Research Paper
`
`Mechanism of cytochrome P450-3A inhibition
`by ketoconazolejphp_1202
`
`214..221
`
`David J. Greenblatta, Yanli Zhaoa, Karthik Venkatakrishnanb,
`Su X. Duana, Jerold S. Harmatza, Sarah J. Parenta, Michael H. Courta
`and Lisa L. von Moltkec
`
`aDepartment of Pharmacology and Experimental Therapeutics, Tufts University School of Medicine and
`Tufts Medical Center, Boston, bMillennium Pharmaceuticals Inc. and cGenzyme Corp., Cambridge, MA,
`USA
`
`Abstract
`Objectives Ketoconazole is extensively used as an index inhibitor of cytochrome P450-3A
`(CYP3A) activity in vitro and in vivo, but the mechanism of ketoconazole inhibition of
`CYP3A still is not clearly established.
`Methods Inhibition of metabolite formation by ketoconazole (seven concentrations from
`0.01 to 1.0 mm) was studied in human liver microsomes (n = 4) at six to seven substrate
`concentrations for triazolam, midazolam, and testosterone, and at two substrate concentra-
`tions for nifedipine.
`Key findings Analysis of multiple data points per liver sample based on a mixed
`competitive–noncompetitive model yielded mean inhibition constant Ki values in the range
`of 0.011 to 0.045 mm. Ketoconazole IC50 increased at higher substrate concentrations,
`thereby excluding pure noncompetitive inhibition. For triazolam, testosterone, and mida-
`zolam a-hydroxylation, mean values of a (indicating the ‘mix’ of competitive and noncom-
`petitive inhibition)
`ranged from 2.1 to 6.3. However,
`inhibition of midazolam
`4-hydroxylation was consistent with a competitive process. Determination of Ki and a based
`on the relation between 50% inhibitory concentration values and substrate concentration
`yielded similar values. Pre-incubation of ketoconazole with microsomes before addition of
`substrate did not enhance inhibition, whereas inhibition by troleandomycin was significantly
`enhanced by pre-incubation.
`Conclusions Ketoconazole inhibition of triazolam a- and 4-hydroxylation, midazolam
`a-hydroxylation, testosterone 6b-hydroxylation, and nifedipine oxidation appeared to be a
`mixed competitive–noncompetitive process, with the noncompetitive component being
`dominant but not exclusive. Quantitative estimates of Ki were in the low nanomolar range for
`all four substrates.
`Keywords competitive inhibition; cytochrome P450-3A; ketoconazole; mechanism-based
`inhibition; noncompetitive inhibition
`
`Introduction
`Ketoconazole is extensively used as an ‘index’ inhibitor of human cytochrome P450-3A
`(CYP3A) isoforms in the academic, regulatory, and drug development communities.[1–3]
`Although ketoconazole has been used in this context for some twenty years, the kinetic
`mechanism of CYP3A inhibition by ketoconazole in human liver microsomal preparations
`still is not clearly established. Published in-vitro studies of ketoconazole inhibition of
`CYP3A have yielded differing and conflicting conclusions as to whether inhibition can be
`explained by competitive, noncompetitive, or mixed competitive–noncompetitive mecha-
`nisms.[4] Accurate assignment of the mechanism of inhibition affects understanding of the
`fundamental nature of the substrate–enzyme and inhibitor–enzyme interactions, as well as
`the interpretation of in-vitro inhibition constants (Ki), the validity of which is dependent on
`the correctness of the underlying model of inhibition.[5]
`We have evaluated the mechanism of CYP3A inhibition by ketoconazole in human liver
`microsomal preparations. To our knowledge this is among the few studies to assess this
`question using CYP3A substrates from different categories, with a sufficient number of
`
`JPP 2011, 63: 214–221
`© 2011 The Authors
`JPP © 2011 Royal
`Pharmaceutical Society
`Received July 27, 2010
`Accepted September 9, 2010
`DOI
`10.1111/j.2042-7158.2010.01202.x
`ISSN 0022-3573
`
`Correspondence: David J.
`Greenblatt, Department of
`Pharmacology and,
`Experimental Therapeutics, Tufts
`University School of Medicine,
`136 Harrison Ave, Boston, MA
`02111, USA.
`E-mail: dj.greenblatt@tufts.edu
`
`214
`
`1
`
`TEVA1024
`
`

`

`Ketoconazole inhibition of CYP3A
`
`David J. Greenblatt et al.
`
`215
`
`substrate and inhibitor concentrations such that enzyme
`kinetic analysis could proceed without a predetermined deci-
`sion as to whether the inhibition mechanism was competitive,
`noncompetitive, or mixed.[6]
`
`Materials and Methods
`Preparation of human liver microsomes
`Liver samples from four individual human donors with no
`known liver disease were provided by the International Insti-
`tute for the Advancement of Medicine (Exton, PA, USA); the
`Liver Tissue Procurement and Distribution System, Univer-
`sity of Minnesota (Minneapolis, MN, USA); or the National
`Disease Research Interchange (Philadelphia, PA, USA). The
`ages of the four donors were: 21, 34, 37, and 62 years. All
`were male and the cause of death was either physical trauma
`or sudden illness.
