Adverse Outcome Pathway (AOP) Informed Modeling of Aquatic

Feb 18, 2016 - This work has been funded in part by the eTOX project, grant agreement ... Jeffries , M. K. S.; Stultz , A. E.; Smith , A. W.; Rawlings...
0 downloads 0 Views 726KB Size
Subscriber access provided by UNIV OSNABRUECK

Article

Adverse Outcome Pathway (AOP) informed modelling of aquatic toxicology - QSARs, read-across and inter-species verification of modes of action Claire Marie Ellison, Przemyslaw Piechota, Judith Clare Madden, Steven Enoch, and Mark Cronin Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.5b05918 • Publication Date (Web): 18 Feb 2016 Downloaded from http://pubs.acs.org on March 2, 2016

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 27

Environmental Science & Technology

1

Adverse Outcome Pathway (AOP) informed modelling of aquatic toxicology - QSARs, read-across

2

and inter-species verification of modes of action

3

Claire M. Ellison, Przemyslaw Piechota, Judith C. Madden*, Steven J. Enoch, and Mark T.D. Cronin

4

School of Pharmacy and Biomolecular Sciences, Liverpool John Moores University, Byrom Street,

5

Liverpool, L3 3AF, England

6

Tel: +44 151 231 2164; Fax: +44 151 2170

7

Email: [email protected]

8

Abstract

9

Alternative approaches have been promoted to reduce the number of vertebrate and invertebrate

10

animals required for assessment of the potential of compounds to cause harm to the aquatic

11

environment. A key philosophy in the development of alternatives is greater understanding of the

12

relevant adverse outcome pathway (AOP). One alternative method is the fish embryo toxicity (FET)

13

assay. Although the trends in potency have been shown to be equivalent in embryo and adult assays, a

14

detailed mechanistic analysis of the toxicity data has yet to be performed; such analysis is vital for a

15

full understanding of the AOP. The research presented herein used an updated implementation of the

16

Verhaar scheme to categorise compounds into AOP informed categories. These were then used in

17

mechanistic (Quantitative) Structure-Activity Relationship ((Q)SAR) analysis

18

descriptors governing the distinct mechanisms of acute fish toxicity) are capable of modelling data

19

from the FET assay. The results show that compounds do appear to exhibit the same mechanisms of

20

toxicity across life stages. Thus this mechanistic analysis supports the argument that the FET assay is

21

a suitable alternative testing strategy for the specified mechanisms, and that understanding the AOPs

22

is useful for toxicity prediction across test systems.

23

1. Introduction

24

The acute aquatic toxicity assessment of chemicals has traditionally been performed on species

25

representing various trophic levels e.g. algae, invertebrates, and juvenile and adult fish from species

26

such as the fathead minnow (Pimephales promelas), rainbow trout (Oncorhynchus mykiss) and

27

Japanese medaka (Oryzias latipes) amongst others. However, legislative mandates such as the

28

Registration, Evaluation, Authorisation, and restriction of Chemicals (REACH) regulation in the

29

European Union have required alternative, non-animal, models to be sought as a replacement for the

30

expensive, time-consuming and ethically questionable in vivo assessment methods. One such

31

alternative assay is the fish embryo toxicity (FET) test1. The FET has a standardised OECD test

32

guideline (number 2362) for the 96hr assay performed with zebrafish (Danio rerio) embryos and,

33

although not explicitly stated in the guideline, can be considered as a suitable assessment method, on 1 ACS Paragon Plus Environment

to show that the

Environmental Science & Technology

34

its own or as part of a strategy, for assessing acute aquatic toxicity. For instance, reasonable

35

correlations (r>0.85) have been observed between the 50% effect concentrations (EC50) measured in

36

the FET and the 50% lethality concentrations (LC50) measured in fish 3-6; although a small number of

37

notable mechanisms are poorly predicted in FET (e.g. neurotoxic compounds which require

38

behavioural analysis of the embryos for toxicity to be observed 6, 7).

39

One of the key philosophies behind the research into alternative methods to animal testing is that of

40

understanding mode of action. Recently, the assessment of modes of action has been incorporated into

41

adverse outcome pathways (AOPs)8. AOPs define a series of key events (KE), and their relationships

42

(key event relationships (KERs)) from an initial exposure, resulting in a molecular initiating event

43

(MIE), through to inducing the adverse outcome (AO)9. The MIE may be described in terms of the

44

chemically defining features of a molecule that control the interaction with the biological

45

macromolecule10, whereas the MIE combined with the required KEs for an AO encompass the

46

biological mode of action8. Understanding the AOP can thus aid in the elucidation of the similarities

47

and differences in the mode of action between species by identifying key uncertainties, and

48

corresponding research gaps, in the biological mechanisms of toxicity11. The rationale for the

49

chemical induction of an MIE is a key aspect of understanding the AOP.

