Genotype-phenotype associations for common CYP3A4 and CYP3A5 variants in the basal and induced metabolism of midazolam in European- and African-American men and women.

Abstract

CYP3A activity in adults varies between individuals and it has been suggested that this has a genetic basis, possibly related to variant alleles in CYP3A4 and CYP3A5 genes. Accordingly, genotype-phenotype associations were investigated under constitutive and induced conditions. Midazolam's systemic and oral clearances, and the erythromycin breath test (ERBT) were determined in 57 healthy subjects: 23 (11 men, 12 women) European- and 34 (14 men, 20 women) African-Americans. Studies were undertaken in the basal state and after 14-15 days pretreatment with rifampin. DNA was characterized for the common polymorphisms CYP3A4*1B, CYP3A5*3, CYP3A5*6 and CYP3A5*7 by direct sequencing, and for exon 21 and exon 26 variants of MDR1 by allele-specific, real-time polymerase chain reaction. In 95% of subjects, the basal systemic clearance of midazolam was unimodally distributed and variability was less than four-fold whereas, in 98% of the study population, oral clearance varied five-fold. No population or sex-related differences were apparent. Similar findings were observed with the ERBT. Rifampin pretreatment markedly increased the systemic (two-fold) and oral clearance (16-fold) of midazolam, and the ERBT (two-fold) but the variabilities were unchanged. No associations were noted between these phenotypic measures and any of the studied genotypes, except for oral clearance and its fold-increase after rifampin. These were related to the presence of CYP3A4*1B and the inversely linked CYP3A5*3 polymorphism, with the extent of induction being approximately 50% greater in CYP3A5*3 homozygotes compared to wild-type subjects. In most healthy subjects, variability in intestinal and hepatic CYP3A activity, using midazolam as an in-vivo probe, is modest and common polymorphisms in CYP3A4 and CYP3A5 do not appear to have important functional significance.

Full text

Page 1

Genotype–phenotype associations for common CYP3A4 and
CYP3A5 variants in the basal and induced metabolism of
midazolam in European- and African-American men and
women
Michael D. Floyd?, Guillermo Gervasini?, Andrew L. Masica?, Gail Mayo,
Alfred L. George Jr, Kolari Bhat, Richard B. Kim and Grant R. Wilkinson
CYP3A activity in adults varies between individuals and it
has been suggested that this has a genetic basis, possibly
related to variant alleles in CYP3A4 and CYP3A5 genes.
Accordingly, genotype–phenotype associations were
investigated under constitutive and induced conditions.
Midazolam’s systemic and oral clearances, and the
erythromycin breath test (ERBT) were determined in 57
healthy subjects: 23 (11 men, 12 women) European- and
34 (14 men, 20 women) African-Americans. Studies were
undertaken in the basal state and after 14–15 days
pretreatment with rifampin. DNA was characterized for the
common polymorphisms CYP3A4?1B, CYP3A5?3,
CYP3A5?6 and CYP3A5?7 by direct sequencing, and for
exon 21 and exon 26 variants of MDR1 by allele-specific,
real-time polymerase chain reaction. In 95% of subjects,
the basal systemic clearance of midazolam was
unimodally distributed and variability was less than four-
fold whereas, in 98% of the study population, oral
clearance varied five-fold. No population or sex-related
differences were apparent. Similar findings were observed
with the ERBT. Rifampin pretreatment markedly increased
the systemic (two-fold) and oral clearance (16-fold) of
midazolam, and the ERBT (two-fold) but the variabilities
were unchanged. No associations were noted between
these phenotypic measures and any of the studied
genotypes, except for oral clearance and its fold-increase
after rifampin. These were related to the presence of
CYP3A4?1B and the inversely linked CYP3A5?3
polymorphism, with the extent of induction being
approximately 50% greater in CYP3A5?3 homozygotes
compared to wild-type subjects. In most healthy subjects,
variability in intestinal and hepatic CYP3A activity, using
midazolam as an in-vivo probe, is modest and common
polymorphisms in CYP3A4 and CYP3A5 do not appear to
have important functional significance. Pharmacogenetics
13:595–606 & & 2003 Lippincott Williams & Wilkins
Pharmacogenetics 2003, 13:595–606
Keywords: CYP3A, midazolam, polymorphisms, CYP3A4?1B, CYP3A5?3,
CYP3A5?6, CYP3A5?7
Divisions of Clinical Pharmacology and Genetic Medicine, Departments of
Pharmacology and Medicine, Vanderbilt University School of Medicine, Nashville,
Tennessee, USA.
Sponsorship: This work was supported in part by USPHS grants GM-31304 and
RR-00095.
Correspondence and requests for reprints to Grant R. Wilkinson, Division of
Clinical Pharmacology, Vanderbilt University School of Medicine, Nashville, TN
37232-6602, USA.
Tel: +1 615 322 3033; fax: +1 615 343 6334;
e-mail: grant.wilkinson@vanderbilt.edu
Received 11 April 2003
Accepted 24 June 2003
Introduction
Enzymes of the CYP3A subfamily are involved in the
metabolism of numerous drugs and a number of endo-
genous compounds, and four members have been
described in humans (http://www.imm.ki.se/CYPalleles/).
CYP3A4 is the major form present in the liver and the
intestinal epithelium and, along with CYP3A5, which is
polymorphically found in these organs as well as some
other tissues, essentially accounts for the collective
catalytic activity of this subfamily in adults. This is
because CYP3A43 does not appear to be present to any
significant extent [1,2], and CYP3A7 is primarily a fetal
enzyme [3] that has markedly less catalytic activity
than either CYP3A4 or CYP3A5 [4].
CYP3A-mediated metabolism is generally characterized
as havingconsiderable
which, in part, reflects its ready modulation by inhibi-
tion and induction, including that associated with
drug–drug interactions and other environmental agents
[5]. A genetic factor(s) is also thought to be constitu-
tively involved [6] although, by contrast to the activity
of several other CYPs [7], that of CYP3A is distributed
in a unimodal fashion [8]. This has led to the search
for, and identification of, single nucleotide polymorph-
isms (SNPs) in the coding and non-coding regions of
both CYP3A4 and CYP3A5 (http://www.imm.ki.se/
CYPalleles/). Most of these appear to be uncommon
(frequency , 1–2%) and/or selectively distributed to
specific populations. However, with CYP3A4, a ?392
A.G transition in the so-called nifedipine specific
inter-individual variability
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
?These individuals contributed equally to the research.
Original article 595
0960-314X & 2003 Lippincott Williams & WilkinsDOI: 10.1097/01.fpc.0000054118.14659.48