`ultracentrifugation;
`by
`prepared
`Microsomes were
`microsomal pellets were suspended in 0.1 m potassium phos-
`phate buffer containing 20% glycerol and were stored at
`-80°C until use.[7]
`
`Preparation of reagents
`incubations and
`Chemical reagents used for microsomal
`HPLC analysis were purchased from commercial sources.
`Triazolam, midazolam,
`testosterone, nifedipine, ketocona-
`zole, and troleandomycin (triacetyloleandomycin) were pur-
`chased from commercial sources or kindly provided by their
`pharmaceutical manufacturers.
`Stock solutions of the drug entities were prepared in
`methanol, and subsequently diluted with methanol as needed.
`Solutions were stored at -18°C.
`
`In-vitro study design
`Four CYP3A substrates were selected for study: triazolam,
`midazolam, testosterone and nifedipine.[8,9] The choice of sub-
`strates was based on categories proposed by Kenworthy
`triazolam a-hydroxylation and
`studies of
`et al.[6] For
`4-hydroxylation, midazolam a-hydroxylation
`and
`4-
`hydroxylation, and testosterone 6b-hydroxylation,
`incuba-
`tions were performed at multiple substrate concentrations. At
`each concentration, additional incubations were done with
`co-addition of ketoconazole at concentrations of 0, 0.01,
`0.025, 0.05, 0.1, 0.25, 0.5 or 1.0 mm. Within this range keto-
`conazole is established as a relatively specific CYP3A inhibi-
`tor. Studies were done with microsomal preparations from
`
`four different human liver samples and all individual incuba-
`tions were performed in duplicate.
`Nifedipine dehydrogenation studies were done at two fixed
`concentrations of 25 and 250 mm. Incubations were performed
`with co-addition of varying concentrations of ketoconazole as
`described above.
`For testosterone as substrate, an additional study was
`conducted to evaluate the effect of 20 min pre-incubation of
`microsomes with ketoconazole, prior to the addition of test-
`osterone (100 mm) on the inhibitory effect of ketoconazole at
`concentrations of 0.05 and 0.25 mm. Troleandomycin was
`used as the positive control in this study. Troleandomycin
`concentrations were 0, 1, 5, 10, 25, and 50 mm.
`
`Incubation procedures
`Incubation mixtures contained 50 mm phosphate buffer, 5 mm
`MgCl2, 0.5 mm nicotinamide adenine dinucleotide phosphate,
`and an isocitrate and isocitric dehydrogenase regenerating
`system.[7,8] Substrates (Table 1) and ketoconazole or trolean-
`domycin, as appropriate, were added to a series of incubation
`tubes. The solvent was evaporated to dryness at 40°C under
`mild vacuum conditions.
`After addition of the incubation mixture components to
`yield a final volume of 250 ml, reactions were initiated by the
`addition of microsomal protein (0.1–0.25 mg/ml). The incu-
`bation duration ranged from 5 to 40 min, depending on the
`intrinsic clearance of the substrate.[8,10–12] All reactions were
`shown to be in the linear range with respect to protein con-
`centration and incubation duration.
`At the appropriate time reactions were stopped by cooling
`on ice and the addition of 100 ml acetonitrile. Internal stan-
`dard was added, the incubation mixture was centrifuged, and
`the supernatant was transferred to an autosampling vial for
`HPLC analysis of metabolites (Table 1). The analytic column
`was stainless steel, 15 cm ¥ 3.9 mm, containing reverse-phase
`C18 Novapak (Waters Associates, Milford, MA, USA)
`(Table 1).
`
`Analysis of data
`For triazolam, midazolam, and testosterone, analyses were
`performed for each individual liver sample using combina-
`tions of reaction velocity (V) at each substrate concentration
`([S]) and inhibitor concentration ([I]). The following equa-
`tion, consistent with mixed competitive–noncompetitive inhi-
`bition, was fitted to the data points by nonlinear regression
`using SAS PROC NLIN:[13–15]
`
`Table 1 Summary of reaction characteristics and chromatographic conditions
`
`Substrate
`
`Product(s)
`
`Internal standard
`
`Mobile phase composition
`
`Ultraviolet absorbance
`wavelength
`
`Triazolam
`
`Midazolam
`
`Testosterone
`Nifedipine
`
`a-OH-triazolam,
`4-OH-triazolam
`a-OH-midazolam,
`4-OH-midazolam
`6b-OH-testosterone
`Oxidized nifedipine
`
`Phenacetin
`
`Phenacetin
`
`Androstenedione
`Diazepam
`
`aPotassium dihydrogen phosphate (pH unadjusted).