50

The modes of action for acute aquatic toxicity have been established through studies on fish behaviour

51

and physiology

52

resultant data14, and more recently on detailed systems biology studies on species from lower taxa

53

such as Daphnia magna15. Fish Acute Toxicity Syndromes (FATS) were derived from measurements

54

of physiological, biochemical and analytical effects separated into discrete mechanisms, or modes, of

55

action12. Building on such knowledge, Verhaar et al16 used fish acute toxicity data to identify clear

56

structural rules associated with a variety of modes or mechanisms of action. The Verhaar scheme

57

utilises 2D chemical structure to classify potential environmental pollutants into one of four categories

58

representing one, or more, mechanisms of action: class 1 (narcosis or baseline toxicity), class 2 (less

59

inert compounds), class 3 (unspecific reactivity) and class 4 (compounds and groups of compounds

60

acting by a specific mechanism). In addition, Russom et al13 used structural classes to assign

61

mechanisms of action to a range of compounds tested on the fathead minnow (Pimephales promelas).

62

This grouping of compounds allows for the development of mechanistically based, local Quantitative

63

Structure Activity Relationship (QSAR) models and also application of the chemical activity

64

principle17.

65

Understanding which Verhaar class, or mode/mechanism of action, a compound belongs to is useful

66

for hazard characterisation, not only identifying compounds predicted to act by narcosis, but also

67

identifying those which may elicit excess toxicity (e.g. compounds in Verhaar classes 3 and 4).

68

Within a mechanistic class it is possible to build high quality (Q)SAR models, based on knowledge of

12, 13

, as well as mechanistic structure-activity relationship (SAR) analysis on the

2 ACS Paragon Plus Environment

Page 2 of 27

Page 3 of 27

Environmental Science & Technology

69

the relevant mechanism/mode of action of toxicity, which are able to estimate relative toxic potency18.

70

These mechanistic models have been shown to provide more transparency and greater statistical

71

performance than equivalent global models18-21. Understanding the mechanism of action also fits well

72

within the AOP framework of toxicity assessment22 and allows inter-species toxicity correlations to be

73

applied within a single mechanism23, 24.

74

The transparency of the Verhaar classification scheme has assisted in its popularity as a hazard

75

characterisation tool and this popularity has in turn led to automated implementations of the scheme

76

becoming widely available. One such implementation is available in the Toxtree software. Toxtree

77

was developed by Ideaconsult Ltd (Sofia, Bulgaria) under the terms of a contract from the European

78

Commission’s Joint Research Centre (JRC). The software encodes several decision trees and

79

classification schemes useful for analysing the potential toxicity hazards of compounds25. The

80

software is freely available (http://Toxtree.sourceforge.net) and version (2.6) includes two forms of

81

the Verhaar decision tree: “Verhaar scheme” and “Verhaar scheme (modified)”. The “Verhaar scheme”

82

is the original implementation of the decision tree based directly on the scheme as it is described by

83

Verhaar et al16. Enoch et al26 assessed the performance of this implementation and suggested possible

84

improvements. These improvements form the basis of “Verhaar scheme (modified)” decision tree.

85

Recent work by Ellison et al27 has shown that the implementation of the “Verhaar scheme (modified)”

86

tree, and hence the suitability of resultant categories for modelling, can be improved further.

87

Specifically, with the use of post-processing filters for polar phenols, reactive aromatic compounds,

88

cyclic non-aromatic hydrocarbons and respiratory uncouplers of oxidative phosphorylation, the

89

positive predictivity of each of the categories was increased by an average of 5%. These filters are

90

available from the authors as a KNIME workflow (www.KNIME.org) for use on the output of

91

Toxtree v2.6.

92

Whilst the Verhaar scheme is accepted for adult fish, it is not known how applicable it may be to

93

other assays such as FET, nor what mechanistic information may be derived from assessing data from

94

such assays. Therefore, the aim of this study was to examine if the mechanistic categories, formed by

95

using the Verhaar scheme classes implemented through the use of Toxtree v2.6 and the Ellison et al

96

KNIME post-processing workflow, are relevant to the AOPs of both juvenile/adult fish and fish

97

embryos. To this end AOP relevant mechanistic (Q)SAR analysis was performed on compounds with

98

measured toxicity in the FET assay and outliers were highlighted. These outliers were of interest as

99

they represent compounds where observed toxicity is in contradiction to that expected according to

100

the mechanism of action for that class, and they may provide useful species specific information.

101

These compounds may either be acting via different mechanisms in the FET assay, have been

102

misclassified by the scheme, or be a misrepresentation of the chemical toxicity due to erroneous data

103

or other effects such as volatility and degradation.

3 ACS Paragon Plus Environment

Environmental Science & Technology

104

2. Methods

105

2.1. Dataset

106

Published experimental results from the FET assay performed using zebrafish (Danio rerio) were

107

manually curated from two literature sources3, 4 into a single dataset. The names, CAS numbers and

108

EC50 values were extracted. The full range of exposure times (24hr – 120hr) were used, which

109

included specimens in both the embryonic and eleutheroembryonic stages of development. Data from

110

all time points were considered equal as it has been previously established that eleutheroembryo and

111

embryo studies generally provided highly similar results4, although some compounds only show

112

significant toxicity at the eleutheroembryo stage and hence it was important to include all data points.

113

The 24hr testing period could provide a lower indication of toxicity because the test duration was

114

insufficient for the compound the reach equilibrium due to the compound’s toxicokinetic properties.

115

However, the small number of compounds (n