 1.8a) assuming single enzyme Michaelis-Menten
`kinetics with a weighting
`factor equal to the reciprocal of
`the observed
`initial rate. No evidence of biphasic kinetics
`was observed
`in Eadie-Hofstee
`plots. Correlation
`analyses
`were performed by linear regression using a commercially
`available statistics program (Statworks 1.2). All results are
`presented as the means of duplicate determinations.
`
`RESULTS
`Kinetics of Metabolite Formation
`
`conditions were developed with HL-02.
`incubation
`Initial
`As expected
`from metabolite
`formation observed
`in ho,
`monodemethylated metabolite
`formed most quickly and ex-
`tensively;
`levels of hydroxylated
`and didemethylated
`metabolites
`remained
`lower throughout
`the observed
`in-
`cubation periods. Product
`formation was linear up to ap-
`proximately
`0.4 mg protein/ml
`and 3 min and was not
`affected by substitution of an NADPH-generating
`system or
`NADPH
`concentrations
`greater than 1 mM. Thus, a pro-
`tein concentration
`of 0.3 mg/mL, an incubation period of 2
`min, and 1 mM NADPH were used routinely for initial rate
`conditions.
`parameter es-
`Table 1 summarizes the Michaelis-Menten
`timates for mifepristone demethylation
`and hydroxylation
`in microsomes
`from HL-01, HL-02, HL-03, and P-lympho-
`blastoid cells expressing CYP3A4. For the microsomes
`from
`the three human
`livers, the mean (*SD)
`apparent K,,, and
`V max values for demethylation were 10.6
`f 3.8 p,M and
`4920 f 1340 pmol/min/mg protein, respectively;
`the corre-
`sponding values for hydroxylation were 9.9 f 3.5 PM and
`610 f 260 pmol/min/mg protein. The microsomes contain-
`
`3
`
`

`

`756
`
`G. R. Jang et al.
`
`TABLE 1. Estimated Michaelis-Menten
`parameters for mifepristone demethylation
`and hydroxylation
`in human liver microsomes 01-03 and in microsomes containing
`recombinant CYP3A4
`and NADPHZYP
`reductase*
`
`Demethylation
`
`Hydroxylation
`
`K,
`14.5
`10.3
`6.9
`7.4
`
`V,,
`3370
`5750
`5640
`1140
`
`V..,zKz,
`232
`558
`817
`154
`
`K,
`13.3
`::;
`
`4.1
`
`V,,,
`310
`800
`720
`110
`
`V-J%
`23
`81
`112
`26
`
`V-K
`ratio
`
`Demethylation/
`Hydroxylation
`
`10.1
`6.9
`7.3
`5.9
`
`HL-01
`HL-02
`HL-03
`CYP3A4
`
`*
`
`RSpf32tlV+
`
`catalyzed the two oxidations
`ing cDNA-expressed CYP3A4
`with similar apparent
`IX,,, but lower V,,,
`values. A com-
`parison of the relative ratio of V,,,/K,,,
`for the two meta-
`bolic pathways revealed a consistent 6- to lo-fold greater
`rate of elimination via demethylation.
`
`Effects of Chemical
`
`Inhibitors on Metabolite Formation
`
`(with
`inhibitors were tested
`competitive
`following
`The
`their CYP
`isoform specificities):
`sulfinpyrazone
`(CYPZC9),
`quinidine
`(CYP2D6),
`progesterone
`and midazolam
`(CYP3A4/5).
`Sulfinpyrazone and quinidine up to concen-
`trations
`of 100 p,M did not
`inhibit mifepristone
`de-
`methylation, while progesterone
`and midazolam over the
`same concentration
`range did so by 77 and 66%, respec-
`tively (Fig. 2). We also attempted
`to evaluate the effects of
`7,8-benzoflavone
`(up to 100 PM), which resulted
`in con-
`centration-dependent
`inhibition
`of metabolism
`(up
`to
`
`for
`is less selective
`78%, data not shown). This compound
`reported
`to
`inhibit
`CYPlA2
`than
`furafylline, has been
`CYP2C9
`[43], and has been found to activate or inhibit
`some CYP3A4
`reactions
`[43-45].