`
`70% buffera (10 mm), 20% acetonitrile,
`10% methanol
`45% buffera (10 mm), 20% acetonitrile,
`35% methanol
`55% water, 45% acetonitrile
`45% water, 55% methanol
`
`220 nm
`
`220 nm
`
`240 nm
`270 nm
`
`2
`
`

`

`216
`
`Journal of Pharmacy and Pharmacology 2011; 63: 214–221
`
`a∞ 5
`
`0
`
`20
`
`10
`
`5
`
`12
`
`100
`
`50
`
`20
`
`10
`
`C50/Ki
`
`5I
`
`2 1
`
`0.1 0.2
`
`0.5
`
`1
`
`2
`
`5
`[S]/Km
`
`10
`
`20
`
`50 100
`
`Figure 1 For a mixed competitive–noncompetitive inhibitor, the rela-
`tionship between the [S]/Km ratio (x-axis) and the IC50/Ki ratio, at varying
`values of a ranging from 1.0 to ‘infinity’. The functions were predicted
`based on equation 4.
`
`⋅[ ]
`max S
`+
`
`(1)
`
`⎞⎠⎟
`
`+ [ ]
`I
`K
`
`i
`
`1
`
`⎛⎝⎜
`
`m
`
`K
`
`⎞⎠⎟
`
`i
`
`V K
`
`[ ] + [ ]
`I
`
`⋅α
`
`1
`
`⎛⎝⎜
`
`S
`
`=
`
`V
`
`Iterated variables were: Vmax, the maximum reaction veloc-
`ity; Km, the Michaels–Menten constant; Ki, the inhibition con-
`stant; and a, a number ⱖ 1.0 indicating the ‘mix’ of
`competitive and noncompetitive mechanisms. In the case of
`testosterone, a sigmoidal (Hill) kinetic pattern was observed
`without and with the addition of ketoconazole. For purposes
`of nonlinear regression analysis of testosterone data, equa-
`tion 1 was empirically modified by addition of an exponent to
`[S] and to S50 (replacing Km).
`For determination of 50% inhibitory concentration (IC50)
`values, reaction velocities with co-addition of inhibitor were
`expressed as a percentage ratio (Rv) of the control velocity
`with no inhibitor present. The relationship of Rv to inhibitor
`concentration ([I]) was analysed by nonlinear regression
`to determine the IC50 value, based on the following
`equation:
`
`=
`
`R
`v
`
`100 1
`
`−
`
`max
`b
`
`b
`
`b
`
`(2)
`
`For triazolam, midazolam and testosterone, ketoconazole
`IC50 values were available at multiple substrate concentra-
`tions. Using mean Km (or S50) values determined as described
`above (eqn 1) as fixed entries in equation 4, this equation was
`fitted to data points ([S] and IC50) by nonlinear regression.
`Iterated variables were a and Ki. For nifedipine, the same
`procedure was applied using the available data points, with
`the Km value from a previously reported study.[8]
`
`Results
`Nonlinear regression based on equation 1 (modified by an
`exponent in the case of testosterone) yielded convergent solu-
`tions (Figures 2 and 3).
`For triazolam, Km values for the 4-OH-triazolam pathway
`the a-OH-triazolam pathway
`for
`exceeded Km values
`(Table 2). This was similar to previous studies from our labo-
`ratory and elsewhere. [8,10,11,17] Ketoconazole inhibition of tria-
`zolam metabolite formation was consistent with the mixed
`competitive–noncompetitive scheme (Figure 2). The mean
`values of a were 2.11 and 2.14 for the a-hydroxylation and
`4-hydroxylation pathways, respectively, indicating a predomi-
`nance of noncompetitive inhibition in the balance.
`Consistent with previous reports, Km values for the mida-
`zolam a-hydroxylation pathway were on average more than
`tenfold lower than for the 4-hydroxylation pathway (Table 2,
`of a-OH-
`inhibition
`Figure 3).[8,11,12,18,19] Ketoconazole
`midazolam formation was explained by the mixed model,
`with a mean a of 6.30. However, inhibition of midazolam
`4-hydroxylation was consistent with a competitive process,
`inasmuch as estimates of a from nonlinear regression were
`very large.
`testosterone 6b-hydroxylation
`As described previously,
`a
`sigmoidal
`(Hill)
`kinetic
`was
`best
`described
`by
`pattern.[5,8,20–25] Inhibition by ketoconazole was consistent with
`the mixed competitive–noncompetitive model
`(Figure 2).
`
`⎞ ⎠⎟
`
`[ ]
`I
`E
`[ ] +
`I
`IC
`
`⎛ ⎝⎜
`
`Iterated variables were: Emax, the maximum degree of inhi-
`bition; IC, the inhibitor concentration producing an Rv value
`of 50% of (100 – Emax); and b, an exponent. The actual IC50
`was calculated as:
`
`IC50
`
`=
`
`IC
`
`(
`
`2
`
`E
`
`max
`
`−
`
`)
`1 1
`
`b
`
`(3)
`
`IC50 values were determined based on mean data
`points across four liver samples at corresponding inhibitor
`concentrations.
