Environment International 73 (2014) 94–103
Contents lists available at ScienceDirect
Environment International
journal homepage: www.elsevier.com/locate/envint
AhR-mediated activities of polycyclic aromatic compound (PAC)
mixtures are predictable by the concept of concentration addition
Maria Larsson a,⁎, John P. Giesy b,c,d,e,f, Magnus Engwall a
a
Man-Technology-Environment Research Centre, School of Science and Technology, Örebro University, SE-701 82 Örebro, Sweden
Department of Veterinary Biomedical Sciences and Toxicological Center, University of Saskatchewan, Saskatoon, Saskatchewan, Canada
Department of Zoology and Center for Integrative Toxicology, Michigan State University, East Lansing, MI, USA
d
Department of Biology and Chemistry, State Key Laboratory in Marine Pollution, City University of Hong Kong, Kowloon, Hong Kong, China
e
School of Biological Sciences, University of Hong Kong, Hong Kong, China
f
State Key Laboratory of Pollution Control and Resource Reuse, School of the Environment, Nanjing University, Nanjing, People's Republic of China
b
c
a r t i c l e
i n f o
Article history:
Received 24 October 2013
Accepted 25 June 2014
Available online 9 August 2014
Keywords:
Polycyclic aromatic hydrocarbons
Concentration addition
Independent action
H4IIE-luc
Additivity
a b s t r a c t
Risk assessments of polycyclic aromatic hydrocarbons (PAHs) are complicated because these compounds exist in
the environment as complex mixtures of hundreds of individual PAHs and other related polycyclic aromatic
compounds (PACs). In this study, the hypothesis that concentration addition (CA) can be used to predict the
aryl hydrocarbon receptor (AhR)-mediated activity of PACs in mixtures containing various combinations of
PACs was tested. AhR-mediated activities of 18 mixtures composed of two to 23 PACs, which included PAHs,
azaarenes and oxygenated PAHs, were examined by the use of the AhR-based H4IIE-luc bioassay. Since greater
AhR-mediated activities have been observed in soils contaminated by PAHs, investigations were done to test
whether soil extract matrix or the presence of non-effect PACs might affect responses of the H4IIE-luc bioassay.
Our results showed that AhR-mediated activities of mixtures of PACs could be predicted by the use of concentration
addition. Additive activities of PACs in multi component mixtures along with the insignificant effect of the soil
matrix support the use of concentration addition in mass balance calculations and AhR-based bioassays in risk
assessment of environmental samples. However, independent action (IA) could not be used to predict the activity
of mixtures of PACs.
© 2014 Elsevier Ltd. All rights reserved.
1. Introduction
Humans and organisms living in the environment are exposed to
complex mixtures of natural and synthetic chemicals. Various matrices
can contain hundreds, even thousands of chemicals, the risk of which
must be assessed. Since it is not possible to identify and quantify all of
the relevant compounds, bioassays such as the H4IIE-luc transactivation
bioassay are often used providing an integrated measure of the potency
of the mixture to cause effects through a particular mechanism of action
or a specific adverse outcome pathway (Andersson et al., 2009). Even
when the mixture composition is characterised, the mechanisms of
potency and interactions between the chemicals are not fully understood (Sanderson and Giesy, 1998).
Polycyclic aromatic hydrocarbons (PAHs) are a group of widespread
organic compounds that have been detected in various media, including
air, water, food, soil, sediment and tissues of humans and wildlife (Chen
et al., 2004; Layshock et al., 2010; Söderström et al., 2005). Because of the
similar sources of origin, organisms can be exposed to PAHs simultaneously with other related compounds, such as azaarenes, oxygenated
⁎ Corresponding author. Tel.: +46 19301370; fax: +46 303166.
E-mail address: maria.larsson@oru.se (M. Larsson).
http://dx.doi.org/10.1016/j.envint.2014.06.011
0160-4120/© 2014 Elsevier Ltd. All rights reserved.
PAHs and methylated PAHs, collectively referred to as polycyclic
aromatic compounds (PACs). Many PACs can cause toxic effects such as
mutagenesis and carcinogenesis in humans and animals (Lundstedt
et al., 2007).
An important pathway concerning toxic effects of PACs is the
cytosolic aryl hydrocarbon receptor (AhR) signal transduction pathway.
Halogenated aromatic hydrocarbons, such as polychlorinated dibenzop-dioxins and furans, polychlorinated biphenyls (PCB) along with a
number of high molecular weight PAHs can bind to and activate the
AhR, which starts the production of a battery of proteins, including
the cytochrome P4501A (CYP1A) (Denison and Heath-Pagliuso, 1998;
Marlowe and Puga, 2005). Induced proteins can alter cellular homeostasis. Moreover, AhR-mediated induction of CYP1 enzymes can
lead to mutation and tumour initiation due to metabolic activation
of several PAHs (Bosetti et al., 2007; Nebert et al., 2000; Trombino
et al., 2000). The relationship between the AhR activation potency
of PAHs and their toxic effects is not clear. For example, PAHs that
strongly activate the AhR have shown to be only weakly carcinogenic
in animal studies (Boström et al., 2002; Machala et al., 2001a). Other
studies have shown good correlations between AhR inducing
capacity and carcinogenic potency (Sjögren et al., 1996; Trombino
et al., 2000).
M. Larsson et al. / Environment International 73 (2014) 94–103
Inhalation of air and ingestion of food are two of the routes through
which humans are exposed to PAHs (Falco et al., 2003; Nwaneshiudu
et al., 2007; Ramirez et al., 2011). Ingestion of soil and dust can be important routes of exposure, especially for young children. PAHs are
well-known contaminants in urban and industrial soils, especially in
abandoned gaswork sites, gas stations and former wood impregnation
facilities with creosote (Nestler, 1974; Šídlová et al., 2009; Thavamani
et al., 2011).
Even though hundreds of PACs often exist in PAH-contaminated environmental samples, risk assessments of PAHs in contaminated soils
are usually based on instrumental quantification of a small number of
individual PAHs, generally the 16 priority PAHs established by the
United States Environmental Protection Agency (USEPA) (Lundstedt
et al., 2003; Musa Bandowe et al., 2010). Traditional methodology of
using chemical analysis of the priority PAHs to determine the degree
of PAH contamination in soils greatly overlooks toxicologically relevant
PAHs or other AhR agonists (Larsson et al., 2013; Lundstedt et al., 2003).
Bioassays are practical techniques to obtain estimates of the toxicities and potential risks of mixtures. Mechanism specific bioassays are
good screening tools for toxic chemicals in different media because
they enable an estimation of the total biological effect of all chemicals
present in the sample with the same mechanism of action (Behnisch
et al., 2001; Richter et al., 1997). Because of the AhR-activating capacity
of many PACs, AhR-based bioassays, like the H4IIE-luc bioassay, can be
used in the analysis of PAH-contaminated samples (Denison et al.,
2002; Larsson et al., 2014). Moreover, the AhR-mediated toxic potency
of single chemicals can be estimated (Villeneuve et al., 2000a). The potency of a single chemical can be quantified compared to 2,3,7,8tetrachlorodibenzo-p-dioxin (TCDD) or another potent compound,
and the relative potency (REP) factor can be used to calculate 2,3,7,8TCDD equivalents (TEQs) (Engwall et al., 1998).
The use of mass balance analysis, i.e. comparison of bioassay derived
TEQs with chemical TEQs based on REPs and measured concentrations,
has shown that chemically identified PAHs often account for only a
fraction of the AhR-mediated activities found in PAH-contaminated
environmental samples (Andersson et al., 2009; Keiter et al., 2008;
Larsson et al., 2013; Villeneuve et al., 2000b). Additionally, additive
behaviour of PAHs in synthetic mixtures has been observed in the
H4IIE-luc assay (Larsson et al., 2012), which indicates that the high
AhR-mediating activities observed in the environmental samples are
due to other unidentified PACs and/or possible synergistic chemical interactions in the complex environmental samples. Sample matrices
might also have an effect on the AhR-mediating activity observed in
samples. For example, dissolved organic matter (DOM) has been suggested to enhance the AhR-mediating activity of a PAH-mixture
(Bittner et al., 2011).
