HSE
Health & Safety
Executive
Inter-individual variability in the interpretation
of biological monitoring guidance values
Prepared by HSL for the
Health and Safety Executive 2004
RESEARCH REPORT 190 HSE
Health & Safety
Executive
Inter-individual variability in the interpretation
of biological monitoring guidance values
George Loizou and Craig Sams
Health & Safety Laboratory
Broad Lane
Sheffield
S3 7HQ
The incorporation of inter-individual differences in the setting of occupational exposure standards such
as biological monitoring guidance values has been studied by investigating the possibility of
transferring in vitro measurements of variability in metabolism to the in vivo situation. The generation of
generic information on the activity of two major enzymes systems considered important in the
metabolism of industrial chemicals, was investigated by using ’probe’ substrates. Variability in the
activities of cytochromes P450 1A2 and 2E1, and the cytosolic glutathione transferases µ and θ were
measured in human hepatocytes, hepatic microsomes and cytosols. Variability in the metabolism of
styrene, a chemical of industrial importance, was also measured for comparison. The ability to use in
vitro data to simulate in vivo variability was investigated by incorporating in vitro data into a PBPK
model for styrene.
This report and the work it describes were funded by the Health and Safety Executive (HSE). Its
contents, including any opinions and/or conclusions expressed, are those of the authors alone and do
not necessarily reflect HSE policy.
HSE BOOKS
© Crown copyright 2004
First published 2004
ISBN 0 7176 2799 3
All rights reserved. No part of this publication may be
reproduced, stored in a retrieval system, or transmitted in
any form or by any means (electronic, mechanical,
photocopying, recording or otherwise) without the prior
written permission of the copyright owner.
Applications for reproduction should be made in writing to: Licensing Division, Her Majesty's Stationery Office, St Clements House, 2-16 Colegate, Norwich NR3 1BQ or by e-mail to [email protected]
ii
1. INTRODUCTION.................................................................................................................... 1
1.1 PBPK Modelling and Biological Monitoring ........................................................... 1
1.2 PBPK Modelling in Chemical Risk Assessment.................................................... 2
1.3 PBPK Modelling of Inter-Individual Variability....................................................... 2
1.4 Incorporating Inter-Individual Variability in the Interpretation of Biological Monitoring Guidance Values............................................................................................. 3
2. MATERIALS AND METHODS ............................................................................................. 7
2.1 Chemicals, reagents and biological materials....................................................... 7
2.2 Enzyme assays and analysis.................................................................................... 8
3. RESULTS.............................................................................................................................. 19
3.1 Variation in enzyme activity...................................................................................... 19
3.2 Validation of the PBPK model .................................................................................. 30
3.3 Incorporation of inter-individual variability .......................................................... 32
4. DISCUSSION ....................................................................................................................... 35
4.1 Variation in enzyme activity...................................................................................... 35
4.2 Incorporating inter-individual variability into PBPK models ........................... 38
5. CONCLUSIONS ................................................................................................................... 41
REFERENCES ......................................................................................................................... 43
iii
iv
1. INTRODUCTION The quantitative estimation of dose following industrial or environmental exposure to chemicals
is of fundamental importance in the protection of the health and safety of workers. However,
there are difficulties in the estimation of dose or ‘body burden’ which are characterised by the
fact that dose often fluctuates considerably over time making any correlation with chronic
disease difficult. Also, the ‘target’ dose, which is the concentration of chemical at the site of
action, cannot be measured directly because biological targets such as organs and tissues are
generally, inaccessible in exposed people. Surrogates of target dose i.e. air exposure
concentration or biological monitoring (BM) must therefore be used. The advantages and
disadvantages of both air and biological monitoring have been discussed (Droz et al., 1991).
Briefly, air monitoring can introduce a large bias into the estimation of the true target tissue
dose. This is because variation in factors such as physical workload, skin absorption, the
partition coefficient of the chemical of interest, and individual anatomical, physiological and
biochemical differences which determine chemical pharmacokinetics, ensure that air monitoring
is, more often than not, a crude indicator of the biologically active dose (Droz et al., 1991).
Wide variations in exposure concentrations due to changes in work procedures and practices
occurring over periods of months to years can be taken into account during air monitoring.
Whereas, short-term within-day and day to day fluctuations are best captured by BM (Droz et
al., 1991). Biological monitoring uses measurements of chemicals or their metabolites, known
as ‘biomarkers’, in biological media such as blood, urine, breath, saliva, and sometimes tissues,
as an index of absorption of chemical (Mason and Wilson, 1999). The strength of BM lies in
the potential to define total absorbed dose from all routes of exposure i.e. inhalation, dermal
penetration and ingestion (Mason and Wilson, 1999). The results are affected by biological
variability and in the past the biological basis of this variability has been characterised as
problematic (Droz et al., 1991). More recently however, the introduction of biologically based
mathematical modelling techniques, has demonstrated the potential to enhance the value of BM
by quantitatively accounting for and explaining the underlying biological variability.
1.1 PBPK Modelling and Biological Monitoring
Many biological factors influence the concentration and delivery of a chemical, or a potentially
more toxic metabolite of a chemical, to the blood, urine, target organ, or other biological
medium. Therefore, by identifying the relationship between a biomarker concentration in an
accessible body fluid and the toxic entity at the target site should improve the application of BM
to risk assessment as well as exposure assessment. Models of varying complexity have been
used to investigate the possible relationships along the exposure, dose-response continuum. The
application of physiologically–based pharmacokinetic (PBPK) models is particularly suited to
interpreting BM data. The ability to accurately predict the quantity, or the ‘dosimetry’, of a
given foreign chemical at the tissue and organ level in the body, following any route of
exposure, is the ‘linchpin’ of a sound risk assessment (Yang et al., 1998). The definition of
target tissue dosimetry can be achieved using PBPK modelling which is a powerful means of
simulating the factors that determine tissue dose, and consequently, correlation with health
effects. A PBPK model is an independent structural model, comprising the tissues and organs
of the body with each perfused by, and connected via, the vascular system. Chemical-specific
kinetic data, which is dependent on organ and tissue solubility, plasma protein binding, and
metabolic rate constants, is overlaid on model predictions. The value of such models is that
they are tools for integrating mechanistic pharmacokinetic and toxicologic information through
their explicit mathematical description of important anatomical, physiological and biochemical
determinants of chemical uptake, disposition and elimination. Model compartments correspond
either to specific, localised anatomical structures such as liver, lung, kidney etc. or to types of
tissues–such as fat, muscle, viscera and bone–that are distributed throughout the body. Actual
1
physiological parameters such as breathing rates, blood flow rates and tissue volumes used in
the models to describe the pharmacokinetic process are obtained from the literature or can be
obtained by experiment. These physiological parameters are coupled with chemical-specific
parameters such as partition coefficients (tissue retention of chemicals related to solubility in
biological media and binding to specific proteins in tissues) and metabolic constants to predict
the dynamics of a compound’s movement through a biological system. The advantage of the
physiologically based model is that by simply using the appropriate physiological, biochemical
and metabolic parameters, the same model can be utilised to describe the dynamics of chemical
transport and metabolism in any species, e.g., mice, rats and humans.
1.2 PBPK Modelling in Chemical Risk Assessment
Quantitative chemical risk assessment requires the definition of a toxicologically relevant
measure of dose known as a ‘dose surrogate’. The term dose surrogate or ‘dose metric’ is used
to describe the different way of expressing a target tissue dose estimate calculated using a PBPK
model. The ability to use a dose surrogate in chemical risk assessment ultimately depends on
knowledge of the relationship between the measure of tissue dose and cellular response
involved in a toxic endpoint. There have been many attempts at incorporating pharmacokinetic
data in risk assessments, which have used a measure of dose, such as peak blood concentration
of parent chemical (Cmax) or total amount metabolised. The underlying principle is the use of the
most up-to-date knowledge on the mode of action of the chemical at the cellular and sub­
cellular level. For example, the neurological effects of short-term exposure to solvents such as
trichloroethylene are rapidly reversible. This suggests that the parent chemical may disrupt
normal lipophilic cellular membrane function via a physicochemical effect. An appropriate
dose metric for such an effect would be the Cmax or AUC for the parent chemical in the brain.
Since tissue-blood partition coefficients are relatively uniform across species, Cmax in the blood
can be used as a surrogate.
1.3 PBPK Modelling of Inter-Individual Variability
It is well known that individuals do not respond in the same way following exposure to the same
concentration of chemical under identical conditions. The broad range of susceptibilities to the
biological effects of chemical exposure is due to the heterogeneity of the human population.
Sources of inter-individual variability include exposure, hereditary factors such as enzyme
polymorphisms and physiology, and environmental factors, such as diet and lifestyle. The most
obvious factors affecting uptake, distribution and clearance of lipophilic volatile organic
chemicals (VOCs) are body–build, respiratory ventilation rate, blood lipid content and
metabolic capability (Monster, 1979; Monster and Houtkooper, 1979; Fiserova-Bergerova et al.,
1980; Sato et al., 1991; Allen et al., 1996; Clewell III and Andersen, 1996). The inclusion of
biologically realistic parameters, which help explain the basis of observed data, also affords the
ability to incorporate and estimate the impact of inter-individual differences on chemical
disposition.
1.3.1 Monte Carlo Analysis
The variability of the parameters, which contribute to human inter-individual variability in
susceptibility to chemical exposure, can be estimated using Monte Carlo analysis. The Monte
Carlo method is a form of uncertainty analysis that allows the propagation of uncertainty
through a model, which results in an estimate of the variance in model output. A PBPK model
is run with parameter values sampled from distributions that reflect the observed variation in
each pharmacokinetic parameter in the human population. For example, the coefficients of
variation for various mixed-function oxidase enzymes such as the isoforms of cytochrome P450
can be based on reported variations (Bois et al., 1996; Thomas et al., 1996). The distributions
of partition coefficients for many compounds have been reported (Gargas et al., 1989) and the
2
distributions of physiological and anatomical parameters are also readily available (ICRP, 1975;
Arms and Travis, 1988; Brown et al., 1997). Each time the model is run with a sampled set of
parameter values, effectively representing a single hypothetical human being, the appropriate
dose metric for the toxicity of interest is output. The process is repeated a large number of
times (typically 500 to 1000) to generate a distribution of the dose metric for a simulated
population.
1.4 Incorporating Inter-Individual Variability in the Interpretation of
Biological Monitoring Guidance Values
The Health and Safety Executive has established a system of non-statutory biological
monitoring guidance values (BMGVs) as an aid in the interpretation of BM data (HSE, 1997).
There are two types of guidance value: Health guidance value (HGV) and Benchmark guidance
value (BGV). HGVs are set at a level at which there is no indication from the scientific
evidence available that the substance being monitored can cause harm. On the other hand,
BGVs are not health based; instead they are achievable levels set at the 90th percentile of
available BM results collected from a representative sample of workplaces with good
occupational hygiene practices (HSE, 1997). Therefore the impact of inter-individual
variability on HGVs, which involves reference to a health effect, is equivalent to variability in
BM in general. However, in the establishment of a BMGV no consideration of inter-individual
variability is given although there is an assumption of exposure to an invariant concentration of
substance by the inhalation or dermal route.
1.4.1 Measurement of Variance
Risk assessment and the setting of occupational exposure standards would ideally be based on
quantitative epidemiological studies, which attempt to define dose-response relationships. As
discussed previously, the characterisation of dose is difficult when based on air monitoring.
This is also primarily the reason for the large numbers of subjects, characteristic of most
epidemiological studies, required to achieve the necessary power (Nieuwenhuijsen et al., 2000).
The use of in vitro metabolism data incorporated into ‘population’ PBPK models may offer an
alternative approach. Human health risk assessments for chemicals may be refined to reflect
variability by characterising the expression of enzymes in the liver, as this organ is the major
site of xenobiotic metabolism. This may be achieved by the quantification and extrapolation of
metabolic rate constants derived in vitro, combined with their use as input parameters in a
PBPK model (Lipscomb and Kedderis, 2002). The determination of in vitro metabolic rate
constants in multiple human liver samples can yield valuable qualitative information on human
variance. However, such data must be put into a realistic context, i.e., the intact human, in order
to yield the most valuable predictions of metabolic differences among humans. For quantitative
metabolism data to be most valuable in risk assessment, they must be tied to human anatomy
and physiology, and the impact of their variance evaluated under real exposure scenarios.
Metabolic constants derived in vitro may be extrapolated to the intact liver, when appropriate
conditions are met (Houston, 1994; Kedderis, 1997; Houston and Kenworthy, 2000; Lipscomb
and Kedderis, 2002). The limiting rate of metabolism, Vmax, can be scaled directly to the
concentration of appropriate enzyme contained in the liver. Environmental, genetic and lifestyle
factors can influence the concentration of the cytochrome P450 enzymes in the liver by
affecting the actual expression of the enzyme in the endoplasmic reticulum of the cell or the
amount of endoplasmic reticulum itself expressed (Lipscomb and Kedderis, 2002). The
endoplasmic reticulum spontaneously forms vesicles, known as microsomes, during tissue
homogenisation, which may be isolated by differential centrifugation during sub-cellular
fractionation. Therefore, the efficiency of recovery of the microsomal fraction has a major
impact on the amount of enzyme activity measured (Houston and Carlile, 1997). The
coefficients of variability of the in vitro metabolic rate constants may be included in a PBPK
3
model to investigate the extent to which human inter-individual variability in enzyme activity
can influence the toxification/detoxification routes of xenobiotic metabolism.
1.4.2 The Inherited Basis of Interindividual Differences in Chemical Metabolism
When an allele of a given gene has the same mutation in at least 1% of the population, the gene
is said to exhibit a polymorphism (Boobis, 2002). Polymorphisms can be divided into two
types. (1) Functional polymorphisms represent a change in the DNA sequence which, results in
the altered expression (phenotype) or function of the protein. The presence of functional
polymorphisms results in more than one phenotype for that gene in a species (Boobis, 2002).