`The relationship between IC50 and substrate concentration
`can be used as an alternative method for determining a and Ki
`for mixed competitive–noncompetitive inhibition.[16] That
`relationship is:
`
`(4)
`
`⎞⎠⎟
`
`⎞⎠⎟
`
`+ [ ]
`1
`+ [ ]
`
`⋅α
`
`⎛⎝⎜
`
`K
`
`i
`
`1
`
`⎛⎝⎜
`
`IC
`
`50
`
`=
`
`S K S K
`
`m
`
`m
`
`This is illustrated in Figure 1. Note that for pure competi-
`tive inhibition (a = ‘infinity’) this reduces to:
`+ [ ]
`
`⎞⎠⎟
`
`S K
`
`m
`
`1
`
`⎛⎝⎜
`
`IC
`
`50
`
`=
`
`K
`
`i
`
`(5)
`
`whereas for pure noncompetitive inhibition (a = 1.0), this
`reduces to:
`
`50 = K
`
`IC
`
`i
`
`(6)
`
`3
`
`

`

`Ketoconazole inhibition of CYP3A
`
`David J. Greenblatt et al.
`
`217
`
`Keto (mM)
`
`0 0
`
`.01
`
`0.025
`
`0.05
`
`0.1
`
`0.25
`0.5
`1.0
`
`Vmax = 3.36
`S50 = 70.5 mM
`Ki = 0.025 mM
`a = 3.15
`
`(b)
`
`3.5
`
`3.0
`
`2.5
`
`2.0
`
`1.5
`
`1.0
`
`0.5
`
`(relative velocity units)
`
`6b-OH-testosterone fromation rate
`
`Keto (mM)
`
`0 0
`
`.01
`
`0.025
`
`0.05
`
`0.1
`
`0.25
`0.5
`1.0
`
`Vmax = 1.02
`Km = 72 mM
`Ki = 0.032 mM
`a = 2.06
`
`(a)
`
`1.0
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
`(relative velocity units)
`
`a-OH-triazolam formation rate
`
`0
`
`100
`
`300
`200
`Triazolam (mM)
`
`400
`
`500
`
`0
`
`100
`
`300
`200
`Testosterone (mM)
`
`400
`
`500
`
`Figure 2 Rates of formation of a-OH-triazolam from triazolam (a) and formation of 6b-OH-testosterone from testosterone (b) in a human liver
`sample. Lines represent the functions determined by nonlinear regression based on equation 1. For testosterone, equation 1 was empirically modified
`by addition of an exponent to account for sigmoidal kinetics.
`
`(a)
`
`0.35
`
`0.30
`
`Vmax = 0.296
`Km = 3.38 mM
`Ki = 0.007 mM
`a = 7.16
`
`(b)
`
`0.21
`
`Vmax = 0.287
`Km = 48.8 mM
`Ki = 0.0175 mM
`
`Keto (mM)
`
`Keto (mM)
`
`0 0
`
`.01
`
`0.025
`
`0.05
`
`0.1
`
`0.25
`0.5
`1.0
`
`0
`
`20
`
`60
`40
`Midazolam (mM)
`
`80
`
`100
`
`0.18
`
`0.15
`
`0.12
`
`0.09
`
`0.06
`
`0.03
`
`(relative velocity units)
`
`4-OH-midazolam formation rate
`
`0 0
`
`.01
`
`0.025
`
`0.05
`
`0.1
`
`0.25
`0.5
`1.0
`
`0
`
`20
`
`60
`40
`Midazolam (mM)
`
`80
`
`100
`
`0.25
`
`0.20
`
`0.15
`
`0.10
`
`0.05
`
`(relative velocity units)
`
`a-OH-midazolam formation rate
`
`Figure 3 Rates of formation of a-OH-midazolam (a) and 4-OH-midazolam (b) from midazolam in a human liver sample. Lines represent the
`functions determined by nonlinear regression based on equation 1.
`
`Table 2 Enzyme kinetic parameters for in-vitro biotransformation of cytochrome P450-3A substrates with co-addition of ketoconazole as inhibitor
`
`Substrate
`
`Product
`
`Triazolam
`
`Midazolam
`
`Testosterone
`Nifedipine
`
`a-OH-triazolam
`4-OH-triazolam
`a-OH-midazolam
`4-OH-midazolam
`6b-OH-testosterone
`Oxidized nifedipine
`
`Km (mm)
`
`63.1 (⫾ 3.8)
`319 (⫾ 45)
`3.34 (⫾ 0.15)
`51.3 (⫾ 8.2)
`69 (⫾ 7)a
`–
`
`Based on equation 1 (mean ⫾ SE, n = 4)
`a
`Ki (mm)
`
`Based on equation 4
`a
`
`Ki
`
`0.031 (⫾ 0.003)
`0.045 (⫾ 0.005)
`0.011 (⫾ 0.004)
`0.019 (⫾ 0.003)
`0.019 (⫾ 0.002)
`–
`
`2.14 (⫾ 0.13)
`2.11 (⫾ 0.43)
`6.30 (⫾ 1.77)
`b
`3.07 (⫾ 0.23)
`–
`
`0.041
`0.052
`0.0083
`0.019
`0.028
`0.072
`
`1.51
`1.41
`8.54
`b
`
`2.11
`4.25
`
`aS50. bValues of a for 4-OH-midazolam formation were very large, consistent with competitive inhibition.