Bioassay analysis of environmental samples gives no information
about concentrations and toxic activities of individual chemicals present
in samples (Sanderson et al., 1996). To increase the knowledge in
mixture assessments, knowledge of single compounds and how they interact with each other must be further investigated. This can be done by
testing single compounds and artificial mixtures with a known composition in bioassays. Based on concentration–response curves of single
compounds, activities of mixtures can be predicted from several models,
like the concentration addition (CA) model or the independent action
(IA) model and then compared with the observed activity in the
known mixtures (Faust et al., 2001). Independent action, also known
as response addition, is based on the assumption that the mixture
components cause a common integrated effect through different mechanisms of action (Backhaus et al., 2000). The CA model is based on the
assumption that all mixture components have a common or similar
mode and mechanism of action. Most studies employ the fixed ratio
design, that is, the ratio between the individual compounds in the mixtures is constant, and the total concentration of the mixtures is varied.
Often, the compounds are mixed in an equivalent-effect concentration,
for example, an effective concentration for 50% (EC50). The advantage of
95
such mixtures is that no discrimination of low-potency compounds
occurs.
Like the approach where TEQ are calculated as the sum of the
products of concentrations of individual compounds and their respective
REPs, the CA and IA models are based on additive behaviour of the
compounds and non-interactions in the mixture. Disagreement in
predicted and observed activities suggests non-additive interactions,
i.e. ultra-additive or infra-additive interactions, of the compounds in
the mixture (Sanderson and Giesy, 1998). The CA model has been
shown to be an accurate reference model of mixtures of chemicals
known to have similar modes of action (Faust et al., 2001; Zhang
et al., 2008). Unlike the REP concept, concentration–response curves
do not have to be parallel.
The objective of the study was to examine whether the concentration addition (CA) model accurately predicts the AhR-mediated mixture
activity at different concentrations and compositions of PACs, including
PAHs, oxy-PAHs and azaarenes, by the use of the H4IIE-luc bioassay.
Synthetic mixtures with various compositions were chosen based on
ranges of concentrations of PACs expected to occur in the environment
(Lundstedt et al., 2003; Musa Bandowe et al., 2010; Šídlová et al., 2009).
A secondary objective was to determine if the matrix, in extracts of soil,
or the presence of non-AhR-active PACs affect the activity of the mixture. The accuracy of the independent action (IA) method, based on
compounds having dissimilar mode of action to predict AhR-activity of
mixtures of PACs was also evaluated.
2. Material and methods
2.1. Chemicals and mixture compositions
Individual standards of 27 PACs were obtained for the present study;
cyclopenta[def]phenanthrene (97%), naphthacene (99.5%), benzo[a]
fluorene (98%), 1-indanone (N99%), 4H-cyclopenta[def]phenanthrenone
(99.5%), 2-methylanthracene-9,10-dione (97%), benzo[a]fluorenone
(99.8%), benz[a]anthracene-7,12-dione (N 98%), anthracene-9,10-dione
(99.8%), naphthacene-5,12-dione (97%), 7H-benz[de]anthracene-7-one
(99%), 1,4-chrysenequinone (93%), quinoline (98%), benzo[h]quinolone
(98%), acridine (N98%), dibenz[a,h]acridine (99.6%), benzo[a]pyrene
(99%), chrysene (99%), indeno[1,2,3-cd]pyrene (99%), benzo[j]fluoranthene (99%), benzo[b]fluoranthene (99%), benzo[k]fluoranthene, dibenz
[a,h]anthracene (99%), naphtho[2,3-a]pyrene (98%), benz[a]anthracene
(98%) dibenz[a,c]anthracene (97.5%), and dibenz[a,j]anthracene (99.8%).
Molecular structures of PACs are presented in Fig. S1. The TCDD standard
had a purity of 99.1%.
Eighteen synthetic mixtures composed of two to 23 individual PACs
were made in dimethyl sulfoxide (DMSO). The mixtures were named
based on the number of compounds included. For example, mixtures
composed of 18 PACs were named mix18. In cases of several mixtures
with the same number of components, a letter was added to the mixture
name. In ten of the mixtures, PACs were mixed in equivalent effect concentrations, so-called equitoxic mixtures (Table 1). Components of the
mixtures were combined in proportion to their EC5 and EC50 values.
That is, concentrations producing 5% or 50% of the maximum induction
caused by TCDD in the bioassay (Altenburger et al., 2000). The
advantage of such mixtures is that no discrimination of low-potency
compounds occurs since all compounds are expected to contribute
equally to the overall effect of the mixture. Since it is unlikely that
PACs occur in the environment in a fixed ratio to their AhR-mediated
effect concentrations, the other eight mixtures were non-equivalent
effect concentrations (Table 2). Various combinations of the compounds
were chosen in the mixtures to estimate the activity and additivity in
mixtures containing solely PAHs, oxy-PAHs or azaarenes and mixtures
containing a combination of the compounds. Different concentrations
were selected to cover concentrations of PACs expected to occur in the
environment. Additionally, a mixture of five PACs that elicited no AhRmediated effect (referred to as non-effect PACs) when each compound
96
M. Larsson et al. / Environment International 73 (2014) 94–103
Table 1
Mixtures of PACs prepared in two equivalent effective concentrations (EC5 and EC50). Chemicals included in each mixture are marked with the EC level of the compound in the mixture.
Compounds
Mixtures
mix18A
1
Benzo[a]fluorene
2
Naphthacene
3
Chrysene
4
Benzo[a]anthracene
5
Benzo[k]fluoranthene
6
Benzo[j]fluoranthene
7
Benzo[b]fluoranthene
8
Benzo[a]pyrene
9
Indeno[1,2,3-cd]pyrene
10
Dibenz[a,c]anthracene
11
Dibenz[a,j]anthracene
12
Dibenz[a,h]anthracene
13
Naphtho[2,3-a]pyrene
16
9,10-Dihydrobenzo[a]pyren-(8H)-none
17
Benzo[a]fluorenone
18
Benz[a]anthracene-7,12-dione
19
7H-benz[de]anthracen-7-one
24
Dibenz[ah]acridine
Total mixture concentrations (μM)
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
34
mix13
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
1.5
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
EC50
2.7
mix11
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
EC5
0.05
mix7A
mix4
EC50
EC50
EC50
EC5
EC5
EC5
EC50
EC50
EC50
EC5
EC5
EC5
EC50
EC50
EC50
EC5
EC5
EC5
EC50
EC50
EC50
EC5
EC5
EC5
EC50
EC50
EC5
EC5
EC50
EC5
EC50
EC5
EC50
EC5
EC50
20
EC5
1.1
1.8
0.43
EC50
EC50
EC50
EC50
EC5
EC5
EC5
EC5
31
1.5
Effective concentration values (EC5 and EC50) are calculated from the individual PAC concentration–response curves (Larsson et al., 2012, 2014).
was tested separately was composed to investigate if non-effect PACs in
a mixture may produce AhR-mediated response. The non-effective PACs
mixture was then mixed 1:1 with an 18-component mixture (mix18A)
to determine whether the presence of non-effective PACs alters potency
of the mixture due to chemical activity of the mixture (Sanderson and
Giesy, 1998). Mixture combinations are listed in Tables 1 and 2 and
effect concentrations in Tables 3 and S1, while proportions (pi) of
individual PACs in the total concentration of mixtures are listed in
Tables S2 and S3.
