The influence of salinity on the effects of Multi-walled carbon nanotubes on polychaetes

Salinity shifts in estuarine and coastal areas are becoming a topic of concern and are one of the main factors influencing nanoparticles behaviour in the environment. For this reason, the impacts of multi-walled carbon nanotubes (MWCNTs) under different seawater salinity conditions were evaluated on the common ragworm Hediste diversicolor, a polychaete species widely used as bioindicator of estuarine environmental quality. An innovative method to assess the presence of MWCNT aggregates in the sediments was used for the first time. Biomarkers approach was used to evaluate the metabolic capacity, oxidative status and neurotoxicity of polychaetes after long-term exposure. The results revealed an alteration of energy-related responses in contaminated polychaetes under both salinity conditions, resulting in an increase of metabolism and expenditure of their energy reserves (lower glycogen and protein contents). Moreover, a concentration-dependent toxicity (higher lipid peroxidation, lower ratio between reduced and oxidized glutathione and activation of antioxidant defences and biotransformation mechanisms) was observed in H. diversicolor, especially when exposed to low salinity. Additionally, neurotoxicity was observed by inhibition of Cholinesterases activity in organisms exposed to MWCNTs at both salinities.


Materials and Methods
Sampling and experimental conditions. H. diversicolor specimens were collected from the Mira channel, a low contaminated area at the Ria de Aveiro lagoon (Portugal) 57,58 . Organisms with similar weight (0.49 ± 0.2 g) were used to prevent differences on biochemical responses.
In the laboratory, specimens were placed in 5 different aquaria (20 L each) for 2 weeks. Aquaria were filled with a mixture of fine and medium sediment from the sampling area (a clean sampling area with sediment median value 1.59; percentage of fines 6.75 ± 0.79; percentage of organic matter content 3.24 ± 0.44), approximately 1/3 of the height of the aquarium, and artificial seawater (salinity 28) set up by the addition of artificial sea salt (Tropic Marin ® Sea Salt) to deionized water, one day prior to utilization. During the acclimation period, organisms were maintained under constant photoperiod of 12 h light: 12 h dark, temperature (18 ± 1 °C), pH 7.9 and constant aeration. To respect species feeding requirements, during this period every two-three days specimens were fed ad libitum with commercial fish food (48.6% protein and 7.7% fat) 59 .
After the acclimation period, organisms were exposed for 28 days to two different salinities (21 and 28-control), each one combined with two different MWCNT concentrations (0.10 and 1.00 mg/L). For each condition, 3 aquaria were used with 10 organisms per aquarium. Each aquarium was set up at the same pH, temperature and aeration conditions as in the acclimation period (see above).
Prior to experiment initiation, the salinity was progressively decreased (1-2 units) every 2 days until testing value was reached (salinity 21). Both MWCNT concentrations and salinity levels were re-established weekly after complete water renewals to ensure the same exposure concentrations during the experiment. To promote stable suspension in the water column 60 , the MWCNTs were sonicated for 1 h using an ultrasound probe at 30 Hz (IKA Labortechnik IKASONIC U50). The added CNTs were homogenously dispersed in the water using one submersible circulation pump per aquarium, which diminishes the possibility that the dynamical equilibrium between gravitational settling and Brownian motion can result in the presence of CNTs near the sediment-water interface 61 .
The concentrations of MWCNT used in this study were 0.10 and 1.00 mg/L prepared from a stock solution of 50 mg/L concentration. Although the Environmentally Relevant Concentrations (ERCs) of CNTs in water (based on a stochastic/probabilistic material-flow computer model) are in the range of µg/L or ng/L 62 , the concentrations tested were chosen considering the PECs of CNTs in aqueous systems reported from the most recent literature 52,53 , projected to approximately 0.001-1000 µg/L. These concentrations have been shown to produce cytotoxic, genotoxic, inflammatory and oxidative stress responses in different organisms 63 The average size distribution and the polydispersity index (PDI) of MWCNT suspensions were analysed in each CNMs exposure condition (control-28 + 0.10 mg/L MWCNTs; control-28 + 1.00 mg/L MWCNTs; 21 + 0.10 mg/L MWCNTs; 21 + 1.00 mg/L MWCNTs) and at different exposure periods: t0: time zero, immediately after medium contamination (dispersed MWCNT concentrations in artificial seawater); t7: water samples collected after 1 week of exposure, before water renewal; t28: water samples collected after fourth week of exposure, before the end of exposure (Table 2).
Measurements were performed on 1000 µL of suspension in three samples per condition, and five analyses per sample performed by dynamic light scattering (DLS), using a DelsaTM NanoC Particle Size Analyzer (Beckman Coulter). Each analysis was carried out by performing 120 acquisitions.