`
`4
`
`

`

`218
`
`Journal of Pharmacy and Pharmacology 2011; 63: 214–221
`
`TST = 50 mM (IC50 = 0.030 mM)
`TST = 500 mM (IC50 = 0.056 mM)
`
`100
`
`80
`
`60
`
`40
`
`20
`
`(% control without inhibitor)
`
`6b-OH-testosterone fromation rate
`
`Triazolam = 50 mM (IC50 = 0.040 mM)
`Triazolam = 500 mM (IC50 = 0.065 mM)
`
`0.01 0.02
`
`0.2
`0.1
`0.05
`Ketoconazole (mM)
`
`0.5
`
`1
`
`0.01 0.02
`
`0.2
`0.1
`0.05
`Ketoconazole (mM)
`
`0.5
`
`1
`
`Nifedipine = 25 mM (IC50 = 0.11 mM)
`Nifedipine = 250 mM (IC50 = 0.30 mM)
`
`100
`
`80
`
`60
`
`40
`
`20
`
`0.01
`
`0.02
`
`0.2
`0.1
`0.05
`Ketoconazole (mM)
`
`0.5
`
`1
`
`(% control without inhibitor)
`
`Oxidized nifedipine formation rate
`
`Midazolam = 10 mM (IC50 = 0.092 mM)
`Midazolam = 100 mM (IC50 = 0.185 mM)
`
`0.01 0.02
`
`0.2
`0.1
`0.05
`Ketoconazole (mM)
`
`0.5
`
`1
`
`100
`
`80
`
`60
`
`40
`
`20
`
`(% control without inhibitor)
`a-OH-triazolam formation rate
`
`100
`
`80
`
`60
`
`40
`
`20
`
`(% control without inhibitor)
`
`a-OH-midazolam formation rate
`
`Figure 4 Inhibition curves for ketoconazole vs four different substrates at low and high concentrations. Points are mean (⫾ SE, n = 4) percentage
`ratios of reaction velocity with coaddition of ketoconazole divided by the control velocity without ketoconazole. Lines are the functions consistent with
`equation 2 determined by nonlinear regression. TST, testosterone.
`
`Values of a fell between 2.7 and 3.7 (mean 3.07), indicating
`the predominance of noncompetitive inhibition in the mixed
`model (Table 2).
`For all four CYP3A substrates, ketoconazole IC50 values
`increased at higher concentrations of substrate (Figure 4).
`Analysis of the relation between IC50 and [S] based on equa-
`tion 4 yielded estimates of Ki and a that were similar to those
`determined from the analysis using equation 1 (Table 2). For
`4-OH-midazolam formation, equation 4 again yielded very
`large estimates of a, indicating competitive inhibition. Keto-
`conazole inhibition of nifedipine oxidation was explained via
`equation 4 as a mixed competitive–noncompetitive process,
`with Ki = 0.072 mm and a = 4.25.
`Inhibition of testosterone 6b-hydroxylation by troleando-
`mycin was clearly enhanced by pre-incubation (Figure 5),
`consistent with mechanism-based inhibition.[26–30] However,
`inhibition
`by
`ketoconazole was
`not
`enhanced
`by
`pre-incubation (Figure 5),
`confirming reversible
`rather
`than mechanism-based
`inactivation
`of CYP3A by
`ketoconazole.[30–33]
`
`Discussion
`
`While ketoconazole is generally recognized as a strong inhibi-
`tor of human CYP3A isoforms, numerous studies published
`over the last two decades have indicated extensive variability
`in quantitative Ki values for ketoconazole vs different CYP3A
`substrates, as well as differences among studies of the same
`substrate.[1,2,4] Ketoconazole Ki may also differ even for two
`parallel metabolic pathways for the same substrate such as
`alprazolam, triazolam, midazolam and terfenadine.[10–12,34,35]
`Quantitative determination of Ki values from in-vitro
`studies is dependent on the assumed or established underlying
`mechanism of inhibition. Any given set of data points will
`yield a different estimate of Ki if the mechanism of inhibition
`is assumed to be competitive or noncompetitive. Therefore it
`is probable that some component of the variability among
`studies in Ki values is explained by differing assignments of
`inhibitory mechanism.
`reporting in-vitro Ki
`Among 51 published studies
`values for ketoconazole inhibition of CYP3A substrate
`
`5
`
`

`

`Ketoconazole inhibition of CYP3A
`
`David J. Greenblatt et al.