`Importantly, while the
`flavone more selectively and potently
`(by -90%)
`inhibits
`CYPlA2
`at low (~10 FM) concentrations
`[43], little inhi-
`bition (< 18%) was observed in our studies at these concen-
`trations
`(data not shown). The
`result
`is therefore more
`consistent with inhibition of CYP3A4
`than of CYPlA2.
`This was confirmed subsequently using furafylline
`(see be-
`low).
`(TAO)
`The effects of the following quasi-irreversible
`or mechanism-based
`inhibitors were evaluated:
`furafylline
`(CYPlA2),
`disulfiram
`(CYPs
`2A6,
`2B6,
`and 2El),
`gestodene
`(CYP3A4/5),
`TAO
`(CYP3A4/5),
`and 17EE
`(CYP3A4).
`The compounds
`specific
`to CYP3A
`enzymes
`markedly
`inhibited both demethylation
`and hydroxylation
`reactions by 70-80%
`(Fig. 3). Moreover,
`17EE and TAO
`inhibited
`the second demethylation
`to the same extent ob-
`
`h .Z .$
`Y
`3
`b
`C
`8
`#
`
`80
`
`60
`
`40
`
`20
`
`01
`0
`
`20
`
`40
`
`60
`
`80
`
`100
`
`ClM
`
`(A), quinie
`(W), progesterone
`FIG. 2. Effects of midazolam
`(A) on mifepristone demeth-
`dine (Cl), and sulfinpyrazone
`ylation. Each data point represents the mean of duplicate
`determinations. Control activity was 1950 pmoUmin/mg
`protein.
`
`0
`
`17EE (300 FM)
`
`TAO (25 FM)
`
`gestodene (20 PM)
`
`disulfiram (30 PM)
`
`furafylline (20 PM)
`
`0
`
`20 40 60 80 100 7%
`control
`activity
`concentration.dependent
`inhibition of
`FIG. 3. Maximal
`mifepristone demethylation
`and hydroxylation
`by quasi*
`irreversible and mechanism-based
`chemical
`inhibitors of
`CYP3A4/5
`(TAO and gestodene), CYP3A4
`(17EE), CYPs
`2A6,2B6,
`and 2El (disulfiram) and CYPlA2
`(furafylline).
`Each bar represents the mean of duplicate measurements.
`Control activities for demethylation and hydroxylation were
`3400 and 530 pmol/min/mg protein, respectively.
`
`4
`
`Apparent K,, V,,, and V,,,/K,,, values are expressed in FM, pmollmmjmg protem, and (ILlminlmg protem,
`hydroxylation
`demethylation
`

`

`Catalysis of Mifepristone Oxidation by CYP3A4
`
`757
`
`(data not shown). Di-
`served for the other two oxidations
`sulfiram and furafylline did not
`inhibit demethylation
`or
`hydroxylation
`significantly
`(Fig. 3). The minor (ll-17%)
`inhibition observed with these compounds was likely due to
`slight
`inhibition of CYP3A4
`at these concentrations
`[43,
`461.
`
`Immunoinhibition
`
`Experiments
`
`inhibited both mifepris-
`strongly
`to CYP3A4
`Antibodies
`tone demethylation
`(-82%)
`and hydroxylation
`(-65%),
`as
`shown in Fig. 4A. We assessed the effects of antibodies
`to
`CYP2C9 because previous work in the rat [31, 32, 35, 361
`had implicated CYP2C enzymes. These antibodies, as well
`as pre-immune sera, had no effect on either biotransforma-
`tion (Fig. 4B).
`
`Correlation Analyses with Relative CYP
`Levels and Rates of Marker Substrate Metabolism
`
`first and second demethylations
`Initial rates of mifepristone
`and hydroxylation
`in human
`liver microsomes HL-A
`through N correlated very well with relative CYP3A
`levels
`(Fig. 5A), with rates of midazolam 4-hydroxylation
`(Fig.
`5B) and l’-hydroxylation
`(Fig. 5C), with rates of erythro-
`mycin N-demethylation
`(Fig. 5D), and with each other
`(Fig. 6). The correlation analyses with rates of midazolam
`hydroxylation depicted
`in panels B and C of Fig. 5 were
`carried out excluding
`the samples known
`to contain
`CYP3A5
`in addition
`to CYP3A4
`(HL-E,
`F, and G).
