Subscriber access provided by Gothenburg University Library
Article
Can toxicokinetic and toxicodynamic modelling be used to understand and predict synergistic interactions between chemicals? Nina Cedergreen, Kristoffer Dalhoff, Dan Li, Michele Gottardi, and Andreas C. Kretschmann Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b02723 • Publication Date (Web): 13 Sep 2017 Downloaded from http://pubs.acs.org on September 13, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 22
Environmental Science & Technology
1
Can toxicokinetic and toxicodynamic modelling be used to understand
2
and predict synergistic interactions between chemicals?
3
Nina Cedergreena*, Kristoffer Dalhoffa, Dan Liab, Michele Gottardia and Andreas C. Kretschmannac
4 5
*Corresponding author, phone: +45 29611743,
[email protected] 6
a
7
University of Copenhagen
8
Thorvaldsensvej 40
9
1871 Frederiksberg C
10
Department of Plant and Environmental Science
Denmark
11 12
b
13
Key Laboratory of Pollution Ecology and Environmental Engineering,
14
Institute of Applied Ecology,
15
Chinese Academy of Sciences,
16
110016, China
Present address:
17 18
c
19
University of Copenhagen
20
Universitetsparken 2
21
2100 Copenhagen Ø
22
Denmark
Toxicology Lab, Department of Pharmacy and Analytical Biosciences
1 ACS Paragon Plus Environment
Environmental Science & Technology
Page 2 of 22
23
ABSTRACT
24
Some chemicals are known to enhance the effect of
25
other chemicals beyond what can be predicted with
26
standard mixture models, such as concentration
27
addition and independent action. These chemicals are
28
called synergists. Up until now, no models exist that
29
can predict the joint effect of mixtures including
30
synergists. The aim of the present study is to develop a
31
mechanistic toxicokinetic (TK) and toxicodynamic
32
(TD) model for the synergistic mixture of the azole fungicide, propiconazole (the synergist), and the
33
insecticide, α-cypermethrin, on the mortality of the crustacean D. magna. The study tests the hypothesis that
34
the mechanism of synergy is the azole decreasing the biotransformation rate of α-cypermethrin and validates
35
the predictive ability of the model on another azole with a different potency: prochloraz. The study showed
36
that the synergistic potential of azoles could be explained by their effect on the biotransformation rate but
37
that this effect could only partly be explained by the effect of the two azoles on cytochrome P450 activity,
38
measured on D. magna in vivo. TKTD models of interacting mixtures seem to be a promising tool to test
39
mechanisms of interactions between chemicals. Their predictive ability is, however, still uncertain.
TOC art
40
2 ACS Paragon Plus Environment
Page 3 of 22
Environmental Science & Technology
41
INTRODUCTION
42
Some chemicals are known to enhance the effect of other chemicals beyond what can be predicted with
43
standard mixture models, such as concentration addition and independent action. These chemicals are called
44
synergists. Up till now no models exist that can predict the joint effect of mixtures including synergists.
45
Azole fungicides have been shown to act as synergists in a range of studies, enhancing the toxicity of
46
pyrethroid insecticides up to 10-50-fold in a range of organisms1-4. Also other pesticide or biocide
47
combinations have shown to induce repeatable synergy in a range of organisms 5. Synergy and antagony
48
between chemicals can occur through a range of mechanisms: First, one chemical can affect the availability
49
of another chemical outside an organism through either precipitation or change in speciation, as it has been
50
demonstrated for metals in mixtures6-8. Secondly, one chemical can affect the uptake rate of another chemical
51
e.g. by enhancing penetration9, or by facilitating availability by enhancing ventilation rates in aquatic
52
organisms10. Thirdly, one chemical can affect the transport of another chemical to its target, as it is often
53
observed in plants11, or a chemical can affect the biological action of other chemicals by either inhibiting or
54
promoting their transformation through interactions with biotransformation enzymes such as cytochrome
55
P450 monooxygenases and esterases12, 13. Finally, chemicals can compete for a common target site or affect
56
the excretion of one another.
57 58
In a review of synergistic interactions, 95 % of all documented pesticide synergies were caused by either
59
azole fungicides or carbamate and organophosphate insecticides, known to inhibit cytochrome P450
60
monooxygenases and/or esterases5. Hence, for pesticide mixtures it seems as if interactions involving
61
biotransformation of other pesticides are the main mechanism behind the observed synergies. Despite of this
62
hypothesis often being cited, very little direct evidence exists proving that inhibition or activation of
63
biotransformation of other chemicals is the single most important mechanism of synergy of azole fungicides,
64
carbamate and organophosphate insecticides3,10,14-16. For azole/pyrethroid interactions, for example,
65
ChalvetMonfrey et al (1996) found that the synergy of prochloraz on deltamethrin in bees could not be
66
explained by effects on biotransformation alone17, but that effects on uptake rates were also likely to take
67
place 16.