`
`219
`
`Without pre-incubation
`With pre-incubation
`
`(b)
`
`40
`
`35
`
`30
`
`25
`
`20
`
`15
`
`10
`
`5
`
`(% control without inhibitor)
`
`6b-OH-testosterone fromation rate
`
`Without pre-incubation
`
`With pre-incubation
` (IC50 = 4.6 mM)
`
`1
`
`2
`
`10
`5
`Troleandomycin (mM)
`
`20
`
`50
`
`Keto = 0.05 mM
`
`Keto = 0.25 mM
`
`(a)
`120
`
`100
`
`80
`
`60
`
`40
`
`20
`
`0
`
`(% control without inhibitor)
`
`6b-OH-testosterone fromation rate
`
`Figure 5 (a) Effect of troleandomycin on the rate of 6b-OH-testosterone formation from testosterone. (b) Mean percentage ratios of reaction
`velocities with coaddition of ketoconazole divided by the control velocity without ketoconazole. For (a): points are mean (⫾ SE, n = 4) percentage
`ratios of reaction velocities with coaddition of troleandomycin divided by the control velocity without troleandomycin. Studies were done without and
`with pre-incubation of troleandomycin with microsomes before addition of the substrate. Dashed line is the function consistent with equation 2
`determined by nonlinear regression. For (b): mean ⫾ SE, n = 4. Ketoconazole was used at 0.05 or 0.25 mm. Studies were done without and with
`pre-incubation of microsomes with ketoconazole.
`
`biotransformation, the majority assigned a fully competitive
`mechanism.[4] In some of these cases, the competitive process
`was assumed or ‘forced’ a priori, without an evaluation of
`whether that mechanism was most appropriate for the data. In
`other studies, the choice of competitive, noncompetitive or
`mixed inhibition was based on the appearance of data when
`plotted after some form of linearizing transformation (typi-
`cally a Dixon plot of the reciprocal of reaction velocity vs
`inhibitor concentration at different substrate concentrations).
`These transformations may provide useful
`information
`(including estimates of Ki), but may be misleading or distor-
`tive since they involve calculations of reciprocals. Data points
`having the smallest numerical values – and therefore subject
`to greater measurement inaccuracy – become inappropriately
`magnified in importance. Small experimental errors in such
`points are thereby expanded, and may bias slopes and inter-
`cepts of reciprocal plots.
`This study evaluated the kinetics of ketoconazole inhibi-
`tion of human CYP3A in vitro, using four different substrates
`to assess the inhibitory mechanism. The pre-incubation study
`with testosterone as substrate verified that inhibition by keto-
`conazole was reversible as opposed to mechanism-based,
`as
`reported previously from our own laboratory and
`elsewhere.[30–33] The finding of increased IC50 values at higher
`concentrations of all four substrates (Figure 4) effectively
`excluded purely noncompetitive inhibition as the underlying
`mechanism, since the IC50 value would be unchanged (and
`equal to Ki) regardless of substrate concentration (eqn 6). For
`triazolam, midazolam, and testosterone, nonlinear regression
`of all data points allowed simultaneous determination of Ki
`and of the ‘mix’ of competitive and noncompetitive processes,
`with no a priori assumptions about the mechanism, and
`without the potential biases of reciprocal transformations. The
`outcome of the analysis demonstrated mixed-mechanism
`inhibition for
`triazolam,
`testosterone
`and midazolam
`a-hydroxylation. The mean values of a were in the range of
`
`2–7, and the Ki values for ketoconazole were in the low
`nanomolar range. It is of interest that the Ki values vs the
`4-OH-triazolam pathway were higher than for the a-OH-
`triazolam pathway. This confirmed our previous observations
`and emphasized that Ki values for a given inhibitor of CYP3A
`may differ even between two parallel metabolic pathways for
`the same substrate.[10,11] Finally, we observed that ketocona-
`zole inhibition of midazolam 4-hydroxylation, as opposed to
`a-hydroxylation, was consistent with a competitive process.
`These phenomena might be explained by cooperativity and
`multisite kinetic behaviour of human CYP3A.[5,20,23,25,36]
`The relation between the IC50 value and substrate concen-
`tration provided an alternative approach to determination of
`the Ki and a-values. Since the IC50 value increased at
`higher substrate concentrations, pure noncompetitive inhibi-
`triazolam a-hydroxylation and
`tion was excluded. For
`testosterone 6b-hydroxylation and mida-
`4-hydroxylation,
`zolam a-hydroxylation, application of equation 4 indicated
`mixed-mechanism inhibition, with a predominance of the
`noncompetitive component. The numerical values of Ki and a
`were similar to those determined from equation 1. Application
`of equation 4 to midazolam 4-hydroxylation indicated com-
`petitive inhibition, as was found using equation 1. Therefore,
`two independent approaches to data analysis yielded similar
`or identical conclusions regarding the mechanism of inhibi-
`tion by ketoconazole, as well as the numerical values of Ki.