`CYP3A5, which is polymorphically
`expressed in only -2O-
`30% of adult human
`livers
`[47, 481,
`is known
`to have
`marked regioselectivity
`for hydroxylation
`of midazolam at
`the
`l’-position
`relative
`to the 4-position
`[49]. For the
`analyses with midazolam
`4-hydroxylation,
`inclusion
`of
`these samples only slightly
`lowered correlation
`coefficients
`for the first and second demethylations
`and hydroxylation
`to 0.97, 0.90, and 0.93 (all P < O.OOl), respectively.
`Inclu-
`sion of the microsomal samples containing CYP3A5
`in the
`analyses with
`l’-hydroxylation
`lowered the respective co-
`efficients more noticeably
`to 0.83, 0.77, and 0.76
`(all P <
`0.001). This
`reflects
`the regioselectivity
`of CYP3A5
`for
`midazolam hydroxylations
`and an apparent
`lack of similar
`regioselectivity
`for oxidations of mifepristone.
`No significant
`correlations were observed between me-
`tabolite formation rates and relative
`immunodetectable
`lev-
`els of CYPs
`lA2,2D6,
`and 2El (r2 range 0.00 to 0.21, mean
`f SD = 0.10 f 0.08, all P > 0.05, data not shown). Addi-
`tionally, no correlations were observed between
`initial
`rates of metabolite
`formation and rates of ethoxyresorufin
`0-deethylation
`(CYPlA2),
`coumarin
`7-hydroxylation
`(CYP2A6),
`S-warfarin 7-hydroxylation
`(CYP2C9),
`S-me-
`phenytoin
`4’-hydroxylation
`(CYP2C19),
`bufuralol
`l’-
`hydroxylation
`(CYP2D6),
`and N-nitrosodimethylamine
`N-
`demethylation
`(CYPZEl)
`(r’ range 0.00
`to 0.28, mean 2
`SD = 0.11 + 0.09, all P > 0.05, data not shown).
`
`140 -
`
`120’
`
`1OOc
`
`80-
`
`60’
`
`40-
`
`20-
`
`“’ . 0 sb .
`
`100
`
`.
`
`’
`150
`
`.
`
`2lio
`
`sera
`
`(pl) / mg protein
`
`0
`
`25
`
`50
`
`75
`
`100
`
`sera
`
`(PI) / mg protein
`
`(0,O)
`FIG. 4. (A) Inhibition of mifepristone hydroxylation
`(A, A) by antibodies to CYP3A4
`(solid
`and demethylation
`symbols) and lack of inhibition by pre-immune
`IgG (open
`symbols).
`(B) Lack of inhibition by antibodies to CYP2C9
`(solid symbols) and lack of inhibition by pre-immune
`IgG
`(open symbols). Each data point is the mean of duplicate
`determinations; control activities for demethylation and hy
`droxylation were 2180 and 240 pmol/min/mg protein, re-
`spectively.
`
`We should note that correlation analyses of rates of first
`and second demethylations
`and hydroxylation with relative
`immunoreactive
`CYP2A6
`levels determined
`previously
`[39-41]
`resulted in r2 values of 0.33,0.45,
`and 0.32, respec-
`tively. The correlations
`for the two demethylations were
`significant
`(P < 0.05) but that for the hydroxylation was
`not (P = 0.06).
`It is likely that these correlations stem from
`an inherent
`reciprocity between relative
`levels of CYP3A
`and CYP2A6
`in this bank of human liver microsomes
`(r* =
`0.41, P < 0.02). With
`this in mind, any CYP3A4-catalyzed
`
`5
`
`1
`

`

`758
`
`4000
`
`3000
`
`2000
`
`1000
`
`n
`“. ~- 0
`
`r 2= 0.93
`
`G. R. Jang et al. B C
`
`loo
`
`200
`
`300
`
`400
`
`n
`“. 0
`
`loo
`
`200
`
`300
`
`400
`
`500
`
`relative CYP3A levels D
`
`rate of midazolam 4-hydroxylation (pmol/min/mg protein)
`
`r *= 0.91
`
`r*= 0.85
`
`aJ
`
`2
`
`0
`0
`
`500
`
`1000
`
`1500
`
`0
`0
`
`200
`
`400
`
`600
`
`800
`
`1000
`
`rate of midazolam rate of erythromycin l’-hydroxylation (pmol/min/mg protein) N-demethylation (pmol/min/mg protein)
`
`(A) and (A) relative
`l?rst (A) and second (0) demethylations and hydroxylation
`FIG. 5. Correlations between mifepristone
`immunodetectable CYP3A
`levels, (B and C) rates of midazolam 4. and l’-hydroxylation,
`and (D) rates of erythromycin
`N-demethylation
`in human liver microsomes
`(all P c 0.001). All initial rates are the means of duplicate determinations.
