Growth Stimulation Effects of Environmentally Realistic Contaminant Mixtures on a Marine Diatom

Abstract

To estimate mixture effects caused by the high number of chemicals simultaneously present in the environment, methods for routine effect assessment of environmentally realistic contaminant mixtures are needed. We repeatedly exposed the marine diatom Phaeodactylum tricornutum to SpeediskTM passive sampler extracts and observed statistically significant growth stimulation up to 6 and 7% for samples from inside and outside the harbor of Zeebrugge, respectively. These effects were found at summed contaminant concentrations (159–166 ng L–1) that were within a 1.1‐ to 2.4‐fold range of those observed in grab water samples taken during sampler deployment. These stimulatory effects were confirmed in 2 independent tests with extracts stored for <1 or 8 mo that had undergone limited sample handling, whereas no effects were observed for extracts that had been stored for 16 mo that had undergone repeated handling (notably repeated freezing and thawing) before biotest spiking. Targeted analysis by ultra‐high performance liquid chromatography was performed to quantify 88 personal care products (n = 8), pesticides (n = 28), and pharmaceuticals (n = 52). Among these compounds, multivariate statistical analysis put forward the β‐blocker atenolol as explaining most of the observed variation in mixture composition between the growth‐stimulating and no effect–causing extracts. However, when tested individually over the entire concentration range present in the extracts, atenolol did not have any effect on P. tricornutum, suggesting that nontargeted substances in the extracts may have contributed to the observed stimulatory effects. Nevertheless, the present study shows that exposure to contaminant mixtures at environmentally realistic concentrations can lead to small but significant growth stimulation effects on the marine diatom P. tricornutum. Environ Toxicol Chem 2019;38:1313–1322. © 2019 SETAC

A margin of safety approach for the assessment of environmentally realistic chemical mixtures in the marine environment based on combined passive sampling and ecotoxicity testing

“…Three additional Speedisks deployed at SC3 HO were extracted and extracts split into two equal parts of 4.9 mL each, while all other extracts were kept undivided before full evaporation of the extraction solvent under a gentle nitrogen stream at 25 °C. The separation of SC3 HO extracts occurred in order to compare a previously applied extraction procedure where extracts were reconstituted in a water-solvent mixture (Moeris et al 2019) with the newly developed one (reconstitution in pure HPLC water). Finally, the…”

“…Based on previous experience (Moeris et al 2019) Speedisk extracts were analyzed chemically and tested for toxicity within maximum four weeks of extraction. Targeted chemical analysis was performed using ultra-high performance liquid chromatography (UHPLC) and focused on the quantification of 89 personal care products (n = 9), pesticides (n = 28) and pharmaceuticals…”

“…Previous research has shown that it is better to predict biotest concentrations based on those chemical concentrations determined in Speedisk extracts in order to maintain a maximum of available information since concentrations in microplate wells conflict with analytical detection limits (Moeris et al 2019).…”

“…On the other hand, adsorption based samplers such as the Speedisks provide excellent enrichment of specific analytes (Jahnke et al 2016). Further, chemical analysis and dosing of toxicity tests can easily be performed with the same set of samples (Moeris et al 2019).…”

“…Additionally, reconstitution of passive sampler extracts is usually performed using the addition of solvents and, therefore, spiking extracts into biotest medium unavoidably also introduces these solvents into the biotest medium. This can be problematic (depending on the concentration of the solvent introduced) and may limit the maximum level of sample enrichment that can be investigated in the biotesting of passive sampler extracts (Moeris et al 2019).…”