Page 2

element (NFSE) of the 59-regulatory region is more
frequently found [9–10], and some evidence suggests
that this SNP may have functional consequences
[9,11–13] (although not all data is supportive of this
notion) [10,14–17]. With CYP3A5, an A.G substitution
within intron 3 (CYP3A5?3C) has been identified and
this is also present in two additional haplotypes
(CYPA5?3A and?3B). Collectively, these are termed
CYP3A5?3 because they all result in cryptic splice sites
leading to a small amount of correctly spliced mRNA
encoding wild-type (CYP3A5.1) enzyme and a variant
forming a truncated protein which is functionally
inactive [18]. This variant is thought to be the genetic
basis for the polymorphic expression of CYP3A5. A
rarer G.A transition in exon 7 (CYP3A5?6) also results
in a splicing defect, and an insertion in exon 11
(CYP3A?7) leads to an early stop codon; therefore,
these variant alleles are null [18]. On the basis of in-
vitro findings, it has been suggested that such alleles,
especially CYP3A5?3, may contribute to the observed
inter-individual variability in overall CYP3A activity in
vivo [18,19]. This is because, in general, CYP3A4 and
CYP3A5havesimilarcatalytic
although the activity of CYP3A5 is often less than that
of CYP3A4 [4], the amount and relative contribution of
this enzyme may be greater than previously thought
[18,19]. If this is the case, it would be expected that
individuals having at least one CYP3A5?3, CYP3A5?6
or CYP3A5?7 allele (heterozygous or homozygous for
the variant) would metabolize CYP3A substrates more
slowly than those with two CYP3A5?1 alleles (homo-
zygous wild-type).
specificitiesand,
A major goal of the present study was to determine the
in-vivo importance of these genetic variants of CYP3A4
and CYP3A5 using midazolam as an in-vivo phenotyp-
ing probe of constitutive CYP3A activity. Because such
activity is readily inducible by, for example, rifampin,
whereas that of CYP3A5 is essentially not [20–22], an
additional objective was to determine any genotype–
phenotype associations related to induction. Because
the level of MDR1 activity has been shown to be a
determinant of in-vitro CYP3A induction [23], potential
genetic relationships were also extended to allelic
variants of this efflux membrane transporter that possi-
bly have functional consequences [24–26].
Methods
Subjects
A total of 57 healthy, non-smoking subjects, aged 18–
55 years andweighing 52–123 kg,
Twenty-three of the subjects were European-American
(11 men) and 34 were African-American (14 men)
residing in middle Tennessee. All study subjects were
non-smokers and were considered healthy on the basis
of medical history, complete physical examination and
routine laboratory tests, including those indicative of
were studied.
hepatic and renal function. Alcohol, medications other
than oral contraceptives, and grapefruit products were
not permitted for 5 days before the study period. The
protocol was approved by the Vanderbilt University
Institutional Review Board-Health Sciences and written
informed consent was obtained from all participants.
Study design
Following an overnight fast, subjects were admitted to
the Vanderbilt General Clinical Research Center (day
1). Midazolam (Versed, Roche Laboratories, Basel,
Switzerland), 2 mg oral syrup (2 mg/ml), and 1 mg intra-
venous [3N15]-midazolam given over 15–30 s were
administered, simultaneously. The subject’s blood pres-
sure, respiratory rate and oxygen saturation were mon-
itored throughout the first 2 h. Blood samples (5 ml)
were collected into ethylenediaminetetraacetic acid
(EDTA)-containing tubes through an intravenous cath-
eter, before and at 5, 15, 30 min and at 1, 2, 3, 4, 5, 6
and 8 h after drug administration. A standard lunch was
served 4–5 h after midazolam dosing, and subjects were
discharged after the final blood sample was collected.
Plasma harvested from blood was stored at ?20 8C until
analysed.
One day after receiving midazolam, the subject returned
to the Clinical Research Center (day 2) for determina-
tion of the erythromycin breath test (ERBT). This
involved the rapid intravenous injection of 3 ?Ci [14C-
N-methyl]-erythromycin (PerkinElmer Life Sciences,
Inc., Boston, Massachusetts, USA) dissolved in 3 ml of
5% dextrose solution. A breath sample was obtained at
10, 20, 30, 40, 50 and 60 min after injection according to
a previously described procedure [27].
For the subsequent 16 days, the subject took oral
rifampin (600 mg) at breakfast time. On day 15, the
study performed originally on day 1 was repeated,
except that the oral dose of midazolam was increased to
25 mg to obtain plasma concentrations similar to those
on day 1. On the following day (day 16), the ERBT
study was repeated.
Sample analysis
Midazolam and its 19-hydroxy metabolite concentrations
in plasma were determined using high-performance
liquid chromatography–tandem mass spectrometry (LC-
MS/MS), as previously described [13]. The amount of
radioactive CO2in the expired air following administra-
tion of labelled erythromycin was measured by a liquid
scintillation counting-based methodology [27].
Pharmacokinetic analysis
The areas under the plasma concentration–time curve
(AUC0–inf) for midazolam were determined by standard
analysis, for both routes of administration, with the
terminal elimination rate constant (k) being estimated
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
596
Pharmacogenetics
2003, Vol 13 No 10

Page 3

by log-linear regression of the last 4–6 data points. The
elimination half-life was obtained from the relationship
t1
clearance values of midazolam were estimated by
statistical moment analysis (equations 1 and 2):
2¼ 0.693/k. The oral (CLO) and intravenous (CLS)
CLS ¼ doseiv=AUCiv
(1)
CLO ¼ CLS=F ¼ doseO=AUCoral
(2)
Genotyping
DNA was isolated from a pre-study blood sample
(10 ml) obtained from each subject using the Qiamp
system (Qiagen Inc., Chatsworth, California, USA).
Direct sequencing was used for the CYP3A genotyping.
Polymerase chain reaction (PCR) was performed with
approximately 30 ng genomic DNA, 0.2 mmol/l of each
deoxyribonucleoside triphosphate, 2.5 U of Amplitaq
DNA polymerase (Perkin Elmer, Branchburg, New
Jersey, USA) and 25 ?mol/l of the specific primer pair
in a total reaction volume of 50 ?l PCR buffer, 10 mmol
TrisHCl, pH 9.2, 3.5 mmol MgCl2, 25 mmol KCl. The
PCR primers were designed to gene regions having the
least similarity to those in other CYP genes at compar-
able locations to the regions of interest (Table 1). DNA
amplification involved an initial denaturation at 958C
for 5 min, followed by 35 cycles of 958C for 15 s,
annealing at 608C for 20 s, and extension at 728C for
40 s. Amplified DNA was purified using the QinQuick
DNA purification system and sequenced using Big
Dye-terminator chemistry (Applied Biosystems, Foster
City, California, USA) and an ABI3700 automated
DNA sequencer (Applied Biosystems). Sequencing was
performed on both strands of the PCR products.
In addition to the 57 subjects in the present study, data
were available from a number of European- and
African-American men who had been previously geno-
typed (n ¼ 15 each group) for CYP3A4?1B [13]. The
subjects in this earlier study all met the same inclusion/
exclusion criteria for the current study and were
investigated using an identical clinical protocol with
midazolam in the basal (day 1) state. Accordingly, their
data was incorporated into the CYP3A4?1B genotype–
phenotype association analysis.
Allele-specific, real-time PCR amplification was em-
ployed to determine the exon 21 and exon 26 geno-
types in the MDR1 gene. Oligonucleotide primers
(Table 2) were designed to include an internal nucleo-
tide mismatch near the 39-terminus of each allele-
specific primer (MWG Biotech Inc., High Point, North
Carolina, USA). Amplicon lengths were 216 bp for
2677G.T or A and 134 bp for 3435C.T. All reactions
were carried out in a total volume of 25 ?l containing a
1:2 dilution of iQTM SYBR Green Supermix (Bio-Rad
Laboratories, Inc., Hercules, California, USA), 160 nmol
of each primer, and 13 ng of genomic DNA that had
been previously digested with the restriction endonu-
clease Xho1 (New England Biolabs, Inc., Beverly,
Massachusetts, USA). Real-time PCR was performed
using an iCycler Thermal Cycler with Optical Module
and analyses were performed using iCycler iQ software
(Bio-Rad Laboratories, Inc.). The amplification program
for 2677G.T or A consisted of one cycle of 958C with
a 4-min hold followed by 27 cycles of 958C with a 30-s
hold and 608C annealing temperature with a 30-s hold.
For 3435C.T, amplification consisted of one cycle of
958C with a 4-min hold followed by 30 cycles of 958C
with a 30-s hold and 588C annealing temperature with a
30-s hold. After amplification, melt-curve analysis was
performed by heating the reaction mixture from 588C
to 988C at the rate of 0.48C/s. Positive controls using
DNA previously determined to be homo- and hetero-
zygous by sequencing and a negative control without
DNA template was run with every assay to assess the
overall specificity.
Statistical analysis
Comparisons of various pharmacokinetic values be-
tween populations and subgroups were analysed para-
metrically using the unpaired Student’s t-test and
ANOVA with Tukey’s multiple comparison test (Prism
3, Graphpad Software, Inc., San Diego, California,
USA). ANOVA along with Scheffe’s test was used to
analyse the genotype–phenotype associations (SPSS,
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Table 1
and CYP3A5 genes
DNA primers used to identify allelic variants in CYP3A4
AmpliconSNPPrimer sequence (forward/reverse)
3A4?1B
?1B
59-TGGGATGAATTTCAAGTATTT TG-39
59-AGGTTTCCATGGCCAAGTCT-39
59-CCTGAGTAACTCACCAGCCCT-39
59-CTCAAGCAACTCACCTGACGA-39
59-GAGTGGCATAGGAGATACCCAC-39
59-TCTAGTTCATTAGGGTGTTGTGACAC-39
59-TACAGCATGGATGTGATTACTG-39
59-AAAGAGAGAAAGAAATAATAGCC-39
59-AAATACTTCACGAATACTATGATCA-39
3A5?3B
H30Y (?3B),
3710-insG (?3B)
?3C
3A5?3C
3A5?6
?6
3A5-exon
11
?7
59-CAGGGACATAATTGATTATCTTTG-39
Table 2
chain reaction amplification of exon 21 and exon 26 SNPs of the
MDR1 gene
Oligonucleotide primers used for real-time polymerase
Polymorphism Primer sequence
Exon 21
2677G
2677T
2677A
2677 Common
Exon 26
3435C
3435T
3435 Common
59-AGTTTGACTCACCTTCCCATC-39
59-AGTTTGACTCACCTTCCCATA-39
59-AGTTTGACTCACCTTCCCATT-39
59-GCTATAGGTTCCAGGCTTGCT-39
59-GTGGTGTCACAGGAAGAGAAC-39
59-GTGGTGTCACAGGAAGAGAAT-39
59-ACTATAGGCCAGAGAGGCTGC-39
CYP3A genotype–phenotype associations Floyd et al.
597

