Cefic-lri Programme | European Chemical Industry Council

ECO43: Improving sediment toxicity testing design and data interpretation for very hydrophobic substances

Principal Investigator

Dr. Michiel Jonker
Institute for Risk Assessment Sciences (IRAS), Utrecht University
Yalelaan 104, 3584 CM Utrecht, the Netherlands
+31 30 253 5338
E-mail: m.t.o.jonker@uu.nl

Collaborators

Michiel T.O. Jonker
Assistant Professor
Institute for Risk Assessment Sciences (IRAS)
Yalelaan 104, 3584 CM Utrecht
m.t.o.jonker@uu.nl
+31 30 253 5338

Albert A. Koelmans
Professor
Dept. of Aquatic Ecology and Water Quality Management Wageningen University
Droevendaalsesteeg 3a, 6708 PB Wageningen
Bart.koelmans@wur.nl
+31 317483201

Joy A. McGrath
Senior Managing Scientist
Exponent
420 Lexington Avenue, Suite 1740, New York 10170
jmcgrath@exponent.com
+1 8452137159

Scientific advisors:

Joop L.M. Hermens; IRAS, Utrecht, the Netherlands
David R. Mount; US-EPA, Deluth, MN, USA

Description

Problem definition

Sediment toxicity testing with very hydrophobic organic chemicals (VHOCs)

Sediment toxicity tests are frequently performed on contaminated field sediments in weight-of-evidence approaches for the purpose of ecological risk assessment (ERA). Also, the tests are often required within the framework of international chemical regulations (e.g. REACH) as part of product risk assessments (Diepens et al, 2017). Several test methods with benthic organisms are available, including assays using Lumbriculus variegatus (survival), Hyallela azteca (survival/growth/ reproduction), and Chironomus spec. (survival/(time to) emergence) (Norberg-King et al, 2006), some of which are standardized (OECD, 2004; ECHA, 2008). Although the toxicity assays generally perform well and deliver useful information on product safety or ecological risks, application of the assays to very hydrophobic organic chemicals (VHOCs), i.e., chemicals with an octanol-water partition coefficient (logKow) above about 6, often results in biased and therefore unreliable data (Redman et al, 2014). Main reasons for this include poor test design and increased susceptibility of these substances to artifacts or confounding factors, i.e., physical, chemical, or biological factors that may alter the outcome of a toxicity test (Rand, 1995). These include e.g. temperature, pH, presence of other toxicants such as NH3, size of the test organism, and exposure duration.

Challenge I: Obtaining reproducible and meaningful dose-response relationships

Due to the very hydrophobic properties of VHOCs, testing these substances is particularly challenging and several additional artifacts potentially can occur. Information on the actual incidence of these artifacts and knowledge on how to circumvent them is however not available.

– VHOCs have very low aqueous solubility and slow dissolution kinetics. If spiked at high concentrations and with insufficient mixing, spiking may result in the formation of a separate phase of pure compound (droplets for liquid substances or crystals for solid compounds). Dissolution kinetics of these so-called non-aqueous phase liquids (NAPLs) and crystals are (very) slow and organisms may not be exposed to the intended dissolved compound concentrations, but to a pure phase instead. In the case of liquid substances, the formed NAPLs may cause ‘fouling’ or ‘oiling’ of the test organisms and thereby false positive responses, as these are ‘physical’ effects, not ‘biochemical’ toxic responses. In the case the test organism is able to avoid the NAPLs (e.g. Lumbriculus) (Muijs et al, 2010) or in the case of crystal formation, the dissolved concentration may be too low to obtain a response or a meaningful dose-response relationship.

– VHOC-sediment equilibration kinetics are slow. Once spiked into sediment, a sorption equilibrium needs to be established between the spiked compound and the sediment. If too short equilibration times are applied, the compound’s binding strength and hence its bioavailability will not represent environmentally-relevant conditions. Redman et al. (2014) suggested that at least 7 days are needed to allow stabilization of threshold effect concentrations. The extent to which test results may be biased by insufficient equilibration is however unclear and may be related to compound hydrophobicity (Redman et al, 2014). Appropriate, intensive mixing may help speeding up the kinetics, but conclusive evidence for this is lacking. Because of the slow equilibration (and dissolution) kinetics, it is furthermore challenging to obtain a homogeneous test matrix. Guidance on how to best obtain a well and realistically equilibrated and homogeneous test sediment is not available.

– Uptake kinetics are slow for VHOCs. Because of the slow toxicokinetics, effects may not be observed, which may stimulate the executer to test at high(er) concentrations. At elevated concentrations however, the sorptive capacity of the sediment may be exceeded, and a separate phase of pure compound (liquid or solid) may be formed, causing the above-described problems.