(2) Nonfunctional polymorphisms describe those changes in the DNA sequence which have no
effect on the expression or activity of the protein (Boobis, 2002). Inherited differences in
individual drug and chemical metabolising enzymes are typically monogenic traits i.e.,
differences in a single gene (Evans and Johnson, 2001). The influence of such a difference on
the pharmacokinetic and pharmacodynamic effects of a xenobiotic is determined by the
importance of these polymorphic enzymes on the activation and deactivation rates of such
substrates (Evans and Johnson, 2001). Polymorphisms known to affect the pharmacokinetics
and pharmacodynamics of xenobiotic chemicals have been described as pharmacogenetic or
toxicogenetic polymorphisms (Boobis, 2002). It is likely that almost every gene involved in
xenobiotic metabolism is subject to polymorphism. Initially, inherited differences were
discovered following clinical observations of marked differences in drug response (Mahgoub et
al., 1977). However, subsequent studies have demonstrated that phenotypic consequences may
be subtle, such as placing an individual at one end or the other of a distribution of xeniobiotic
metabolism phenotypes, instead of conferring a complete deficiency of encoded enzyme (Evans
and Johnson, 2001). Thus, inactivating polymorphisms can be broadly categorised into two
groups, those that confer complete or near-complete loss of activity of encoded proteins and
those that confer more subtle changes in function via modest changes in expression, regulation,
stability, or catalytic activity, but without complete loss of function (Evans and Johnson, 2001).
1.4.3 Demonstration of Approach Using Styrene as a Model Chemical
Many industrial and environmental chemicals, including carcinogens, may be metabolically
activated and deactivated. Styrene (ST) is such a chemical. In humans, more than 80% of
inhaled ST is taken up and oxidized (activated) by cytochrome P450 2E1 (CYP2E1) and P450
2B6 (Nakajima et al., 1994) to styrene-7,8-oxide (SO) which exists in two enantiomeric forms,
R-SO and S-SO (Delbressine et al., 1981). The enantiomeric forms of SO may be further
hydrolysed (deactivated) by microsomal epoxide hydrolase (mEH) to phenylethylene glycol
with further oxidation by alcohol dehydrogenase and aldehyde dehydrogenase leading
eventually to the urinary metabolites, mandelic acid (MA) and phenylglyoxylic acid (PGA). A
minor route of SO deactivation in humans (about 1% of inhaled ST) is conjugation with
glutathione (GSH) which eventually leads to the appearance in the urine of phenylhydroxyethyl
mercapturic acids (PHEMAs) (Hallier et al., 1995). SO is a known non-genotoxic (Sarangapani
et al., 2002) carcinogen in mice, therefore the disposition of SO following ST metabolism is a
key determinant of the carcinogenic potential of ST. Large differences have been observed in
the rates of activation and deactivation processes in vitro using rat, mouse and human tissues.
Differences in activation/deactivation (A/D) ratios in rats and mice have been shown to
correlate with the risk of cancer from inhalation of butadiene, a chemical metabolised by the
same enzymes as ST. Therefore, the concept of an A/D ratio may be applicable to ST toxicity.
Individuals who have a high A/D ratio (rapid activation and slow deactivation) and are exposed
to high concentrations of ST may have a higher risk of cancer.
The aim of this study was to use ST as a model chemical to:
4
·
Determine in vitro the coefficients of variation of two important routes of xenobiotic
metabolism and their use in a PBPK model to simulate inter-individual variability in vivo.
·
Investigate how inter-individual variability can be used to interpret and set BMGVs.
A number of in vitro techniques employing commercially available human hepatic microsomes,
cytosol and cryopreserved hepatocytes were to be explored for their use in the generation of
kinetic data.
The specific objectives were to:
· determine the coefficient of variation and specific activity of cytochromes P450 2E1 & 1A2
in a battery of human microsomes,
· determine the coefficient of variation and specific activity of GST µ & ș in a battery of
human hepatic cytosol preparations,
· determine the coefficient of variation and specific activity of GST µ & ș in a battery of
human blood samples,
· measure the in vitro rate of oxidative biotransformation of styrene in human microsomes at
physiological pH and temperature,
· measure the in vitro rate of oxidative biotransformation of styrene in human hepatocytes
and GSH conjugation of styrene in the same preparation at physiological pH and
temperature,
· calibrate and validate existing human PBPK models for styrene using available human data,
· to scale up in vitro biokinetic data to in vivo and input into human PBPK models,
· test models with scaled biokinetic data against human data,
· incorporate measured coefficients of variation for metabolic parameters and perform Monte
Carlo uncertainty analysis on PBPK output,
· Propose a range of variation about mean BMGV value for ST.
5
6
2. MATERIALS AND METHODS
2.1 Chemicals, reagents and biological materials
Chlorzoxazone, 6-hydroxychlorzoxazone, 3-cyano-7-ethoxycoumarin and 3-cyano-7hydroxycoumarin (Ultrafine, Manchester, UK). NADP+, glucose-6-phosphate, magnesium
chloride, glucose-6-phosphate dehydrogenase and perchloric acid containing phenacetin
(Sigma-Aldrich, Poole, UK). Styrene (inhibited with 10-15 ppm 4-tert-butylcatechol) styrene
glycol (Sigma Aldrich Gillingham, UK). 1-chloro-2,4-dinitrobenzene (CDNB); 1,2-epoxy-3-(4nitrophenoxy)-propane (ENPP); and trans-4-phenyl-but-3-en-2-one (tPBO) (Acros Organics,
Loughborough, UK). All other chemicals used were of the highest available grade.
Human hepatic microsomes, cytosols, and cryopreserved hepatocytes (Tables 1, 2 and 3) in the
form of liverbeads (LiverbeadsTM) and SoftCell recovery medium were obtained from a
commercial supplier (TCS Cellworks Ltd., Botolph Claydon, UK distributors for Biopredic,
Rennes, France). Two ampoules (batch L1991) containing 4-4.5 % 106 hepatocytes prepared
from a 238g male Sprague Dawley rat with a reported viability of m 85% were obtained from
the same supplier and used initially to practice the procedures before using human hepatocytes.
Table 1
Human hepatic microsome demographics
Microsome
identification
Age
(years)
Sex
Disease status
MIC259002
MIC259006
37
53
M
M
Steatosis
Myocardial
infarction
MIC259007
52
M
MIC259009
74
MIC259018
Total CYP content
(pmol P450/mg
microsomal
protein)
663
434
Microsomal
protein
concentration
(mg/ml)
31.1
31.7
Angioma
539
24.8
M
Liver metastasis
629
34.7
F
-
434
15.3
MIC259015
not
known
62
F
Liver metastasis
659
20.4
MIC259021
45
F
Angioma
571
14.3
MIC259011
59
F
Gall bladder
adenocarcinoma
556
17.2
MIC259016
49
M
-
430
16.3
MIC259017
not
known
F
-
410
19.3
7
Table 2
Human hepatic cytosol demographics
Cytosol identification
Age
Sex
Disease status
I006HL
62
F
Liver metastasis
I019HL
45
F
Angioma
I016HL
52
M
Angioma
53
M
Myocardial infarction
CYT259006‡
CYT259007‡
52
M
Angioma
CYT259008
64
M
Liver metastasis
74
M
Liver metastasis
CYT259009‡
CYT259010
40
F
Benign tumor
‡
cytosol preparations came from the same donors listed in Table 1
Liverbeads
TM
Protein content
(mg/ml)
42.7
36.3
35.7
30.2
28.4
16.3
45.7
15.8
Table 3
immobilized human hepatocytes demographics
Hepatocyte
Identification
Age
Sex
Disease status
LVB005001
L1545
L1247
L1239
HUMA136
HUMA048
77
50
45
43
58
71
M
M
F
F
F
F
Liver metastasis
Liver metastasis
Carcinoma
Benign tumour
Liver metastasis
Liver metastasis
Viability reported by supplier
%
0h
24 h
79
71
82
73
82
73
82
76
86
75
77
63
2.2 Enzyme assays and analysis
2.2.1 Microsomal chlorzoxazone hydroxylation
Duplicate samples were prepared in 0.5 ml 50 mM phosphate buffer (pH 7.4) containing 0.5
mg/ml microsomal protein, 1.3 mM NADP+, 3.3 mM glucose-6-phosphate, 5 mM magnesium
chloride, 0.4 U/ml glucose-6-phosphate dehydrogenase and chlorzoxazone in the concentration
range 0 - 600 µmol/l. Incubations were initiated by addition of chlorzoxazone. Samples were
incubated for 10 minutes at 37˚C in a water bath. The reaction was stopped by addition of 100
µl 6% perchloric acid containing phenacetin as internal standard and mixed by vortexing. Tubes
were centrifuged for 5 minutes at 13000 rpm in a bench-top centrifuge. The supernatant was
transferred into a glass, screw–cap tube and extracted with 2 ml diethyl ether. The ether fraction
was evaporated to dryness under nitrogen and samples were reconstituted in 150 µl 50mM
ammonium acetate.
Standards of 6-hydroxychlorzoxazone were prepared in phosphate buffer at 0 - 5 µg/ml (0 - 26.7
µM) and extracted in 2 ml diethyl ether. The ether fraction was evaporated to dryness under
nitrogen and samples were reconstituted in 150µl 50mM ammonium acetate. Samples were
analysed by HPLC using a Hewlett Packard 1050 fitted with a diode array detector operating at
wavelengths of 287 and 246 nm. 6-Hydroxychlorzoxazone was quantified using 287 nm. A
guard column filled with Bondapak C18 was used. Separation was performed on a Speroclone
ODS(2) 5 µm column (250 mm x 4.6 mm I.D., Phenomenex, Macclesfield, UK) at ambient
temperature. The mobile phase composition was ammonium acetate buffer (0.1 M); acetonitrile;
tetrahydrofuran (70:24.5:5.5) at a flow rate of 1 ml / min.
8
Calibration curves were constructed using the peak height ratios of 6-hydroxychlorzoxazone to
internal standard. Least-squares linear regression analysis (MS Excel software) was used to
determine the slope, intercept and correlation coefficient and to quantify the unknowns. Quality
control, was performed by analysing a 0.5 µg/ml (2.7 µM) standard after every 10 samples. The
acceptance criterion was ± 15% of nominal concentration (2.3 - 3.1 µM).
KM and Vmax for microsomal hydroxylation of chlorzoxazone were determined using a non­
linear regression one-site binding curve fit (GraphPad Prism version 3.00 for Windows,
GraphPad Software, San Diego California USA).
2.2.2 Hepatocytic chlorzoxazone metabolism
LiverbeadsTM are made of freshly isolated hepatocytes that are immobilised in sodium alginate
to form beads approximately 1 mm in diameter (Fremond et al., 1993), which are then frozen
and stored in liquid nitrogen. Microencapsulation protects the cells from freeze-thaw damage
during cryopreservation and subsequent thawing for use. LiverbeadsTM therefore have potential
as an easily transportable and convenient source of hepatocytes for use in in-vitro
pharmacotoxicology. It has previously been shown that isolated hepatocytes retain a wide
variety of hepatic functions when immobilised in alginate, and many of these functions are well
retained following cryopreservation (Guyomard et al., 1996). LiverbeadsTM metabolised probe
substrates at comparative rates to conventional hepatocyte cultures (Hammond et al., 1999).
The probe substrate chlorzoxazone was selected for in-vitro incubations with rat liverbeadsTM in order to: · Gain experience in handling them and because liver metabolism of this compound is
relatively simple and well documented.
· Compare rates of chlorzoxazone metabolism from in-vitro studies with published data.
LiverbeadsTM were handled according to the protocol supplied by Biopredic. An ampoule of
liverbeadsTM was immersed in a water bath at 37˚C until almost completely thawed. Beads were
then removed using a pipette with the tip cut off to avoid damage to the beads and placed in 10
ml SoftCell recovery medium in a 25 ml Sterilin container. Diluted cells were incubated for 2
minutes at room temperature, gently shaking the tube by hand 3 or 4 times during this period,
after which the beads were left to sediment. This typically took 10 - 20 minutes. The
supernatant was removed and 5 ml SoftCell Hep medium was added and the beads rinsed by
gently shaking. This step was repeated to wash the beads. Cell viability was measured using
trypan blue exclusion. Approximately 50 µl drop of bead suspension was removed using a
pipette with a cut down tip and placed onto a 35 mm Petri dish. Excess medium was removed
from the beads. 100 µl trypan blue diluted to 0.05% in SoftCell Hep medium was added and left
at room temperature for 1 minute. Beads were observed under a microscope at %100
magnification. Dead cells take up the dye and appear blue while live cells exclude the dye. The
percentage of live cells was estimated by averaging the number of live and dead cells counted
from 3 different beads (one focus per bead).
Liverbeads were seeded into multiwell plates. Either 24– or 48– well plates were used. All
medium was removed from sedimented cells and 5.4 ml of SoftCell Hep medium was added for
48–well plates. A bead suspension of 0.5 ml was added to each well whilst continually swirling
to keep the beads in suspension. Beads were left to sediment for a few minutes and 300 µl of
medium was removed from each well to give a final bead concentration of approximately 1000
beads / ml. The floor of each well should be almost covered (80–90%) with beads. LiverbeadsTM
were left to pre-incubate for 1 hour in an incubator at 37˚C. Beads were washed by adding 200
µl SoftCell Hep medium to each well, mixing, then removing 200µl medium after allowing
9
beads to sediment. Substrate was prepared at 100-fold the desired final concentration in suitable
solvent. For chlorzoxazone, acetonitrile was used and final concentration in the incubation
mixture ranged from 0–590 µmol/l. Substrate solution (2 µl) was then added to each well and
plates were incubated for 1 hour at 37OC. At the end of the incubation, 100 µl of incubation
medium was removed and internal standard was added. Samples were extracted and analysed
using the same method previously described for microsomes (Section 2.2.1).
2.2.3 3-Cyano-7-ethoxycoumarin O-deethylation
A novel method was developed for measurement of CYP1A2 activity in microsomes using a
continuous fluorimetric assay developed on a Cobas Fara centrifugal analyser (Roche, Welwyn
Garden City, UK). The substrate 3-cyano-7-ethoxycoumarin (CEC) is de–ethylated to 3–cyano–
7–hydroxycoumarin, which is fluorescent at neutral pH values, making it suitable for a
continuous assay. The method described below was based on an original assay described by
White (White, 1988).