To investigate if the matrix of soil might affect the AhR-mediated
activities of mixtures of PACs, soil extracts in n-hexane from extraction
of an agricultural soil were added to three of the PAC mixtures. Before
exposing cells to these mixtures the n-hexane was vaporised. A mixture
containing extract of the agricultural soil was also prepared in DMSO.
2.2. H4IIE-luc bioassay
AhR-mediated activities for the PAC-mixtures were measured by the
use of the H4IIE-luc assay, as previously described in Larsson et al.
(2012). This transactivation assay, based on a rat hepatoma cell line, is
stably transfected with a luciferase reporter gene under control of the
dioxin response enhancer (DRE) of the AhR-mediated signal transduction pathway (Murk et al., 1996).
Concentration–response curves were constructed for the mixtures
using a sigmoidal dose–response (four-parameter logistic equation)
(Eq. (1)).
3. Calculations
3.1. Prediction of activities of mixtures
Activities of mixtures were predicted by the use of the concentration addition (CA) model or the independent action (IA) model on
the basis of concentration–response curves of single PAHs, oxygenated PAHs and azaarenes, which have been reported elsewhere
(Larsson et al., 2012, 2014). Data from two to four experiments
were used for calculation of the predicted activity of mixtures. To
make the response curves comparable, responses to PACs were
normalised to the mean of the maximum efficacy observed for their
corresponding TCDD standard. The mean solvent control response
was subtracted from both TCDD standard and PAC responses prior
to conversion to get the responses scaled from 0% to 100% of TMI.
Concentration–response curves used in the mixture predictions
had a mean maximum induction of luciferase activity N50% of TMI.
Effect concentrations (ECx) of individual PACs were defined relative
to the effects of the TCDD concentration used as the reference compound,
for example EC50 is the concentration producing an effect of 50% of the
maximum effect of TCDD.
Because the composition of the mixtures is known, the concentrations
of each PAC can be expressed as a fraction of the total concentration.
Concentration addition is expressed mathematically as
ECx;mix ¼
EðcÞ ¼ min þ
ð max− minÞ
1 þ 10ð LogEC50−cÞslope
n
X
pi
ECx;i
i¼1
!−1
ð2Þ
ð1Þ
where: min and max are the minimal and maximal observed effect plateaus, respectively, c is the concentration of the test mixture, EC50 is the
concentration of the test mixture yielding 50% effect of TCDD maximum
induction (TMI). The curve-fitting was performed by the use of the
GraphPad Prism® 5.0 software.
Only plates with a relative standard deviation of b16% within the
triplicate wells of each concentration of TCDD and PACs, a TCDD
maximum induction factor N6 and a TCDD EC50 value between 8 and
18 pM were included in the study. Limit of detection (LOD) was calculated as the mean luciferase activity of DMSO control triplicates + 3
times standard deviation (SD).
where: n is the number of mixture constituents, ECx,mix is the concentration of the mixture provoking x% effect, ECx,i is the concentration of
the ith mixture component provoking x% effect when applied singly,
and pi is the fraction of the ith component in the mixture
(Berenbaum, 1985).
Concentrations giving 10–100% mixture effects were calculated in
steps of 5% and the concentration/effect coordinates were plotted and
analysed using the Eq. (1) (GraphPad Prism® 5 software) to give a
predicted concentration–response curve.
The alternative IA model is a common approach used for prediction
of the mixture toxicity of compounds with diverse modes of action.
Response addition is defined by Eq. (3).
M. Larsson et al. / Environment International 73 (2014) 94–103
97
Table 2
Mixtures of PACs prepared in non-equivalent effective concentrations. Chemicals included in each mixture are marked with the EC level of the compound in the mixture.
Compounds
Mixtures
mix2A
1
Benzo[a]fluorene
2
Naphthacene
3
Chrysene
4
Benzo[a]anthracene
5
Benzo[k]fluoranthene
6
Benzo[j]fluoranthene
7
Benzo[b]fluoranthene
8
Benzo[a]pyrene
9
Indeno[1,2,3-cd]pyrene
10
Dibenz[a,c]anthracene
11
Dibenz[a,j]anthracene
12
Dibenz[a,h]anthracene
13
Naphtho[2,3-a]pyrene
14
2-Methylanthracene-9,10-dione
15
Naphthacene-5,12-dione
16
9,10-Dihydrobenzo[a]pyren-(8H)-none
17
Benzo[a]fluorenone
18
Benz[a]anthracene-7,12-dione
19
7H-benz[de]anthracen-7-one
20
1,4-Chrysenequinone
21
1-Indanone
22
Anthracene-9,10-dione
23
4H-cyclopenta[def]phenanthrenone
24
Dibenz[ah]acridine
25
Acridine
26
Benzo[h]quinoline
27
Quinoline
Total mixture concentrations (μM)
mix5A
mix3
mix5Ba
EC30
EC31
EC22
EC40
EC56
EC30
mix7B
EC17
EC27
EC24
EC35
EC23
EC30
EC48
EC50
EC58
mix18B
mix18C
mix23a
EC31
EC41
EC37
EC35
EC28
EC46
EC41
EC38
EC43
EC39
EC38
EC41
EC39
EC15
EC22
EC15
EC12
EC17
EC29
EC25
EC18
EC26
EC19
EC16
EC22
EC20
EC31
EC41
EC37
EC35
EC28
EC46
EC41
EC38
EC43
EC39
EC38
EC41
EC39
EC32
EC30
EC37
EC30
EC9
EC9
EC14
EC7
EC32
EC30
EC37
EC30
EC21
nd (2.5)
nd (2.0)
nd (2.5)
EC38
3.4
nd (4.0)
nd (4.0)
32
nd (5.0)
nd (4.0)
nd (5.0)
EC98
EC80
EC38
20.2
21.5
13.7
nd (8.0)
nd (8.0)
30
17.6
17
Effective concentration levels (ECx) are calculated from the individual PAC concentration–response curves (Larsson et al., 2012, 2014).
nd = no detectable AhR-mediated response observed for the single PACs within the concentration range tested (Larsson et al., 2014).
a
Molar concentrations (μM) are bracketed for non-effect PACs.
n
Eðcmix Þ ¼ 1−∏ ð1−Eðci ÞÞ
ð3Þ
i¼1
Where: ci denotes the concentrations of the ith mixture component,
E(ci) its corresponding effect, and E(cmix) the overall effect caused by the
total concentration of the mixture (cmix) (Bliss, 1939).
Table 3
AhR-mediated potency of single PACs; effect concentrations (EC5 and EC50) and slopes of
the concentration–response curves.
1
2
3
4
5
6
7
8
9
10
11
12
13
16
17
18
19
20
21
22
23
24
25
26
27
Compounds
EC50 (nM)
EC5 (nM)
Slope
Benzo[a]fluorene
Naphthacene
Chrysene
Benzo[a]anthracene
Benzo[k]fluoranthene
Benzo[j]fluoranthene
Benzo[b]fluoranthene
Benzo[a]pyrene
Indeno[1,2,3-cd]pyrene
Dibenz[a,c]anthracene
Dibenz[a,j]anthracene
Dibenz[a,h]anthracene
Naphtho[2,3-a]pyrene
9,10-Dihydrobenzo[a]pyren-(8H)-none
Benzo[a]fluorenone
Benz[a]anthracene-7,12-dione
7H-benz[de]anthracen-7-one
1,4-Chrysenequinone
1-Indanone
Anthracene-9,10-dione
4H-cyclopenta[def]phenanthrenone
Dibenz[ah]acridine
Acridine
Benzo[h]quinoline
Quinoline
730
14
390
800
4.8
60
79
320
120
25
21
23
120
7800
4200
10,800
8400
2800
nd
nd
nd
4
1600
nd
nd
5.5
0.015
8.8
28
0.007
0.060
0.038
4.6
0.97
0.049
0. 57
0. 18
0. 10
400
180
310
580
34
nd
nd
nd
0.013
570
nd
nd
0.435
0.522
0.736
0.834
0.311
0.496
0.388
0.588
0.465
0.672
0.662
0.515
0.571
0.877
0.733
0.673
0.988
0.697
nd
nd
nd
0.493
0.945
nd
nd
Concentration/effect coordinates were plotted and analysed by the
use of (Eq. (1)) a sigmoidal response (four-parameter logistic equation)
curve fitting (GraphPad Prism® 5 software) to give a predicted
concentration–response curve.