`
`to corre-
`in these microsomes would be expected
`reaction
`late weakly with CYP2A6
`levels. Indeed, similar weak but
`significant
`correlations with CYP2A6 were also observed
`for
`initial
`rates of erythromycin N-demethylation
`(r2 =
`0.46, P < 0.01) and midazolam 4-hydroxylation
`(r2 = 0.35,
`P c 0.03),
`both well established marker activities
`of
`CYP3A4.
`Furthermore,
`the weak correlations
`for mifepris-
`tone demethylations were inconsistent with the lack of cor-
`relation observed with rates of coumarin 7-hydroxylation
`(all P > 0.05 as noted above) and the lack of inhibition by
`disulfiram at concentrations
`that have been shown to in-
`hibit CYP2A6 by >70%
`[46].
`
`DISCUSSION
`
`lines of evidence were ob-
`In this work, complementary
`tained that collectively
`support CYP3A4 as the major CYP
`catalyzing mifepristone demethylations
`and hydroxylation
`in human
`liver microsomes.
`Chemical
`and
`immuno-
`inhibition of CYP3A4
`resulted in significant
`inhibition of
`mifepristone metabolism, which was further
`confirmed
`through correlation
`analyses. Furthermore,
`a recombinant
`form of CYP3A4,
`like the metabolizing microsomal
`en-
`zyme, appeared to oxidize preferentially
`to the demethylat-
`ed derivative
`(as evidenced by a higher V,_,/K,,, relative to
`
`6
`
`

`

`Catalysis of Mifepristone Oxidation by CYP3A4
`
`759
`
`0
`
`100 200 300 400 500 600
`
`04
`0
`
`1000
`
`8
`
`2Ow 3000 rate of second demethylation (pmol/min/mg protein)
`
`0
`
`loo 200 300 400 500 600 rate of hydroxylation (pmol/min/mg protein) FIG. 6.
`
`Respective correlations between initial rates of mifepristone metabolite formation
`determinations)
`in human liver microsomes
`(all P < 0.001).
`
`(reported as the means of duplicate
`
`of other
`inhibition
`Conversely,
`that of hydroxylation).
`CYP
`isoforms had no effect on mifepristone metabolism,
`which was again consistent with the results of correlation
`analyses. Thus, unlike that reported
`in the rat [31, 32, 35,
`361, enzymes of the CYPZC and CYP2B subfamilies do not
`appear to be involved
`in mifepristone metabolism
`in hu-
`mans.
`The weak but significant correlations observed between
`rates of mifepristone demethylation
`and relative
`immuno-
`reactive CYP2A6
`levels illustrate
`the need to supplement
`data from correlation analyses with that from other lines of
`experimentation. An inherent weakness in performing cor-
`relation analyses
`is the possibility of apparent but likely
`artifactual relationships
`in a particular bank of microsomes.
`A very small role of CYP2A6
`in mifepristone metabolism
`cannot be ruled out completely, but is refuted by the ob-
`served
`lack of biphasic kinetics
`(at concentrations
`up to
`50-fold
`those observed
`in ho),
`lack of correlation with
`CYP2A6 marker activity, and a lack of inhibition by disul-
`firam. Moreover,
`the very high
`levels of significance
`of
`
`levels and activity probes, inte-
`correlations with CYP3A4
`results, argue clearly for a prin-
`grated with the inhibition
`important)
`role of CYP3A4. This
`cipal (and thus clinically
`importance of evaluating
`the en-
`demonstrates
`the critical
`tire body of evidence
`to reach a conclusion about the prin-
`cipal CYP catalyzing the metabolism of a drug.