Sample buffer was prepared by dissolving 4.2 mg NADPH and 20.3 mg magnesium chloride in
10 ml 0.15 M phosphate buffer (pH 7.4). CEC substrate solution was prepared by adding 45 µl
stock CEC (23.3 mM in DMSO) to 455 µl DMSO. Microsomes were diluted in 0.1 M
phosphate buffer (pH7.4) to an approximate microsomal protein concentration of 4 mg/ml.
Incubations were performed at 37˚C and the reaction was started by addition of CEC. Final
concentrations in the incubation mixture were 30 µM CEC, 0.4 mM NADPH, 8.5 mM
magnesium chloride and 20 µg microsomal protein in a total volume of 350 µl. Production of
the metabolite 3–cyano–7–hydroxycoumarin was followed for 10 minutes using an excitation
wavelength of 408 nm and measuring emission at 450 nm. The assay was calibrated using a
factor, which was determined by measuring the fluorescence of a 53 µM solution of 3–cyano–
7–hydroxycoumarin measured in the same way as a normal sample.
Enzyme kinetic data in human liver microsomes was generated by measuring CEC de–ethylase
activity at different substrate concentrations. The data was analysed as described above for
chlorzoxazone.
2.2.4 Styrene biotransformation
Standard incubation mixtures for the microsomal metabolism of styrene contained the following
components: 0.05M phosphate buffer (pH 7.4); 0.5 mg/ml microsomal protein; 1.3 mM
NADPH; 3.3 mM glucose-6-phosphate; 5 mM magnesium chloride; and 0.4 mg/ml glucose-6phosphate dehydrogenase in a total volume of 0.5 ml. After 3 minutes pre-incubation in a water
bath at 37°C, reactions were initiated by addition of 5µl substrate in acetonitrile and mixed by
inversion. Acetonitrile has been previously shown to cause the least inhibition to cytochrome
P450 activities of the commonly used solvents (Chauret et al., 1998) and does not inhibit
CYP2E1 activity at a concentration of 1%. Substrate was prepared in acetonitrile at 100 times
the desired final concentration in the incubation mixture. Final substrate concentrations in
microsomal incubations were 5-5000 µM for styrene. Duplicate incubations were performed at
each substrate concentration. Microsomal incubations were carried out at 37°C for 20 minutes.
Metabolite production was determined to be linear with respect to both microsomal protein
concentration and incubation time under these conditions at high and low substrate
concentrations. After 20 minutes, incubations were stopped by addition of 50 µl 3M sulphuric
acid, cooled on ice and centrifuged to precipitate proteins. Supernatants were removed for
analysis of metabolite production. The styrene metabolite, styrene glycol was quantified by
HPLC (250 mm x 4.6 mm ID column packed with Speroclone ODS(2) (Phenomenex,
Macclesfield, UK); mobile phase, 10% acetonitrile in water containing 0.25% orthophosphoric
acid; flow rate, 1 ml/min; detection wavelength, 200 nm).
10
KM and Vmax values for microsomal metabolism of styrene were determined using non-linear,
two-site binding regression analysis (GraphPad Prism version 3.00 for Windows, GraphPad
Software, San Diego CA, USA).
2.2.5 Variation in glutathione transferase activity
CDNB is a general GST substrate widely used for quantifying GST activity in tissue extracts. In
a previous study, ENPP activity was not detectable in human liver cytosol (Schulz et al., 2000),
however in rats ENPP is predominantly metabolised by GST Theta (GSTș). tPBO is a specific
substrate for GST Mu (GSTµ).
Cytosols prepared from individual human livers were obtained from TCS Cellworks Ltd. and
were characterised for total protein concentration (Table 2). Spectrophotometric assays to
determine GST activity of human liver cytosols towards CDNB, ENPP, and tPBO were
developed on a Cobas Fara centrifugal analyser (Roche, Welwyn Garden City, UK). Incubation
conditions were based on those previously described by Mannervik and Widersten (Mannervik
and Widersten, 1995) and are outlined in Table 4. All incubations were performed in 0.1 M
sodium phosphate buffer (pH 6.5) at 37˚C. Cytosols were pre-incubated for 2 minutes with
respective substrates before the reaction was initiated by addition of glutathione. Incubation
conditions were chosen so that metabolite production, i.e., the respective glutathione conjugates
of the three substrates, was linear with respect to protein concentration and incubation time.
Table 4
Incubation conditions used for measuring human liver cytosol GST activity towards
individual substrates
Substrate
CDNB
ENPP
tPBO
Substrate
concentration
range (mM)
0.03 - 5
0.005 - 0.5
0.015 - 0.3
GSH
concentration
(mM)
1
5
0.25
Protein
concentration
(µg/ml)
8
200
200
Incubation
time (min)
5
6
8
Detection
wavelength
(nm)
340
360
290
2.2.6 In vitro to in vivo scaling of metabolic rate constants
In order to use in vitro metabolic rate constants to produce quantitative information they must
be input within a biologically realistic framework such as a PBPK model. Scaling of in vitro to
in vivo was achieved by assuming a hepatic microsomal protein yield of 35 mg g liver-1 (Brian
Houston, personal communication) which is similar to the 32.3!2.3 reported by Pacifici et al.,
(Pacifici et al., 1988) although somewhat lower than the 60.1 mg g liver-1 reported by Carlile et
al., (Carlile et al., 1997). However, the value reported by Brian Houston, which is to be
published in the peer reviewed literature, is considered more reliable as it has been determined
from extensive studies conducted over the past few years. Thus, the in vitro rate constants
reported here would be multiplied by the microsomal protein yield to give the number of moles
per gram liver. This value can then be multiplied by the weight of the liver to give a V max per
min per liver. Multiplying by 60 converts this value to moles per hour per liver.
Scaling of cytosolic glutathione transferases was conducted in a similar way except that a
cytosolic protein yield of 46.5 mg per g liver was used (Pacifici et al., 1988).
11
2.2.7 Presentation of variability of Michaelis-Menten constants
It is pertinent to emphasise an important point about the form in which variation in metabolic
rate constants should be presented. The best measure of enzyme efficiency is Kcat/KM,
(Guengerich et al., 1999) where Kcat (min-1) is the intrinsic catalytic activity of an enzyme,
K cat
KM
=
mol min -1 (mg enzyme protein)-1
mol enzyme (mg microsomal protein)-1
This parameter is determined by measuring the actual enzyme concentration using an
immunochemical assay (Guengerich et al., 1999). However, measurement of enzyme
concentration can be difficult, and so, the next best measure of enzyme efficiency, CLint (l h-1 g
liver-1) is most often used. Therefore, the correct way is to express the variability as intrinsic
clearance, CLint, which is the ratio Vmax/KM of any enzyme, when substrate concentrations are
low and Vmax when substrate concentrations are saturating. Reporting the variation in Vmax or
KM, determined in a number of different preparations, in isolation of each other, may in fact
give an erroneous measure of variation. This is because Vmax and KM are inter-related
parameters i.e., KM cannot be determined without Vmax. This means that the lowest KM value
cannot be matched with the highest Vmax , and vice versa, which would give more extreme CLint
values than actually measured. However, a more fundamental flaw in doing this would be the
fact that one would be using constants derived from different graphs.
2.3 The PBPK model for Styrene and styrene oxide
The PBPK model for ST and SO described by Csanády et al. (Csanády et al., 1994) and
Tornero-Velez and Rappaport (Tornero-Velez and Rappaport, 2001) were modified to include
uptake of SO by inhalation and the changes in regional blood flow to organs and tissues which
occur with changes in work rate (WR). Further, the production, distribution and excretion of
mandelic (MA) and phenyglyoxylic (PGA) acids were also added. A diagram of the model is
shown in Figure 1. The systemic circulation distributes ST and SO to the liver (the only
metabolising organ), fat, muscle and the richly perfused tissues. All the organs and tissues were
assumed to be homogeneous, well–mixed compartments with human physiological and
biochemical parameters listed in Table 5. The chemical–dependent physicochemical and
biochemical constants used in the model are shown in Table 6. Simulation of the changes in
regional blood flow occurring at different WRs was achieved by incorporating the appropriate
equations for each organ or tissue (Table 5). The latter equations were obtained by fitting a
trend line to the changes in fraction of cardiac output perfusing a given organ or tissue against
WR in watts (W). These data were adapted from Table 3, p.109 from Fiserova-Bergerova
(1983) (Fiserova-Bergerova, 1983) and shown in Figure 2. The distribution of MA and PGA to
the liver, fat, muscle and richly perfused tissue compartments only were described in the sub­
models for these metabolites. Exhalation was assumed to be negligible because the saturated
vapour pressures of MA and PGA are very low, therefore a lung compartment and blood:air
partition coefficients were considered unnecessary. The liver, fat, muscle and richly perfused
tissue partition coefficients for the MA and PGA sub-models were predicted using the tissue­
composition based algorithm of Baláž and Lukácová (Baláž and Lukácová, 1999). This
algorithm requires Log P (octanol:water) values as inputs. Calculated Log P values for both MA
and
PGA
were
obtained
from
the
SRC
on-line
Log
P
calculator
(http://esc.syrres.com/interkow/kowdemo.htm). The proportions of MA and PGA measured in
the urine relative to total ST metabolism and the first-order urinary excretion rates for the initial
rapid phase (0-20 hr for MA and 0-50 hr for PGA) were obtained from Guillemin and Berode
(Guillemin and Berode, 1988).
12
Styrene
Air
Styrene Oxide
Air
A dashed border indicates
steady state
Phenylglyoxylic acid
QP
Alveolar
PBST
Richly perfused
PRST
Fat
PFST
Muscle
PMST
Liver
PLST
QP
QC
QR
QF
QM
Alveolar
PBSO
Richly perfused
PRSO
Fat
PFSO
Muscle
PMSO
QL
Liver
PLSO
RPT
QC
Fat
Muscle
QR
Liver
QF
Zero order
production of
glutathione
Mandelic acid
RPT
QM
Fat
QL
ST® SO
SO ® ST(OH)2
Liver
GSH
Muscle
Liver
First-order
elimination of
glutathione
ST(OH)SG
Figure 1. Schematic of the styrene, styrene oxide, mandelic and phenylglyoxylic acid PBPK model
13
Table 5. ¶
Work rate induced changes in physiological parameters used in the PBPK model
Physiological parameters
Mathematical relationship describing change in rate due to work
rate (WR)
Values for standard man (70 kg)
Rest
50W
Cardiac output (QC) (l/h)
( QC = QCC ´ BW 0.81 )
QCC =
(- 0.0001´WR ) + (0.0202 ´WR ) + (3.652 ´WR) + 350.4
(70 )
350.4
571
Alveolar ventilation (QP) (l/h)
( QP = QPC ´ BW 0.74 )
QPC =
(- 0.0005 ´WR ) + (0.0852 ´WR ) + (13.544 ´WR) + 347.9
(70 )
347.9
1175.6
17.1
39.3
QLC = (6 ´ 10-6 ´ WR 2 ) - (0.0023 ´ WR ) + 0.2597
90.9
91.2
QTC = (- 2 ´ 10-5 ´ WR 2 ) + (0.0058 ´ WR ) + 0.2514
88.1
280.6 QRC = 1 - (QFC + QLC + QTC )
154.2
159.9 Organ and tissue Perfusion rates
Fat (QF) (l/h)
( QF = QFC ´ BW )
Liver (QL) (l/h)
( QL = QLC ´ BW )
Muscle tissue (QT) (l/h)
( QT = QTC ´ BW )
Richly perfused (QR) (l/h) ( QR = QRC ´ BW )
3
2
0.81
3
2
0.74
QFC = (- 4 ´ 10-6 ´ WR 2 ) + (0.0006 ´ WR ) + 0.0489
Compartment volumes as a fraction of body weight (BW) Fat
0.19
Liver
0.026
¶
Richly perfused
Muscle
Adapted from (Fiserova-Bergerova, 1983)
14
0.05
0.62
Table 6
Physiological and biochemical parameters and partition coefficients used in the model
Enzyme
Symbol
Value (mean±SD)
Population
Distributiona
Units
(high
Vmaxmo
0.0024±0.001b
U
mmol/h/g liver
(low
KMmo
Vmax2
0.014±0.006b
0.0069±0.002b
U
U
mmol/l
mmol/h/g liver
KM2
Vmaxeh
KMih
KMapp
VmaxGST
KMGSH
KMSO
0.691±0.20b
0.0045±0.0025c
0.001±0.0002d
0.01±0.002d
0.028±0.0084e
0.1±0.02d
2.5±0.05d
U
U
U
U
L
U
U
mmol/l
mmol/h/g liver
mmol/l
mmol/l
mmol/h/g liver
mmol/l
mmol/l
CGSH0
5.9f
N
mmol/l
KdGSH
0.2±0.06d
U
l/h
MA
KMA
0.074-0.177
U
/h
PGA
KPGA
0.058-0.076
U
/h
MA
PMA
0.05-0.56
U
PGA
PPGA
0.05-0.33
U
Cytochrome
affinity)
Cytochrome
affinity)
P450
P450
Epoxide hydrolase
Glutathione transferase
Initial concentration
cytosolic GSH
of
Elimination rate constant
for GSH turnover
Urinary excretion ratesI
Proportion of MetabolismI
Partition coefficients
Blood:air
Fat:blood
Liver:blood
Richly
pefused:blood
Muscle:blood
STh
48
SOh
93.8
2.71
2.60
2370g
6.1
2.6
2.6
1.96
1.5
a
MAJ
1.99
1.03
1.03
1.33
1.33
1.06
1.15
(N=normal, U=uniform)
Measured in vitro c
Coefficient of variation taken from (Fennell and Brown, 2001) d
Coefficient of variation taken from (Tornero-Velez and Rappaport, 2001) e
Variability measured in vitro using 1-chloro-2,4-dinitrobenzene f
Coefficient of variation not available g
(Tornero-Velez and Rappaport, 2001)
h
(Csanády et al., 1994)
I
(Guillemin and Berode, 1988)
b
15
PGAJ
6.96
J
Calculated (Baláž and Lukácová, 1999)
Fraction of Cardiac
Output
0.75
0.50
0.25
0.00
0
50
100
150
200
Work Rate
(Watts)
Figure 2. Change in rate of perfusion with changing work rate. The symbols represent the
fraction of cardiac output perfusing various organs and tissues (adapted from (FiserovaBergerova, 1983)) as follows: (Ŷ) (QFC) fat compartment, (ɰ) (QLC) liver, (ż) (QTC)
muscle compartment.
