The European technical report on aquatic effect-based monitoring tools under the water framework directive

The Water Framework Directive (WFD), 2000/60/EC, requires an integrated approach to the monitoring and assessment of the quality of surface water bodies. The chemical status assessment is based on compliance with legally binding Environmental Quality Standards (EQSs) for selected chemical pollutants (priority substances) of EU-wide concern. In the context of the mandate for the period 2010 to 2012 of the subgroup Chemical Monitoring and Emerging Pollutants (CMEP) under the Common Implementation Strategy (CIS) for the WFD, a specific task was established for the elaboration of a technical report on aquatic effect-based monitoring tools. The activity was chaired by Sweden and co-chaired by Italy and progressively involved several Member States and stakeholders in an EU-wide drafting group. The main aim of this technical report was to identify potential effect-based tools (e.g. biomarkers and bioassays) that could be used in the context of the different monitoring programmes (surveillance, operational and investigative) linking chemical and ecological status assessment. The present paper summarizes the major technical contents and findings of the report.


Introduction
The Water Framework Directive (WFD), 2000/60/EC, [1] requires an integrated approach to the monitoring and assessment of the quality of surface water bodies in the European Union. The assessment of ecological status takes into account the effects at the population and community levels, based on the use of specific indices and ecological quality ratios. The chemical status assessment is based on compliance with legally binding Environmental Quality Standards (EQSs) for 53 selected chemical pollutants (priority substances) of EU-wide concern [2].
Chemical analysis generally requires a priori knowledge about the type of substances to be monitored as, for technical and economic reasons, it is not possible to analyse, detect and quantify all substances that are present in the aquatic environment. Even for the thousands of unique substances registered under Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) it would be highly challenging to perform a chemical monitoring programme on the aquatic environment. Furthermore, to estimate the risk of effects related to the large number of substances that are present and detected in the environment (including pollutants of emerging concern, metabolites and transformation products), it would be necessary to develop a very large number of assessment criteria (EQS). Such assessment criteria for chemicals are generally developed substance by substance, based on laboratory studies, and usually do not consider the consequences of simultaneous exposure to multiple chemicals [3] occurring in the environment, possibly giving rise to cumulative effects [4].
In the mandate for 2010 to 2012 of the European subgroup Chemical Monitoring and Emerging Pollutants (CMEP) under the WG "Chemical Aspects" of the Common Implementation Strategy (CIS) for the WFD, a specific task was foreseen for the elaboration of a technical report on effect-based tools [5]. The activity was chaired by Sweden and co-chaired by Italy and progressively involved several Member States and stakeholders in an EU-wide drafting group (47 experts). According to the mandate from the CMEP, the aim of the report was to identify potential effect-based tools (e.g. bioassays, biomarkers and ecological indicators) that could be used in the context of the different monitoring programmes (surveillance, operational and investigative) linking the chemical and ecological status assessment.

Technical report
The technical report on aquatic effect-based monitoring tools (Additional file 1) [6] aims at presenting the state of the art of aquatic effect-based monitoring tools and describing how these tools can help EU Member States to make monitoring programmes more efficient (including reduction of monitoring costs). The report further contains specific sections on the use of such tools in marine systems such as the Regional Seas Conventions and the Marine Strategy Framework Directive (MSFD) 2008/56/EC. The MSFD has foreseen the use of effect-based tools: in particular, the indicators related to Descriptor 8 (contaminants and pollution effects) of the MSFD should include effects from hazardous substances on ecosystem components.
For reasons of clearness, the tools described in the report are categorised into three main groups, primarily depending on the type of monitoring approach used: 1) Bioassays, both in vitro and in vivo, which measure the toxicity of environmental samples under defined laboratory conditions, on cellular or individual levels, respectively. 2) Biomarkers, i.e. biological responses at the cellular or individual levels, measured in field-exposed organisms.
3) Ecological methods, measuring changes observed at higher biological organisation levels, i.e. the population and/or community.
In Europe, several of the tools described in the report are already used for both marine and limnic applications [7,8]. Biomarkers are included in the monitoring programmes of Regional Seas Conventions to detect the presence of substances or combinations of substances not previously identified as a concern and to identify regions of decreased environmental quality. Bioassays are used for example to support risk assessment and management of contaminated sediments and provide decision support for reducing the release of toxic substances into the environment (e.g. in the evaluation of dredged sediments that are considered for sea disposal and whole effluent assessments in the permitting process). They are also used in broad screening of different pollutant sources (such as sewage treatment plant effluents). Other applications include, for example, alarm systems directly triggering control measures (e.g. closing drinking water intakes). Effect-based tools support also the ecotoxicological characterisation and classification of hazardous wastes in the context of the Waste Framework Directive (2008/98/EC). Specific sections of the report are dedicated to EDA/TIE approaches and OMICS.
The report was approved by the CMEP subgroup in