`The four CYP3A substrates used in this study were based
`on the hypothetical categories originally proposed by Ken-
`worthy et al.[6] The conclusion from that study was that probe
`substrates could be divided into two major groups, one of
`which included triazolam and midazolam, and another which
`included testosterone. Nifedipine was described as not falling
`into either of the discrete groups. Those authors made the
`point that there was a high degree of consistency among
`CYP3A substrates in terms of extensive inhibition caused by
`established strong CYP3A inhibitors, and minimal inhibition
`
`6
`
`

`

`220
`
`Journal of Pharmacy and Pharmacology 2011; 63: 214–221
`
`caused by weak inhibitors. We have reported a very high
`degree of correlation of IC50 values for 22 different CYP3A
`inhibitors using midazolam and testosterone as CYP3A sub-
`strates (see Figure 2.2 of Volak et al.[15]) based on data pub-
`lished by Obach et al.[37] A similar point has been made in
`other publications.[20,38,39] In this study, ketoconazole was dem-
`onstrated to be a strong inhibitor of CYP3A – with Ki or IC50
`values in the low nanomolar range – regardless of the hypo-
`thetical category of the substrate. Although current Food and
`Drug Administration guidelines recommend the use of more
`than one CYP3A substrate for in-vitro studies of CYP3A
`inhibition, our findings suggested that a single CYP3A index
`substrate may be sufficient for in-vitro characterization of a
`potential CYP3A inhibitor.[40] There appeared to be low prob-
`ability of acquiring unique information from study of one or
`more additional CYP3A substrates.
`
`Conclusions
`Although ketoconazole is a strong CYP3A inhibitor regard-
`less of the specific substrate, there are nonetheless variations
`among substrates in the quantitative potency of ketoconazole
`inhibition even under standardized in-vitro study conditions.
`The mechanism of inhibition, though reversible, in general
`cannot be characterized as fully competitive or fully noncom-
`petitive. This study has suggested that a mixed competitive–
`noncompetitive mechanism was likely to be applicable. The
`noncompetitive component appeared dominant
`in most
`instances, but the quantitative ‘mix’ of competitive and non-
`competitive processes will vary among substrates and even
`between parallel biotransformation pathways for the same
`substrate.
`
`Declarations
`Conflict of interest
`The Author(s) declare(s) that they have no conflicts of interest
`to disclose.
`
`Funding
`This research received no specific grant from any funding
`agency in the public, commercial, or not-for-profit sectors.
`
`References
`
`1. Venkatakrishnan K et al. Effects of the antifungal agents on
`oxidative drug metabolism in humans: clinical relevance. Clin
`Pharmacokinet 2000; 38: 111–180.
`2. Greenblatt DJ et al. The CYP3 family. In: Ioannides C, ed.
`Cytochrome P450: Role in the Metabolism and Toxicology of
`Drugs and Other Xenobiotics. Cambridge, UK: Royal Society of
`Chemistry, 2008; 354–383.
`3. Greenblatt DJ, von Moltke LL. Clinical studies of drug-drug
`interactions: design and interpretation. In: Pang KS et al., eds.
`Enzyme and Transporter-Based Drug-Drug Interactions:
`Progress and Future Challenges. New York: Springer, 2010;
`625–649.
`4. Greenblatt DJ et al. Sources of variability in ketoconazole inhi-
`bition of human cytochrome P450 3A in vitro. Xenobiotica 2010;
`40: 713–720.
`
`5. Houston JB, Galetin A. Modelling atypical CYP3A4 kinetics:
`principles and pragmatism. Arch Biochem Biophys 2005; 433:
`351–360.
`6. Kenworthy KE et al. CYP3A4 drug interactions: correlation of
`10 in vitro probe substrates. Br J Clin Pharmacol 1999; 48:
`716–727.
`7. von Moltke LL et al. Alprazolam metabolism in vitro: studies of
`human, monkey, mouse, and rat liver microsomes. Pharmacol-
`ogy 1993; 47: 268–276.
`8. Patki KC et al. In vitro metabolism of midazolam, triazolam,
`nifedipine, and testosterone by human liver microsomes and
`recombinant cytochromes P450: role of CYP3A4 and CYP3A5.
`Drug Metab Dispos 2003; 31: 938–944.
`9. Galetin A et al. Multisite kinetic analysis of interactions between
`prototypical CYP3A4 subgroup substrates: midazolam,
`test-
`osterone, and nifedipine. Drug Metab Dispos 2003; 31: 1108–
`1116.
`10. von Moltke LL et al. Triazolam biotransformation by human
`liver microsomes in vitro: effects of metabolic inhibitors, and
`clinical confirmation of a predicted interaction with ketocona-
`zole. J Pharmacol Exp Ther 1996; 276: 370–379.
`11. Perloff MD et al. Midazolam and triazolam biotransformation in
`mouse and human liver microsomes: relative contribution of
`CYP3A and CYP2C isoforms. J Pharmacol Exp Ther 2000; 292:
`618–628.
`12. von Moltke LL et al. Midazolam hydroxylation by human liver
`microsomes in vitro: inhibition by fluoxetine, norfluoxetine, and
`by azole antifungal agents. J Clin Pharmacol 1996; 36: 783–791.
`13. Segel IH. Enzyme Kinetics. New York: John Wiley and Sons,
`1975.