The changes in cardiac output (QCC) and alveolar ventilation (QPC) rates were calculated in
a similar manner (data not shown). To ensure the model did not violate mass balance the flow
rate to the richly perfused compartment (QRC) was set equal to the difference between the
cardiac output and the flows to the remaining tissues (Table 5). The fitting of trend lines was
conducted in Microsoft EXCEL 97 SR-2 (Microsoft Corporation). The model was coded in
Berkeley Madonna version 8.0.1 (http://www.berkeleymadonna.com/) for initial visualisation
and exercising. For Monte Carlo analysis the model was coded in MCSIM version 4.2.0
(http://www.ibiblio.org/pub/gnu/mcsim/).
2.3.1 Model validation
The PBPK model was validated using published measured in vivo data on ST and SO
pharmacokinetics. The data sets for ST were taken from Ramsey and Young and Wigaeus et
al (Ramsey and Young, 1978; Wigaeus et al., 1983) and for SO from Korn et al (Korn et al.,
1994). The latter were occupational exposure data where time-weighted average exposure
concentrations of ST in ambient air of 10 to 73 ppm were measured during periods of 248-325
min (4.13-5.42 h). In order to generate model predictions for this data, various exposure
concentrations ranging from 10-73 ppm over a period of 4.13 and 5.42 h were simulated.
Linear regression analysis of the predicted data was performed in GraphPad Prism version 3
(San Diego, CA 92121 USA).
Urinary excretion data for MA and PGA were obtained from internal HSL data, some of
which was published in the peer-reviewed literature (Wilson et al., 1983). The MA data was
obtained from 4 volunteers, aged 45, 45, 39 and 43 weighing 68.9, 64.4, 75 and 86 kg
exposed to 50 ppm styrene in the HSL controlled atmosphere exposure facility for 6 h. Urine
samples were taken every hour during and following exposure for up to 14 h. Data were
expressed as mol per hour. The PGA data was obtained from 2 volunteers, aged 45 and 43
weighing 64.4 and 86 kg exposed to 50 ppm styrene in the HSL controlled atmosphere
16
exposure facility for 10 h. Urine samples were taken every hour during and following
exposure for up to 19 h. Data were expressed as mol per hour.
2.3.2 Monte Carlo analysis
The PBPK model was modified for Monte Carlo analysis. The anatomical, physiological,
biochemical and physicochemical parameters comprising the PBPK model were assigned
distributions (Table 6). The basic PBPK model structure was modified to permit re-sampling
of model parameters from within the assigned distributions. Each run essentially generates a
hypothetical human being exposed to the same exposure conditions where model output
should reflect inter–individual variability.
The variation in metabolic rate constants reported as mean±standard deviation (SD) (Table 6)
were determined as follows: the high and low affinity cytochrome P450 rate constants were
measured in vitro as reported in Section 2.2.4. The SD for EH was calculated by multiplying
the Vmaxeh value of 0.0045 mmol h-1 g-1 by the in vitro CV of 56% measured by Fennell and
Brown (Fennell and Brown, 2001). The SD for KMih and KMapp was calculated by multiplying
the values reported by Csanády et al (Csanády et al., 1994) by the CVs reported by TorneroVelez and Rappaport (Tornero-Velez and Rappaport, 2001). The SD for VmaxGST was
calculated by multiplying the value of 0.028 mmol h-1 g-1 reported by Csanády et al (Csanády
et al., 1994) by the CV determined in vitro using 1-chloro-2,4-dinitrobenzene. In the absence
of appropriate information on allelic frequencies and the corresponding expression levels of
enzyme polymorphisms, uniform distributions were used, in most cases, in order to capture
the possibility that some individuals will have little or no enzyme activity. In such cases the
uniform distribution was set from 0 to mean+SD.
Variation in the urinary excretion of MA and PGA, the two major metabolites of ST, during
and after constant exposure to 50 ppm ST for 6 and 10 h in the HSL controlled atmosphere
exposure facility was investigated.
Sensitivity analysis of the model determined that Vmaxmo, for the high affinity CYP450
metabolism (probably CYP2E1) was the most important parameter in the activation of ST,
The
and Vmaxeh, the most important parameter in the deactivation of SO.
activation/deactivation (A/D) ratio is the rate of activation of a parent chemical to a toxic
entity versus the rate of deactivation of the toxin to a non-toxic entity. In this case it is the
ratio of the intrinsic clearance of activation versus the intrinsic clearance of deactivation at
low exposure concentrations, which changes to Vmaxmo versus (Vmaxeh + VmaxGST) as exposure
concentrations approach saturation of the ezymes involved. Therefore, the A/D ratio was
expressed as the net clearance of SO. This was obtained by calculating the intrinsic
clearances (CLint) i.e., Vmax/KM for each enzyme system described in the model. Therefore,
Clint for the bioactivation of ST to SO by both high and low affinity CYP450 activity
(Clmo=Vmaxmo/KMmo+Vmax2/KM2), SO to phenylethylene glycol (Cleh) and the gutathione
conjugation of SO (Clgst) were calculated by importing the model parameters obtained from
the Monte Carlo analysis into Microsoft Excel 2000 where the appropriate calculations were
effected. The net clearance, Clmo-(Cleh+Clgst) was plotted against area-under-the-curves of
venous SO concentrations (SOAUC). This would allow the investigation of the frequency and
consequences of various A/D ratios with regard to SOAUC, and therefore, potentially the
numbers of people at higher risk in any given size of simulated cohort. In this exercise each
run essentially generates a hypothetical human being exposed to 50, 100, 200 ppm ST for 8 h
per day, 5 days per week with a work load ranging from 40-100 W. A work rate of 50 W is
about equivalent to a moderate stroll. 3000 runs were performed generating a hypothetical
population of 3000 different people. The dose metric chosen as an output from the model was
the area-under-the-curve of venous concentration of SO after an 8 h daily exposure, 5 days a
week (SOAUC).
17
18
3. RESULTS
3.1 Variation in enzyme activity
3.1.1 Kinetics of chlorzoxazone 6-hydroxylation in human liver microsomes
Chlorzoxazone was incubated with human liver microsomes from 10 individuals and the
production of 6-hydroxychlorzoxazone measured as an estimate of cytochrome P450 2E1
activity. Operating under linear conditions with respect to incubation time and microsomal
protein concentration, the formation of 6-hydroxychlorzoxazone was readily detectable at
substrate concentrations of 3 µM. Rate of formation against substrate concentration is shown
in Figure 3.
nmol 6-OHCZX/
min/mg protein
5
4
3
2
1
0
0
100 200 300 400 500 600 700
Chlorzoxazone (m mol/l)
Figure 3. Hydroxylation of chlorzoxazone by 10 human liver microsomes (mean ± SEM;
n=3).
In the 10 individual human liver microsomes used in this study, the mean KM was 45 µM
(range 25–73; CV = 36%) and mean Vmax was 1.536 nmol 6-hydroxychlorzoxazone minute-1
mg microsomal protein-1(range 0.690 - 4.833; CV = 83%). Inter-assay variability was
assessed using results obtained for QC samples. Mean QC concentration over 10 runs was 2.9
µmol l-1 (SD = 0.35) (Table 7). The method CV was calculated to be 12%. The variation in
Vmax was 7.03 fold whereas CLint varied 3.36 fold (Table 7).
3.1.2 Kinetics of chlorzoxazone 6-hydroxylation in rat cryopreserved
hepatocytes
Percentages of live cells recovered from thawed rat LiverbeadsTM ranged from 20 to 65%.
Typically, cell viabilities obtained were around 60%. One experiment gave very low viability
(20%) and the results from this preparation were discounted.
It was assumed that using the Biopredic protocol, 48 well plates contained 333,333 cells per
well since one vial containing 4 million hepatocytes was sufficient for 12 wells. This number
was multiplied by the percentage of live cells measured by trypan blue exclusion in order to
give the number of viable cells per well. Typically, this number was approximately 200,000
live cells per well. 0.2 ml incubation mixtures in each well consisted of 0.05 ml beads and
0.15 ml medium. From the concentration of 6-hydroxychlorzoxazone (6-OHCZX) measured
in each sample of incubation medium, the amount of 6-OHCZX in 0.15 ml was calculated.
This was the amount of 6-OHCZX produced per hour by the number of live cells in the well.
19
6-OHCZX
-1
(nmoles 100,000 viable hepatocytes
-1
hour )
This concentration was corrected to give the amount of 6-OHCZX nmoles (100,000
hepatocytes)-1 hour-1. Rates of metabolism of chlorzoxazone to 6-hydroxychlorzoxazone by
rat LiverbeadsTM are presented in figure 4.
2.5
2.0
1.5
62% viability
20% viability
1.0
65% viability
0.5
0.0
0
100
200
300
400
500
600
700
-1
Chlorzoxazone (mmol l )
Figure 4. Metabolism of chlorzoxazone by rat LiverbeadsTM in 3 separate incubations (mean ±
SEM; n = 2 or 3).
6-OH-CZX
-1
(nmoles 100,000 viable hepatocytes
-1
hour )
Data from one of these incubations were fitted to a non-linear regression model to predict
Michaelis-Menten kinetic parameters in Figure 5. Using this curve fit, the predicted value for
KM was 38 µmol l-1 and Vmax was 1.63 nmoles 6-OHCZX (100,000 hepatocytes)-1 hour-1.
2.0
1.5
1.0
0.5
0.0
0
100
200
300
400
500
600
-1
Chlorzoxazone (m mol l )
Figure 5. Metabolism of chlorzoxazone by rat LiverbeadsTM fitted to the Michaelis-Menten
equation. Individual data points are shown.
20
Table 7
Summary of variation in metabolic rate constants and parameters for chlorzoxazone and styrene biotransformation
Microsomal
Preparation
MIC259011
MIC259017
MIC259006
MIC259018
MIC259007
MIC259021
MIC259015
MIC259009
MIC259016
MIC259002
Mean
SD
Variation (% fold)
Chlorzoxazone hydroxylation
KM
Vmax
(nmol/min/mg)
(µmol/l)
0.6872
34.76
0.8374
39.01
0.8447
34.9
0.913
24.61
1.016
54.48
1.087
39.19
1.245
73.39
1.27
42.6
2.618
39.24
4.833
72.7
1.53
45.5
1.28
16.3
7.03
-
CLint
(l/min)
1.98%10-5
2.15%10-5
2.42%10-5
3.71%10-5
1.86%10-5
2.77%10-5
1.69%10-5
2.98%10-5
6.67%10-5
6.65%10-5
3.29%10-5
1.87%10-5
3.36
Styrene high affinity metabolism
Vmax
KM
CLint
(nmol/min/mg)
(µmol/l)
(l/min)
1.043
9.77
10.68%10-5
0.5927
8.71
6.80%10-5
0.9664
22.8
4.23%10-5
1.172
12.52
9.36%10-5
0.7538
10.75
7.01%10-5
0.9175
9.295
9.87%10-5
1.118
12.79
8.74%10-5
1.265
16.5
7.67%10-5
2.483
23.13
10.74%10-5
4.929
59.95
8.22%10-5
1.524
18.6
8.33%10-5
1.301
15.5
2.00%10-5
2.57
2.53
21
Styrene low affinity metabolism
Vmax
KM
CLint
(nmol/min/mg)
(µmol/l)
(l/min)
2.551
602.9
4.23%10-6
2.659
590.9
4.5%10-6
4.709
583.9
8.06%10-6
3.279
787.2
4.16%10-6
5.013
700
7.16%10-6
3.06
484.2
6.32%10-6
2.485
601.6
4.13%10-6
3.689
696.8
7.16%10-6
2.435
1170
2.08%10-6
1.097
3203
3.42%10-7
3.098
942.5
4.63%10-6
1.153
816.4
2.29%10-6
4.57
23.5
3.1.3 Kinetics of 3–cyano–7–ethoxycoumarin O–de–ethylation in human liver
microsomes
3-Cyano-7-ethoxycoumarin
-1
O-deethylase activity (pmol min mg
-1
protein )
Contrary to rat liver microsomes, human liver microsomes appeared to exhibit product
inhibition at 30 µmol l-1 3–cyano–7–ethoxycoumarin. Therefore, this concentration was
omitted from kinetic calculations. Rates of formation of 3–cyano–7–hydroxycoumarin against
substrate concentration for 10 individual human liver microsome preparations are shown in
Figure 6.
1500
1250
1000
750
500
250
0
0.0
2.5
5.0
7.5
10.0
12.5
15.0
17.5
3-Cyano-7-ethoxycoumarin (m mol/l)
Figure 6. O–de–ethylation of 3–cyano–7–ethoxycoumarin by 10 human liver microsomes
(mean ± SEM; n=3).
The mean KM for the 10 individual liver microsome preparations was 8.2 µmol l-1 (range 5.2 10.4; CV = 19%) and the mean Vmax was 1168 pmol 3–cyano–7–ethoxycoumarin min-1 mg
microsomal protein-1 (range 536–2213; CV = 42%). Variation in Vmax was 4.13 fold and CLint
was 3.45 fold (Table 8).