`con-
`It was reported recently
`that CYP3A7, heretofore
`sidered fetal liver specific, was detected at the protein and
`mRNA
`levels in endometrium
`(of pregnant and nonpreg-
`nant women) and placenta
`[50]. CYP3A4
`and CYP3A5
`were not detected
`in these
`tissues. CYP3A7
`shares some
`substrate specificity with CYP3A4 and is known to oxidize
`one steroid, dehydroepiandrosterone
`3-sulfate, at an appar-
`ently greater rate [34, 51, 521. The expression of CYP3A7
`in these extrahepatic
`tissues was variable, but seemed
`to
`increase during
`the menstrual
`cycle and with gestation
`length. When used as an abortifacient, mifepristone derives
`its effect primarily through antagonism of receptors
`in the
`endo- and myometrium. We hypothesize
`that instances of
`non-response
`to mifepristone when used in this capacity
`
`7
`
`

`

`target tissue
`could be related in part to CYP3A7-mediated,
`in response
`metabolism of the compound. Such differences
`to mifepristone
`could not be attributed
`to differences
`in
`drug or metabolite plasma levels or levels of al-acid glyco-
`protein (to which it is highly bound) [42]. It seems plausible
`to also suggest that this isoform may influence
`the efficacy
`of antiprogestins when they are used as contraceptives
`or
`for endometriosis.
`Given
`the numerous and promising potential uses of
`mifepristone
`(and other antiprogestins),
`the finding
`that
`CYP3A4
`is its major metabolizing enzyme in human
`liver
`suggests
`the
`likelihood
`of drug-drug
`interactions
`subse-
`quent to long-term administration of the compound. This is
`notably exemplified by the
`implications
`for its potential
`anticancer uses, since several current antineoplastic
`agents
`are also CYP3A4
`substrates. Knowledge of this, combined
`with
`its potential
`for
`inhibiting
`P-glycoprotein
`in uiuo,
`could lead to more rational and effective use of this com-
`pound.
`
`Institutes of Health
`in part, by National
`This work was supported,
`Grants GM 26691 and GM 07175 and by a fellowship from Johnson
`&? Johnson-American
`Foundation
`for Pharmaceutical Education
`(to
`G. R. J.).
`
`steroid RU486. Pharma-
`in humans. Acta Obster Gy-
`
`References
`Ulmann A, Teutsch G and Philibert D, RU 486. Sci Am 262:
`1.
`42-48,
`1990.
`0, Antiprogesterone
`Heikinheimo
`cokinetics
`and receptor binding
`necol Stand 69: 357-358,
`1990.
`Donaldson M, Dorflinger L, Brown SS and Benet LZ (Eds.),
`Clinical Applications of Mifepristone
`(RU 486) and Other Anti-
`progestins. National Academy Press, Washington, DC, 1993.
`Glasier A and Baird DT, Post-ovulatory
`contraception.
`Bail-
`lieres Clin Obstet Gynaecol 4: 283-291,
`1990.
`of Yuzpe
`Webb AMC, Russell
`J and Elstein M, Comparison
`regimen, danazol, and mifepristone
`(RU486)
`in oral post-
`coital contraception.
`Bw
`305: 927-931,
`1992.
`Batista MC, Cartledge TP, Zellmer AW, Merino MJ, Axiotis
`C, Loriaux DL and Nieman LK, Delayed endometrial matu-
`ration
`induced by daily administration
`of the antiprogestin
`RU 486: A potential new contraceptive
`strategy. Am J Obstet
`Gynecol 167: 60-65,
`1992.
`of endo-
`Interruption
`Greene KE, Kettel LM and Yen SSC,
`metrial maturation without hormonal changes by an antipro-
`gesterone during the first half of luteal phase of the menstrual
`cycle: A contraceptive
`potential.
`Fertil St&l 58: 338-343,
`1992.
`Schneider MR, Michna H, Nishino Y and el Etreby MF,
`Antitumor
`activity and mechanism of action of different an-
`tiprogestins
`in experimental
`breast cancer models.
`.I Steroid
`Biochem Mol Biol 37: 783-787,
`1990.
`Bakker GH, Setyono-Han
`B, Portengen H, De Jong FH, Foe-
`kens JA and Klijn
`JG, Treatment
`of breast cancer with dif-
`ferent antiprogestins:
`Preclinical
`and clinical
`studies. _I Steroid
`Biochem Mel Biol 37: 789-794,
`1990.