68 69
Synergists acting on biotransformation pathways could potentially be screened by using in vitro or in vivo
70
assays for the effect of chemicals on different metabolic enzymes12,18,19. But even if a chemical is known to
71
inhibit certain enzymes, the size of the potential synergistic interactions and its development over time,
72
cannot be quantified with any existing model approach20. A possible tool to test mechanisms of synergy and
73
ultimately to predict the size of synergy over time are toxicokinetic (TK) and toxicodynamic (TD) models21.
74
TK models predict uptake and elimination of chemicals over time and TD models predict the development of
75
effects over time as a function of the modelled or measured internal chemical concentrations21,22. Mixture
3 ACS Paragon Plus Environment
Environmental Science & Technology
Page 4 of 22
76
toxicity calculations using TKTD models have been proposed, but only for non-interacting chemicals with
77
similar molecular target sites23. TKTD models for interactive chemicals could be a tool to test hypotheses on
78
the mechanism of interaction. If they are successful, they may also be used to predict the size of synergistic
79
interactions under different exposure scenarios.
80 81
The aim of the present study is therefore three-fold: First, we wish to build and parameterise a full TKTD
82
model for the synergistic interactions between the azole fungicide propiconazole (the synergist) and the
83
pyrethroid insecticide α-cypermethrin on the mortality of the crustacean D. magna (Figure 1), when
84
propiconazole is present at one constant concentration. Secondly, we wish to test the different hypotheses
85
concerning the mechanism of synergy (effects on uptake versus the effect on biotransformation rate), and
86
thirdly, we will validate the model assumptions in terms of synergistic interactions with a synergist with a
87
similar mode of action but different potency, the azole fungicide prochloraz. To confirm the hypothesis that
88
the azoles induce synergy through interference with biotransformation, the model is parameterised to a
89
dataset with variable propiconazole or prochloraz concentrations, and the modelled effect on
90
biotransformation rate is compared with cytochrome P450 activity inhibition measured in vivo.
91
92 93
Figure 1. The figure shows a conceptual model of the toxicokinetic (left side) and toxicodynamic (right side) processes
94
of a pyrethroid insecticide and an azole fungicide and their interactions in Daphnia magna, symbolised by the large
95
square. Inside the daphnid, there are two targets for the pesticides: the sodium channel (grey half-circle) which is the
96
target site of the pyrethroid, and P450 enzymes (grey circle), the main target site for the azoles. The pyrethroids act as
4 ACS Paragon Plus Environment
Page 5 of 22
Environmental Science & Technology
97
substrates for the P450 enzymes when they are biotransformed (green circle). The state variables, describing how the
98
amounts of pesticide in the different locations change over time, are given in blue circles for the pyrethroids and by red
99
squares for the azoles, and are all described by differential equations given in the text. The rate constants, describing TK
100
processes, are given next to the solid arrows denoting the specific processes (values are given in Table 1), while
101
parameters relating damage to mortality, assuming GUTS-SD (Equation 8 and 9), are given next to the grey dashed
102
arrows. The synergistic interaction is proposed to occur when azoles bind to P450 enzymes, making them unavailable
103
for pyrethroid biotransformation, thereby decreasing the rate by which the P450 enzymes can biotransform the
104
pyrethroid (red dashed arrow). The alternative hypothesis for synergy, where azoles affect pyrethroid uptake rates, is
105
denoted by a red dotted arrow. Mechanisms that are neglected in the first versions of the model, but which might be of
106
importance, such as direct effects of the azoles on daphnia mortality, or effects on uptake rates due to the α-
107
cypermethrin damage done to daphnia mobility, are denoted by dotted grey arrows.
108
METHODS
109
Theory The conceptual model is shown in Figure 1. It is initially assumed that the uptake of the pyrethroid
110
will follow a first order kinetic uptake and elimination model including a biotransformation rate constant,
111
km_pyr, describing the rate by which the pyrethroid is biotransformed in the organism.
112 113
_ ( )
= _ ∗ _ () − _ ∗ _ () − _ ∗ _ ()
(1)
114 115
In this equation the change in internal pyrethroid concentrations, Cin_pyr, in the daphnids over time is
116
described as a function of the uptake rate, kin_pyr, the excretion rate, kout_pyr, the biotransformation rate, km_pyr,
117
and the external pyrethroid concentration in the water, Cw_pyr. Phase I biotransformation is, in the case of
118
pyrethroids, believed mainly to be governed by P450 monooxygenases 24, though esterases may also play a
119
substantial role 24-26. As pyrethroids are very hydrophobic (log Kow = 6.94 at pH 7 for α-cypermethrin),
120
sorption to the daphnid exoskeleton could also be a process of quantitative significance. It is given in Figure
121
1 as Csorp_pyr, and the change over time of Csorp_pyr is proposed to be described with first order kinetics using a
122
sorption specific uptake and elimination rate constant, ksorp_pyr and kdesorp_pyr, respectively:
123 124
_ ( )
= _ ∗ _ () − _ ∗ _ ()
(2)
125 126
The toxicokinetics of the azoles are described with a simplification of Equation 1, only including an uptake
127
and elimination rate constant, kin_az and kout_az20. The elimination rate constant in this equation therefore
128
describes the sum of all biotransformation processes and efflux of the parent azole compound:
129 130
_ ( )
= _!" ∗ () − ∗ _!" ()
(3)
131 5 ACS Paragon Plus Environment
Environmental Science & Technology
Page 6 of 22
132
To test the hypothesis that the synergy is caused by the effect of the azole on the biotransformation rate, it
133
was initially assumed that both pyrethroid uptake, kin_pyr, and excretion, kout_pyr, were independent of the
134
presence of the azole, and that the only effect of the azole was on km_pyr. As azoles bind to the catalytic site of
135
the P450 enzymes, thereby prohibiting binding of the pyrethroid for biotransformation, we assume
136
competitive inhibition of P450 enzymes by the azoles27. This means that the presence of azoles will decrease
137
the amount of active P450 enzymes with a fraction depending on the internal azole concentration. This
138
fraction is given by the parameter s. The parameter s can be defined by the ratio of the biotransformation rate
139
constant km_pyr with and without co-exposure to the azole under steady state conditions.