22
Table 8
Summary of variation in metabolic rate constants and parameters for 3-cyano-7ethoxycoumarin O-deethylation
Microsomal
Preparation
MIC259016
MIC259007
MIC259017
MIC259011
MIC259006
MIC259015
MIC259002
MIC259018
MIC259009
MIC259021
Mean
SD
Variation (% fold)
Vmax
(nmol/min/mg)
0.5328
0.9309
0.8338
1.139
1.172
1.167
0.7813
1.141
1.723
2.201
1.16
0.483
4.13
KM
(µmol/l)
6.936
10.25
8.088
9.459
9.433
9.273
5.178
6.498
8.663
8.302
8.208
1.57
-
CLint
(l/min)
7.68%10-5
9.08%10-5
1.03%10-4
1.20%10-4
1.24%10-4
1.26%10-4
1.51%10-4
1.76%10-4
1.99%10-4
2.65%10-4
1.43E-04
5.68E-05
3.45
3.1.4 Biotransformation of styrene in human liver microsomes
Styrene was incubated with human liver microsomes from 10 individuals and the production
of styrene glycol measured as an estimate of cytochrome P450 2E1 activity. Operating under
linear conditions (5-5000 µmol l-1, for 20 minutes, 0.5 mg ml-1 protein) with respect to
incubation time and microsomal protein concentration, the formation of styrene glycol was
readily monitored. Rate of formation against substrate concentration is shown in Figure 7. In
vitro ST kinetics were described using a 2-binding site hyperbola equation consistent with
high and low affinity metabolism. Microsomal preparations MIC259002 and MIC259016
were more satisfactorily described using a single-binding site equation. MIC259002 gave
KM=59.95 µmol l-1 and Vmax=4.929 nmol min-1 mg-1 and MIC259016 gave KM=23.13 µmol l-1
and Vmax= 2.483 nmol min-1 mg-1.
Styrene glycol
-1
(pmol min mg microsomal
-1
protein )
6000
5000
4000
3000
2000
MIC259002
1000
MIC259016
0
0
500
1000
1500
2000
2500
5000
-1
Styrene (m mol l )
Figure 7. Hydroxylation of styrene by 10 human liver microsomes (mean ± SEM; n=3).
The Vmax values for CHZ and high affinity ST biotransformation determined in the same
microsomal preparations were matched and plotted in order to investigate potential
correlation (Figure 8).
23
Styrene high affinity Vmax
(nmol/min/mg)
6
5
4
3
2
1
0
0
1
2
3
4
5
6
Chlorzoxazone hydroxylation
Vmax (nmol/min/mg)
Figure 8 Correlation of Vmax for CHZ hydroxylation and high affinity ST biotransformation
(r2 =0.9732 Sy.x=0.2259).
However, if the highest Vmax values for both CHX and ST biotransformation are excluded
then the correlation deteriorates (Figure 9).
Styrene high affinity Vmax
(nmol/min/mg)
2.0
1.5
1.0
0.5
0.0
0.00
0.25
0.50
0.75
1.00
1.25
Chlorzoxazone hydroxylation Vmax
(nmol/min/mg)
1.50
Figure 9 Correlation of Vmax for CHZ hydroxylation and high affinity ST biotransformation:
Exclusion of two highest values. (r2 =0.1625, Sy.x=0.2194).
High affinity styrene versus chlorzoxazone CLint also shows no significant correlation (Figure
10).
24
-1
-1
Styrene CLint (l h g liver)
0.3
0.2
0.1
0.0
0.000
0.025
0.050
0.075
0.100
-1
0.125
0.150
-1
Chlorzoxazone CLint (l h g liver)
Figure 10. Correlation of CLint for the high affinity component of styrene biotransformation
with the CLint for the biotransformation of the CYP2E1 probe substrate, chlorzoxazone (r2
=0.1070, Sy.x=0.04220).
It interesting to note that the Eadie-Hofstee plots for the determination of Vmax for ST
biotransformation in the microsomal samples giving the two highest values suggest that the
contribution to metabolism is predominantly from a single enzyme, whereas those of the other
microsomal samples indicate that there are at least two. A linear decrease in rate of reaction
versus clearance indicates that one enzyme is operating MIC259002 (Figure 11b), whereas
MIC259007 (Figure 11a) shows two different phases, indicating a contribution from two
enzymes to metabolism.
25
a) MIC259007
v (rate of reaction)
-1
-1
(pmol min mg )
7500
5000
2500
0
0
10
20
30
40
50
60
v/[ST] ( ´ 10-6 l min-1)
b) MIC259002
v (rate of reaction)
-1
-1
(pmol min mg )
7500
5000
2500
0
0
10
20
30
40
50
60
70
v/[ST] (´ 10-6 l min-1)
Figure 11. Eadie-Hofstee plots for the rate of production of styrene glycol from styrene in a)
human microsomal preparation MIC259007 which is representative of 8 from 10 of the
microsomal preparations and b) MIC259002 which represents the one of the two preparations
with very high activity.
26
Although the main focus of the study was on cytochrome P450 2E1, which is an inducible
enzyme and considered to be the high affinity component of ST metabolism, the MichaelisMenten rate constants for the low affinity component were also determined. It is interesting
to note that the variation in CLint measured for the low affinity component was 23.5, which is
higher than that of 2.53 observed for the high affinity component (Table 7). The enzyme(s),
which comprise the low affinity component of ST metabolism, are not known at this time
although preliminary results indicate a contribution from more than one enzyme.
3.1.5 Kinetics of glutathione transferase
3.1.5.1 Glutathione transferase activity towards 1-chloro-2,4-dinitrobenzene
The mean Vmax of glutathione transferase activity to CDNB was 441.1!186.5 nmol min-1 mg
cytosolic protein-1 (range 156.9-757 nmol min-1 mg cytosolic protein-1; CV=42.3%) (Figure
12). The mean KM was 399.4!241.7 µmol l-1. Vmax varied 3.93 and CLint 4.94–fold (Table 9).
800
CDNB-glutathione
(nmol/min/mg protein)
700
600
500
400
300
200
100
0
0
1000
2000
3000
4000
5000
6000
CDNB (mM)
Figure 12. Glutathione conjugation with broad-spectrum substrate CDNB. The mean rate of
glutathione conjugation was measured in 8 human cytosolic preparations.
3.1.5.2 Glutathione transferase-ș activity towards 1,2-epoxy-3-(4nitrophenoxy)-propane
The mean catalytic activity of GST-ș activity to ENPP was 3265.51!1427 pmol min-1 mg
cytosolic protein-1 (range 1977-5846 pmol min-1 mg cytosolic protein-1; CV=43.7%) (Figure
13). The mean KM was 19.8!8 µmol l-1. Vmax varied 2.96 and CLint 6.61–fold (Table 9).
27
7000
ENPP-glutathione
pmol /min/ mg protein
6000
5000
4000
3000
2000
1000
0
0
100
200
300
400
500
600
ENPP (mM)
Figure 13. GST-ș activity. GST-ș activity was measured by determining the rate of
glutathione conjugation with ENPP. The mean rate of glutathione conjugation was measured
in 8 human cytosolic preparations.
3.1.5.3 Glutathione transferase activity towards trans-4-phenyl-but-3-en-2-one
The mean catalytic activity of GST-µ activity to tPBO was 1609.4!1861 pmol min-1 mg
cytosolic protein-1 (range 18.24-4966 pmol min-1 mg cytosolic protein-1; CV=115.7%) (Figure
14). The mean KM was 125.1!107.7 µmol l-1. Vmax varied 272 and CLint 141–fold (Table 9)
(Figure 14).
tPBO-glutathione
pmol/min/mg protein
4000
3000
2000
1000
0
0
50
100
150
200
250
300
350
tPBO (mM)
Figure 14. GST-µ activity. GST-µ activity was measured by determining the rate of
glutathione conjugation with tPBO. The mean rate of glutathione conjugation was measured
in 8 human cytosolic preparations.
28
Table 9
Summary of variation in metabolic rate constants and parameters for glutathione conjugation
Cytosolic
Preparation
CYT6B
CYT16
CYT19
CYT10
CYT9
CYT7
CYT8
CYT6
Mean
SD
Variation (% fold)
1-chloro-2,4-dinitrobenzene
(CDNB)
KM
CLint
Vmax
(nmol/min/mg)
(µmol/l)
(l/min)
439.8
893.6
4.92%10-4
500.9
518.2
9.67%10-4
370.8
343.3
1.08%10-4
304.7
215.6
1.41%10-4
528
359.7
1.47%10-4
616.9
398.9
1.55%10-3
757
401.3
1.89%10-3
156.9
64.46
2.43%10-3
441.13
399.4
1.41%10-3
186.5
241.7
5.91%10-4
3.93
4.94
1,2-epoxy-3-(4-nitrophenoxy)-propane
(ENPP)
Vmax
KM
CLint
(nmol/min/mg)
(µmol/l)
(l/min)
5.846
13.64
4.29%10-4
2.99
22.54
1.33%10-4
2.388
14.61
1.63%10-4
1.977
30.45
6.49%10-4
5.183
32.69
1.59%10-4
2.599
18.86
1.38%10-4
2.567
13.86
1.85%10-4
2.574
11.76
2.19%10-4
3.27
19.8
1.86%10-4
1.43
8.05
1.08%10-4
2.96
6.61
29
trans-4-phenyl-but-3-en-2-one
(tPBO)
Vmax
KM
CLint
(nmol/min/mg) (µmol/l)
(l/min)
0.01824
62.06
2.94%10-7
1.872
95.91
1.95%10-5
0.1201
133
9.03%10-7
0.1036
54.33
1.91%10-6
0.129
11.96
1.1%10-5
2.235
158.8
1.41%10-5
3.431
364.7
9.41%10-5
4.966
120.1
4.13%10-5
1.61
125.1
1.23%10-5
1.86
107.7
1.36%10-5
272
141
3.2 Validation of the PBPK model
The PBPK model output was very similar to that reported by Csanády et al (Csanády et al.,
1994). Model output for simulation of ST and SO kinetics under controlled experimental and
occupational exposure conditions are shown in figures 15a & b and 15c & d, respectively.
Styrene in arterial blood
(mol L-1)
1.0×10 -04
1.0×10 -05
1.0×10 -06
1.0×10 -07
0
2
4
6
Time
1.0×10 -05
-1
Styrene in venous blood (mol L )
Figure 15a. Concentration-time course of ST in arterial blood of male volunteers (mean body
weight of 69 kg) during and after exposure to 69 ppm ST for 2 h during light physical exercise
(50 W work load) (Wigaeus et al., 1983). The symbols represent the mean of seven subjects.
7.5×10 -06
5.0×10 -06
2.5×10 -06
0.0×10 -00
0
10
20
30
40
50
Time
Figure 15b. Concentration-time course of ST in venous blood of male volunteers (mean body
weight of 83 kg) during and after a 6 h inhalation exposure to 80 ppm ST (Ramsey and Young,
1978). The symbols represent the mean of four subjects.
30
Styrene-7,8-oxide in venous
blood (mg/l)
7.5
5.0
2.5
0.0
0
25
50
75
100
125
Styrene in ambient (ppm)
Figure 15c. Correlation of styrene-7,8-oxide concentration in venous blood with styrene
concentration in ambient air. ST concentrations in ambient air were given as time-weighted
averages. Venous blood concentrations of styrene-7,8-oxide were determined at the end of the
exposure period. Each dot represents a measured value for a single person (Korn et al., 1994).
The solid line represents the model predictions at various reported time points and ambient
exposure concentrations. The broken lines represent the 95% confidence interval for the
predicted mean values of SO (See Materials and Methods section 2.3.1 Model validation).
Styrene oxide in venous blood
-1
(predicted) (mg l )
10.0
7.5
5.0
2.5
0.0
0
1
2
3
4
5
6
Styrene oxide in venous blood (measured) (m g l-1)
31
7
Figure 15d. Correlation of measured versus predicted styrene-7,8-oxide concentration in venous
blood. Venous blood concentrations of styrene-7,8-oxide were determined at the end of the
exposure period. Each dot represents a measured value for a single person (Korn et al., 1994).
The solid line was calculated by regression analysis through the origin: y = (0.8313 ± 0.07096)x,
Sy.x = 0.8002. The broken lines represent the 95% confidence interval for the predicted mean
values of SO (See Materials and Methods section 2.3.1 Model validation).
3.3 Incorporation of inter-individual variability
The urinary excretion data for MA and PGA, the two main metabolites of ST biotransformation,
were expressed as mole per hour excreted rather than moles per mol creatinine. The reason for
this is because creatinine normalisation, more often than not, increases rather than decreases
variability in urinary data. Figures 16a and 16b show that experimentally determined values for
urinary excretion of both MA and PGA fall within the envelope bounded by the predicted
maximum and minimum rates of excretion.
The venous concentration of SO versus the A/D ratio is shown in figure 17. Positive ratios or
differences correspond to the range where the rate of activation of ST to SO exceeds the rate of
deactivation of SO. It is interesting to note that the higher venous concentrations of SO
observed in some people could not solely be explained by the absolute value of the A/D
difference.
Rate of urinary excretion of mandelic
acid (mol h-1 )
1.0×10 -03
7.5×10 -04
5.0×10 -04
2.5×10 -04
0.0×10 -00
0
2
4
6
8
10
12
14
16
18
20
Time
Figure 16a. Urinary excretion of Mandelic Acid. Four volunteers were exposed to 50 ppm ST
for 6 h in the HSL controlled atmosphere exposure facility. Urine samples were taken every
hour up to 14 h after the beginning of exposure. The symbols represent measured
concentrations of MA and the solid lines are model simulations.
32
Rate of urinary excretion of
phenylglyoxylic acid (mol h-1 )
4.0×10 -04
3.0×10 -04
2.0×10 -04
1.0×10 -04
0.0×10 -00
0
2
4
6
8
10
12
14
16
18
20
Time
Figure 16b. Urinary excretion of Phenylglyoxylic Acid. Two volunteers were exposed to 50
ppm ST for 10 h in the HSL controlled atmosphere exposure facility. Urine samples were taken
every hour up to 19 h after the beginning of exposure. The symbols represent measured
concentrations of MA and the solid lines are model simulations.
