An adaptable mesocosm platform for performing integrated assessments of nanomaterial risk in complex environmental systems
Physical chemists, (micro)biologists, and ecologists need to conduct meaningful experiments to study the environmental risk of engineered nanomaterials with access to relevant mechanistic data across several spatial and temporal scales. Indoor aquatic mesocosms (60L) that can be tailored to virtually mimic any ecosystem appear as a particularly well suited device. Here, this concept is illustrated by a pilot study aimed at assessing the distribution of a CeO2 based nanomaterial within our system at low concentration (1.5mg/L). Physico chemical as well as microbiological parameters took two weeks to equilibrate. These parameters were found to be reproducible across the 9 mesocosm setup over a 45 day period of time. Recovery mass balances of 115 18% and 60 30% of the Ce were obtained for the pulse dosing and the chronic dosing, respectively. This demonstrated the relevance of our experimental approach that allows for adequately monitoring the fate and impact of a given nanomaterial.
Engineered nanomaterials (ENMs) have become a fast growing economic sector. As a consequence of the many debates concerning their safety, efforts are developed at international and national levels to develop a code of ethics for a safe and responsible development of ENMs1. A sustained growth of the nanotechnology industry will rely heavily on the characterization of risks to the environment (water and soil resources, trophic transfers, biodiversity) and human health that may be posed by ENMs. To date, investigation of the roles of nano scale objects in the context of evolutionary change, environmental disturbance, ecosystem structure and function are limited2, 3. Moreover, current strategies to assess the environmental safety of ENMs are based on classical ecotoxicology approaches4, which are not always adequate for ENMs. Indeed, while the hazard is extensively investigated, little attention is paid to the exposure to ENMs despite its pivotal role in understanding their environmental risks. Extent and mode of exposure to ENMs is controlled by a number of parameters including aggregation state and sorption of (in)organic substances, redox as well as ecological factors such ecological feeding type and trophic transfer potential5. There is an abundant literature about the effects of all these parameters taken separately. However, for a robust characterization of the exposure, the complex interplay between these parameters in real ecosystems needs to be considered.
Mesocosms are experimental systems designed to simulate ecosystems6 and are an invaluable tool for addressing the complex issue of exposure during nano ecotoxicological testing. This experimental strategy has already been used to study the behavior or impacts of ENMs7, 8, 9, 10. A broad diversity of mesocosms design exists in term of dimension, location (indoor, outdoor) and ecosystem simulation type (estuarine, aquatic freshwater, and terrestrial)11, 12, 13, 14. A common factor of all these studies is that mesocosms are considered as a small portion of the natural environment that is brought under controlled conditions. In our study we define a mesocosm as an experimental design which is (i) self sustaining once set up and acclimation without any additional input of nutrients or resources, and (ii) that allows controlling all (or the maximum of) input and output parameters to draw a real life mass balance whatever its dimension or location. Mesocosms has already been applied to trace the transfer of gold ENMs from the water column in an estuarine food web7. Clams and biofilms were observed to accumulate most of gold on a per mass basis. The long term (18 month) distribution and transformation of silver ENMs was also studied in outdoor mesocosms mimicking freshwater emergent wetlands10. Silver sulfidation was demonstrated in the terrestrial soils and subaquatic sediments and a high body burden of Ag was measured in mosquito fish and chironomids10. Another study used indoor estuarine mesocosms to monitor the leaching of Ag from consumer products incorporating ENMs over 60 days. The investigations described in the literature involve rather large facilities (tank size 120L and above), to reproduce the assimilative capabilities of a larger natural system. However, the need for multiple replicas in biological studies limits the practicality of large mesocosms due to obvious limitations in space and cost. ENMs runoff rain or vent loading, or a continuous point source discharge as wastewater treatment plant or industrial discharge).
In this study, we conceived a laboratory scale mesocosm facility to serve as a platform for investigating ENM exposure and impacts. We opted for modular, small size (60L), indoor mesocosms. aggregation, settling, mass balance, trophic transfer, biotransformation, oxidative stress, microbial diversity) under environmentally meaningful conditions. This experimental design can accommodate several types of ecosystems such as lotic, lentic, estuarine, or lagoon environments, without requiring expensive and/or cumbersome infrastructures. This versatile tool can then be used by a large community of physical chemists, (micro)biologists, and ecologists to study the exposure and impacts of ENMs (low doses, chronic contamination) as well as the mechanistic concepts at various temporal and spatial scales. Here we show that, with the adequate methodology, using small sized mesocosms is an approach as robust as using large(r) systems.