`14. Venkatakrishnan K et al. Human drug metabolism and the cyto-
`chromes P450: application and relevance of in vitro models.
`J Clin Pharmacol 2001; 41: 1149–1179.
`15. Volak LP et al. In vitro approaches to anticipating clinical drug
`interactions. In: Li AP, ed. Drug-Drug Interactions in Pharma-
`ceutical Development. Hoboken, NJ: John Wiley & Sons, 2008;
`75–93.
`16. Brown HS et al. Use of isolated hepatocyte preparations for
`cytochrome
`P450
`inhibition
`studies:
`comparison with
`microsomes for Ki determination. Drug Metab Dispos 2007; 35:
`2119–2126.
`17. Schrag ML, Wienkers LC. Triazolam substrate inhibition: evi-
`dence of competition for heme-bound reactive oxygen within the
`CYP3A4 active site. Drug Metab Dispos 2001; 29: 70–75.
`18. Ghosal A et al. Inhibition and kinetics of cytochrome P4503A
`activity in microsomes from rat, human, and cDNA-expressed
`human cytochrome P450. Drug Metab Dispos 1996; 24: 940–
`947.
`19. Gorski JC et al. Regioselective biotransformation of midazolam
`by members of the human Cytochrome P450 3A (CYP3A) sub-
`family. Biochem Pharmacol 1994; 47: 1643–1653.
`20. Galetin A et al. CYP3A4 substrate selection and substitution in
`the prediction of potential drug-drug interactions. J Pharmacol
`Exp Ther 2005; 314: 180–190.
`21. Hosea NA et al. Elucidation of distinct ligand binding sites for
`cytochrome P450 3A4. Biochemistry 2000; 39: 5929–5939.
`22. Lin Y et al. Substrate inhibition kinetics for cytochrome P450-
`catalyzed reactions. Drug Metab Dispos 2001; 29: 368–374.
`23. Kenworthy KE et al. Multisite kinetic models for CYP3A4:
`simultaneous activation and inhibition of diazepam and testoster-
`one metabolism. Drug Metab Dispos 2001; 29: 1644–1651.
`24. Harlow GR, Halpert JR. Analysis of human cytochrome P450
`3A4 cooperativity: construction and characterization of a site-
`directed mutant that displays hyperbolic steroid hydroxylation
`kinetics. Proc Natl Acad Sci USA 1998; 95: 6636–6641.
`
`7
`
`

`

`Ketoconazole inhibition of CYP3A
`
`David J. Greenblatt et al.
`
`221
`
`25. Ueng Y-F et al. Cooperativity in oxidations catalyzed by cyto-
`chrome P450 3A4. Biochemistry 1997; 36: 370–381.
`26. Obach RS et al. Mechanism-based inactivation of human cyto-
`chrome P450 enzymes and the prediction of drug-drug interac-
`tions. Drug Metab Dispos 2007; 35: 246–255.
`27. Obach RS et al. Inactivation of human Cytochrome P450
`enzymes and drug-drug interactions. In: Pang KS et al., eds.
`Enzyme- and Transporter-Based Drug-Drug Interactions:
`Progress and Future Challenges. New York: Springer, 2010;
`473–492.
`28. Bertelsen KM et al. Apparent mechanism-based inhibition
`of human CYP2D6 in vitro by paroxetine: comparison with
`fluoxetine and quinidine. Drug Metab Dispos 2003; 31: 289–
`293.
`29. Fontana E et al. Cytochrome P450 enzymes mechanism based
`inhibitors: common sub-structures and reactivity. Curr Drug
`Metab 2005; 6: 413–454.
`30. von Moltke LL et al. Potent mechanism-based inhibition of
`human CYP3A in vitro by amprenavir and ritonavir: comparison
`with ketoconazole. Eur J Clin Pharmacol 2000; 56: 259–261.
`31. Weemhoff JL et al. Apparent mechanism-based inhibition of
`human CYP3A in-vitro by lopinavir. J Pharm Pharmacol 2003;
`55: 381–386.
`32. Perloff ES et al. Atazanavir: effects on P-glycoprotein transport
`and CYP3A metabolism in vitro. Drug Metab Dispos 2005; 33:
`764–770.
`
`33. Sekiguchi N et al. Prediction of drug-drug interactions based on
`time-dependent inhibition from high throughput screening of
`cytochrome P450 3A4 inhibition. Drug Metab Pharmacokinet
`2009; 24: 500–510.
`34. von Moltke LL et al. Inhibitors of alprazolam metabolism in vitro:
`effect of serotonin-reuptake-inhibitor antidepressants, ketocona-
`zole and quinidine. Br J Clin Pharmacol 1994; 38: 23–31.
`35. von Moltke LL et al. In vitro prediction of the terfenadine-
`ketoconazole pharmacokinetic interaction. J Clin Pharmacol
`1994; 34: 1222–1227.
`36. Niwa T et al. Heterotropic cooperativity in oxidation mediated
`by Cytochrome P450. Curr Drug Metab 2008; 9: 453–462.
`37.