Venous concentration of SO
(mol L-1 t) (AUC)
1.0×10 -06
7.5×10 -07
0
50
100
200
5.0×10 -07
2.5×10 -07
0.0×10 -00
-1.50 -1.25 -1.00 -0.75 -0.50 -0.25
0.00
0.25
0.50
Activation/Deactivation Ratio
Clmo-(Cleh+Clgst)
Figure 17. Venous concentration of styrene oxide versus the Activation/Deactivation ratio for
styrene metabolism. Each symbol represents the model prediction for a single person. 3000
iterations, simulating a cohort of 3000 people, were performed at 0, 50, 100 and 200 ppm
exposure concentrations to ST with a work rate varying from 40-100 W. The vertical dotted
line at 0 along the x axis corresponds to the point at which the rate of activation is equal to the
rate of deactivation. A positive A/D ratio reflects the situation where the rate of production of
SO exceeds the rate of removal. Calculation of A/D ratios is described in section 2.3.2 Monte
Carlo analysis.
33
34
4. DISCUSSION
4.1 Variation in enzyme activity
4.1.1 Cytochrome P450 2E1
CYP2E1 is constitutively expressed in the liver, where the highest concentration is found in the
centrilobular region, specifically in the four to five layers of hepatocytes surrounding the
terminal hepatic venules (Ingelman-Sundberg et al., 1988). In general, CYP2E1 accounts for
approximately 8% of the total P450 in liver (Shimada et al., 1994). CYP2E1 is an inducible
enzyme with a complex regulation governing variability (Johansson et al., 1990). CYP2E1
activities, as measured by the rate of chlorzoxazone hydroxylation in this study, were similar to
previously reported values in the literature. Peter et al (Peter et al., 1990) reported Vmax values
between 0.58 and 2.3 nmol 6-hydroxychlorzoxazone/min/mg microsomal protein and an
apparent KM of 39 ± 7 µmol/l (mean ± SD) in human liver microsomes from 14 individuals. In
the 10 individual human liver microsomes used in this study, two individuals had significantly
higher Vmax values than the other eight. This would be expected as CYP2E1 is highly inducible
(Kadlubar and Guengerich, 1992). Importantly, the lowest rate of chlorzoxazone hydroxylation
measured in this study (0.69 nmol/min/mg) was very close to the lowest rate reported by Peter
et al (Peter et al., 1990). This would be expected assuming both of these rates correspond to
uninduced enzyme expression. Iyer and Sinz (Iyer and Sinz, 1999) reported a 4.8 fold variation
in the Vmax for CYP2E1 using p-nitrophenol as probe substrate which compares favourably with
variation in Vmax and CLint calculated from Peter et al (Peter et al., 1990) which were 5.36 and
3.75 fold respectively, and with the 7.03 and 3.36, respectively obtained in this study (Table 7).
The highest activity measured and presented in Figure 3 was obtained with microsome
preparation MIC259002, which came from a donor with steatosis; a condition characterised by
fatty liver, and known to induce the activity of CYP2E1. Although the disease status of the
other microsomal preparation with high CYP2E1 (MIC259016) activity was not reported, it is
likely also to have come from a donor with a condition which induced this enzyme (Table 1).
The inter-batch CV of the method is acceptable. However, the value of 12% only accounts for
analytical variation. Some degree of experimental error would also be expected in the
microsomal incubations, although the use of a standard protocol should minimise these errors.
This is supported by the low standard errors of replicate data points. Knowing the total error
(incubation and analytical) would be a useful parameter to help interpret variability of the
kinetic parameters measured. An estimate of this error could be achieved by replicating the
incubation of microsomes from the same batch with a fixed substrate concentration.
Cell viabilities obtained from rat LiverbeadsTM were lower than manufacturers specification of
at least 70%. Hammond et al (Hammond et al., 1999) reported cell viabilities between 74 and
95% in their studies using LiverbeadsTM. They also reported wider variability in viability of
LiverbeadsTM compared to freshly isolated cells. This is probably largely due to day-to-day
variations in the thawing and processing of the LiverbeadsTM. Thawing and dilution of
LiverbeadsTM needs to be as rapid as possible to prevent damage to the cells by the
cryopreservant, dimethyl sulphoxide. Typical cell viability of thawed LiverbeadsTM was 60–
65% in this study. With the exception of one preparation, viability was fairly constant (>60%).
Therefore, LiverbeadsTM appear to be suitable for use in future metabolism studies. Cell
viability may improve with experience of handling LiverbeadsTM.
In a comparative study of chlorzoxazone hydroxylation across a range of species using liver
microsomes (Court et al., 1997), Michaelis-Menten kinetic parameters were similar for both
human and rat. It this study, human microsomal chlorzoxazone hydroxylation was shown to
35
have a KM of 45 µmol/l and a Vmax of 1.536 nmol/min/mg microsomal protein. It may therefore
be expected that similar results would arise using the same incubation conditions for rat liver
microsomes. Given this assumption, KM for chlorzoxazone hydroxylation by rat liverbeads was
38 µmol/l. Since this reaction is catalysed by CYP2E1, then it is purely microsomal. Therefore
it is reasonable to expect that KM should be similar in both microsomes and intact hepatocytes.
This is supported by the experimental data. Similarly, Vmax for chlorzoxazone hydroxylation
could be expected to be similar in both microsomes and intact hepatocytes. However, the two
experimental systems cannot be directly compared because Vmax is expressed in different units.
However, the use of scaling factors allows data from the two systems to be compared by scaling
each Vmax to a common term, based on liver weight. Using scaling factors determined by Carlile
et al (Carlile et al., 1997) the two in-vitro systems are compared in Table 10.
Table 10
Comparison of Vmax determined experimentally for chlorzoxazone hydroxylation in
human liver microsomes and rat liverbeads.
Human liver
microsomes
Vmax
1.536
nmol min-1 mg microsomal
protein-1
Rat
liverbeads
1.63
nmol 6-OHCZX 100,000
hepatocytes-1 hour-1
a
Scaling factors taken from (Carlile et al., 1997).
Scaling factora
60.1
mg protein-1
g liver-1
Corrected Vmax
5539
nmol hour-1
g liver-1
109%106
hepatocytes g liver-1
1777
nmol hour-1
g liver-1
This demonstrates that Vmax for chlorzoxazone determined in rat liverbeads was of similar
magnitude to Vmax determined in human liver microsomes. Scaling factors have a significant
impact on the final value obtained. (A value of 60.1 mg protein per g liver was used in this case,
however, further work on microsomal protein recovery during subcellular fractionation and
differential centrifugation indicates values of 35-45 mg microsomal protein per g liver are more
common (B Houston, personal communication)). It would be preferable to use scaling factors
determined in-house for our own incubation systems, however this crude method of comparison
still demonstrates that Vmax for chlorzoxazone hydroxylation calculated from rat liverbeads is
approximately what would be expected. This finding further supports the use of cryopreserved
hepatocytes in the form of liverbeads in in-vitro metabolism studies.
Cryopreserved hepatocytes, in the form of LiverbeadsTM, have been successfully used in this
study. They are a convenient source of hepatocytes and are the only practical one since the
scarcity and unpredictable availability of fresh liver prevents routine preparation of fresh
hepatocytes. However, the process of alginate encapsulation and cryopreservation is fairly new
and there is therefore limited data comparing cryopreserved hepatocytes to freshly prepared
hepatocytes. In particular, it is important to determine whether cryopreservation alters
expression or activity of metabolic enzymes and also if preservation or recovery media contain
compounds that could inhibit metabolic enzyme activity. A few studies have attempted to
investigate the effects of cryopreservation (Guyomard et al., 1996; Hammond et al., 1999; Li et
al., 1999; Rialland et al., 2000), however this is an area which requires substantial further
research.
36
4.1.2 Cytochrome P450 1A2
CYP1A2 expression is largely restricted to the liver, where it represents, on average, about 12
% of total hepatic P450 content (Shimada et al., 1994). It is involved in the metabolism of a
large number of environmental and dietary chemicals, in some cases converting them into toxic
or carcinogenic intermediates (Guengerich, 1988; Guengerich and Shimada, 1991; Guengerich
et al., 1999). CYP1A2 is inducible by many environmental agents such as polycyclic aromatic
hydrocarbons (Sesardic et al., 1990), polyhalogenated aromatic hydrocarbons (Lake et al.,
1996), dioxins (Xu et al., 2000) and indoles contained in cruciferous vegetables (Schrenk et al.,
1998). Evidence for genetic polymorphism in human CYP1A2 has been presented based upon
in vivo caffeine N3–demethylation pharmacokinetics (Kadlubar et al., 1992; Lang et al., 1994).
Some of the 40-fold variation observed may be related to the regulation of expression levels
(Guengerich et al., 1999).
The kinetics of 3–cyano–7–ethoxycoumarin O–de–ethylation was investigated in human liver
microsomes as a measure of CYP1A2 variability. The KM measured in this study is close to the
KM reported by Crespi et al (Crespi et al., 1997) of 3.5 µmol/l using cDNA expressed CYP1A2.
Unfortunately, Vmax data reported by Crespi et al (Crespi et al., 1997) could not be converted to
the units used in this study for comparison. A 4– and 3.45–fold variation in Vmax and CLint,
respectively, were measured in this study, which differ markedly from the 180–fold in vitro
variation (Sesardic et al., 1988) and 6–fold in vivo variation (Schrenk et al., 1998) in CYP1A2–
dependent activity reported elsewhere. The choice of substrate however does determine the
extent of variation in activity measured. For example, in non-smokers CYP1A2–dependent
caffeine metabolic ratio varies 6–fold (Schrenk et al., 1998), whereas the CLint of phenacetin
varies approximately 10–fold (Kahn et al., 1985). Also, substantial intra–individual variation of
the CYP1A2 urinary metabolic ratio has also been noted (Nordmark et al., 1999).
4.1.3 Styrene biotransformation
The major biotransformation pathway of ST is via oxidation to SO which is present in two
enantiomeric forms (R–SO and S–SO). Many in vitro studies have been conducted to examine
the enzymatic processes involved in this process. The in vitro conversion of ST to SO is
catalysed by CYP2E1 (Guengerich et al., 1991; Shimada et al., 1994; Kim et al., 1996)
although there is evidence that other isoforms, particularly CYP2B6, are also involved at higher
concentrations (Nakajima et al., 1994; Kim et al., 1997) including, more recently, CYP2F1
(Cruzan et al., 2002). The data obtained in this study are consistent with previous reports
which indicate that there are other isoenzymes of CYP P450 contributing to ST metabolism e.g.,
a 2–binding site hyperbola equation to fit the data (Figure 7) and the Eadie–Hofstee plots
showing the difference between one and two-enzyme activity (Figure 11). Attempts to correlate
ST high affinity Vmax with CHZ Vmax showed that in most cases (8 from 10) there was a very
weak correlation (Figure 9). The two microsomal preparations with the highest Vmax showed
very good agreement with CHZ metabolism (Table 7), which together with the Eadie–Hofstee
plot (Figure 11b), suggests that CYP2E1 content and contribution to ST biotransformation,
relative to other CYP isoforms, was high. This result supports the view that the measurement of
variability in metabolism should be compound specific i.e., variability measured using a probe
substrate cannot be transferred to another CYP2E1 substrate (Guengerich et al., 1999; Gentry et
al., 2002). This is because most substrates are metabolised by more than one enzyme whereas a
probe substrate may be metabolised predominantly by a single enzyme. It is interesting to note
that the KM values for low–affinity ST biotransformation for microsomal preparations
MIC259016 and MIC259002 were very much higher, and the CLint values lower, than in the
other 8 preparations suggesting that the amount of low-affinity isoenzymes available to
contribute to ST metabolism was lower in these microsomes.
37
4.1.4 Kinetics of glutathione transferase
Glutathione transferases (GSTs) function as detoxification enzymes and are generally
considered to play a prominent role in cellular defence against electrophilic chemical species, of
endogenous as well as xenobiotic origins. Human GSTs can be divided into two main
categories: membrane-bound and soluble GSTs, with the latter often referred to a cytosolic
GSTs although nuclear and mitochondrial forms are included in the soluble category. The
soluble GSTs are subdivided into classes based on sequence similarities that reflect evolutionary
branches from an ancestral protein (Mannervik, 1985). The classes are designated by the names
of Greek letters: a (alpha), l (mu), o (pi), h (theta) etc., and are abbreviated A, M, P, T, and so
on. However, all GST polymorphisms so far reported are limited to the soluble GSTs and it is
for this reason that this category has received most interest.
Human tissues, like the rat, show differential expression of the multiple soluble forms of GST.
Most tissues have different isoenzyme profiles. In the liver, alpha class GSTs, in particular
GST A1–1, are predominant and may represent 2-3% of total cytosolic protein (van Ommen et
al., 1990; Rowe et al., 1997). GST M1–1 and GST T1–1 are also major hepatic forms, in
individuals expressing these isoenzymes (Sherratt et al., 1998; Thier et al., 1998; Landi, 2000;
Sherratt et al., 2002; Ross and Pegram, 2003). GST P1–1, which is present in most human
tissues, is not expressed in normal adult hepatocytes, and its presence in the liver appears to be
restricted to epithelial cells such as those of the bile duct (Kano et al., 1987; Cowell et al., 1988;
Islam et al., 1989; Morrow et al., 1989).
GST M1–1, GST M3–3, GST T1–1 AND GST P1–1 have well-established polymorphisms in
the human population. GST M1–1 and GST T1–1 both have a frequently occurring null
phenotype (Warholm et al., 1980; Board, 1981; Pemble et al., 1994). Approximately half the
human population is homozygous for the GSTM1 null allele and consequently does not express
the GST M1–1 protein. The GSTT1 null phenotype is due to the deletion of the complete
GSTT1 gene (Pemble et al., 1994).
In this study three glutathione transferase substrates were used to measure the in vitro variation
in activity of these enzymes. However, all three substrates are preferentially metabolised by
various GST enzymes to varying degrees. CDNB is considered to be a non-specific substrate
for alpha-, mu-, and pi-class GSTs (Takamatsu and Inaba, 1994) whereas tPBO is more specific
for GST M1–1 and ENPP for GST T1–1 (Hayes and Pulford, 1995). The results listed in Table
9 show that measured variability is dependent upon substrate used, which presents difficulties
when trying to draw general conclusions. CDNB and ENPP show similar variation where the
first reflects the activity of alpha, mu and pi class GSTs and the second is believed to be more
specific for GST T1–1. The activities measured using tPBO suggests that about 50% of the
samples came from GST M1 null phenotypes, which is consistent with reported frequencies in
various ethnic groups (58% in Chinese, 52% in English, 48% in Japanese and 43% in French
nationals) (Hayes and Pulford, 1995). This finding supports previous results with other enzyme
systems, which suggest that reliable measurement of variability can only be compound specific.