3.2. Observed activity of mixtures
Concentration–response curves were calculated for each mixture by
the use of a four-parameter logistic equation (Eq. (1)). In order to make
the response curves comparable, the mixture responses were normalised against the mean of the maximum response observed by the
TCDD standard. The mean solvent control response was subtracted
from both TCDD standard and mixture responses prior to conversion
to get responses scaled from 0% to 100% of TCDD maximum induction
(TMI).
4. Results and discussion
4.1. Predictability of activities of mixtures of PACs by the CA model
Based on previous results of additive behaviour of PAHs in the
AhR-based, H4IIE-luc bioassay (Larsson et al., 2012), the accuracy
of the CA model to predict activity of mixtures composed of additional polycyclic aromatic compounds, such as oxy-PAHs and azaarenes,
was investigated. Oxy-PAHs are less potent AhR agonists than their
unsubstituted, analogous PAHs, while azaarenes exhibited potencies
in the H4IIE-luc bioassay that are similar or even greater than their
analogous PAHs (Larsson et al., 2014; Machala et al., 2001b;
Sovadinová et al., 2006). However, in assessments of potential
hazards of PACs in the environment contributions of numerous compounds and/or mixture interactions to the overall AhR-mediated
mixture are of importance.
98
M. Larsson et al. / Environment International 73 (2014) 94–103
Fig. 1. Observed and predicted AhR-mediated activity in mixtures of PACs mixed in their equivalent effective concentrations (EC50) at 24 h of exposure in the H4IIE-luc bioassay. The observed data are based on two experiments consisting of triplicates at each concentration. Predicted data calculated with concentration addition (CA) or independent action (IA) are based
on two to four experiments for each of the individual compounds with triplicate exposures for each concentration. Error bars represent standard deviations, and dashed lines give the 95%
confidence intervals of the observed curve. When a mixture contains n PACs in the ratio of their individual EC50 values, there is one point on the concentration–response curve (marked
with an arrow), where the concentrations of all single constituents equal just 1/n of these individual EC50 values.
All 18 PAC mixtures tested in the present study elicited AhRmediated activities as determined in the H4IIE-luc bioassay (Figs. 1
and 2). Several mixtures caused super induction, which are efficacies
that are greater than the maximum efficacy of TCDD that was used as
the reference chemical. Super induction has also been observed in
H4IIE-luc bioassay by individual PAHs, oxy-PAHs, or mixtures of PAHs
in extracts of environmental matrices. That has also been observed in
other bioassays, for example by estrogens in the ERLUX bioassay
(Evans et al., 2012; He et al., 2011; Larsson et al., 2012).
Based on relative prevalence of each compound in the mixtures and
their individual concentration–response curves (Figs. 3 and 4), activities
of mixtures were predicted by the use of the CA model. Individual concentration–response curves used were determined in previous studies
(Larsson et al., 2012, 2014). Maximum efficacies of predicted concentration–response curves varied among compositions of mixtures, because
several of the individual PACs did not elicit full concentration–response
curves within the tested concentrations. For mathematical reasons,
predictions are limited to the maximum effect observed for the
compound with least induction of luciferase activity in each mixture. The
CA model calculates concentrations of mixtures of compounds that produce a predetermined effect, and these predictions are based on the effect
concentration of each compound that will singly produce the same effect.
Mixture interactions were studied by comparing the observed
activities of mixtures with activities predicted by the use of the CA
model. Concentration–response curves are displayed in Figs. 1 and 2.
In general, the absolute differences between predicted and observed
mixture activities were quite small. Predicted EC50 values observed in
the bioassay for the two mixtures composed of three to four oxy-PAHs
were approximately a factor of 1.4 (mix3) to 1.6 (mix4) greater than
the observed EC50 values (Table 4). Otherwise the CA model tended to
overestimate the toxicity to some extent in the mixtures containing
seven to 13 compounds mixed in equivalent effect concentrations
(EC50). The greatest difference between observed and predicted potency was shown in the mixture of 7 PAHs, mix7A. The observed EC50 was
4.2-fold greater than the predicted EC50, compared to 2.7 times greater
in mix11, and 1.4 times greater in mix13. The predicted potency in
mix7B, composed of 7 oxy-PAHs, was overestimated in the lesser effect
range of the curve and underestimated at greater concentrations. Additive potencies were observed for mix5A, which was composed of five
oxy-PAHs, and in the binary azaarene mixture, mix2 (indicated by overlapping 95% confidence intervals, Table 4). Additivity was also observed
for the mixtures containing 18 PACs. These findings were in agreement
with those of a previous study, in which a strictly additive response of
a mixture containing 15 PAHs was observed (Larsson et al., 2012). The
mixture that contained 23 PACs including five non-effect PACs also exhibited a strictly additive activity.
Comparing mixtures composed of only PAHs, oxy-PAHs or
azaarenes, almost additive interactions were observed for mixtures
composed of oxy-PAHs. A tendency to infra-additive interactions was
observed for mixtures containing only PAHs. As indicated by the nonoverlapping 95% confidence interval, the observed EC50 values were
significantly greater than predicted in the PAH mixtures (Table 4).
Since supra-additive interactions for mixtures of two to four PAHs
have been reported previously (Larsson et al., 2012), it seems likely
that mixtures of PACs might cause infra- and supra-additive responses.
This is more observable for mixtures composed of fewer compounds. In
M. Larsson et al. / Environment International 73 (2014) 94–103
99
Fig. 2. Observed and predicted AhR-mediated activity in mixtures of PACs mixed in non-equivalent effective concentrations at 24 h of exposure in the H4IIE-luc bioassay. The observed data
are based on two experiments consisting of triplicates. Predicted data calculated with concentration addition (CA) are based on two to four experiments for each of the individual compounds with triplicate exposures for each concentration. Error bars represent standard deviations, and dashed lines give the 95% confidence intervals of the observed curve.
multicomponent mixtures both supra- and infra-additive effects could
occur simultaneously among various constituents, but the overall effect
might be simple additivity if the effects cancel each other (Kortenkamp
et al., 2009; Warne and Hawker, 1995). This is a possible explanation for
why the predictive power of the CA model for multicomponent
mixtures was better than it was for mixtures with fewer compounds.
Fig. 3. Concentration–response curves for the 13 polycyclic aromatic hydrocarbons
(PAHs). Identity of the compounds is given in Tables 1 and 2 (Larsson et al., 2012).
Fig. 4. Concentration–response curves for the 14 oxygenated polycyclic hydrocarbons
(oxy-PAHS) and azaarenes. Identity of the compounds is given in Tables 1 and 2 (Larsson
et al., 2014).
100
M. Larsson et al. / Environment International 73 (2014) 94–103
Table 4
Predicted and observed effect concentrations (EC50) of selected PAC mixtures (μM) and 95% confidence intervals (CI).