Effect-based tools in the WFD
The WFD mandates three monitoring programmes: Surveillance monitoring aimed to supplement and validate an impact analysis, support efficient and effective design of future monitoring programmes and assess longterm changes in natural conditions and changes resulting from anthropogenic activity. Monitoring is performed at least once every management cycle (usually every 6 years).
Operational monitoring aimed to establish the status of water bodies identified as being at risk of failing to meet the WFD environmental objectives and assess any changes in the status resulting from the programme of measures.
Investigative monitoring aimed to determine reasons for exceedances or predicted failure to achieve environmental objectives if the reasons are not already known and to determine the magnitude and impacts of accidental pollution.
As with all other components of a monitoring programme, it is important to assess the suitability of different effect-based tools against the specified objectives of the monitoring programmes [9]. The suitability of any particular approach must be evaluated in terms of method, cost, practicality and capability to provide information that can be translated into management practices useful for achieving the monitoring programme objectives.
As it has been already evident from previous CIS guidance documents, it is possible to identify several objectives for the use of effect-based tools in a WFD context, some of them are mentioned below: As screening tools, in the framework of the pressures and impacts assessment to aid in the prioritisation of water bodies. To establish early warning systems. To prioritise further studies in areas that are not identified as being at risk because they are located far from known local sources. To take the effects of chemical mixtures or chemicals that are not analysed into account (e.g. to support investigative monitoring where causes of a decline of specific species are unknown). To provide additional support in water and sediment quality assessment, though not as a replacement for conventional chemical and ecological monitoring under the WFD.
Effect-based tools are particularly suitable as part of investigative monitoring programmes, for which the regulatory requirements are determined less formally. However, as with any investigative monitoring, the optimum set of tools varies on a case-by-case basis. The optimal approach will frequently involve several effectbased tools as well as chemical analysis, as illustrated by several of the case studies described in the Appendix section of the report. To optimize cost-effectiveness, it is often wise to make use of the same samples for both chemical and effect-based analyses.

Bioassays in vitro
The use of in vitro assays is increasing for ethical reasons in order to comply with the regulations on animal experimentation. They measure effects at the subcellular level, such as receptor activation and DNA damage, rather than investigate cells or tissues of organisms [10] exposed in the field (as it is the case with biomarkers), the effects are studied in cells after exposure to environmental samples [11]. An advantage of this approach is that in vitro bioassays can often be performed on many different matrixes (such as concentrated extracts of surface water, sediment or pore water samples, biological tissues, passive samplers and effluents). Additional advantages are that only small amounts of sample are generally needed and exposure time is short compared to the time needed for an in vivo assay to detect a response.
In most cases, in vitro assays are considered highly sensitive, because they measure effects at a low organisational level. Many in vitro assays are suitable for screening and high-throughput/automated applications and can be added to the analytical tool package at comparatively lower costs (especially if taking into account the number of substances they respond to). Table 1 presents some in vitro assays that were nominated for monitoring purposes in a Swedish workshop [12]. The table also includes information about the mode of action the assay responds to.
The use of in vitro assays is also common outside Europe; for example, they are largely used in the context of the U.S. A. Programme US TOX 21. For some in vitro bioassays, results are expressed in chemical equivalents, comparing the response induced by the sample to that induced by the reference chemical (positive control). However, before comparing such bioassay results to water quality criteria developed for single chemicals or a number of specific chemicals, it is necessary to consider that the assay can respond to different combined substances that have the same mode of action, e.g. via receptor activation. Generally, for a mixture of agonistic substances with the same mode of action, the biological signal is higher than for a single substance which makes in vitro assays highly suitable as screening tools for environmental samples. They integrate effects of all substances with the same mode of action, e.g. estrogen receptor binding. As such integrative detection tools, in vitro assays are also able to quantify and distinguish agonistic and antagonistic effects. Although many in vitro bioassays can be used on any matrix/extract, some are more suited for the assessment of certain matrices than others, in part because they have been so far only validated for certain uses but also because relevant substances eliciting certain types of responses are primarily found in certain compartments. When using in vitro bioassays, we must be aware that we use highly specific systems to detect chemicals interacting with a specific cellular receptor. These cellular events have been shown to be an important link in the development of specific adverse effects on higher levels of biological organisation. However, it is obvious that simplified in vitro systems do not cover the complexity. Some drawbacks with in vitro tools are that, as opposed to in vivo bioassays and biomarkers, the systems studied are highly simplified when compared to the complexity of whole organisms. Thus, the potential interactions between different receptors, cells and organs are not detected.
In vitro assays are suitable and sometimes necessary when it is needed to conduct follow-up studies using biomarkers. In comparison to most biomarkers, they can easily be used to track local pollution sources by sampling water and sediment in a pollution gradient or effluents from suspected point sources. In vitro assays are also valuable in effect-directed analysis (EDA)/toxicity identification evaluation (TIE) approaches to identify toxic fractions and provide guidance for the identification of causative agents.