Page 4

version 11.5 for Windows, SPSS Inc., Chicago, Illinois,
USA). Based on previous studies using an identical
study protocol and analytical methodology in Eur-
opean-Americans [13,28], who were assumed to be
homozygous for CYP3A4?1A and CYP3A5?3, sample
sizes of 16 and 22 were estimated to have 80% and 90%
power at Æ ¼ 0.05, respectively, to detect an inter-
genotypic difference of 25% in the systemic clearance
of midazolam and 35% in its oral clearance. Correlations
between parameters were examined by unweighted
linear regression. In all tests, P , 0.05 was considered
to be statistically significant.
Results
The systemic clearance of midazolam before rifampin
pretreatment (day 1) in the 57 study subjects exhibited
an approximately five-fold range (Table 3), which was
somewhat skewed towards higher values (Fig. 1).
Significantly, however, 95% (54/57) of the clearance
values varied by less than four-fold (150–521 ml/min)
(Fig. 1). Excluding the three outliers with values great-
er than 3 SD from the mean of the major subpopulation
(three European women), significant differences in
systemic clearance were not present according to popu-
lation (European- versus African-American) or sex,
whether analysed independently or dependently of
population.
Following oral administration, the clearance of midazo-
lam was approximately three-fold higher than after
intravenous dosing (Table 3) and, with the exception of
one subject (an African-American woman) exhibiting
the highest value, inter-individual variability was five-
fold (Fig. 1). Again, after exclusion of the single outlier,
no population or sex-related differences were apparent
in the oral clearance of midazolam, except in women.
The oral clearance of midazolam in African-American
women was statistically smaller than those of European
origin (964 ? 243 versus 1231 ? 434 ml/min, P ¼ 0.035)
even though they were heavier (71.9 ? 13.8 versus
62.1 ? 5.6 kg, P ¼ 0.018). Only approximately one-third
of the orally administered dose of midazolam was
systemically bioavailable and this value (F) varied 4.5-
fold (Table 3).
Linear relationships were observed between the esti-
mated AUC for midazolam and its measured plasma
concentration at 4 h following both intravenous and oral
administration (Fig. 2). After dose-normalization, the
ratio of the two 4-h plasma levels (oral/intravenous) was
also significantly related to the estimated oral bioavail-
ability of midazolam (F) based on conventional AUC
determination (equations 1 and 2) [F ¼ 0.01 + (0.74 3
ratio), r ¼ 0.96, P , 0.0001 for the day 1 and day 15
combined data].
Approximately 1–3% of administered radioactivity was
excreted in the breath during the first 60 min following
intravenous administration of radio-labelled erythromy-
cin (Table 3). The ERBT value was lower in African-
Americans as a group and in men relative to comparable
European-Americans (1.59 ? 0.54 versus 2.07 ? 0.58%
dose, P , 0.004 and 1.44 ? 0.41 versus 2.10 ? 0.71,
P , 0.008, respectively). However, no other differences
were noted by population or sex. Additionally, no linear
relationship was present between the systemic clear-
ance of midazolam and the ERBT value (r ¼ 0.19,
P ¼ 0.15).
Pretreatment with rifampin for 2 weeks increased the
oral clearance of midazolam to a far greater extent
(15.8-fold, range 7.1–38.4) than the drug’s systemic
clearance (1.9-fold, range 1.1–3.6); the ERBT value
also increased (2.1-fold, range 1.1–4.7) (Table 3). As on
day 1, African-American women had a lower oral
clearance (20 942 ? 7811 versus 14 148 ? 5552 ml/min,
P ¼ 0.009). Although the mean change in the systemic
clearance of midazolam was similar to that of the
ERBT value, no linear relationship was present be-
tween these two measures (r ¼ 0.21, P ¼ 0.135). The
extent of change in the clearance values of midazolam
and the ERBT appeared to be inversely related to the
parameter’s baseline value; however, with the excep-
tion of the ERBT (r ¼ 0.68; P , 0.0001), the correla-
tions were weak (r ¼ 0.32–0.46).
Genotyping for CYP3A4?1A/CYP3A4?1B, CYP3A5?1/
CYP3A5?3/CYP3A5?6 was successful in all cases and,
for CYP3A5?7 and the determined SNPs in exons 21
and 26 of MDR1, all but three to four subjects were
satisfactorily characterized (Table 4). Population differ-
ences in allele frequencies (q) were present. For
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Table 3
subjects [23 European- (11 men, 12 women) and 34 African-
Americans (14 men, 20 women)]
Pharmacokinetic parameters of midazolam in 57 health
Mean ? SDRange
Basal (day 1)
Systemic clearance
(ml/min)
(ml/min/kg)
Oral clearance
(ml/min)
(ml/min/kg)
Oral bioavailability (F), %
1-h ERBT, % dose
Rifampin-induced (days 15/16)
Systemic clearance
(ml/min)
(ml/min/kg)
Oral clearance
(ml/min)
(ml/min/kg)
Oral bioavailability (F), %
1-h ERBT, % dose
386 ? 105
5.54 ? 2.04
150–793
2.61–15.20
1142 ? 485
16.64 ? 8.91
36.5 ? 9.3
1.79 ? 0.68
436–3373
6.07–64.8
12.9–60.4
0.77–3.10
709 ? 195
10.08 ? 3.25
197–1119
3.42–20.58
17 058 ? 7220
243.7 ? 108.9
4.92 ? 2.8
3.59 ? 1.03
2667–39 167
41.3–479.4
1.7–18.7
1.41–5.78
598
Pharmacogenetics
2003, Vol 13 No 10

Page 5

example, with CYP3A4?1B, the frequency of the var-
iant was 4% in European-Americans but 66% in Afri-
can-Americans. By contrast, the allele frequency of
CYP3A5?3 was higher (87%) in European-Americans
than in African-Americans (28%); all subjects were
homozygous wild-type for C3705 and the 3709G inser-
tion variant in the CYP3A5?3B haplotype. The CYP
3A?6 allele was present relatively infrequently and
present only in African-Americans (16%). Similarly, the
2713 insT variant in CYP3A5?7 was rare, only being
found in four heterozygous African-Americans. With
MDR1, the 2677T variant as in exon 21 was present in
45.6% of European-Americans but only 9.7% in Afri-
can-Americans; no 2677A variant was found. In the case
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
95% of population: 4-fold variability
98% of population: 5-fold variability
Oral clearance
n ? 57
400030002000 1000
Midazolam clearance, ml.min?1
0
0
5
10
15
20
Number of subjects
Midazolam clearance, ml.min?1
800 700600 5004003002001000
0
5
10
15
20
Number of subjects
Systemic clearance
n ? 57
(a)(b)
Fig. 1
Frequency distributions of midazolam clearance following (a) intravenous and (b) oral administration in 23 European- and 34 African-Americans.
Oral administration (2 mg)
9876543210
4-hr midazolam plasma level, ng.ml?1
4-hr midazolam plasma level, ng.ml?1
0
10
20
30
40
50
60
70
80
90
100
Oral AUC, ng.ml?1. hr
109876543210
0
25
50
75
100
125
150
Intravenous AUC, ng.ml?1.hr
Intravenous administration (1 mg)
(a)(b)
Fig. 2
(a) Relationship between the 4-h plasma level and AUC of midazolam following 1 mg intravenously (4-h level ¼ 9.18 + 11.00 AUC, r ¼ 0.950,
P , 0.0001); pre- (m) and post- (j) treatments with rifampin on day 1 and day 15 are combined. (b) Similar relationship after 2 mg oral midazolam
in the basal (day 1) state (4-h level ¼ 9.70 + 8.18 AUC, r ¼ 0.914, P , 0.0001).
CYP3A genotype–phenotype associations Floyd et al.
599