Mixture
mix18A
mix13
mix11
mix7A
mix4
mix2
mix5A
mix3
mix7B
mix18B
mix18C
mix23
Observed
Predicted by CA
Predicted by IA
Mean (μM)
95% CI
Mean (μM)
95% CI
Mean (μM)
95% CI
1.630
0.321
1.500
0.185
4.65
0.326
2.871
5.068
3.980
1.489
1.555
3.127
1.499, 1.775
0.279, 0.370
1.323, 1.702
0.166, 0.206
4.280, 5.033
0.252, 0.422
2.347, 3.536
4.692, 5.527
3.738, 4.240
1.253, 1.784
1.431, 1.685
2.736, 3.589
1.880
0.221
0.546
0.044
7.131
0.330
2.836
7.104
3.560
1.880
1.895
3.744
1.774, 1.990
0.210, 0.233
0.511, 0.573
0.043, 0.046
6.928,7.340
0.329, 0.331
2.770, 2.904
7.091, 7.116
3.446, 3.678
1.774, 1.993
1.784, 2.012
3.461, 4.047
0.112
0.0011
0.217
0.030
6.995
0.078, 0.164
0.00067, 0.0016
0.164, 0.289
0.021, 0.046
5.877, 8.415
4.2. Comparison of the CA and IA models for predicting AhR-mediated
activities in PAC mixtures
In contrast to the CA concept, IA assumes different target sites and
dissimilar mechanisms of action of all mixture components. The CA
model was expected to predict better than the IA model, because the
mechanisms through which AhR-agonists operate is well defined and
specific in the H4IIE-luc bioassay, which responds to all compounds
that activate the AhR.
To assess accuracy of predictions of the two models their
concentration–response curves were compared (Fig. 1). The best agreement between the two models (overlapping 95% confidence intervals
for EC50) was observed for a four-component mixture, mix4, but predictions of both models were 1.3 times greater than the EC50 of the observed
curve (Table 4). Both the CA- and IA-model tended to overestimate the
potency of mixtures, mix7A and mix11. Based on predicted EC50 values
for mix7A and mix11, the IA-model predicted a mixture potency that
was 1.4 to 2.5 times greater than the potency predicted by the CA
model. However, the accuracy of predictions made by the use of the IAmodel was better than that of the CA-model for lesser concentrations of
mix7 and mix11. Based on a comparison of predicted EC50 values, of
mix13 and mix18A, the IA-model predicted a mixture potency that was
17- to 200-fold greater than the potency predicted by the CA-model.
Shallower slopes (less than 1) of the concentration–response curves
for most single PACs (Table 3) might be the reason for the poorer
predictability of activities of the 13- and 18-component mixtures by
the IA-model. Less steep slopes of the mixture components would generate greater predictions by the IA-model (Backhaus et al., 2004; Brosche
and Backhaus, 2010; Drescher and Boedeker, 1995). Differences between
activities predicted by the IA- and CA-models were the greatest in the
13-component mixture, composed of only unsubstituted PAHs.
Concentration-curves for PAHs were less steep (slope = 0.3 to 0.8)
than those of the oxy-PAHs or azaarenes (slope = 0.5 to 1) (Table 3).
Calculation of ratios between individual EC50 and EC5 values indicated
the steepness of the concentration–response curves. The ratio varied
between 28 and 2000 for the PAHs and between 20 and 44 for the
oxy-PAHs. Lower slopes of PAHs concentration–response curves might
be an explanation for the difference between mixture activities of
mix13 predicted by the IA- and CA-model (Backhaus et al., 2000,
2004). The IA-model failed to accurately predict the potencies of all
mixtures as determined by the H4IIE-luc assay.
Since the predictability of the CA model is independent of the slopes
and PACs commonly have concentration–response curves with less
steep slopes, the CA-model would be the more appropriate model for
prediction of activities of mixtures. This conclusion is supported by the
better agreement of the experimental data with the additive expectations
by the CA-model. Predictability of activities of mixtures is not restricted
to a certain mixture ratio or effect concentration, because similarity between observed and predicted AhR mediated activity in the H4IIE-luc
bioassay was independent of the mixture composition. Additive
interactions were observed in both equitoxic- and non-equitoxic
mixtures. Accuracy of predictions of multi-component PAH-mixtures by
the use of the CA-model or the REP-concept (Larsson et al., 2012) along
with the predictions of PAC-mixtures by the CA-model in the present
study confirm the suitability of using the concentration-addition model
in predictions of the activity of PAC mixtures, and the use of the TEQ
approach based on REPs in mass-balance calculations. In this study
TCDD was used as a reference chemical, but the conclusions would
have been similar if, instead, benzo[a]pyrene had been used as a reference.
4.3. Mixture activity at lower effective concentrations of single PACs
Generally, when mixtures contain compounds at low effective concentrations full concentration–response curves are not produced. In
those cases point estimates can be made (Faust et al., 2001). In an
equitoxic mixture that contains “n” numbers of PACs in the ratio of their
individual EC50 values, there is one point on the concentration–response
curve, where the concentrations of all single constituents equal 1/n
fraction of their individual EC50 value. The total concentration for
those individual EC50 values is marked with an arrow in each mixture
(Fig. 1) and corresponding effects are presented (Fig. 5).
The total concentration in the 18-component mixture produces a total
combined effect of 50%. This is exactly equal to the prediction by the use of
the CA-model, which predicted a total effect of 50%. The total effect predicted by the use of the IA-model was almost a factor of two greater
(98%). Thus, the CA-model predicted greater effect for mix13, mix11 and
mix7 than the observed combined effects, with a deviation of 12, 15 and
18%, respectively. A lower effect was predicted by CA in mix4 and the
deviation was 19% compared to the observed total combined effect.
Similar point estimations can be performed for the equitoxic mixtures composed of PACs based on EC5 concentrations. Because the
total mixture concentrations were equal to the sum of the individual
EC5 concentrations of PACs, total mixture concentrations were used
(Tables 1 and 3). The total combined effect was significant for all mixtures, despite the low total concentrations, 0.05 to 1.5 μM. Observed
combined effects of individual EC5 values were 28 and 18% in mix18A
and mix13, respectively, and the corresponding prediction of total
effects by the CA-model were 28 and 20%, respectively, which are
considered to be good predictions. The predictions made by the use of
the IA-model were a factor of two greater (Fig. 5). Thus, both models
predicted greater effect than the combined effect observed in mix13,
mix11 and mix4.
4.4. Effect of the soil matrix or non-effect PACs on the mixture activity
Because PAHs often account for only a portion of the AhR-mediated
effects of PAH-contaminated environments, effects of a soil matrix
might have on the total activity of mixtures of PACs was investigated.
M. Larsson et al. / Environment International 73 (2014) 94–103
101
Fig. 5. Comparison of the total effects of PAC mixtures. (A) Concentration of individual mixture components equal to individual EC5 values. (B) Concentrations of individual components
equal to 1/n of individual EC50 values. When a mixture contains n PACs in the ratio of their individual EC50 values, there is one point on the concentration–response curve, where the
concentrations of all single constituents equal just 1/n of these individual EC50 values.
An equal equivalent amount of an extract of an agricultural soil was
added to three 18-component mixtures (mix18A, mix18B and 18C),
which had similar compositions of PACs, but different total concentrations (34, 17 and 3.4 μM, respectively). The soil mixture caused significant AhR-mediated activity in the H4IIE-luc bioassay when tested
alone, but the concentration–response curve was shifted to the right
on the x-axis compared to the 18-component mixtures (Fig. 6). The
efficacy of the mix18A was reduced from 175% to 120% in the presence
of soil, but besides that, the concentration–response curves were comparable (as indicated by overlapping 95% confidence of the curves).
Also the concentration–response curves for the other two mixtures
(mix18B and mix18C) with and without the presence of soil were comparable. However, a slight effect of the soil was observed at the lesser concentration (mix18C), which increased the potency of the mixture to some
Fig. 6. Effects of soil extract matrix or non-effect PACs on luciferase activity in H4IIE-luc cells by different AhR agonistic PAC mixtures. Comparison of concentration–response curves of
mixtures at different concentrations with or without the presence of soil matrix (A–C) and of a mixture with or without the presence of five non-effect PACs (D) at 24 h of exposure in
the H4IIE-luc bioassay. The data are based on two experiments consisting of triplicates. Error bars represent standard deviations.