140 141
$%_ (& ' )
#=$
(4)
% ( & ' )
142 143
For a variable internal concentration of azoles, we expect s to vary according to the internal concentration of
144
the azole, cin_az, following a sigmoidal function. We here describe the relationship with a log-logistic two-
145
parameter model, where IC50 is the internal azole concentration inhibiting the biotransformation rate of the
146
pyrethroid by 50% and b is the slope parameter of the curve:
147
# = )*(&
) _
(5)
/,-. )/
148
We choose to use internal azole concentrations rather than scaled damage (Equation 7) to describe s, as it can
149
be measured experimentally. We recognise that the presence of the pyrethroid may also affect the activity of
150
P450 monooxygenases. However, as the pyrethroid acts as a substrate for the P450 enzymes rather than as an
151
inhibitor28, and in addition is expected to occur at much lower internal concentrations compared to the
152
azoles, we assume the quick catalytic biotransformation action will not significantly affect the total pool of
153
P450 catalytic sites available. Hence, we chose not to include the pyrethroid’s effect on P450 activity in the
154
presented model
155
The toxicodynamic part of the model describes the relation between internal pyrethroid concentrations and
156
observed mortality. All azole concentrations included in the studies of synergy are chosen not to affect
157
daphnid mortality (< EC10 2). Hence, mortality was assumed to depend on internal pyrethroid concentrations
158
alone. Toxicodynamic parameters for the azoles are, however, inserted in Figure 1, and given in Table 1, as
159
they have been determined in a previous publication20, and will be used in the combined TKTD-model
160
(Figure 1). Internal pyrethroid concentrations were related to mortality by including a damage stage,
161
assuming that the pyrethroid insecticide induces some undefined damage with a rate, kdam_pyr, proportional to
162
the internal pyrethroid concentration, and that the damage can be repaired by a rate, kdr_pyr, proportional to
163
the size of the damage.
6 ACS Paragon Plus Environment
Page 7 of 22
164
Environmental Science & Technology
0 ( )
= !_ ∗ _ () − _ ∗ 1 ())
(6)
165
This is analogue to how internal chemical concentrations depend on external concentrations over time. In
166
this case, however, we cannot measure damage directly. Pyrethroids inhibit the sodium channels of the
167
nerves 28, which will lead to a range of biochemical disruptions in the organism, which ultimately leads to
168
immobilisation and death. Because damage can rarely be measured directly, the authors of Jager et al. (2011)
169
introduced the concept of scaled damage, D*, which is proportional to the actual (but undefined) damage
170
level, and has the unit of an internal concentration21. This is done by dividing Equation 6 with the ratio of
171
damage accrual and damage repair, kdam_pyr/kdr_pyr, thereby getting:
172
0 ∗ ( )
= _ ∗ (_ () − 1 ∗ ())
(7)
173
The parameter kdr_pyr can be determined from the time course of survival of the test organisms. How damage
174
relates to survival can be determined in two ways, representing extreme cases: One is assuming stochastic
175
death above a certain damage threshold (the GUTS-SD model), the other is assuming that the organisms in
176
the trial die, when they have exceeded an individual threshold for damage (the GUTS-IT model). For
177
derivation, discussion and testing of the two assumptions we refer to: Jager et al. (2011) and Ashauer et al.
178
(2011, 2015 and 2016) 21,29-31. Here we present the equations used to link scaled damage to survival under the
179
assumption of stochastic death. For the individual threshold implementation and test, see SI B. For GUTS-
180
SD, hazard to the organism, Hpyr, takes place when the damage increases above a certain threshold defined
181
by zpyr. Above zpyr, hazard increases proportionally with the damage with a rate defined by the killing rate
182
kk_pyr:
183 184
2 ( )
= $_ ∗ 345(0, (1 ∗ () − 8 )
(8)
185 186
The survival probability as a function of time Spyr(t) is calculated from the hazard, adding the background
187
hazard (hb), derived from observations of control mortality:
188 189
9 () = : ;(2 ( )*