Page 6

of exon 26, the allele frequency of the 3435T variant
was 47.8% and 6.3% in European- and African-Amer-
icans, respectively. With the exception of CYP3A5?6,
the distribution of genotypes was consistent with
Hardy–Weinberg equilibrium. Because of these allelic
frequency differences, the various genotypes were dis-
proportionately distributed in the two populations: the
homozygous variant groups for CYP3A4?1B and CYP-
3A5?6 being African-Americans and that for CYP3A5?3
being essentially European-American in origin, with
the reverse distribution for the homozygous wild-type
genotypes (Table 4).
Because no consistent sex- or population-related differ-
ences in the systemic clearances of midazolam were
apparent, investigation of possible genotype–pheno-
type associations was based on all the combined data
from both populations. No associations were apparent
in that similar basal systemic and oral clearance values
and their distributions were observed in homozygous
wild-type, heterozygous and homozygous variant sub-
groups regardless of the CYP3A allele of interest (Table
5, Fig. 3). The rarity of CYP3A5?7 precluded formal
genotype–phenotype analysis, but the data from the
four heterozygote individuals did not appear to be
different from those of the homozygotes (Table 5). In
the case of CYP3A4?1B, this lack of association was
present regardless of whether the previously collected
data [13] was included or not. Furthermore, none of the
various CYP3A genotypes appeared to be associated
with the ERBT value (Table 5, Fig. 3). A similar
finding was noted with the two MDR1 genotypes (data
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Table 4
according to population (values in parenthesis indicate number
reported elsewhere [13])
Number of genotypically characterized subjects
Gene
European-
AmericansAfrican-AmericansTotal
CYP3A4?1B
A/A
A/G
G/G
CYP3A5?3
A/A
A/G
G/G
CYP3A5?6
G/G
G/A
A/A
CYP3A5?7
wt/wt
wt/mut
mut/mut
MDR1 exon 21
G/G
G/T
T/T
MDR1 exon 26
C/C
C/T
T/T
21 (15)
2 (0)
0 (0)
4 (1)
15 (4)
15 (10)
41
21
25
0
6
16
17
16
23
18171
23 2649
0
0
5
3
5
3
232750
0
0
4
0
4
0
5 2631
19 154
134
5 19
10
24
2414
437
Table 5
Phenotypic values for the pharmacokinetics of midazolam disposition and the 1-h erythromycin breath test (ERBT) value according to CYP3A genotype
CYP3A4?1B
CYP3A5?3
CYP3A5?6
CYP3A5?7
Genotype
AA
AG
GG
AA
AG
GG
GG
GG
AA
wt/wt
wt/mut
Number of subjects
41
21
25
16
23
18
50
3
4
50
4
Baseline/control
Systemic clearance (ml/min)
417 ? 123
374 ? 75
347 ? 91
353 ? 89
375 ? 68
430 ? 142
395 ? 101
430 ? 142
339 ? 37
391 ? 107
379 ? 68
Oral clearance (ml/min)
1177 ? 591
1119 ? 334
1110 ? 459
1162 ? 443
1048 ? 317
1244 ? 670
1157 ? 507
1244 ? 670
1040 ? 289
1144 ? 489
1370 ? 514
ERBT (% dose)
2.02 ? 0.68
1.68 ? 0.61
1.58 ? 0.63
1.60 ? 0.49
1.74 ? 0.57
2.13 ? 0.73
1.86 ? 0.64
2.13 ? 0.73
1.86 ? 0.53
1.86 ? 0.62
1.30 ? 0.020
Rifampin-induced
Systemic clearance (ml/min)
764 ? 188
710 ? 164
618 ? 216
625 ? 176
709 ? 185
786 ? 185
739 ? 176
786 ? 185
528 ? 94
709 ? 193
614 ? 66
Systemic clearance (fold-increase)
1.93 ? 0.62
1.93 ? 0.44
1.79 ? 0.59
1.78 ? 0.47
1.93 ? 0.54
1.97 ? 0.66
1.95 ? 0.57
1.97 ? 0.66
1.55 ? 0.19
1.87 ? 0.53
1.67 ? 0.37
Oral clearance (l/min)
18.86 ? 6.76
16.73 ? 6.76
14.43 ? 7.12
14.38 ? 6.99
16.38 ? 6.85
20.30 ? 7.03þ17.83 ? 6.85
20.30 ? 7.03þ11.09 ? 5.46
17.18 ? 7.18
17.32 ? 8.98
Oral clearance (fold-increase)
17.8 ? 6.9
15.6 ? 6.0
12.5 ? 6.4?
12.0 ? 6.8
16.6 ? 7.0
18.0 ? 5.2þþ
16.5 ? 6.6
18.0 ? 5.2þþ
11.0 ? 5.6
16.2 ? 6.5
9.7 ? 8.7
ERBT (% dose)
3.83 ? 0.88
3.60 ? 1.01
3.11 ? 1.18
3.50 ? 0.63
3.43 ? 1.00
4.06 ? 0.92
3.69 ? 0.95
4.06 ? 0.92
3.45 ? 0.41
3.61 ? 0.82
3.66 ? 0.79
ERBT (fold-increase)
2.06 ? 0.72
2.29 ? 0.70
2.19 ? 0.87
2.36 ? 0.76
2.11 ? 0.72
2.06 ? 0.78
2.14 ? 0.70
3.03 ? 2.13
1.92 ? 0.33
2.11 ? 0.77
2.81 ? 0.32
P-values between homozygous genotypes (?0.055,þ0.054,þþ0.033).
600
Pharmacogenetics
2003, Vol 13 No 10

Page 7

not shown). Also, no genotype–phenotype associations
were found with respect to either the induced systemic
clearance and ERBT values (Fig. 4) or the fold-changes
in the clearance parameters (Fig. 5). On the other hand,
the oral clearance of midazolam in the induced state
appeared to be associated in an inverse fashion with
the CYP3A4?1B and CYP3A5?3 genotypes, although
the differences were not quite statistically significant
because of data variability (Table 5, Fig. 4). However,
in the case of the fold-increase in oral clearance, this
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Systemic clearance
800
700
600
500
400
300
200
100
0
Oral clearance
Midazolam clearance (ml/min)
Midazolam clearance (ml/min)
ERBT (% 14C dose exhaled in 1 h)
AA
CYP3A4*1B
AG GGAA GA GG
CYP3A5*3
GG GA AA
CYP3A5*6
AA
CYP3A4*1B
AG GGAA GA GG
CYP3A5*3
GG AG AA
CYP3A5*6
AA
CYP3A4*1B
AG GGAA AG GG
CYP3A5*3
GG AG AA
CYP3A5*6
4000
3000
2000
1000
0
4
3
2
1
0
ERBT
Fig. 3
Genotype–phenotype associations between the erythromycin breath
test and the clearances of midazolam and CYP3A4?1B, CYP3A5?3
and CYP3A5?6 in the basal (day 1) state.
Systemic clearance
1250
1000
750
500
250
0
AA AG GG
CYP3A4*1B
AA AG GG
CYP3A5*3
GG GA AA
CYP3A5*6
AA AG GG
CYP3A4*1B
AA AG GG
CYP3A5*3
GG GA AA
CYP3A5*6
AA AG GG
CYP3A4*1B
AA AG GG
CYP3A5*3
GG AG AA
CYP3A5*6
Midazolam clearance (ml/min)
Oral clearance
p ? 0.055
40000
30000
20000
10000
0
Midazolam clearance (ml/min)
ERBT
7
6
5
4
3
2
1
0
ERBT (% 14C dose exhaled in 1 h)
Fig. 4
Genotype–phenotype associations between the erythromycin breath
test and the clearances of midazolam and CYP3A4?1B, CYP3A5?3
and CYP3A5?6 following pretreatment with rifampin (day 15/16).
CYP3A genotype–phenotype associations Floyd et al.
601