Of course, this may well constitute a major obstacle to the determination of generic enzyme
variability data for use in chemical risk assessment.
4.2 Incorporating inter-individual variability into PBPK models
The incorporation of human inter-individual variation in anatomical, physiological, biochemical
and physicochemical parameters has been achieved in this study using Monte Carlo sampling
linked to PBPK modelling. The form of the data required is the mean and standard deviation of
each parameter, which implies that much existing data can be used in this approach. The results
indicate that the measurement of variation in in vitro parameters may adequately reflect the
variation observed in in vivo data, although more simulations of in vivo data than are presented
38
in this study are required to further support this observation. However, the power of PBPK
modelling has been typically demonstrated by the integration of such a wide range of
parameters i.e., biological, chemical and statistical into a single platform i.e., a PBPK model.
Specifically, this process represents a means of encapsulating as much of the existing
knowledge of a given chemical, in this case styrene, and using it to understand, quantitatively
and qualitatively, the behaviour of the chemical in the body. Altogether, this should provide a
more transparent and justifiable approach to chemical risk assessment. However, the study also
highlighted a problem with the form in which some biological monitoring data is expressed, in
particular the rate of urinary excretion of a marker expressed against the rate of urinary
excretion of creatinine. Normalisation of urinary excretion data against creatinine has always
been controversial and is dependent on the assumption that the rate of urinary creatinine
excretion is constant. However, in reality, urinary creatinine excretion is variable to a
considerable degree (Harris et al., 2000; LeBeau et al., 2002). HSL routinely measures urinary
creatinine and most individual data varies up to 50%. Therefore, a parameter such as creatinine
would introduce variability into a PBPK model. The consequence of this would be to obscure
the variability of other parameters, which determine the pharmacokinetics of the chemical under
study, and make the quantification of variance difficult.
PBPK models can be exercised in many different ways, such as, simulating different work
patterns, constant and pulsatile exposure concentrations, and the effects of exercise. In this
study we also investigated the effects of variation in the activation and deactivation enzymes of
ST metabolism with different work rates. Although this was predominantly a theoretical
exercise the assumptions made were realistic. It is reasonable to assume that different
individuals will have different metabolic capacities governed by the phenotypic expression of
relevant enzymes. One extreme may be those individuals with highly active CYP2E1 and low
mEH activity; the opposite extreme may be low CYP2E1 activity and high mEH, with most
people having A/D ratios in between. Individuals with high CYP2E1 and low mEH activity
would rapidly transform ST to SO but slowly remove SO. This would imply that such
individuals would be more susceptible to ST toxicity. Figure 16 shows that relatively few
people, less than 100 in 3000, would have a reduced potential to deactivate SO even after
exposure to 200 ppm ST for 8 hr per day 5 days per week. This figure also indicates that the
A/D difference, which is based on CLint, alone cannot explain the causes underlying the higher
venous SO concentrations predicted in these few individuals. The absolute values of Vmax better
correlate with the ability to metabolise chemicals at the higher exposure concentrations, which
is consistent with classical enzyme behaviour. Exposure concentrations of 100-200 ppm are
very high and unlikely to be encountered in the work place in developed countries. A more
accurate model is being developed which can estimate the concentrations of SO in the cells of
the lung where tumours are observed following chronic exposure of mice to ST. In addition, the
model will be capable of incorporating specific distributions and expression levels of enzyme
polymorphisms according to known allelic frequencies. Such a model could be used to interpret
epidemiological data.
39
40
5. CONCLUSIONS Commercially available human hepatic microsomes and cytosols are ideal for the determination
of variation in chemical metabolism. Commercially available cryopreserved human hepatocytes
show promise although further development is required in thawing and cell viability protocols
in order to improve reproducible yields of viable cells. Variability measured using probe
substrates cannot be transferred to other chemicals because probe substrates are by definition
almost always exclusively metabolised by one enzyme, whereas other chemicals often have
contributions from more than one enzyme. Enzyme activity should be expressed as intrinsic
clearance, CLint, at sub-saturating and Vmax, at saturating, substrate concentrations. Variability in
the urinary excretion of phenylglyoxylic and mandelic acids in human volunteers exposed to
styrene was successfully simulated using in vitro metabolism data in a PBPK model. However
expression of urinary biomarkers against creatinine increases, rather than decreases, variability,
which confounds biological monitoring strategies. Finally, in vitro data were used in a Monte
Carlo approach to generate a population PBPK model for describing the variability in the ratio
of the activation of styrene to the suspect, non-genotoxic carcinogen styrene oxide, relative to
the deactivation of styrene oxide. Population models such as this could help improve
epidemiological studies.
41
42
REFERENCES Allen, B. C., Covington, T. R., and Clewell, H. J. (1996). Investigation of the impact of
pharmacokinetic variability and uncertainty on risks predicted with a pharmacokinetic
model for chloroform. Toxicology 111, 289-303.
Arms, A. D., and Travis, C. C. (1988). Reference physiological parameters in pharmacokinetic
modeling. Office of Health and Environmental Assesment, U.S. EPA, Washington, D.C.
NTIS PB88-196019.
Baláž, t., and Lukácová, V. (1999). A model-based dependence of the human tissue/blood
partition coefficients of chemicals on lipophilicity and tissue composition. Quantitative
Structure-Activity Relationships 18, 361-368.
Board, P. G. (1981). Biochemical genetics of glutathione-S-transferase in man. American
Journal of Human Genetics 33, 36-43.
Bois, F. Y., Jackson, E. T., Pekari, K., and Smith, M. T. (1996). Population toxicokinetics of
benzene. Environmental Health Perspectives 104 Suppl 6, 1405-1411.
Boobis, A. R. (2002). Genetic variation in the human population. In Variability and
Susceptibility in Human Response to Occupational Exposure to Chemicals in the UK
(C. Courage, L. Levy, J. Hughes, and J. Emeny, Eds.), pp. 92-99. Institute for
Environment and Health, Leicester.
Brown, R. P., Delp, M. D., Lindstedt, S. L., Rhomberg, L. R., and Beliles, R. P. (1997).
Physiological parameter values for physiologically based pharmacokinetic models.
Toxicology and Industrial Health 13, 407-484.
Carlile, D. J., Zomorodi, K., and Houston, B. J. (1997). Scaling factors to relate drug metabolic
clearance in hepatic microsomes, isolated hepatocytes and the intact liver: Studies with
induced livers involving diazepam. Drug Metabolism and Disposition 25, 903-911.
Chauret, N., Annick, G., and Nicoll-Griffith, D. A. (1998). Effect of common organic solvents
on in vitro cytochrome P450-mediated metabolic activities in human liver microsomes.
Drug Metabolism and Disposition 26, 1-4.
Clewell III, H. J., and Andersen, M. E. (1996). Use of physiologically based pharmacokinetic
modeling to investigate individual versus population risk. Toxicology 111, 315-329.
Court, M. H., Von Moltke, L. L., Shader, R. I., and Greenblatt, D. J. (1997). Biotransformation
of chlorzoxazone by hepatic microsomes from humans and ten other mammalian
species. Biopharmaceutics & Drug Disposition 18, 213-226.
Cowell, I. G., Dixon, K. H., Pemble, S. E., Ketterer, B., and Taylor, J. B. (1988). The structure
of the human glutathione S-transferase pi gene. Biochemical Journal 255, 79-83.
Crespi, C. L., Miller, V. P., and Penman, B. W. (1997). Microtiter plate assays for inhibition of
human, drug-metabolizing cytochromes P450. Analytical Biochemistry 248, 188-190.
Cruzan, G., Carlson, G., Johnson, K., Andrews, L., Banton, M., Bevan, C., and Cushman, J.
(2002). Styrene Respiratory Tract Toxicity and Mouse Lung Tumors Are Mediated by
CYP2F-Generated Metabolites. Regulatory Toxicology and Pharmacology 35, 308.
Csanády, G. A., Mendrala, A. L., Nolan, R. J., and Filser, J. G. (1994). A physiologic
pharmacokinetic model for styrene and styrene-7,8-oxide in mouse, rat and man.
Archives of Toxicology 68, 143-157.
Delbressine, L. P., Van Bladeren, P. J., Smeets, F. L., and Seutter-Berlage, F. (1981).
Stereoselective oxidation of styrene to styrene oxide in rats as measured by mercapturic
acid excretion. Xenobiotica 11, 589-594.
Droz, P. O., Berode, M., and Wu, M. M. (1991). Evaluation of concomitant biological and air
monitoring. Applied Occupational and Environmental Hygiene 6, 465-474.
Evans, W. E., and J ohnson, J. A. (2001). Pharmacogenomics: the inherited basis for
interindividual differences in drug response. Annual Review of Genomics and Human
Genetics 2, 9-39.
43
Fennell, T. R., and Brown, C. D. (2001). A physiologically based pharmacokinetic model for
ethylene oxide in mouse, rat, and human. Toxicology and Applied Pharmacology 173,
161-175.
Fiserova-Bergerova, V. (Ed.) (1983). Modeling of inhalation exposure to vapors: uptake,
distribution and elimination. CRC Press, Boca Raton.
Fiserova-Bergerova, V., Vlach, J., and Cassady, J. C. (1980). Predictable individual differences
in uptake and excretion of gases and lipid soluble vapours simulation study. British
Journal of Industrial Medicine 37, 42-49.
Fremond, B., Malandain, C., Guyomard, C., Chesne, C., Guillouzo, A., and Campion, J. P.
(1993). Correction of bilirubin conjugation in the Gunn rat using hepatocytes
immobilized in alginate gel beads as an extracorporeal bioartificial liver. Cell
Transplant 2, 453-460.
Gargas, M. L., Burgess, R. J., Voisard, D. E., Cason, G. H., and Andersen, M. E. (1989).
Partition coefficients of low-molecular weight volatile chemicals in various liquids and
tissues. Toxicology and Applied Pharmacology 98, 87-99.
Gentry, P. R., Hack, C. E., Haber, L., Maier, A., and Clewell, H. J., 3rd (2002). An approach for
the quantitative consideration of genetic polymorphism data in chemical risk
assessment: examples with warfarin and parathion. Toxicological Sciences 70, 120-139.
Guengerich, F. P. (1988). Roles of cytochrome P-450 enzymes in chemical carcinogenesis and
cancer chemotherapy. Cancer Research 48, 2946-2954.
Guengerich, F. P., Kim, D. H., and Iwasaki, M. (1991). Role of human cytochrome P-450 IIE1
in the oxidation of many low molecular weight cancer suspects. Chemical Research in
Toxicology 4, 168-179.
Guengerich, F. P., Parikh, A., Turesky, R. J., and Josephy, P. D. (1999). Inter-individual
differences in the metabolism of environmental toxicants: cytochrome P450 1A2 as a
prototype. Mutation Research 428, 115-124.
Guengerich, F. P., and Shimada, T. (1991). Oxidation of toxic and carcinogenic chemicals by
human cytochrome P-450 enzymes. Chemical Research in Toxicology 4, 391-407.
Guillemin, M. P., and Berode, M. (1988). Biological monitoring of styrene: a review. American
Industrial Hygiene Association Journal 49, 497-505.
Guyomard, C., Rialland, L., Fremond, B., Chesne, C., and Guillouzo, A. (1996). Influence of
alginate gel entrapment and cryopreservation on survival and xenobiotic metabolism
capacity of rat hepatocytes. Toxicology and Applied Pharmacology 141, 349-356.
Hallier, E., Goergens, H. W., Karels, H., and Golka, K. (1995). A note on individual differences
in the urinary excretion of optical enantiomers of styrene metabolites and of styrene­
derived mercapturic acids in humans. Archives of Toxicology 69, 300-305.
Hammond, A. H., Garle, M. J., and Fry, J. R. (1999). A preliminary comparison of liverbeadsTM
with a conventional rat hepatocyte culture preparation: some aspects of xenobiotic
metabolism and related toxicity. Alternatives To Laboratory Animals 27, 111-119.
Harris, S. A., Purdham, J. T., Corey, P. N., and Sass-Kortsak, A. M. (2000). An evaluation of
24-hour urinary creatinine excretion for use in identification of incomplete urine
collections and adjustment of absorbed dose of pesticides. American Industrial Hygiene
Association Journal 61, 649-657.
Hayes, J. D., and Pulford, D. J. (1995). The glutathione S-transferase supergene family:
regulation of GST* and the contribution of the isoenzymes to cancer chemoprotection
and drug resistance. Critical Reviews in Biochemistry and Molecular Biology 30, 445­
600.
Houston, J. B. (1994). Relevance of in vitro kinetic parameters to in vivo metabolism of
xenobiotics. Toxicology in Vitro 8, 507-512.
Houston, J. B., and Carlile, D. J. (1997). Prediction of hepatic clearance from microsomes,
hepatocytes, and liver slices. Drug Metabolism Review 29, 891-922.
44
Houston, J. B., and Kenworthy, K. E. (2000). In vitro-in vivo scaling of CYP kinetic data not
consistent with the classical Michaelis-Menten model. Drug Metabolism and
Disposition 28, 246-254.
HSE (1997). Biological Monitoring in the Workplace: A guide to its practical application to
chemical exposure. Health and Safety Executive.
ICRP (1975). Report of the Task Group on Reference Man. Pergamon Press, New York.
Ingelman-Sundberg, M., Johansson, I., Penttila, K. E., Glaumann, H., and Lindros, K. O.
(1988). Centrilobular expression of ethanol-inducible cytochrome P-450 (IIE1) in rat
liver. Biochemical and Biophysical Research Communications 157, 55-60.