DesignAn experiment performed in this platform involves 2 phases: the acclimation and equilibration of the mesocosms, and the exposure phase of the experiment, per se. The initial setup consists in introducing the sediment and filling the tank (750 200 600mm) with water. The water and the sediment are chosen to be close to the solution chemistry of the natural ecosystem of interest. Macro and micro organisms are collected from the same natural environment.
During phase I, the particles suspended by the addition of water are given time to settle, the pH, conductivity, O2, and redox potential stabilize, and the primary producers develop. pH, T, turbidity, ammonia). primary, secondary consumers) and the water pumps are turned on. The selected organisms are involved in a real food web and have different habitats and ecological functions in the ecosystems. The density of organisms is adjusted as a function of the natural environment. The duration of the acclimation depends on biological features of the species as growth rate, metabolism activity, life cycle duration.
Phase II corresponds to the ENMs exposure period and ecotoxicity test. It can be either a single pulse or multiple dosing experiments. The treatments are distributed between each mesocosm as a function of the picoplankton and algal abundances. To avoid any variation in term of primary production, the triplicates are selected to have average concentrations of picoplankton and algae as close as possible between the different treatments.
Sampling and analysesSeveral outlet scarpe gucci physico chemical, microbial, and biological analyses can be performed to assess both the exposure and impacts of ENMs on a designated trophic link. pH, temperature, Eh). metal concentration, number of colloids, picoplankton and algae concentrations) require sampling. Using gucci borse outlet a small mesocosm gucci borse outlet setup, water, superficial sediments, cores, picoplankton, algae, and macro invertebrates can be sampled with any desired periodicity. During sampling, special attention must be given (i) to avoid disturbing the sediment and water column properties, and (ii) to keep micro organism densities and ENMs concentrations constant.
The distribution of the ENMs or their degradation by products is assessed by measuring their concentration in the water, sediment, biota, etc. ICP MS or ICP AES). When necessary, the dissolution of ENMs can be measured separately by placing sealed dialysis bags in the mesocosms.
A thorough characterization of the speciation, (bio)transformation, bio(distribution) of the ENMs in the water, sediment, or biota can be performed using X ray, IR, Raman spectroscopies, Nuclear magnetic resonance, as well as electron or X ray based microscopy and tomography. Such an experimental design also allows monitoring the mechanisms of toxicity at the sub individual scale on the micro , macro organisms as well as on microbial communities. For instance, oxidative stress18 can be assessed using ecotoxicological markers and ecophysiological processes19, 20, 21, 22.
DesignA pilot study was conducted to assess the evolution of the distribution of a CeO2 based ENM that is included on the OECD list for ENMs requiring (eco)toxicological testing23. Citrate coated CeO2 nanoparticles (8nm of hydrodynamic diameter) sold (Nanobyk 3810, Byk24) as long term UV stabilizers were used in this work25.
The study proceeded through phase I and II with respective exposure periods of 17, and 28 days (Fig. 1a). Two contamination scenarios were simulated. The response of the mesocosms to a single mass addition (pulse dose) of 69mg to achieve an initial concentration of 1.5mg/L of CeO2 ENMs at day 0 was compared to that of resulting from 11 smaller doses (chronic doses) of 5.2mg administered 3 times per week during 4 weeks to achieve a final concentration at day 28 of 1.4mg/L of CeO2 ENMs (Fig. 1a).
The mesocosm platform was configured to simulate a pond ecosystem using an invertebrate species (Planorbarius corneus L., 1758, commonly named ramshorn snail) and a natural inoculum coming from a non contaminated pond (43.34361 N, 6.259663 E, and 107 m above sea level). This pond is part of the protected Natura 2000 Reserve Network. single pulse versus multiple dosing) condition and the control. Each mesocosm is made of monolithic glass panels of 12 mm thick. Five holes ( 15mm) drilled at mid height of the large panels are connected to a pump using silicon tubes (Fig. 1b). The mesocosms were filled with 5 8cm of artificial sediment made of 84 5% (dry weight) of quartz (grain size: 60% from 0.05 0.2mm, and 40% from 0.2 2mm), 15 5% of kaolinite, and 1% of CaCO3 (adapted from26). Three hundred g (water saturated weight) of a natural inoculum collected in the pond was sieved at 0.2mm and laid at the surface of the artificial sediment (1 mm thick). This natural inoculum contained CaCO3, SiO2, and clay minerals (see supporting information) and introduced the primary producers into the mesocosms. Each mesocosm contained 55L of the commercialized natural Volvic, and eleven adults P. corneus (3 1cm of diameter). Volvic is a French commercialized natural water with the following composition: pH 7, 11.5mg/L Ca2+, 13.5mg/L Cl, 71mg/L HCO3, 8mg/L Mg2+, 6.3mg/L NO3, 6.2mg/L K+, 11.6mg/L Na+. The mesocosms were operated 4 days to acclimate the invertebrates to the experimental conditions before the introduction of the CeO2 ENMs suspensions. This few days of acclimation period were based on previous studies working with micro and macro invertebrates in mesocosms27, 28, 29.