Page 8

was significantly different between the homozygous
CYP3A5?1 and CYP3A5?3 groups (P ¼ 0.033) and was
almost significant (P ¼ 0.055) between the homozygous
wild-type and homozygous CYP3A4?1B groups (Table
5, Fig. 5). Trends consistent with a gene–dose effect
were also apparent (Figs 4 and 5).
Discussion
Human CYP3A is involved in the oxidation of many
drugs and inter-individual variability is usually consid-
ered to be marked. For example, a recent survey
indicates an over 10-fold range in the reported systemic
clearance of midazolam [29] and even larger (. 30-fold)
differences are present in in-vitro measured hepatic
activity [15]. Drug–drug interactions and other environ-
mental factors resulting in inhibition or induction are a
major contributor to such variability [5]. However, it is
unclear to what extent other factors reflecting possible
genetic differences are involved. A major finding of this
study is that constitutive CYP3A activity, as measured
by metabolism of midazolam, exhibited only a modest
degree of variability in most (. 95%) medication-free,
healthy subjects regardless of whether the in-vivo probe
drug was administered intravenously (four-fold) or
orally (five-fold), reflecting
patic and collective intestinal/hepatic activity, respec-
tively. Somewhat greater activity was present in some
individuals which increased the range of observed
activity, especially after oral administration, but even
then the in-vivo variability was less than 10-fold and
such individuals represented only a small proportion of
the population. Such relatively small variability has
been noted in similar studies in African-Americans,
Mexican and Japanese using the same clinical protocol
and inclusion/exclusion criteria [13,30,31]. A recent
he-
compilation of data obtained in 17 independent studies,
mostly in healthy European-Americans, also found
inter-individual variability to be only four-fold in
approximately 85% of the subjects [8,32]. Thus, it
appears that, by contrast to the general perception,
constitutive in-vivo CYP3A activity varies only mod-
estly, although, occasional outliers exist for reasons that
remain unknown. One reason for this much smaller
range of inter-individual variability than is generally
considered may reflect the use of relatively small doses
of midazolam employed in the present study. At higher
doses, especially following oral administration where
dose-dependency of the oral bioavailability of midazo-
lam has been reported [33], higher variability might
occur.
In addition to inter-individual variability, another long-
standing concern regarding CYP3A is whether its
activity is different between men and women. As
recently summarized [34–36], the evidence is equivocal
and confounded by the fact that many of the studied
CYP3A substrates are also transported by MDR1. How-
ever, midazolam is not a substrate for MDR1 [37] and
therefore its clearance presumably only reflects CYP3A-
mediated metabolism. Thus, the current observations
are consistent with other studies [38–40] indicating that
CYP3A activity per se is not sex-related. However, this
conclusion does not exclude the possibility that modest
sex differences in the disposition of other CYP3A
substrates may exist, when factors other than the
enzyme’s intrinsic activity are involved [35–36]. Also,
the present findings in European- and African-Ameri-
can subjects, along with those in other populations such
as Mexicans [30] and Japanese [31], collectively suggest
that dosage regimen of drugs that are CYP3A substrates
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
40
30
20
10
0
GGAGAA
CYP3A5*6
Fold-increase in midazolam
oral clearance
NS
40
30
20
10
0
AAGAGG
CYP3A5*3
Fold-increase in midazolam
oral clearance
P ? 0.033
40
30
20
10
0
AAAGGG
CYP3A4*1B
Fold-increase in midazolam
oral clearance
P ? 0.055
Fig. 5
Genotype–phenotype associations between the fold-increase (day 15/day 1) in the erythromycin breath test and the clearances of midazolam and
CYP3A4?1B, CYP3A5?3 and CYP3A5?6 following pretreatment with rifampin.
602
Pharmacogenetics
2003, Vol 13 No 10

Page 9

are unlikely to require modification across major popu-
lation groups throughout the world, at least from a
metabolic standpoint.
When used as an in-vivo CYP3A phenotypic trait, the
clearance of midazolam requires multiple blood sam-
pling over several hours. Because of the inconvenience
and effort needed, more limited sampling strategies
have been studied. For example, a strong correlation
was found between the plasma concentration of mid-
azolam at 4 h and the respective AUC obtained from a
full plasma concentration–time profile [8]. The present
study confirms this observation (Fig. 2), and provides
additional data to further support the use of a 4-h
plasma level as a phenotypic trait measure. Thus,
combining the current data obtained in 57 additional
subjects (2 3 57 after intravenous dosing) to that pre-
viously collated [8] yields the following, forced through
the origin, relationships: after 1 mg midazolam intra-
venously, AUCiv¼ 14.3 (4-h plasma level) and, after
2 mg midazolam orally, AUCoral¼ 12.1 (4-h plasma
level). Because of the dose-dependency of the clear-
ance of midazolam following oral administration [33],
caution should be taken when extending these relation-
ships to other doses.
In recent years, considerable effort has been made to
identify SNPs in both CYP3A4 and CYP3A5 genes with
the hope that such variation might provide insight into
a possible molecular mechanism(s) to account for inter-
individual variability in CYP3A activity. With CYP3A4,
most of the SNPs appear to be relatively uncommon
(, 1–2% frequency) and possibly selectively present in
certain populations [32] (http://www.imm.ki.se/CYPal
leles/). It is therefore unlikely that these variants
significantly contribute in a major fashion to the
variability in activity. By contrast, CYP3A4?1B invol-
ving a ?392A.G transition occurs more often; in
African-Americans, the allele frequency has been re-
ported to be 35–67% whereas, in Caucasians, it is
significantly smaller, ranging from 4–10% [32]. The
present findings are consistent with these frequencies.
Epidemiological studies suggest that CYP3A4?1B is
associated with an increased risk of developing prostate
cancer in African-Americans [9,11] and, in addition,
treatment-related leukaemia following administration of
chemotherapeutic agents that are metabolized by
CYP3A [12]. In both cases, it has been speculated that
the variant allele might lower CYP3A expression result-
ing in impaired metabolism of testosterone or cancer
chemotherapeutic agents, respectively. However, more
recently, the association of CYP3A4?1B with treatment-
related leukaemia has not been confirmed [17] and it
has also been suggested that the reported associations
may not be causal, but rather an artefact resulting from
population stratification [16]. This conclusion is consis-
tent with several reports that have failed to demon-
strate any effect of CYP3A4?1B on in-vitro gene
expression [14,15]. Similarly, studies performed in vivo
with orally administered CYP3A substrates such as
midazolam, as in the present study, and nifedipine
[10] have failed to observe differences in drug dis-
position between homozygous CYP3A4?1A/?1A and
CYP3A4?1B/?1B individuals. No difference was noted
either using the erythromycin breath test as a pheno-
typic in-vivo probe for hepatic CYP3A4 activity in these
two genotypic groups [10]. However, a 30% lower
systemic clearance of midazolam was noted in another
study in European- and African-American men (n ¼ 15,
each group) and a gene–dose effect was present [13].
However, the current data from an additional 23
European-Americans and 34 African-Americans, either
alone or combined with the original 30 subjects do not
confirm this earlier finding. Thus, the initial observa-
tion probably reflects a chance statistical finding (Type
I error) and CYP3A4?1B does not lead to reduced gene
expression and lower enzyme activity in vivo. However,
there is a caveat. Because of the high allelic frequency
of CYP3A4?1B in African-Americans, the prevalence of
homozygous CYP3A4?1A/?1A subjects in this popu-
lation is very low. Accordingly, comparison of the two
homozygous populations strongly reflects a racial di-
chotomy (Table 4); therefore, it is possible that the
results from a population with common ancestry might
demonstrate a genotype–phenotype difference. How-
ever, this would be a difficult study to perform and the
limited data in African-Americans homozygous for
CYP3A4?1A compared to the two other genotypes is
not suggestive in this respect.
Until recently, CYP3A5 has been largely neglected with
respect to its role and importance in CYP3A-mediated
metabolism. However, this enzyme may be more
important than previously considered because the
abundance of CYP3A5 has been suggested to be greater
thanpreviouslyconsidered
CYP3A5 was reported to represent 46–85% of total
CYP3A content in certain livers and, similarly, in the
intestinal epithelium [19]. By contrast, CYP3A5 tran-
scripts, on average, were only found to account for
approximately 2% of the total CYP3A mRNA pool
obtained from human livers, and even in those with the
highest CYP3A5 expression, reflecting CYP3A5.1A het-
erozygosity, the contribution was only 20% [41]. Similar
findings were more recently reported along with the
observation that CYP3A5 only contributed 2% of over-
all hepatic CYP3A protein and was not a contributing
factor in the variability of CYP3A catalytic activity in
Caucasians [42]. On the other hand, genetic variants
have been associated with the level of CYP3A5 expres-
sion. For example, with CYP3A5?3, catalytically active
enzyme is only minimally encoded [18,19], thus, the 19-
hydroxylation of midazolam was found to be higher in
hepatic microsomes expressing one or more CYP3A5?1
[18,19].Forexample,
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
CYP3A genotype–phenotype associations Floyd et al.
603