102
M. Larsson et al. / Environment International 73 (2014) 94–103
extent. This result would be expected due to the lesser concentration of
PACs in that mixture. The observed reduction in efficacy of the mix18A
and mix18B might be due to sorption of the PACs to organic material in
the mixture, which would decrease concentrations of soluble compounds
in the cell medium and thereby the effect of the PACs. Another explanation
can be competitive AhR ligands in soils that are less efficient agonists than
the PAHs. If this reduction in effect really results in a decrease of risk is
unknown, because the effect is still greater than TMI, i.e. superinduction
exists, in the mixtures (mix18A and 18B) containing soil. Moreover, the
potency of the concentration–response curves is unaffected of the soil.
The amount of soil tested in this study; 81 mg soil/ml cell medium is
within the range (4 to 150 mg soil/ml medium) that is commonly extracted for the use in the H4IIE-luc analysis of soils, depending on the
degree of PAH-contamination. Lesser AhR-mediated potency in the agricultural soil agrees with previous H4IIE-luc bioanalysis of agriculture
soils (Larsson et al., 2013; Šídlová et al., 2009). The soil had a negligible
effect on the mixture potency of PACs. No significant differences (overlapping 95% confidence intervals) in potency were shown between the
concentration-curves of the mixtures (mix18A and mix18B), with or
without soil.
AhR-mediated activity of a mixture containing five non-effect PACs
(mix5B) was tested in the H4IIE-luc bioassay, which showed no significant AhR-mediated activity (Fig. 6). Additionally, the mixture was
mixed 1:1 with an 18-component mixture (mix18A), to see if the
presence of non-effect PACs altered the activity of PAC mixtures. The
concentration–response curve was compared with the concentration–
response curve of mix18B, containing the 18 PACs but at half the
concentration of mix18A. If the potency was not affected by addition
of the non-effect PACs (mix5B), the mixture was suggested to produce
the same concentration–response relationship as mix18B. As indicated
by the overlapping 95% confidence intervals, no significant difference
in concentration–response curves with or without addition of noneffect compounds (mix5B) to the 18-component mixture was observed
(Fig. 6). These findings suggest that these PACs (4H-benzo[h]quinolone,
quinolone, cyclopenta[def]phenanthrenone, 1-indanone and anthracene9,10-dione), although present in a sample, will not produce any AhRmediated activity or alter the overall mixture activity.
Complexities of mixtures of PACs in the environment will result in
exposure to PACs with dissimilar AhR activating potency, depending
on the molecular structure of the compounds. Even though a compound
lacks the capacity to activate a receptor, it might have agonistic or
synergistic effects on other compounds in the mixture (Sanderson and
Giesy, 1998). For example, the presence of DOM can result in reduced
AhR-mediated activity due to sorption of hydrophobic compounds to
the DOM structure or enhanced activity due to AhR-active or facilitate
effects of DOM (Bittner et al., 2011). The results of the present study
indicate that AhR-mediated activity observed in some soils contaminated
with PAHs is not due to enhanced activity by the soil matrix, since the
effect was reduced in greater concentrations of PACs. However, only
three mixtures were tested in this study containing the same soil extract
and it cannot be ruled out that the composition in other soils may affect
the AhR-mediated potency of the soils.
5. Conclusions
The results of the study presented here demonstrated that it is possible to predict AhR-mediated activities of mixtures in the H4IIE-luc
assay by the use of concentration addition. Accuracy was the best for
multicomponent mixtures and independent of the mixture composition. Thus, responses were strictly additive, in mixtures composed of
PACs in equivalent- or non-equivalent effect concentrations. These results strongly support the use of concentration addition in assessment
of hazard and mass-balance calculations for mixtures of PACs in samples
from the environment, such as soils. In contrast independent action
seems to be an inappropriate model for predictions of mixture activity
in PAC mixtures. Results presented here are an important step in the
evaluation of AhR-based bioassay analysis of complex PAC-contaminated
samples. Additive behaviour of PACs along with the insignificant effect
of the soil matrix support the suggestion that numerous PACs are
responsible for the high AhR-mediated activities observed in the
H4IIE-luc analysis of environmental samples. Questions remain regarding the identity of chemicals, their mechanisms of action, interactions,
and whether the total AhR-mediated activating potency of complex
PAC-mixtures corresponds to an adverse biological effect. This is of importance since results from this study indicate that exposure even for a
few PACs at relatively low concentrations can cause AhR-mediated
effects.
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.envint.2014.06.011.
Acknowledgements
This work was financially supported by the KK Foundation (the
Knowledge Foundation). Prof. Giesy was supported by the Canada
Research Chairs programme, a Visiting Distinguished Professorship in
the Department of Biology and Chemistry and State Key Laboratory in
Marine Pollution, City University of Hong Kong, the 2012 “High Concentration Foreign Experts” (#GDW20123200120) programme, funded by
the State Administration of Foreign Experts Affairs, the P.R. China to
Nanjing University and the Einstein Professor Program of the Chinese
Academy of Sciences.
References
Altenburger R, Backhaus T, Boedeker W, Faust M, Scholze M, Grimme LH. Predictability of
the toxicity of multiple chemical mixtures to Vibrio fischeri. Mixtures composed of
similarly acting chemicals. Environ Toxicol Chem 2000;19:2341–7.
Andersson E, Rotander A, von Kronhelm T, Berggren A, Ivarsson P, Hollert H, et al. AhR
agonist and genotoxicant bioavailability in a PAH-contaminated soil undergoing
biological treatment. Environ Sci Pollut Res 2009;16:521–30.
Backhaus T, Altenburger R, Boedeker W, Faust M, Scholze M, Grimme LH. Predictability of
the toxicity of a multiple mixture of dissimilarly acting chemicals to Vibrio fischeri.
Environ Toxicol Chem 2000;19:2348–56.
Backhaus T, Faust M, Scholze M, Gramatica P, Vighi M, Grimme LH. Joint algal toxicity of
phenylurea herbicides is equally predictable by concentration addition and independent action. Environ Toxicol Chem 2004;23:258–64.
Behnisch PA, Hosoe K, Sakai S. Bioanalytical screening methods for dioxins and dioxinlike compounds — a review of bioassay/biomarker technology. Environ Int 2001;
27:413–39.
Berenbaum MC. The expected effect of a combination of agents, the general solution. J
Theor Biol 1985;114:413–31.
Bittner M, Macikova P, Giesy JP, Hilscherova K. Enhancement of AhR-mediated activity of
selected pollutants and their mixtures after interaction with dissolved organic
matter. Environ Int 2011;37:960–4.
Bliss CI. The toxicity of poisons applied jointly. Ann Appl Biol 1939;26:585–615.
Bosetti C, Boffetta P, La Vecchia C. Occupational exposures to polycyclic aromatic hydrocarbons, and respiratory and urinary tract cancers: a quantitative review to 2005.
Ann Oncol 2007;18:431–46.
Boström CE, Gerde P, Hanberg A, Jernström B, Johansson C, Kyrklund T, et al. Cancer risk
assessment, indicators, and guidelines for polycyclic aromatic hydrocarbons in the
ambient air. Environ Health Perspect 2002;110(Suppl. 3):451–88.
Brosche S, Backhaus T. Toxicity of five protein synthesis inhibiting antibiotics and their
mixture to limnic bacterial communities. Aquat Toxicol 2010;99:457–65.
Chen B, Xuan X, Zhu L, Wang J, Gao Y, Yang K, et al. Distributions of polycyclic aromatic
hydrocarbons in surface waters, sediments and soils of Hangzhou City, China.
Water Res 2004;38:3558–68.