Temperature, pH, conductivity, redox potential, and dissolved O2 were measured every 5min at mid height of the water column using multi parameter probes (Odeon Open X) and at the water/sediment interface (up to 10mm below surficial sediment) and mid height of the sediment using platinum tipped redox probes10, 30. A day/night cycle of 10h/14h was applied using full spectrum light (Viva light T8 tubes 18W), and room temperature was kept constant.
All data were analyzed for normality and homoscedasticity using the Kolmogorov Smirnov test and Levene test31. In low sample size comparisons, differences between groups were analyzed using a Mann Whtiney U test31.
Phase I: acclimation and equilibrationDuring this phase of acclimation and equilibration, values for each key parameter (pH, conductivity, and dissolved O2 concentration) and their associated variance between the 9 mesocosms were recorded and calculated. The condition that was set for ending phase I (equilibration) was the reduction by 50% (compared to the beginning of phase I) of the standard deviation corresponding to the average pH, conductivity, and dissolved O2 concentration between the 9 mesocosms. At the end of phase I, pH of 0.5, T of 0.4C, O2 of 2.5mg/L, and C of 20S/cm were reached (Fig. 2). The conductivity (C) monotonically increased from 250 to 330S/cm due to refilling with mineral water to compensate for evaporation. The pH stabilized around 7.9 0.1 (at 20 25C) and the dissolved O2 concentration reach 8 0.2mg/L (90%), which is close to the natural pond water. Both pH and dissolved O2 concentration underwent diurnal variations of 0.35 pH units, and 0.7mg/L respectively over the day night cycle. The redox potential in the water column was stable and positive (between +220 and +250mV) during phase I, whereas in the sediment the redox potential exhibited a positive negative inversion (down to 330mV) two days after the filling with water. This indicates that anoxic conditions prevailed in the sediment (Fig. 2).
A significant decrease (p 0.05) of the picoplankton concentration (from 105 to 104cells/mL) and algae concentration below 105algae/mL were observed in the water column at the end of phase I (see supporting information). This decrease was concomitant with the settling of suspended particles and a corresponding decrease in concentration from 9 105 to 3 105particle/mL (see supporting information). In contrast, at the surface of the sediment, picoplankton and algae concentrations slightly increased during phase I from 106 to 107cells/mL and from 105 to 106algae/mL respectively. This is attributed to the development of the picoplankton and algae coming from the natural inoculum at the sediment surface. Pictures of microorganisms that developed in the water column as single cells or small flocs, and as biofilms in the sediment are given in supporting information. Before the introduction of organisms, bacteria formed biofilms of 100m depth aggregating on clays, quartz particles, and algae, which are primer producers for non phototrophic bacteria.