Page 10

alleles compared to those in which only CYP3A4 was
present (i.e. CYP3A5?3/?3). Also, the correlation be-
tween CYP3A activity and protein content was signifi-
cantly improved when both CYP3A4 and CYP3A5
contents were included compared to the consideration
of only one of these [19]. Other, but much less frequent
variants such as CYP3A5?6 also result in reduced ex-
pression of encoded protein and, collectively, such
variants provide a molecular basis for the polymorphic
expression of the enzyme.
Measurement of the major 19-hydroxylation and minor
4-hydroxylation pathways for midazolam metabolism
provides an excellent phenotypic trait because CYP3A4
and CYP3A5 both mediate these reactions which
account for most of the metabolism of the drug [43,44].
Moreover, the enzymes’ catalytic activities have been
found to be equivalent using recombinant enzymes [4]
and, in the case of enzymes purified from human livers,
the activity of CYP3A5 was found to be higher than
that for CYP3A4 [44]. Therefore, in spite of the
difficulties of accurately optimizing in-vitro assay condi-
tions comparable to those in vivo [45], it is highly likely
that assessment of the clearance of midazolam in vivo
reflects the combined activities of both CYP3A4 and
CYP3A5.
The second important finding from this study is that
the disposition of midazolam, after either oral or intra-
venous administration, was not different in subjects
carrying the essentially null allelic variants of CYP3A5.
The absence of any genotype–phenotype association
was most apparent with the CYP3A5?3 allele because
of its relatively high frequency. This finding is consis-
tent with a recently reported investigation with mid-
azolamin Taiwanese subjects
heterozygous or homozygous for CYP3A5?3 [46], and
recent in-vitro findings [41,42]. Although the numbers
of subjects with the CYP3A5?6 and CYP3A5?7 allelic
variants were much smaller than with CYP3A5?3, again
no genotype–phenotype associations were discernible.
Therefore, given the earlier described caveat related to
population differences in allelic frequency, it does not
appear that CYP3A5 activity is an important contributor
to overall CYP3A activity in vivo, at least for midazolam
after administration of the relatively low doses used in
this study. Whether this conclusion applies to higher
doses or other CYP3A substrates requires further in-
vestigation. The in vitro–in vivo discordancy possibly
reflects factors associated with the stability of the two
isoforms when obtained post mortem from human livers
and their subsequent storage and handling [47].
whowere either
CYP3A activity is readily inducible through pregnane X
receptor (PXR)-mediated transcriptional up-regulation
[48]. Rifampin is a particularly potent PXR agonist [49]
and its ability to induce CYP3A [50] and modulate both
the erythromycin breath test [51–53] and the metabo-
lism of midazolam [55–57] is well-established. The
systemic clearance of midazolam was increased by
approximately two-fold and the erythromycin breath
test was enhanced to a similar extent. This degree of
induction and its approximately four-fold range in
inter-individual variability is similar to that in earlier
reports [51–54]. An eight-fold greater induction was
noted with respect to the oral clearance of midazolam
following rifampin pretreatment and, again, the asso-
ciated five-fold range in variability between subjects is
also consistent with previous studies [55–56]. Route-
dependent induction has also been observed with other
CYP3A inducers such as St John’s wort [28,57] and
reflectsthe additionalup-regulation
CYP3A and its major role in determining the bioavail-
ability of midazolam after oral administration [13,58].
ofintestinal
The inducibility of CYP3A5 is far less than that of
CYP3A4 [20–22]. Accordingly, the extent of induction
might be expected to be related to CYP3A5 genotype
in that individuals with comparable overall CYP3A
activity but different amounts of CYP3A5 (CYP3A5.1/1
versus CYP3A5.3/3) would respond differently. Inter-
estingly, the absolute value and the fold-increase in the
oral clearance of midazolam were greater in homozy-
gous CYP3A5?3 individuals compared to homozygous
wild-type subjects, and trends consistent with a gene–
dose effect were apparent. A similar but inverse rela-
tionship was also noted with respect to CYP3A4?1B
which probably reflects the incomplete genetic linkage
between these variants of the two alleles. No such
relationships were noted with respect to the fold-in-
crease in the systemic clearance of midazolam which
probably reflects the considerably smaller extent of
induction of hepatic CYP3A activity compared to that
in the intestine. However, it must be noted that the
difference in the extent of change in oral clearance was
relatively modest (approximately 50%); again consistent
with a minor role of CYP3A5 in overall CYP3A activity.
Furthermore, these differences were not noted when
using the 4-h plasma concentration of midazolam as a
trait measure, probably because of its less robust nature
compared to clearance of the drug along with the
relative small changes observed.
The extent to which rifampin induces CYP3A is likely
to be determined by the intracellular concentration of
the inducing agent in the intestine and liver. Impor-
tantly, this is dependent on MDR1 localized in the
apical membrane domain of the enterocyte and hepato-
cyte because rifampin is an MDR1 substrate. There-
fore, its accumulation in such cells inversely reflects the
efflux transporter’s functional activity: the higher the
activity, the lower the extent of induction [23]. Also, in
mice the ERBT has been shown to be determined by
the amount of functional mdr1 [59]. Recently, a num-
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
604
Pharmacogenetics
2003, Vol 13 No 10