Denison MS, Heath-Pagliuso S. The Ah receptor: a regulator of the biochemical and
toxicological actions of structurally diverse chemicals. Bull Environ Contam Toxicol
1998;61:557–68.
Denison MS, Pandini A, Nagy SR, Baldwin EP, Bonati L. Ligand binding and activation of
the Ah receptor. Chem Biol Interact 2002;141:3–24.
Drescher K, Boedeker W. Assessment of the combined effects of substances, the relationship between concentration addition and independent action. Biometrics 1995;51:
716–30.
Engwall M, Naf C, Broman D, Brunstrom B. Biological and chemical determination of
contaminant levels in settling particulate matter and sediments — a Swedish river
system before, during, and after dredging of PCB-contaminated lake sediments.
Ambio 1998;27:403–10.
Evans RM, Scholze M, Kortenkamp A. Additive mixture effects of estrogenic chemicals in
human cell-based assays can be influenced by inclusion of chemicals with differing
effect profiles. PLoS One 2012;7:e43606.
Falco G, Domingo JL, Llobet JM, Teixido A, Casas C, Muller L. Polycyclic aromatic hydrocarbons in foods, human exposure through the diet in Catalonia, Spain. J Food Prot 2003;
66:2325–31.
M. Larsson et al. / Environment International 73 (2014) 94–103
Faust M, Altenburger R, Backhaus T, Blanck H, Boedeker W, Gramatica P, et al. Predicting
the joint algal toxicity of multi-component s-triazine mixtures at low-effect concentrations of individual toxicants. Aquat Toxicol 2001;56:13–32.
He G, Tsutsumi T, Zhao B, Baston DS, Zhao J, Heath-Pagliuso S, et al. Third-generation Ah
receptor-responsive luciferase reporter plasmids: amplification of dioxin-responsive
elements dramatically increases CALUX bioassay sensitivity and responsiveness.
Toxicol Sci 2011;123:511–22.
Keiter S, Grund S, van Bavel B, Hagberg J, Engwall M, Kammann U, et al. Activities and
identification of aryl hydrocarbon receptor agonists in sediments from the Danube
river. Anal Bioanal Chem 2008;390:2009–19.
Kortenkamp A, Backhaus T, Faust M. State of the art report on mixture toxicity. Available
http://ec.europa.eu.eu/environment/chemicals/pdf/report_Mixture%20toxicity.pdf,
2009.
Larsson M, Orbe D, Engwall M. Exposure time-dependent effects on the relative potencies
and additivity of PAHs in the Ah receptor-based H4IIE-luc bioassay. Environ Toxicol
Chem 2012;31:1149–57.
Larsson M, Hagberg J, Rotander A, Bavel B, Engwall M. Chemical and bioanalytical characterisation of PAHs in risk assessment of remediated PAH-contaminated soils. Environ
Sci Pollut Res 2013;20:8511–20.
Larsson M, Hagberg J, Giesy JP, Engwall M. Time-dependent changes in relative potency
factors (REPs) for PAHs and their derivatives in the H4IIE-luc assay. Environ Toxicol
Chem 2014;33:943–53.
Layshock JA, Wilson G, Anderson KA. Ketone and quinone-substituted polycyclic aromatic
hydrocarbons in mussel tissue, sediment, urban dust, and diesel particulate matrices.
Environ Toxicol Chem 2010;29:2450–60.
Lundstedt S, Haglund P, Oberg L. Degradation and formation of polycyclic aromatic compounds during bioslurry treatment of an aged gasworks soil. Environ Toxicol Chem
2003;22:1413–20.
Lundstedt S, White PA, Lemieux CL, Lynes KD, Lambert IB, Oberg L, et al. Sources, fate, and
toxic hazards of oxygenated polycyclic aromatic hydrocarbons (PAHs) at PAHcontaminated sites. Ambio 2007;36:475–85.
Machala M, Vondracek J, Blaha L, Ciganek M, Neca J. Aryl hydrocarbon receptor-mediated
activity of mutagenic polycyclic aromatic hydrocarbons determined using in vitro reporter gene assay. Mutat Res Genet Toxicol Environ 2001a;497:49–62.
Machala M, Ciganek M, Bláha L, Minksová K, Vondracek J. Aryl hydrocarbon receptormediated and estrogenic activities of oxygenated polycyclic aromatic hydrocarbons
and azaarenes originally identified in extracts of river sediments. Environ Toxicol
Chem 2001b;20:2736–43.
Marlowe JL, Puga A. Aryl hydrocarbon receptor, cell cycle regulation, toxicity, and tumorigenesis. J Cell Biochem 2005;96:1174–84.
Murk AJ, Legler J, Denison MS, Giesy JP, Van De Guchte C, Brouwer A. Chemical-activated
luciferase gene expression (CALUX): a novel in vitro bioassay for Ah receptor active
compounds in sediments and pore water. Fundam Appl Toxicol 1996;33:149–60.
Musa Bandowe BA, Shukurov N, Kersten M, Wilcke W. Polycyclic aromatic hydrocarbons
(PAHs) and their oxygen-containing derivatives (OPAHs) in soils from the Angren industrial area, Uzbekistan. Environ Pollut 2010;158:2888–99.
Nebert DW, Roe AL, Dieter MZ, Solis WA, Yang Y, Dalton TP. Role of the aromatic
hydrocarbon receptor and [Ah] gene battery in the oxidative stress response, cell
cycle control, and apoptosis. Biochem Pharm 2000;59:65–85.
103
Nestler FHM. Characterization of wood-preserving coal-tar creosote by gas–liquid chromatography. Anal Chem 1974;46:46–53.
Nwaneshiudu OC, Autenrieth RL, McDonald TJ, Donnelly KC, Degollado ED, Abusalih AA.
Risk of polycyclic aromatic hydrocarbon (PAH) exposure from ingested food, the
Azerbaijan case study. J Environ Sci Health B 2007;42:201–9.
Ramirez N, Cuadras A, Rovira E, Marce RM, Borrull F. Risk assessment related to atmospheric polycyclic aromatic hydrocarbons in gas and particle phases near industrial
sites. Environ Health Perspect 2011;119:1110–6.
Richter CA, Giesy JP, Tieber VL, Denison MS. An in vitro rainbow trout cell bioassay for aryl
hydrocarbon receptor-mediated toxins. Environ Toxicol Chem 1997;16:543–50.
Sanderson JT, Giesy JP. Functional response assays in wildlife toxicology. In: Meyers RA,
editor. Encyclopedia of environmental analysis and remediation. New York: John
Wiley and Sons Inc.; 1998. p. 5272–97.
Sanderson JT, Arts J, Brouwer A, Froese KL, Giesy JP. Comparison of Ah receptor-mediated
luciferase and ethoxyresorufin-O-deethylase induction in H4IIE cells: implications for
their use as bioanalytical tools for the detection of polyhalogenated aromatic hydrocarbons. Toxicol Appl Pharmacol 1996;137:316–25.
Šídlová T, Novák J, Janošek J, Anděl P, Giesy JP, Hilscherová K. Dioxin-like and endocrine
disruptive activity of traffic-contaminated soil samples. Arch Environ Con Tox 2009;
57:639–50.
Sjögren M, Ehrenberg L, Rannug U. Relevance of different biological assays in assessing
initiating and promoting properties of polycyclic aromatic hydrocarbons with respect
to carcinogenic potency. Mutat Res 1996;358:97–112.
Söderström H, Hajšlová J, Kocourek V, Siegmund B, Kocan A, Obiedzinski MW, et al. PAHs
and nitrated PAHs in air of five European countries determined using SPMDs as passive samplers. Atmos Environ 2005;39:1627–40.
Sovadinová I, Bláha L, Janošek J, Hilscherová K, Giesy JP, Jones PD, et al. Cytotoxicity and
aryl hydrocarbon receptor-mediated activity of N-heterocyclic polycyclic aromatic
hydrocarbons: structure–activity relationships. Environ Toxicol Chem 2006;25:
1291–7.