Bioassays in vivo
In vivo bioassays are tests where whole living organisms (including bacteria) are exposed to environmental samples, such as surface water, sediment, wastewater, dredge material or extracts from such samples. Tests are performed in the laboratory or, less frequently, in the field ("in situ" bioassays) [13].
The "endpoint" is the type of effect that is measured in a toxicity test, and some examples that are frequently used in this context are mortality, immobilization, fertilisation rate, hatching rate, embryo development, effects on growth of individuals (e.g. weight), effects on growth of populations (e.g. number of individuals), metabolic or physiological changes, reduced swimming activity, bioluminescence and specific molecular/biochemical responses. Table 2 reports the most common in vivo bioassays applied by certain Member States within aquatic monitoring programmes.
In general, in vivo bioassays are broad spectrum assays, i.e. an in vivo bioassay responds to a variety of substances and different types of toxicity. An example is the "fish embryo acute toxicity (FET) test," recently adopted by the OECD N.236 of 26 July 2013, that is based on individual exposure of eggs to evaluate the embryotoxicity of samples with the aim to detect contaminants (relevant for the WFD) such as industrial chemicals, pesticides, pharmaceuticals and biocides.
Nevertheless, it is important to underline that the evaluation of toxic effects is based on the responses observed in several species, because they can exhibit intrinsic differences in terms of sensitivity to various chemicals; it also depends on the endpoints measured in the test [14]. Both short-and long-term in vivo bioassays should preferably be carried out using at least three species from different taxonomic groups and trophic levels (primary producer, decomposer/saprophytic, detritivore/filter feeder and consumer). A battery of ecotoxicological tests should have sufficient sensitivity and discriminatory power and respond to as many contaminant groups as possible.