Islam, M. Q., Platz, A., Szpirer, J., Szpirer, C., Levan, G., and Mannervik, B. (1989).
Chromosomal localization of human glutathione transferase genes of classes alpha, mu
and pi. Human Genetics 82, 338-342.
Iyer, K. R., and Sinz, M. W. (1999). Characterization of Phase I and Phase II hepatic drug
metabolism activities in a panel of human liver preparations. Chemical and Biological
Interactions 118, 151-169.
Johansson, I., Lindros, K. O., Eriksson, H., and Ingelman-Sundberg, M. (1990). Transcriptional
control of CYP2E1 in the perivenous liver region and during starvation. Biochemical
and Biophysical Research Communications 173, 331-338.
Kadlubar, F. F., Butler, M. A., Kaderlik, K. R., Chou, H. C., and Lang, N. P. (1992).
Polymorphisms for aromatic amine metabolism in humans: relevance for human
carcinogenesis. Environmental Health Perspectives 98, 69-74.
Kadlubar, F. F., and Guengerich, F. P. (1992). Inducibility of human cytochromes P-450
primarily involved in the activation of chemical carcinogens. Chemosphere 25, 201­
204.
Kahn, G. C., Boobis, A. R., Brodie, M. J., Toverud, E. L., Murray, S., and Davies, D. S. (1985).
Phenacetin O-deethylase: an activity of a cytochrome P-450 showing genetic linkage
with that catalysing the 4-hydroxylation of debrisoquine? British Journal of Clinical
Pharmacology 20, 67-76.
Kano, T., Sakai, M., and Muramatsu, M. (1987). Structure and expression of a human class pi
glutathione S-transferase messenger RNA. Cancer Research 47, 5626-5630.
Kedderis, G. L. (1997). Extrapolation of in vitro enzyme induction data to humans in vivo.
Chemical and Biological Interactions 107, 109-121.
Kim, H., Wang, R. S., Elovaara, E., Raunio, H., Pelkonen, O., Aoyama, T., Vainio, H., and
Nakajima, T. (1997). Cytochrome P450 isozymes responsible for the metabolism of
toluene and styrene in human liver microsomes. Xenobiotica 27, 657-665.
Kim, R. B., Yamazaki, H., Chiba, K., O'Shea, D., Mimura, M., Guengerich, F. P., Ishizaki, H.,
Shimada, N., and Wilkinson, G. R. (1996). In vivo and in vitro characterization of
CYP2E1 activity in Japanese and Caucasians. Journal of Pharmacology and
Experimental Therapeutics 279, 4-11.
Korn, M., Gfrorer, W., Filser, J. G., and Kessler, W. (1994). Styrene-7,8-oxide in blood of
workers exposed to styrene. Archives of Toxicology 68, 524-527.
Lake, B. G., Charzat, C., Tredger, J. M., Renwick, A. B., Beamand, J. A., and Price, R. J.
(1996). Induction of cytochrome P450 isoenzymes in cultured precision-cut rat and
human liver slices. Xenobiotica 26, 297-306.
Landi, S. (2000). Mammalian class theta GST and differential susceptibility to carcinogens: a
review. Mutation Research 463, 247-283.
Lang, N. P., Butler, M. A., Massengill, J., Lawson, M., Stotts, R. C., Hauer-Jensen, M., and
Kadlubar, F. F. (1994). Rapid metabolic phenotypes for acetyltransferase and
cytochrome P4501A2 and putative exposure to food-borne heterocyclic amines increase
the risk for colorectal cancer or polyps. Cancer Epidemiology and Biomarkers and
Prevention 3, 675-682.
45
LeBeau, M. A., Christenson, R. H., Levine, B., Darwin, W. D., and Huestis, M. A. (2002).
Intra- and interindividual variations in urinary concentrations of endogenous gamma­
hydroxybutyrate. Journal of Analytical Toxicology 26, 340-346.
Li, A. P., Gorycki, P. D., Hengstler, J. G., Kedderis, G. L., Koebe, H. G., Rahmani, R., de
Sousas, G., Silva, J. M., and Skett, P. (1999). Present status of the application of
cryopreserved hepatocytes in the evaluation of xenobiotics: consensus of an
international expert panel. Chemical and Biological Interactions 121, 117-123.
Lipscomb, J. C., and Kedderis, G. L. (2002). Incorporating human interindividual
biotransformation variance in health risk assessment. Science of the Total Environment
288, 13-21.
Mahgoub, A., Idle, J. R., Dring, L. G., Lancaster, R., and Smith, R. L. (1977). Polymorphic
hydroxylation of Debrisoquine in man. Lancet 2, 584-586.
Mannervik, B. (1985). The isoenzymes of glutathione transferase. Advances in Enzymology and
Related Areas of Molecular Biology 57, 357-417.
Mannervik, B., and Widersten, M. (1995). Human glutathione transferases: classification, tissue
distribution, structure, and functional properties. In Advances in Drug Metabolism in
Man (G. M. Pacifici, and G. N. Fracchia, Eds.), pp. 407-459. European Commission,
Luxembourg.
Mason, H., and Wilson, K. (1999). Biological Monitoring: The role of toxicokinetics and
physiologically based pharmacokinetic modeling. American Industrial Hygeine
Association Journal 60, 237-242.
Monster, A. C. (1979). Difference in uptake, elimination and metabolism in exposure to
trichloroethylene, 1,1,1-trichloroethane, and tetrachloroethylene. International Archives
of Occupational and Environmental Health 42, 311-317.
Monster, A. C., and Houtkooper, J. M. (1979). Estimation of individual uptake of
trichloroethylene, 1,1,1-trichloroethane and tetrachloroethylene from biological
parameters. International Archives of Occupational and Environmental Health 42, 319­
323.
Morrow, C. S., Cowan, K. H., and Goldsmith, M. E. (1989). Structure of the human genomic
glutathione S-transferase-pi gene. Gene 75, 3-11.
Nakajima, T., Elovaara, E., Gonzalez, F. J., Gelboin, H. V., Raunio, H., Pelkonen, O., Vainio,
H., and Aoyama, T. (1994). Styrene metabolism by cDNA-expressed human hepatic
and pulmonary cytochromes P450. Chemical Research in Toxicology 7, 891-896.
Nieuwenhuijsen, M. J., Toledano, M. B., and Elliott, P. (2000). Uptake of chlorination
disinfection by-products; a review and a discussion of its implications for exposure
assessment in epidemiological studies. Journal of Exposure Analysis and
Environmental Epidemiology 10, 586-599.
Nordmark, A., Lundgren, S., Cnattingius, S., and Rane, A. (1999). Dietary caffeine as a probe
agent for assessment of cytochrome P4501A2 activity in random urine samples. British
Journal of Clinical Pharmacology 47, 397-402.
Pacifici, G. M., Franchi, M., Bencini, C., Repetti, F., Di Lascio, N., and Muraro, G. B. (1988).
Tissue distribution of drug-metabolizing enzymes in humans. Xenobiotica 18, 849-856.
Pemble, S., Schroeder, K. R., Spencer, S. R., Meyer, D. J., Hallier, E., Bolt, H. M., Ketterer, B.,
and Taylor, J. B. (1994). Human glutathione S-transferase theta (GSTT1): cDNA
cloning and the characterization of a genetic polymorphism. Biochemical Journal 300,
271-276.
Peter, R. M., Böcker, R., Beaune, P. H., Iwasaki, M., Guengerich, F. P., and Yang, C. S. (1990).
Hydroxylation of chlorzoxazone as a specific probe for human liver cytochrome P­
450IIE1. Chemical Research in Toxicology 3, 566-573.
Ramsey, J. C., and Young, J. D. (1978). Pharmacokinetics of inhaled styrene in rats and
humans. Scandinavian Journal of Work and Environmental Health suppl. 2, 84-91.
46
Rialland, L., Guyomard, C., Scotte, M., Chesne, C., and Guillouzo, A. (2000). Viability and
drug metabolism capacity of alginate-entrapped hepatocytes after cryopreservation. Cell
Biology and Toxicology 16, 105-116.
Ross, M. K., and Pegram, R. A. (2003). Glutathione Transferase Theta 1-1-Dependent
Metabolism of the Water Disinfection Byproduct Bromodichloromethane. Chemical
Research in Toxicology 16, 216-226.
Rowe, J. D., Nieves, E., and Listowsky, I. (1997). Subunit diversity and tissue distribution of
human glutathione S-transferases: interpretations based on electrospray ionization-MS
and peptide sequence-specific antisera. Biochemical Journal 325, 481-486.
Sarangapani, R., Teeguarden, J. G., Cruzan, G., Clewell, H. J., and Andersen, M. E. (2002).
Physiologically based pharmacokinetic modeling of styrene and styrene oxide
respiratory-tract dosimetry in rodents and humans. Inhalation Toxicology 14, 789-834.
Sato, A., Endoh, K., Kaneko, T., and Johanson, G. (1991). A simulation study of physiological
factors affecting pharmacokinetic behaviour of organic solvent vapours. British Journal
of Industrial Medicine 48, 342-347.
Schrenk, D., Brockmeier, D., Morike, K., Bock, K. W., and Eichelbaum, M. (1998). A
distribution study of CYP1A2 phenotypes among smokers and non-smokers in a cohort
of healthy Caucasian volunteers. European Journal of Clinical Pharmacology 53, 361­
367.
Schulz, T. G., Wiebel, F. A., Thier, R., Neubert, D., Davies, D. S., and Edwards, R. J. (2000).
Identification of theta-class glutathione S-transferase in liver cytosol of the marmoset
monkey. Archives of Toxicology 74, 133-138.
Sesardic, D., Boobis, A. R., Edwards, R. J., and Davies, D. S. (1988). A form of cytochrome
P450 in man, orthologous to form d in the rat, catalyses the O-deethylation of
phenacetin and is inducible by cigarette smoking. British Journal of Clinical
Pharmacology 26, 363-372.
Sesardic, D., Pasanen, M., Pelkonen, O., and Boobis, A. R. (1990). Differential expression and
regulation of members of the cytochrome P450IA gene subfamily in human tissues.
Carcinogenesis 11, 1183-1188.
Sherratt, P. J., Manson, M. M., Thomson, A. M., Hissink, E. A., Neal, G. E., van Bladeren, P. J.,
Green, T., and Hayes, J. D. (1998). Increased bioactivation of dihaloalkanes in rat liver
due to induction of class theta glutathione S-transferase T1-1. Biochemical Journal 335,
619-630.
Sherratt, P. J., Williams, S., Foster, J., Kernohan, N., Green, T., and Hayes, J. D. (2002). Direct
comparison of the nature of mouse and human GST T1-1 and the implications on
dichloromethane carcinogenicity. Toxicology and Applied Pharmacology 179, 89-97.
Shimada, T., Yamazaki, H., Mimura, M., Inui, Y., and Guengerich, F. P. (1994). Interindividual
variations in human liver cytochrome P450 enzymes involved in the oxidation of drugs,
carcinogens and toxic chemicals: studies with liver microsomes of 30 Japanese and 30
Caucasians. Journal of Pharmacology and Experimental Therapeutics 270, 414-423.
Takamatsu, Y., and Inaba, T. (1994). Inter-individual variability of human hepatic glutathione
S-transferase isozymes assessed by inhibitory capacity. Toxicology 88, 191-200.
Thier, R., Delbanco, E. H., Wiebel, F. A., Hallier, E., and Bolt, H. M. (1998). Determination of
glutathione transferase (GSTT1-1) activities in different tissues based on formation of
radioactive metabolites using 35S-glutathione. Archives of Toxicology 72, 811-815.
Thomas, R. S., Bigelow, P. L., Keefe, T. J., and Yang, R. S. H. (1996). Variability in biological
exposure indices using physiologically based pharmacokinetic modeling and Monte
Carlo simulation. American Industrial Hygiene Association Journal 57, 23-32.
Tornero-Velez, R., and Rappaport, S. M. (2001). Physiological modeling of the relative
contributions of styrene-7,8-oxide derived from direct inhalation and from styrene
metabolism to the systemic dose in humans. Toxicology Sciences 64, 151-161.
47
van Ommen, B., Bogaards, J. J., Peters, W. H., Blaauboer, B., and van Bladeren, P. J. (1990).
Quantification of human hepatic glutathione S-transferases. Biochemical Journal 269,
609-613.
Warholm, M., Guthenberg, C., Mannervik, B., von Bahr, C., and Glaumann, H. (1980).
Identification of a new glutathione S-transferase in human liver. Acta chemica
Scandinavica. Series B: Organic chemistry and biochemistry 34, 607-621.
White, I. N. H. (1988). A continuous fluorometric assay for cytochrome P-450-dependent mixed
function oxidases using 3-cyano-7-ethoxycoumarin. Analytical Biochemistry 172, 304­
310.
Wigaeus, E., Löf, A., Bjurström, R., and Nordqvist, M. B. (1983). Exposure to styrene: Uptake,
distribution, metabolism and elimination in man. Scandinavian Journal of Work and
Environmental Health 9, 479-488.
Wilson, H. K., Robertson, S. M., Waldron, H. A., and Gompertz, D. (1983). Effect of alcohol on
the kinetics of mandelic acid excretion in volunteers exposed to styrene vapour. British
Journal of Industrial Medicine 40, 75-80.
Xu, L., Li, A. P., Kaminski, D. L., and Ruh, M. F. (2000). 2,3,7,8 Tetrachlorodibenzo-p-dioxin
induction of cytochrome P4501A in cultured rat and human hepatocytes. Chemical and
Biological Interactions 124, 173-189.
Yang, R. S. H., Thomas, R. S., Gustafson, D. L., Campain, J., Benjamin, S. A., Verhaar, H. J.
M., and Mumtaz, M. M. (1998). Approaches to developing alternative and predictive
toxicology based on PBPK/PD and QSAR modeling. Environmental
Health
Perspectives 106, 1385-1393.
48
Printed and published by the Health and Safety Executive
C30 1/98
Printed and published by the Health and Safety Executive
C1.10
02/04
ISBN 0-7176-2799-3
RR190
£10.00
9 78071 7 627998