Thavamani P, Megharaj M, Krishnamurti GSR, McFarland R, Naidu R. Finger printing of
mixed contaminants from former manufactured gas plant (MGP) site soils: implications to bioremediation. Environ Int 2011;37:184–9.
Trombino AF, Near RI, Matulka RA, Yang S, Hafer LJ, Toselli PA, et al. Expression of the aryl
hydrocarbon receptor/transcription factor (AhR) and AhR-regulated CYP1 gene transcripts in a rat model of mammary tumorigenesis. Breast Cancer Res Treat 2000;63:
117–31.
Villeneuve DL, Khim JS, Kannan K, Falandysz J, Blankenship AL, Giesy JP. Relative potencies of individual polychlorinated naphthalenes to induce dioxin-like responses in
fish and mammalian in vitro bioassays. Arch Environ Contam Toxicol 2000a;39:
273–81.
Villeneuve DL, Blankenship AL, Giesy JP. Derivation and application of relative potency estimates based on in vitro bioassay results. Environ Toxicol Chem 2000b;19:2835–43.
Warne MS, Hawker DW. The number of components in a mixture determines whether
synergistic and antagonistic or additive toxicity predominate, the funnel hypothesis.
Ecotoxicol Environ Saf 1995;31:23–8.
Zhang YH, Liu SS, Song XQ, Ge HL. Prediction for the mixture toxicity of six organophosphorus pesticides to the luminescent bacterium Q67. Ecotoxicol Environ Saf 2008;71:
880–8.
SUPPLEMENTARY DATA
Chemicals and mixture compositions
Table S1. Concentrations of single PACs in different mixtures prepared in non-equivalent effective
concentrations (µM).
Mixture
Compounds
mix 2A
mix 5A
mix 3
mix 5B
mix 7B
mix 18B
mix 18C
mix 23
1
Benzo[a]fluorene
2
Naphthacene
3
Chrysene
4
Benzo[a]anthracene
0.40
0.080
0.40
5
Benzo[k]fluoranthene
0.0024
0.00048
0.0024
6
Benzo[j]fluoranthene
0.030
0.0060
0.030
7
Benzo[b]fluoranthene
0.039
0.0079
0.039
8
Benzo[a]pyrene
0.16
0.032
0.16
9
Indeno[1,2,3-cd]pyrene
0.061
0.012
0.061
10
Dibenz[a,c]anthracene
0.012
0.0025
0.012
11
Dibenz[a,j]anthracene
0.010
0.0021
0.010
12
Dibenz[a,h]anthracene
0.012
0.0023
0.012
13
Naphtho[2,3-a]pyrene
0.062
0.012
0.062
14
2-Methylanthracene-9,10-dione
15
Naphthacene-5,12-dione
1.2
16
9,10-dihydrobenzo[a]pyren-(8H)-none
5.4
2.69
3.9
0.78
3.9
17
Benzo[a]fluorenone
5.0
2.52
2.1
0.42
2.1
18
Benz[a]anthracene-7,12-dione
5.0
2.52
5.4
1.1
5.4
19
7H-Benz[de]anthracen-7-one
4.20
4.2
0.84
4.2
20
1,4-Chrysenequinone
21
1-Indanone
5.0
2.5
22
Anthracene-9,10-dione
4.0
2.0
23
4H-cyclopenta[def]phenanthrenone
5.0
2.5
24
Dibenz[ah]acridine
25
Acridine
26
Benzo[h]quinoline
27
Quinoline
Total mixture concentrations (µM)
3.4
2.42
1.9
1.54
8.4
4.9
0.37
0.073
0.37
0.0071
0.0014
0.0071
0.20
0.039
0.20
1.68
0.3
0.0020
0.00048
0.0020
20.2
8.0
4.0
8.0
20.5
21.5
13.7
30
4.0
17.6
17
3.4
32
Table S2. The concentration ratios (pi) in percent for mixtures composed of PACs in equivalent effective
concentrations (EC5 or EC50).
mix18
Compounds
EC50
1
Benzo[a]fluorene
2.2
2
Naphthacene
0.042 0.0001
3
Chrysene
1.2
4
Benzo[a]anthracene
2.3
5
Benzo[k]fluoranthene
6
Benzo[j]fluoranthene
7
EC5
Mix13
mix11
EC50
EC5
27
11
0.52
0.03
0.58
15
18
1.9
1.9
29
57
3.9
0.014 0.0004
0.18
0.01
0.18
0.004
2.2
0.12
Benzo[b]fluoranthene
0.23
0.003
2.9
8
Benzo[a]pyrene
0.93
0.31
9
Indeno[1,2,3-cd]pyrene
0.36
0.065
0.37
EC50
EC5
EC5
35
23
92
2.5
46
6.6
0.024 0.0006
0.28
0.002
0.08
0.39
0.004
4.5
0.009
12
9.4
1.55
0.41
18
1.08
4.5
2.0
0.6
0.09
7.0
0.23
1.3
0.042
10 Dibenz[a,c]anthracene
0.073 0.003
0.92
0.10
11 Dibenz[a,j]anthracene
0.060 0.038
0.76
1.2
0.1
0.05
12 Dibenz[a,h]anthracene
0.068 0.012
0.85
0.37
0.11
0.02
13 Naphtho[2,3-a]pyrene
0.36
4.5
0.21
38
35
53
27
0.007
16 9,10-dihydrobenzo[a]pyren-(8H)-none 23
26
17 Benzo[a]fluorenone
12
12
18 Benz[a]anthracene-7,12-dione
32
20.2
19 7H-Benz[de]anthracen-7-one
25
38
24 Dibenz[ah]acridine
0.012 0.0001
mix7
EC50
0.2
0.0002
mix4
EC50
EC5
25
27
13
12
35
21
27
40
Table S3. The concentration ratios (pi) in percent for mixtures composed of PACs in non-equivalent effective
concentrations.
Mixture
Compounds
mix2A
mix5A
mix3
mix5B
mix7B
mix18B
mix18C
mix23
1
Benzo[a]fluorene
2.2
2.2
1.1
2
Naphthacene
0.04
0.04
0.02
3
Chrysene
1.2
1.16
0.62
4
Benzo[a]anthracene
2.4
2.4
1.2
5
Benzo[k]fluoranthene
0.01
0.01
0.008
6
Benzo[j]fluoranthene
0.18
0.18
0.09
7
Benzo[b]fluoranthene
0.23
0.23
0.12
8
Benzo[a]pyrene
0.93
0.93
0.43
9
Indeno[1,2,3-cd]pyrene
0.36
0.36
0.19
10
Dibenz[a,c]anthracene
0.07
0.07
0.04
11
Dibenz[a,j]anthracene
0.06
0.06
0.03
12
Dibenz[a,h]anthracene
0.07
0.07
0.04
13
Naphtho[2,3-a]pyrene
0.36
0.36
0.19
14
2-Methylanthracene-9,10-dione
15
Naphthacene-5,12-dione
5.4
16
9,10-dihydrobenzo[a]pyren-(8H)-none
25
15
23
23
12
17
Benzo[a]fluorenone
24
14
12
12
6.5
18
Benz[a]anthracene-7,12-dione
24
14
32
32
17
19
7H-Benz[de]anthracen-7-one
24
25
25
13
20
1,4-Chrysenequinone
21
1-Indanone
17
7.8
22
Anthracene-9,10-dione
13
6.3
23
4H-cyclopenta[def]phenanthrenone
17
24
Dibenz[ah]acridine
1.4
25
Acridine
99
26
Benzo[h]quinoline
27
13
27
Quinoline
27
13
25
9.6
14
8.8
61
23
14
7.8
0.012
0.012
0.006