Page 11

ber of SNPs in the MDR1 gene have been identified, in
particular transitions 2677G.T in exon 21 [24,25] and
3435C.T in exon 26 [25,26]. The functional conse-
quences of such allelic variants remains controversial
[60] and, even when functional associations with these
allelic variants have been observed, the magnitude of
the intergenotypic differences have been quite modest.
Nevertheless, it was of interest to investigate possible
genotype–phenotype associations between the MDR1
variants and the extent of change in the clearance of
midazolam following rifampin treatment. However, no
such associations were apparent with respect to the oral
or intravenous clearances of midazolam or the changes
induced by rifampin pretreatment. In the latter case, it
is not possible to discriminate between a true lack of an
MDR1 effect from the possibility that, at the rifampin
dosage used, induction was maximal and therefore
small changes in the intracellular concentration of the
drug would not change the degree of induction. Also, it
should be noted that the number of homozygous
variant subjects was quite small.
The major aim of this study was to determine whether
common variants in the CYP3A4 and CYP3A5 genes
have in-vivo functional consequences for the metabo-
lism of a prototypic substrate. It was hoped that such
results would provide insight into the putative genetic
factors involved in the presumed large inter-individual
variability in CYP3A activity. However, not only were
the variabilities in the oral and systemic clearance of
midazolam found to be relatively modest, but also
CYP3A-mediated metabolism did not appear to be
associated with the CYP3A4?1B, CYP3A5?3, CYP3A5?6
and CYP3A5?7 genetic polymorphisms. This observa-
tion does not preclude the possibility that other SNPs
may affect catalytic activity, especially in outlier indivi-
duals. Nevertheless, it raises the question of the
mechanistic basis for the apparently high heritability of
CYP3A activity [6] and the possible role of genetic
variants in regulatory factors such as PXR [61,62] and
other determinants.
Acknowledgements
The authors thank Reshma Desai and Amanda Ingram
for their technical assistance in determining the CYP3A
and MDR genotypes, respectively.
References
1 Westlind A, Malmebo S, Johansson I, Otter C, Andersson TB, Ingelman-
Sundberg M, et al. Cloning and tissue distribution of a novel human
cytochrome P450 of the CYP3A subfamily, CYP3A43. Biochem Biophys
Res Commun 2001; 281:1349–1355.
2 Gellner K, Eiselt R, Hustert E, Arnold H, Koch I, Haberl M, et al. Genomic
organization of the human CYP3A locus: identification of a new, inducible
CYP3A gene. Pharmacogenetics 2001; 11:111–121.
3 Komori M, Nishio K, Kitada M, Shiramatsu K, Muroya K, Soma M, et al.
Fetus-specific expression of a form of cytochrome P-450 in human livers.
Biochemistry 1990; 29:4430–4433.
4Williams JA, Ring BJ, Cantrell VE, Jones DR, Eckstein J, Ruterbories K,
et al. Comparative metabolic capabilities of CYP3A4, CYP3A5, and
CYP3A7. Drug Metab Dispos 2002; 30:883–891.
Thummel KE, Wilkinson GR. In vitro and in vivo drug interactions
involving human CYP3A. Ann Rev Pharmacol Toxicol 1998; 38:
389–430.
O¨zdemir V, Kalow W, Tang BK, Paterson AD, Walker SE, Endrenyi L,
et al. Evaluation of the genetic component of variability of CYP3A4
activity: a repeated drug administration method. Pharmacogenetics 2000;
10:373–388.
Bertilsson L. Geographical/interracial differences in polymorphic drug
oxidation. Current state of knowledge of cytochromes P450 (CYP) 2D6
and 2C19. Clin Pharmacokinet 1995; 29:192–209.
Lin YS, Lockwood GF, Graham MA, Brian WR, Loi CM, Dobrinska MR,
et al. In-vivo phenotyping for CYP3A by a single-point determination of
midazolam plasma concentration. Pharmacogenetics 2001; 11:781–791.
Paris PL, Kupelian PA, Hall JM, Williams TL, Levin H, Klein EA, et al.
Association between a CYP3A4 genetic variant and clinical presentation
in African-American prostate cancer patients. Cancer Epidemiol Biomar-
kers Prev 1999; 8:901–905.
Ball SE, Scantina J, Kao J, Ferron GM, Fruncillo R, Mayer P, et al.
Population distribution and effects on drug metabolism of a genetic
variant in the 59 promoter region of CYP3A4. Clin Pharmacol Ther 1999;
66:288–294.
Rebbeck TR, Jaffe JM, Walker AH, Wein AJ, Malkowicz SB. Modification
of clinical presentation of prostate tumors by a novel genetic variant in
CYP3A4. J Natl Cancer Inst 1998; 90:1225–1229.
Felix CA, Walker AH, Lange BJ, Williams TM, Winick NJ, Cheung NK,
et al. Association of CYP3A4 genotype with treatment-related leukemia.
Proc Natl Acad Sci 1998; 95:13176–13181.
Wandel C, Witte JS, Hall JM, Stein CM, Wood AJJ, Wilkinson GR.
CYP3A activity in African American and European men: population
differences and functional effect of the CYP3A4?1B 59-promoter region
polymorphism. Clin Pharmacol Ther 2000; 68:82–91.
Spurdle AB, Goodwin B, Hodgson E, Hopper JL, Chen X, Purdie DM,
et al. The CYP3A4?1B polymorphism has no functional significance and
is not associated with risk of breast or ovarian cancer. Pharmacogenetics
2002; 12:355–366.
Westlind A, Lo ¨fberg L, Tindberg N, Andersson TB, Ingelman-Sundberg
M. Interindividual differences in hepatic expression of CYP3A4: relation-
ship to genetic polymorphisms in the 59-upstream regulatory region.
Biochem Biophys Res Commmun 1999; 259:201–205.
Kittles RA, Chen W, Panguluri RK, Ahaghotu C, Jackson A, Adebamowo
CA, et al. CYP3A4-V and prostate cancer in African Americans: casual or
confounding association because of population stratification? Hum Genet
2002; 110:553–560.
Blanco JG, Edick MJ, Hancock ML, Winick NJ, Dervieux T, Amylon MD,
et al. Genetic polymorphisms in CYP3A5, CYP3A4 and NQO1 in
children who developed therapy-related myeloid malignancies. Pharmaco-
genetics 2002; 12:605–611.
Kuehl P, Zhang J, Lin Y, Lamba J, Assem M, Schuetz J, et al. Sequence
diversity in CYP3A promoters and characterization of the genetic basis of
polymorphic CYP3A5 expression. Nature Genet 2001; 4:383–391.
Lin YS, Dowling AL, Quigley SD, Farin FM, Zhang J, Lamba J, et al.
Co-regulation of CYP3A4 and CYP3A5 and contribution of hepatic and
intestinal midazolam metabolism. Mol Pharmacol 2002; 62:162–172.
Wrighton SA, Ring BJ, Watkins PB, VandenBraden M. Identification of a
polymorphically expressed member of the human cytochrome P-450III
family. Mol Pharmacol 1989; 36:97–105.
Schuetz EG, Schuetz JD, Strom SC, Thompson MT, Fisher RA, Molowa
DT, et al. Regulation of human liver cytochromes P-450 in family 3A in
primary and continuous culture of human hepatocytes. Hepatology 1993;
18:1254–1262.
Rae JM, Johnson MD, Lippman ME, Flockart DA. Rifampin is a selective,
pleiotropic inducer of drug metabolism genes in human hepatocytes:
studies with cDNA and oligonucleotide expression arrays. J Pharmacol
Exp Ther 2001; 299:849–857.
Schuetz EG, Schinkel AH, Relling MV, Schuetz JD. P-glycoprotein:
a major determinant of rifampicin-inducible expression of cytochrome
P4503A in mice and humans. Proc Natl Acad Sci 1996; 93:4001–4005.
Cascorbi I, Gerloff T, Johne A, Meisel C, Hoffmeyer S, Schwab M, et al.
Frequency of single nucleotide polymorphisms in the P-glycoprotein drug
transporter MDR1 gene in white subjects. Clin Pharmacol Ther 2001;
69:169–174.
Kim RB, Leake BF, Choo EF, Dresser GK, Kubba SV, Schwarz UI, et al.
Identification of functionally variant MDR1 alleles among European Amer-
icans and African Americans. Clin Pharmacol Ther 2001; 70:189–199.
Hoffmeyer S, Burk O, von Richter O, Arnold HP, Brockmo ¨ller J, Cascorbi
I, et al. Functional polymorphisms of the human multidrug-resistance
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
CYP3A genotype–phenotype associations Floyd et al.
605