Biomarkers
Biomarkers are molecular, biochemical, cellular and physiological indicators of contaminant stress measured in organisms resident or exposed in situ in a specific location. They are used in the monitoring programmes of Regional Seas Conventions and more recently in the MSFD to identify the impact from substances or combinations of substances not previously identified to be of concern, study trends and identify regions of decreased environmental quality [15]. Contrary to bioassays but similar to the ecological/community-based tools, biomarkers are analysed on field-exposed, usually resident, organisms. The sampling step is therefore primarily focused on the organisms that should to be examined. However, active monitoring based on caged organisms can also be used to measure biomarkers.
It is important to detect deleterious effects due to chemicals before significant effects at the population level occur. Damage at the population and ecosystem level can take a long time to repair. For certain trophic levels, recolonisation may take much longer than the 6-year management cycles considered in the WFD and MSFD. Ecological tools or indices are not predictive of damage, whereas several biomarkers can be used as early warning systems because they can detect effects caused by chemical substances and other environmental stress at an early stage.
Biomarkers are frequently divided into two different categories, depending on the number of substances/ groups of substances they are known to respond to: General (integrative) biomarkers that respond to several classes of toxic substances and, frequently, also to other types of stressors. Specific biomarkers that respond primarily to only a few/groups of/substances. For example, imposex is considered to be a highly specific effect, responding primarily to organic tin compounds such as tributyltin (TBT), whereas lysosomal stability is a more general biomarker for cellular stress. Both general and specific biomarkers can be useful, depending on the monitoring goals and the level of prior knowledge regarding the type of contaminants present at the given location. Because general biomarkers respond to several classes of compounds, they cover more substances and are therefore valuable in identifying areas of concern in environments exposed to complex mixtures of stressors. Specific biomarkers are, for example, valuable in second tier assessments to detect and identify the effects of specific types of substances in impacted locations. Table 3 presents biomarkers used in the integrated monitoring approach proposed by International Council for the Exploration of the Sea (ICES) and the Regional Seas Conventions.
Biomarkers of exposure allow statements about the quality and/or quantity of exposure, whereas with biomarkers of effect statements about effects and the health status of exposed organisms can be made.
Exposure biomarkers, such as ethoxyresorufin-Odeethylase (EROD), can provide a sensitive indication of cellular changes at the enzyme level, which often represent the first warning signals of environmental disturbance. EROD can be used to detect exposure to classes of organic pollutants such as co-planar polychlorinated biphenyls (PCB), polycyclic aromatic hydrocarbons (PAH), planar dibenzodioxins (PCDD) and dibenzofurans (PCDF). Metallothioneins, peroxisomal enzymes (e.g. acyl CoA oxidase) and inhibition of acetylcholinesterase activity are other more or less specific exposure responses towards trace metals, organic chemicals and organophosphate pesticides, respectively. On the other hand, biomarkers of effect indicate the occurrence of various forms of molecular to cellular/tissue alterations, although the health-related effects may differ in terms of toxicological and ecological relevance. Some effect biomarkers detect effects at early stages (such as genetic changes), whereas others, such as imposex, are related to later stages from a population risk perspective.

Ecological methods
The assessment criteria (primarily based on values of biodiversity indices) selected as the biological quality elements of the WFD do not respond in a specific way to the effects of hazardous substances. Specific tools for the assessment of hazardous substances applied in the context of WFD monitoring are extremely rare. Therefore, ecological status related to hazardous substances is generally based on individual pollutant concentrations, which may or may not be consistent with the ecological quality status assessment.
A new approach within biomonitoring is to consider the ecological role of communities, based on their functional, rather than structural composition, through the identification of species traits [16]. The resistance and resilience characters of individual taxa determine the response of communities to disturbance. Undisturbed communities display a diversity of species traits, whereas the communities downstream of a pollution source consist of those species that have a suite of traits which convey tolerance to the new conditions. Species that do not have these traits cannot survive [17]. The advantage of using functional traits instead of taxonomic composition (an entirely structural approach) of communities is bound to the a priori predictable response of traits to individual stressors, because each selection pressure affects different traits.

Spear
The SPEcies At Risk (SPEAR) bioindicator index, based on biological traits, has been shown to be highly sensitive to particular groups of contaminants such as pesticides and relatively independent of confounding factors [18]. The index measures the proportion of sensitive (SPEAR) and less sensitive (SPEnotAR, "SPEcies not At Risk") species and is expressed as a percentage.
The SPEAR concept is applicable for the assessment of the effects of pesticides on invertebrate communities in rivers but not lakes, coastal areas or temporary streams. Sampling must be performed in early summer (around June and July), not too long after peak pesticide application. Sensitivity data and information on other relevant traits for the taxa are included within the database used for the SPEAR online calculator. So far, validation studies have been performed in Finland, Germany, Sweden, France, Spain, Czech Republic and Australia. Nonetheless, there is a need for further validation before the SPEAR index can be used on a routine basis and as part of the WFD classification. In particular, the baseline sensitivity and variability of the method need to be assessed.