`Lin MF, Kawachi MH, Stallcup MR, Grunberg SM and Lin
`FF, Growth
`inhibition
`of androgen-insensitive
`human pros-
`tate carcinoma
`cells by a 19-norsteroid derivative agent, mife-
`pristone. Prostate 26: 194-204,
`1995.
`J, Sadun A,
`Grunberg
`SM, Weiss MH, Spitz
`IM, Ahmadi
`Russell CA, Lucci L and Stevenson LL, Treatment
`of unre-
`
`2.
`
`3.
`
`4.
`
`5.
`
`6.
`
`7.
`
`8.
`
`9.
`
`10.
`
`11.
`
`12.
`
`13.
`
`15.
`
`16.
`
`17.
`
`18.
`
`19.
`
`20.
`
`21.
`
`22.
`
`23.
`
`24.
`
`25.
`
`26.
`
`27.
`
`28.
`
`sectable meningiomas with the antiprogesterone
`pristone.
`.J Neurosurg 74: 861-866,
`1991.
`Lamberts SW, Tanghe HL, Avezaat CJ, Braakman R, Wijn-
`gaarde R, Koper JW and de Jong H, Mifepristone
`(RU 486)
`treatment of meningiomas.
`J Neural Neurosurg Psychiatry 55:
`486-490,
`1992.
`Kettel LM, Murphy AA, Morales AJ and Yen SS, Clinical
`efficacy of the antiprogesterone
`RU486
`in the treatment
`of
`endometriosis
`and uterine
`fibroids. Hum Reprod 9(Suppl
`1):
`116-120,
`1994.
`14.
`Lecureur V, Fardel 0 and Guillouzo A, The antiprogestatin
`drug RU 486 potentiates
`doxorubicin
`cytotoxicity
`in multi-
`drug resistant cells through
`inhibition
`of P-glycoprotein
`func-
`tion. FEBS Lett 355:
`187-191,
`1994.
`Gruel DJ, Zee MC, Trotter
`J and Bourgeois S, Reversal of
`multidrug resistance by RU 486. Cancer Res 54: 3088-3091,
`1994.
`van der Lely A, Foeken K, van der Mast R and Lamberts SW,
`Rapid reversal of acute psychosis
`in the Cushing
`syndrome
`with the cortisol-receptor
`antagonist mifepristone
`(RU 486).
`Ann Intern Med 114: 143-144,
`1991.
`FP and Lapenson DP,
`Waxman DJ, Attisano C, Guengerich
`Human liver microsomal
`steroid metabolism:
`Identification
`of
`the major microsomal
`steroid hormone 6P-hydroxylase
`cyto-
`chrome P-450 enzyme. Arch Biochem Biophys 263: 424-436,
`1988.
`Waxman DJ, Lapenson DP, Aoyama T, Gelboin HV, Gonza-
`lez FJ and Korzekwa K, Steroid hormone hydroxylase
`speci-
`ficities of eleven cDNA-expressed
`human cytochrome P45Os.
`Arch Biochem Biophys 290:
`160-166,
`1991.
`I’, Wolff T
`Guengerich
`FP, Martin MV, Beaune PH, Kremers
`and Waxman DJ, Characterization
`of rat and human
`liver
`microsomal
`cytochrome
`P-450
`forms
`involved
`in nifedipine
`oxidation,
`a prototype
`for genetic polymorphism
`in oxidative
`drug metabolism.
`.I Biol Chem 261: 5051-5060,
`1986.
`Kronbach T, Fischer V and Meyer UA, Cyclosporine metabo-
`lism in human
`liver: Identification
`of a cytochrome P-450111
`gene family as the major cyclosporine-metabolizing
`enzyme
`explains
`interactions
`of cyclosporine with other drugs. Clin
`Pharmacol Ther 43: 630-635,
`1988.
`Sattler M, Guengerich
`FP, Yun C-H, Christians U and Sew-
`ing K-F, Cytochrome
`P-450 3A enzymes are responsible
`for
`biotransformation
`of FK506 and rapamycin
`in man and rat.
`Drug Metab Dispos 20: 753-761,
`1992.
`Kronbach T, Mathys D, Umeno M, Gonzalez FJ and Meyer
`UA, Oxidation
`of midazolam and triazolam by human
`liver
`cytochrome P450IIIA4. Mel Pharmacol 36: 89-96,
`1989.
`Imaoka S, Enomoto K, Oda Y, Asada A, Fujimo