At the end of phase I, the similarity of microbial community compositions was assessed by pyrosequencing of the 16S rRNA gene (see supporting information). Microbial diversity in the mesocosms was considered in terms of richness and phylogenetic distance. The number of OTUs (Operational Taxonomic Units)33 and the Chao134 estimator best characterize the microbial richness, while the Bray Curtis dissimilarity describes the phylogenetic distance between two samples. Table 1 summarizes diversity indices that describe water and benthic compartments of the mesocosms at the end of phase II. Based on Kruskall Wallis test, there was no significant difference (p 0.05) among the medians of observed OTUs and Chao1 metrics between A, B, and C in the water column or in the sediment. Thus the richness values for the limnetic or the sediment compartments of the three sets of mesocosms were not statistically different. Microbial phylogenetic diversity was examined at the phylum level. Fig. 3 shows the percentage distribution of OTUs in phyla, which accounted for more than 1%. The Bray Curtis dissimilarity indices were low ( showing that the diversity of the mesocosms was similar. Proteobacteria, dominated in the water column (86%), with Proteobacteria as main class (76%), followed by the Actinobacteria (3.8%). Both of them often prevail in freshwater microbial communities35, 36. As expected, the diversity in sediment was higher than in the water column37. Sequences assigned to Proteobacteria were more abundant in water column ( than in sediments (46%) while sequences assigned to Actinobacteria were more abundant in sediment (14%) than in water column (3.8%). Low abundance phyla included Acidobacteria (5.9%) and Cyanobacteria (5.4%), Chloroflexi (3.7%), Firmicutes (3%), Bactereoidetes (2.6%), Planctomycetes (1.6%), and Nitrospira (1.1%). These phyla are generally enriched in freshwater sediments38 and pond freshwater sediments39. All concluded that prior to addition of ENMs, the mesocosms were reasonably similar and environmentally relevant as they reliably showed similar patterns in microbial communities composition.
Phase II: CeO2 ENMs contaminationDuring phase II, pH, redox potential, total organic carbon, dissolved O2, picoplankton, and algae concentrations were stable and not different between control and contaminated mesocosms. The recovery mass balance of the total Ce was determined in the sediment, water and organisms. After 28 days, 115 18% of the Ce injected (for the pulse dosing) and 60 30% of the Ce (for the multiple dosing) were recovered (Table 2). These results were not statistically different (p 0.05). 89.2 5% to 99.2 0.2% of the Ce recovered was found in the surficial sediments (for multiple and single dosing respectively), 10.8 5% to 0.8 0.2% in the water column, and in P. corneus (Table 2). As a function of the contamination scenario, different distributions of Ce were observed over time (Fig. 4). Following the single pulse dosing, the total Ce concentration in the water column decreased to 226 86g/L at day 7, 96 25g/L at day 14, and 14 3g/L at day 28. However, multiple dosing of 5.2mg of Ce 3 times per week, maintained the concentration of Ce in the water column around 50 10g/L. It should be noted that the sampling for Ce dosing was performed before the next ENMs injection. Sedimentation rate constants (ks) of Ce in the water column were obtained over time. A ks pulse of 0.16 per day was calculated from data on the pulse dosing while the multiple dosing yielded a more complex behavior. The multiple dosing in fact consisted borse gucci 2011 of a series of smaller pulse inputs of ENMs to the system. Initially, a higher rate constant (ks multiple1 of 0.34 per day) characterizes the sedimentation of the first input of ENMs as calculated from the data following the first injection, which was treated mathematically as the response to a pulse input. A value of ks multiple2 equal to 0.44 per day is calculated from the longer term trend in Ce concentrations (
Our pilot study demonstrated that the physical chemical conditions during 45 days were reproducible between the 9 mesocosms. Two weeks were necessary to reach a state of equilibrium with anoxic conditions in the sediments, the sedimentation of the particles, and the homogenization of the microbial community composition. The 14 day stabilization time is consistent with earlier mesocosm studies for estuarine ecosystems40, 41. At the end of the contamination, recovery mass balances were about 115 18% of the Ce (pulse dosing) and 60 30% of the Ce (multiple dosing) which is in agreement with 84% of recovery observed in ref. 7 or 68 76% obtained in ref. 10.
This pilot study also highlighted that the exposure of the organisms (benthic vs. planktonic) will strongly depend on the contamination scenario. After a pulse dosing, the ENMs aggregated in 1 week as evidenced by the decreased of the total Ce concentration in the water column (to 15g/L) and concomitantly increased in the surficial sediments (to 540mg/kg). In contrast, after multiple dosing, total Ce concentration in the water column remained almost constant (50 10g/L) while slightly increasing in the sediment (100mg/kg). Sedimentation appears to have favored the ingestion of Ce by P. corneus since at day 28, 104 75mg/kg and 60 40mg/kg (dry weight of digestive gland) were assimilated following a pulse versus multiple dosing. Sedimentation of Ce is related to the homo or heteroaggregation kinetics of the ENMs. In our pilot study, the initial number of colloidal particles (clay, bacteria, etc.) was low (105particles/mL). Hence, the addition of 1mg/L of ENMs corresponding to 6 105particles/mL could lead to both homoaggregation of CeO2 ENMs and to their heteroaggregation with other particles. water quality and depth, sediment mineral