Pict
Pollution-induced community tolerance (PICT) has been suggested as a sensitive tool to track changes in community function (and therefore indicative of structural changes) that can be attributed to toxic substances. The PICT approach was developed by Blanck and Wängberg [19]. The approach relies on the assumption that sensitive components of the exposed community (species, genotypes or phenotypes) will be replaced by more tolerant ones during exposure, thus leading to an increase of community tolerance. PICT is measured by a functional test that detects the consequences of selection pressures. Tolerance development, for example, can be measured as a shift in the effect concentration (usually EC 50 ) that is obtained with a short-term toxicity test based on an ecophysiological endpoint. Such an endpoint is preferably related to community metabolism (photosynthesis, respiration, protein synthesis, nucleic acid synthesis etc.). In recent years, PICT combined with the transplantation of periphyton communities has been suggested as a promising tool to identify impaired sites by detecting an induced tolerance after transplantation. In situ PICT assays using transplanted communities have been suggested as a promising tool that can link ecological and chemical status in the WFD context [20]. The PICT approach has the disadvantage that cannot be used to assess the risks for long-lived organisms with complex life cycles (e.g. insects and vertebrates).

EDA and TIE
While effect-based monitoring indicates hazards due to chemical contamination and provides information on toxicological endpoints of concern, tools are required to identify causes and elucidate links between exposure and effects. EDA and TIE are integrated biological and chemical approaches which aim to identify those compounds in an environmental or technical sample (water, soil, sediment, air, food, consumer product and technical mixture) that cause a biological response. Both approaches combine biotesting, physicochemical fractionation and chemical analysis in a sequential procedure. However, the philosophy behind both approaches is slightly different [21]. The TIE approach has its origin in whole effluent testing, which focuses on the question, whether an effluent will cause adverse effects on aquatic organisms when emitted to the environment. In the case that effects are detected in whole organisms under realistic exposure conditions, TIE should help to characterise and identify the cause of the measured effect. Thus, TIE applies in vivo biotesting and avoids extraction and preconcentration steps as far as possible.
EDA is based on the understanding that environmental samples may contain thousands of mostly organic chemicals and that only a fraction of them can be analysed by chemical target analysis. EDA takes a biological effect (typically observed by effect-based monitoring) as the basis to narrow down the huge amount of possible chemical substances and aims to direct chemical analysis to those compounds that contribute significantly to a measurable effect. Thus, in EDA, bioassays are considered as tools to sensitively detect chemicals with similar biological targets or modes of action. The focus of EDA is on unravelling the contamination with organic toxicants representing the most complex group of chemicals. Similar to chemical analysis, there are no restrictions with respect to extraction or pre-concentration. Since the isolation and identification of individual toxicants out of thousands of components in typical environmental mixtures often demand for large numbers of fractions, high-throughput tools are preferred. In addition, the identification of unknown toxicants is very much supported by information on the mode of action. Both criteria are often met best with in vitro assays, although small-scale in vivo assays may be helpful, too [22]. The sample or an extract thereof is tested with the bioassays of choice depending on the objective of the study. If effects are detectable, the mixture is fractionated according to the physico-chemical properties of the components. The fractions are tested with the same bioassays for prioritisation according to effects. The mixture may undergo several fractionation steps to further reduce complexity. The components of active fractions are identified and quantified by chemical analytical means. Depending on the objective of the study, in a final confirmation step, the contribution of the identified candidate compound to the measured effect should be quantified or estimated in order to exclude that major contributors have been overlooked.
The major components of EDA are (i) separation including extraction, clean up and fractionation, (ii) biotesting, (iii) chemical analysis including computational tools for structure elucidation and (iv) confirmation.
EDA is a tool for investigative monitoring at selected sites of particular interest or with conspicuous effects. EDA aids in linking ecological status to contamination, to establish cause-effect relationships and to target mitigation measures. Although providing enormous progress over present target chemical monitoring, a general limitation of EDA is the requirement to pre-select toxicological endpoints. The combination of integrating whole organism tests with in vitro test batteries applying sufficient pre-concentration reduces the risk to overlook important effects and thus toxicants.

OMICS
The recent advances in DNA sequencing and characterisation of genomes have opened up a range of new possibilities. A particular field of molecular studies within biology is called "Omics" and refers to high-throughput molecular profiling technologies, such as genomics, metagenomics, transcriptomics, proteomics, metabolomics and metabonomics. The suffix "-ome" refers to the collection of all genes or gene products such as the genome, proteome or metabolome, respectively. A study of all or a very large number of these genes would fall under the definition of omics. Omics and bioinformatics tools can e.g. be used to: Develop molecular biomarkers of exposure as early signals to predict effects (that at a later stage could have an impact on physiological level and further on at population level).
Provide information about the mode of action (MOA) of chemicals, i.e. the mechanism of toxicity; in turn, reducing the uncertainties involved in chemical risk assessment by providing, for example, a basis for extrapolation of the effects across species. Integrate MOA data with a deleterious outcome and in this way, aid towards understanding the impact on the ecosystem instead of just on single organism or species. Distinguish the site of origin of organisms, based on the transcriptomics changes in organisms coming from different locations.