Challenge II: Exposure quantification

The traditional way of risk assessment and management of contaminated sediments is based on total, solvent-extractable concentrations of sediment-associated chemicals (Ortega-Calvo et al, 2015). However, within the scientific environmental community it is generally accepted that this approach does not lead to a realistic assessment of actual risks (Parkerton et al, 2014). Therefore, several methods for measuring the ‘bioavailable’ concentration of chemicals have been developed during the past decades. Among these methods, partitioning-based, non-depletive extraction methods with polymers (“passive sampling methods”) are considered the best developed and having the most solid scientific basis (Mayer et al, 2014). Through passive sampling, the freely dissolved concentration (Cfree) of a chemical in sediment pore water is determined, which is the driving force behind accumulation and toxicological effects in organisms (Lydy et al, 2014). When trying to understand, explain, or model sediment toxicity tests with VHOCs, data on Cfree are therefore crucial. Passive sampling to determine Cfree of moderately hydrophobic chemicals is well-developed and has successfully been applied in many cases (Lydy et al, 2014); however, the technique is more challenging for VHOCs, due to kinetic issues and particularly the difficulties in determining the required passive sampler-water partition coefficients (Jonker, 2015, 2016; Booij et al, 2016). These are only available nowadays for selected PAHs and PCBs and practical guidance or standardized protocols for their determination, as well as for measuring Cfree are not yet available.

Challenge III: Modeling

A simple, but for many nonpolar organic chemicals effective way of explaining and modeling sediment toxicity data across sediments and organisms is provided by a modelling framework combining the Equilibrium Partitioning (EqP) Theory (Di Toro et al, 2000a, b; Burgess et al, 2013) and the Target Lipid Model (TLM) (Di Toro et al, 2000a). The EqP Theory assumes that chemicals present in sediment, porewater, and organism are in a thermodynamic equilibrium, characterized equilibrium (sediment-water, organism-water) constants and that toxicity is related to Cfree. The TLM assumes that organism lipid is the target site for toxic effects, as characterized by a critical target lipid body burden (CTLBB). The modelling framework has been applied successfully to a wide range of chemicals to explain the toxicity in water and sediments (Redman et al, 2014). However, for VHOCs, reliable sediment toxicity data are scarce and validation of the approach has been challenging (Redman et al, 2014). On the one hand, uncertainty in the modelling results may be explained by data of insufficient quality (i.e., biased by the above-described artifacts); on the other hand, the modelling framework may not be valid. After all, it may not be possible to reach equilibrium conditions for VHOCs during the time course of toxicity tests due to their slow kinetics (Diepens et al, 2015; Sidney et al, 2016). High-quality data based on intelligent testing design are therefore first of all needed to investigate the validity of the EqP-TLM approach. If the modelling framework appears unsuitable for VHOCs, toxicokinetic(/toxicodynamic) modelling may be needed (Ashauer et al, 2007; Jager et al, 2005), although such an approach is not preferable, due to the many required input data (kinetic constants).

Objectives:

The above demonstrates that several important knowledge gaps exist in the field of VHOC sediment chemistry and toxicity testing. Testing is very challenging, and biased, unreliable data are often obtained. This hampers realistic product and ecological assessments and may lead to improper or even unwanted management of contaminated sediments and products. Therefore, there is an urgent need for intelligent testing design, practical guidance, and standard protocols for VHOC testing. The general objective of the proposed research is to improve sediment toxicity testing design, performance, and data interpretation for VHOCs and to develop guidance on both aspects in order to maximize realism and value of future testing and thereby to support product and ecological risk assessment. To satisfy this objective the project will:

1. Perform a critical literature review on VHOC sediment toxicity data; and design, based on the review results, tiered experimental work;
2. Develop, test, and compare different VHOC spiking methods and develop standardized protocols for the preferred approach(es);
3. Set up exposure quantification methods (passive sampling) for the VHOCs and develop standardized experimental protocols;
4. Evaluate the impacts of VHOC-specific test design parameters/confounding factors on results of sediment toxicity assays and demonstrate means to circumvent or deal with these. Factors that will be investigated include upper limit test concentrations, sediment-VHOC equilibration time before test initiation, exposure duration, and possibility of organism fouling;
5. Evaluate the applicability of the EqP-TLM modelling approach to VHOCs and propose modifications or alternatives if necessary;
6. Provide recommendations and develop guidance for sediment toxicity tests with VHOCs based on the project results and international expert opinions.

Advancing the state of the science

Although sediment toxicity testing has a long tradition, currently there are several gaps in our knowledge on the behavior and bioavailability of VHOCs in toxicity assays and in sediments in general. Moreover, there is a clear lack of high-quality methods and protocols for spiking, handling, and exposure assessment of VHOCs. The proposed project aims at addressing these important gaps and will increase our knowledge on processes involved in VHOC behavior and toxicity in sediments. Specifically, it will obtain novel mechanistic knowledge on the (concentration-dependent) phase distribution of liquid and solid VHOCs in sediments, toxicokinetics of VHOCs (based on Cfree and internal concentrations), the impact of VHOC-sediment contact time on bioavailability and toxicity, possibilities and the effects on assay responses of fouling with neat substances, and the applicability of the EqP-TLM approach for VHOCs. Furthermore, it will obtain practical knowledge on aspects such as the optimal way of VHOC spiking and homogenization; as well as on passive dosing and passive sampling of VHOCs in sediments, including kinetics. There is practically no experience in the scientific community with passive dosing in sediments. Having standardized protocols and being able to perform high-quality toxicity assays by circumventing confounding factors (Van der Heijden et al, 2016), will facilitate future high-quality work on VHOCs in sediments and soils, which thereby will further advance the state of the science on risk assessment of VHOCs.

Timeline: April 2018 > April 2021

LRI funding: €398361

Cefic-Lri Programme Responsible Care

© Copyright 2017 Cefic | European Chemical Industry Council. All rights reserved | Terms and Conditions of Use