Determination of toxicity by compound and species characteristics at different scales (individual and population)

Anna-Maija Nyman, PhD project (2/2010 – 5/2013), EAWAG, Zürich, Switzerland

Current occupation of the fellow: Helmholtz Centre for Environmental Research – UFZ, Germany / University of Eastern Finland, Finland
Contact: anna-maija.nyman[at] or anna-maija.nyman[at]


The aim of the study was to link mechanistically pesticide exposure to lethal effects observed in three aquatic invertebrate species. The effects were investigated at individual level, including toxicokinetic processes (uptake, elimination, chemical distribution, biotransformation) and toxicodynamic processes (e.g. damage recovery). Laboratory experiments were conducted and the data were used to calibrate toxicokinetic-toxicodynamic (TKTD) models, which can be:

a) used alone to predict the individual level effects in different exposure regimes instead of evaluating the toxicity by fixed LC50 concentrations

b) combined with individual based models (IBMs) to simulate toxicity within the population model in a more mechanistic manner.


Lymnaea stagnalis is an air-breathing freshwater snail with simple nervous system. Photo: Anna-Maija Nyman


Species in the study:

Gammarus pulex

– Gammarus fossarum

– Lymnaea stagnalis


Gammarus pulex. Photo: Anna-Maija Nyman


TKTD models and fluctuating pesticide concentrations

Currently in chemical risk assessment only the initial peak concentration from fate models is used to compare the exposure with effect levels (i.e. EC or LC50 values). TKTD models enable predicting pesticide effects in varying exposure scenario, e.g. in fluctuating exposure profiles. The models take into account chemical concentration in organisms and internal damage caused by the chemical. TKTD models can be calibrated using available acute toxicity data which are also used to derive the LC50 values. Thus tools and data for more accurate chemical risk assessment do already exist – however, they are not used.

We investigated calibration data requirements for TKTD models in a case study of Gammarus pulex exposed to the fungicide propiconazole (Nyman et al. 2012). We showed that toxicokinetic experiments and modeling toxicokinetics explicitly does not bring more accuracy in predicting survival. In addition, acute toxicity data might be sufficient for predicting survival in multiple pulse exposures.


Do individual level processes (i.e. toxicokinetics and toxicodynamics) matter if we are looking at the populations?

In the EU legislation, invertebrate populations are protected, not individuals. Does it matter then which kind of model we are using to predict effects at the individual level? Are dose-response models sufficient enough or do we need to use TKTD models which take into account detailed chemical uptake and damage caused by the chemical?

We combined a dose-response model and a TKTD model with an individual based population model (IBM) for Gammarus pulex and compared the population recovery times when the different models were used to simulate the toxicity (Galic et al. 2013). The results showed that the dose-response model mostly underestimated the effects at the population level, especially in case of compounds with delayed and irreversible effects, i.e. diazinon and chlorpyrifos. When 4-d LC50 concentration was used for these compounds, populations went to extinct within 20 days using the TKTD model while dose-response model killed 50% of the individuals and the population recovery was reached within 380 days. The difference between these two models is that the dose-response model does not consider the effects after the exposure is terminated, i.e. the surviving individuals are again in the state as before the exposure. This means that compounds which have irreversible effects must be modeled with TKTD models which take into account internal concentrations and damage which can cause mortality also after the exposure period.


Modelling the effects of starvation using a TKTD model

Sublethal effects can play a major role in determining the population level effects of pesticides. For instance some neurotoxic chemicals, such as neonicotinoid imidacloprid, have the potential to indirectly cause lethality in aquatic invertebrate populations at low, sublethal concentrations by impairing movements and thus feeding (Nyman et al. 2013). This was shown in our study where we investigated feeding activity, lipid content, immobility, and survival of the aquatic arthropod Gammarus pulex under exposure to imidacloprid. In addition to experiments where we exposed the animals to imidacloprid, we also performed a starvation experiment without exposure to the chemical. We used these experiments to calibrate a multiple stressor toxicokinetic-toxicodynamic modeling approach: direct effects on survival were calibrated with pulsed treatment data and the effects of starvation with starvation experiment with no chemical exposure. With the model we showed that both starvation and other toxic effects of imidacloprid play a role for determining mortality in constant exposure to the insecticide imidacloprid.


TKTD models to investigate interspecies variation  in sensitivity to pesticides

In ecological risk assessment of chemicals, one major difficulty is interspecies variation in the organism responses to chemical exposure. In the first tier of risk assessment, the toxicity data of standard test species, such as Daphnia magna, are used together with assessment factors to generate a predicted no effect concentration for the whole ecological community. The assessment factors are used to cover uncertainty when extrapolating from e.g. a species to other species. However, the use of these factors is not based on data produced scientifically but they are rather used as rules of thumb.

Using TKTD models and comparing differences in model parameters, we might achieve a better understanding of the interspecies and interchemical variation in terms of toxic effects. It might enable to pinpoint the processes that might cause differences in responses to toxicants. The differences in sensitivity can be caused by toxicokinetics, i.e. uptake, elimination, distribution, biotransformation, or by toxicodynamics, e.g. absence of sites of toxic action. We studied the causes of interspecies variation by conducting a detailed study on the toxicokinetics (bioaccumulation experiments, chemical distribution analysis) and the toxicity (endpoint survival) of the pesticides diazinon, propiconazole and imidacloprid in Gammarus pulex, Gammarus fossarum and Lymnaea stagnalis (Nyman et al., manuscripts in progress).

Investigation of interspecies variation in sensitivity to pesticides.





















Galic N, Ashauer A, BavecoH, Nyman A-M, Barsi A, Thorbek P, Bruns E, van den Brink P (2013): The relevance of toxicokinetic and toxicodynamic processes for the population recovery of Gammarus pulex after exposure to pesticides – Environmental Toxicology and Chemistry (in press)

Nyman A-M, Hintermeister A, Schirmer K, Ashauer R (2013) The insecticide imidacloprid causes mortality of the freshwater amphipod Gammarus pulex by interfering with feeding behavior. PLoS ONE 8(5):e62472

Nyman A-M, Schirmer K, Ashauer R (2012) Toxicokinetic-toxicodynamic modelling of survival of Gammarus pulex in multiple pulse exposures to propiconazole: model assumptions, calibration data requirements and predictive power. Ecotoxicology 21: 1828-1840.

Nyman A-M, Schirmer K, Ashauer R: Towards understanding the differences in sensitivity to pesticides among invertebrate species – (manuscript in progress)


Supervisor: Roman Ashauer (Eawag)
Co-Supervisors: Paul van den Brink (Wageningen University), Kristin Schirmer (Eawag)
Associated partner: Bayer