Genomics-DNA microarray applications
Genomics can be indicative of the susceptibility of an organism for a certain chemical or group of chemicals [23] and are more frequently also been used for assessment of complex environmental samples, such as sediments [24]. A DNA microarray is a glass slide or a nylon membrane on which part of the organism's gene sequences (probes) is spotted or synthesised. Normally, complementary DNA (cDNA) is made using reverse transcriptase from RNA in the sample to be analysed. Then, the cDNA is thereafter hybridised to the array. After scanning and image analysis, the RNA abundance (amount of RNA molecules bound to the complementary probes on the microarray) is analysed and the relative gene expression of the treated sample can be compared to the untreated control.

Next-generation sequencing
The development of DNA sequencing technology four decades ago was a major scientific hallmark and opened the doors for several breakthrough achievements in all areas of biology. Next-generation sequencing (NGS) is a more recent technology, also named second-generation sequencing (SGS) and has been commercially available since 2004.
Compared to the first-generation capillary electrophoresis (CE)-based Sanger sequencing, NGS has increased sequencing speed, the throughput to millions of sequencing experiments on fragmented DNA run in parallel, has reduced sequencing costs per base pair in some cases more than 10,000 fold. NGS platforms enable a wide variety of applications including the study of the genome or transcriptome of any organism.

RNA-seq
RNA-seq is a recently developed approach, extending the high-throughput sequencing to the profiling of the transcriptome. Instead of capturing transcript molecules by molecular hybridisation, as on microarrays, RNA-seq directly sequences the transcripts present in a sample. Transcript sequences are then mapped back to a reference genome and counted to assess the expression level of that gene or genomic region [25].

Metagenomics
In recent years, a tremendous increase in DNA sequencing capacity, combined with an unprecedented drop in price per obtained nucleotide sequence, has made it possible to study the functional elements of a microbial ecosystem at the levels of the actual genes responsible for these functions. Such sequencing studies of the total DNA content of an environment are generally referred to as metagenomics [26] and could be used to pinpoint the genes or species that cause, for example, a PICT response. A possible relevant use of metagenomics is to monitor the presence and abundance of antibiotic resistance genes in the environment.

Proteomics
Proteomics is the large-scale study of proteins, particularly their structures and functions [27]. After genomics and transcriptomics, high-throughput proteomics is considered the next step in the study of biological systems. It is much more complicated than genomics mostly because while an organism's genome is more or less constant, the proteome differs from cell to cell and from time to time. This is because distinct genes are expressed in distinct cell types. This means that even the basic set of proteins which are produced in a cell needs to be determined.

Metabolomics
Within metabolomics, the endogenic metabolic profile of an organism is studied. The metabolites that are studied can be considered to be the result of the ongoing metabolic activity of the cells. To measure metabolites is considered advantageous since it is well known that metabolites are formed at an early stage of environmental stress [28].

Conclusions
The topic "effect-based tools" highly ranked in the CIS science-policy interface report [29] elaborated on the basis of inputs from the Working Group "Chemicals" of the WFD". The new mandate 2013 to 2015 of the Working Group, approved by the Water Directors, has foreseen the continuation of the activity on effect-based tools, in particular, in relation to the detection and evaluation of effects caused by mixtures of pollutants. This activity will be strongly linked to the work of the WG Ecostat and the implementation needs of the MSFD. Furthermore, a project has been planned with the aim to evaluate the use of bioanalytical methods for the detection of the pharmaceuticals 17α-ethinylestradiol (EE2) and 17βestradiol (E2) which are included in the WFD "Watch List" of emerging pollutants foreseen by the Directive 2013/39/EU. This technical report, elaborated in close collaboration with the scientific community, can already be considered to provide important support to the managers, assessors and local operators involved in the analysis and monitoring of surface water bodies.