Page 12

gene: multiple sequence variations and correlation of one allele with P-
glycoprotein expression and activity in vivo. Proc Natl Acad Sci 2000;
97:3473–3478.
Kinirons MT, O’Shea D, Downing TE, Fitzwilliam AT, Joellenbeck L,
Groopman JD, et al. Absence of correlations among three putative in vivo
probes of human cytochrome P4503A activity in young healthy men. Clin
Pharmacol Ther 1993; 54:621–629.
Dresser GK, Schwarz UI, Wilkinson GR, Kim RB. Coordinate induction of
both CYP3A and MDR1 by St. John’s wort in healthy subjects. Clin
Pharmacol Ther 2003; 73:41–50.
Rogers JF, Rocci ML Jr, Haughey DB, Bertino JS Jr. An evaluation of the
suitability of intravenous midazolam as an in vivo marker for hepatic
cytochrome P4503A activity. Clin Pharmacol Ther 2003; 73:153–158.
Poland RA, Lin K-M, Nuccio C, Wilkinson GR. Cytochrome P4502E1
and 3A activities do not differ between Mexicans and European Amer-
icans. Clin Pharmacol Ther 2002; 72:288–293.
Tateishi T, Watanabe M, Nakura H, Asoh M, Shirai H, Mizorogi Y, et al.
CYP3A activity in European American and Japanese men using midazo-
lam as an in vivo probe. Clin Pharmacol Ther 2001; 69:333–339.
Lamba JK, Lin YS, Schuetz EG, Thummel KE. Genetic contribution to
variable human CYP3A-mediated metabolism. Adv Drug Deliv Rev 2002;
54:1271–1294.
Bornemann LD, Min BH, Crews T, Rees MMC, Blumenthal HP, Colburn
WA, et al. Dose dependent pharmacokinetics of midazolam. Eur J Clin
Pharm 1985; 29:91–95.
Harris RZ, Benet LZ, Schwartz JB. Gender effects in pharmacokinetics
and pharmacodynamics. Drugs 1995; 50:222–239.
Meibohm B, Beierle I, Derendorf H. How important are gender differences
in pharmacokinetics? Clin Pharmacokinet 2002; 41:329–342.
Cummins CL, Wu C-Y, Benet LZ. Sex-related differences in the clearance
of cytochrome P450 3A4 substrates may be caused by P-glycoprotein.
Clin Pharmacol Ther 2002; 72:474–489.
Kim RB, Wandel C, Leake B, Cvetkovic M, Fromm MF, Dempsey PJ,
et al. Interrelationship between substrates and inhibitors of human
CYP3A and P-glycoprotein. Pharm Res 1999; 16:408–414.
Holazo AA, Winkler MB, Patel IH. Effects of age, gender and oral
contraceptives on intramuscular midazolam pharmacokinetics. J Clin
Pharmacol 1998; 28:1040–1045.
Greenblatt DJ, Abernethy DR, Locniskar A, Harmatz JS, Limjuco RA,
Shader RI. Effect of age, gender, and obesity on midazolam kinetics.
Anesthesiology 1984; 61:27–35.
Thummel KE, O’Shea D, Paine MF, Shen DD, Kunze KL, Perkins JD, et al.
Oral first-pass elimination of midazolam involves both gastrointestinal and
hepatic CYP3A-mediated metabolism. Clin Pharmacol Ther 1996;
59:491–502.
Koch I, Weil R, Wolbold R, Bro ¨ckmoller J, Hustert E, Burk O, et al.
Interindividual variability and tissue-specificity in the expression of
cytochrome P450 3A mRNA. Drug Metab Dispos 2002; 30:
1108–1114.
Westlind-Johnsson A, Malmebo S, Johansson A, Otter C, Andersson TB,
Johansson I, et al. Comparative analysis of CYP3A expression in human
liver suggests only a minor role for CYP3A5 in drug metabolism. Drug
Metab Disp 2003; 31:755–761.
Kronbach T, Mathys D, Umeno M, Gonzalez FJ, Meyer UA. Oxidation of
midazolam and triazolam by human liver cytochrome P450IIIA4. Mol
Pharmacol 1989; 36:89–96.
Gorski JC, Hall SD, Jones DR, VandenBranden M, Wrighton SA.
Regioselective biotransformation of midazolam by members of the human
cytochrome P450 3A (CYP3A) subfamily. Biochem Pharmacol 1994;
47:1643–1653.
Maenpaa J, Hall SD, Ring BJ, Strom SC, Wrighton SA. Human
cytochrome P450 3A (CYP3A) mediated midazolam metabolism: the
effect of assay conditions and regioselective stimulation by Æ-naphtho-
flavone, terfenadine and testosterone. Pharmacogenetics 1998; 8:
137–155.
Shih P-S, Huang J-D. Pharmacokinetics of midazolam and 19-hydroxymi-
dazolam in Chinese with different CYP3A5 genotypes. Drug Metab Disp
2002; 30:1491–1496.
Rekka E, Evdokimova E, Eeckhoudt S, Labar G, Calderon PB. Role of
temperature on protein and mRNA cytochrome P450 3A (CYP3A)
isozymes expression and midazolam oxidation by cultured rat precision-
cut liver slices. Biochem Pharmacol 2002; 64:633–643.
Goodwin B, Redinbo MR, Kliewer SA. Regulation of CYP3A gene
transcription by the pregnane X receptor. Annu Rev Pharmacol Toxicol
2002; 42:1–23.
Goodwin B, Hodgson E, Liddle C. The orphan human pregnane X
receptor mediates the transcriptional activation of CYP3A4 by rifampicin
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
through a distal enhancer module. Mol Pharmacol 1999; 56:
1329–1339.
Ged C, Rouillon M, Pichard L, Combalbert J, Bressot N, Bories P. The
increase in urinary excretion of 6ß-hydroxycortisol as a marker of human
hepatic cytochrome P450IIIA induction. Br J Clin Pharmac 1989;
28:373–387.
Watkins PB, Murray SA, Winkelman LG, Heuman DM, Wrighton SA,
Guzelian PS. Erythromycin breath test as an assay of glucocorticoid-
inducible liver cytochromes P-450. Studies in rats and patients. J Clin
Invest 1989; 83:688–697.
Gharaibeh MN, Gillen LP, Osborne B, Schwartz JI, Waldman SA. Effect
of multiple doses of rifampin on the [14C N-methyl] erythromycin breath
test in healthy male volunteers. J Clin Pharmacol 1998; 38:492–495.
Polk RE, Brophy DF, Israel DS, Patron R, Sadler BM, Chittick GE, et al.
Pharmacokinetic interaction between amprenavir and rifabutin or rifampin
in healthy males. Antimicrob Agents Chemother 2001; 45:502–508.
Eeckhoudt SL, Desager JP, Robert AR, Leclercq I, Verbeeck RK,
Horsmans Y. Midazolam and cortisol metabolism before and after CYP3A
induction in humans. Int J Clin Pharmacol Ther 2001; 39:293–299.
Backman JT, Olkkola KT, Neuvonen PJ. Rifampin drastically reduces
plasma concentrations and effects of oral midazolam. Clin Pharmacol
Ther 1996; 59:7–13.
Backman JT, Kivisto ¨ KT, Olkkola KT, Neuvonen PJ. The area under the
plasma concentration-time curve for oral midazolam is 400-fold larger
during treatment with itraconazole than with rifampicin. Eur J Clin
Pharmacol 1998; 54:53–58.
Wang Z, Gorski JC, Hamman MA, Huang S-M, Lesko LJ, Hall SD. The
effects of St. John’s wort (Hypericum perforatum) on human cytochrome
P450 activity. Clin Pharmacol Ther 2001; 70:317–326.
Gorski JC, Jones DR, Haehner-Daniels BD, Hamman MA, O’Mara EM,
Hall SD. The contribution of intestinal and hepatic CYP3A to the
interaction between midazolam and clarithromycin. Clin Pharmacol Ther
1998; 64:133–144.
Lan L-B, Dalton JT, Schuetz EG. Mdr1 limits CYP3A metabolism in vivo.
Mol Pharmacol 2000; 58:863–869.
Kim RB. MDR1 single nucleotide polymorphisms: multiplicity of haplo-
types and functional consequences. Pharmacogenetics 2002; 12:
425–427.
Zhang J, Kuehl P, Green ED, Touchman JW, Watkins PB, Daly A, et al.
The human pregnane X receptor: genomic structure and identification
and functional characterization of natural allelic variants. Pharmacoge-
netics 2001; 11:555–571.
Hustert E, Zibat A, Presecan-Siedel E, Eiselt R, Mueller R, Fuss C, et al.
Natural protein variants of pregnane X receptor with altered transactiva-
tion activity toward CYP3A4. Drug Metab Disp 2001; 29:1454–1459.
50
51
52
53
54
55
56
57
58
59
60
61
62
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
606
Pharmacogenetics
2003, Vol 13 No 10