MWCNTs in sediments.
Thermogravimetric analysis (TGA) has been used as innovative method to assess the presence of MWCNT aggregates in the sediments exposed to MWCNT dispersions. To the best of our knowledge no reports describing the use of TGA for such purpose are reported in the literature. TGA may represent an effective method to detect the presence of MWCNTs in the sediments since it records the weight loss of materials upon heating and can distinguish the contribution given by each component of a mixture by applying the derivative operation to the thermogravimetric curve (DTGA analysis). To this aim the degradation behaviour of MWCNTs is unique and might be easily distinguished from that of both inorganic and organic background 66 .
The detection of MWCNTs in sediments has been carried out on 50 mg samples by thermogravimetric analysis (TGA) by using a Mettler Toledo Star-system TGA/SDTA 851e apparatus, with air as the purge gas (60 mL/ min) at a scan rate of 10 °C/min. Samples were conditioned for 90 min. at 150 °C prior to analysis in order to remove water.
Physiological parameters. Mortality. Mortality was determined in H. diversicolor by subtracting the total number of dead individuals at the end of the exposure period (28 days) by the total number of individuals used at the start of the experiment, within each exposure conditions: (control-28 + 0.00 mg/L MWCNTs; control-28 + 0.10 mg/L MWCNTs; control-28 + 1.00 mg/L MWCNTs; 21 + 0.00 mg/L MWCNTs; 21 + 0.10 mg/L MWCNTs; 21 + 1.00 mg/L MWCNTs). Mortality was expressed as percentage.

Property
Unit Value Biochemical parameters. The whole body of frozen organisms (3 per aquarium, 9 per condition) was individually pulverized with liquid nitrogen, divided in 0.2 g aliquots, and used for biochemical analyses. Extractions were performed with specific buffers to determine: (i) energy reserves and metabolic capacity (protein (PROT) content, glycogen (GLY) content, electron transport system (ETS) activity); (ii) activity of antioxidant (superoxide dismutase (SOD); catalase (CAT); glutathione peroxidase (GPx)) and biotransformation (Glutathione S-transferases, GSTs) enzymes; (ii) indicators of cellular damage (lipid peroxidation (LPO) levels, reduced (GSH) and oxidized (GSSG) glutathione content) and iv) neurotoxicity (cholinesterases (ChEs) (Acetylcholinesterase (ATChI-ChE) and Propionylcholinesterase (PTChI-ChE) enzymes activity) markers. Biochemical analyses were performed twice for each sample and parameter (18 measurements per condition). The methodologies used to obtain each specific biomarker are described in details in De Marchi et al. 67,68 . Data analysis. Results of MWCNT characterizations, PROT and GLY contents, ETS activity, SOD, GPx, CAT and GSTs activities, LPO levels, GSH/GSSG, and ChE activities, were submitted to hypothesis testing using permutational multivariate analysis of variance with PERMANOVA + add-on in PRIMER v6. A one-way hierarchical design was followed in this analysis. The pseudo-F p-values in the PERMANOVA main tests were evaluated in terms of significance. Significant differences were observed using main test and consequently pairwise comparisons were performed. Values lower than 0.05 were considered as significantly different. The null hypotheses tested were: (i) for each biomarker and for each salinity level, no significant differences existed between MWCNT exposure concentrations or, on the contrary, there is a concentration-dependent toxicity in H. diveriscolor exposed to MWCNTs material under both salinities; (ii) for each biomarker and for each MWCNTs exposure concentration, no significant differences exist between salinity levels or, on the contrary, salinity shifts altered the toxicity of the MWCNTs material as well as the sensitivity of H. diversicolor exposed to this contaminant. In figures, different letters represent significant differences among exposure concentrations for each salinity level, with lower case letter for conditions maintained at 28 and upper-case latter for conditions maintained at 21. Significant differences between salinities for each MWCNTs concentration are represented with an asterisk.

Method of Measurement
A matrix gathering all markers responses per exposure concentration was used to calculate a Euclidean distance similarity matrix. This similarity matrix was simplified through the calculation of the distance among centroids matrix based on the exposure concentrations, which was then submitted to ordination analysis, performed by Principal Coordinates (PCO). Pearson correlation vectors of biochemical descriptors (correlation > 0.75) were provided as supplementary variables and superimposed on the PCO graph.

Results
MWCNTs characterization. Water characterization. The water samples obtained after 7 days (d7) and 28 days (d28) of exposure to MWCNTs were almost transparent solutions containing whitish coarse material susceptible to settling and precipitation. In order to assess the dimension of the dispersed material DLS analysis was carried out by using ad hoc experimental conditions such as the vigorous and reproducible agitation of each sample before the analysis and the record of a large number of acquisitions during the measurements.
The mean diameter and polydispersity index values, reported in Table 2 were obtained through the cumulant method, which represents the most effective processing tool to reliably compare the DLS data of polydisperse samples and monitor their time evolution. DLS analysis of the samples taken at d7 and d28 did not evidence the presence of particles whose mean dimension was higher than that of the controls indicating the absence of materials deriving from dispersed MWCNTs.
Sediment characterization. TGA analysis showed the presence of MWCNT aggregates in sediments exposed to MWCNT dispersions stating the feasibility of sediment experimental samples ( Fig. 2A). DTGA analysis evidenced the presence of a negative peak at 600 °C relevant to the maximum rate of degradation of MWCNTs. The thermal characterization was carried out on samples exposed to higher concentration of MWCNT dispersions (1.00 mg/L) and compared to that of the sediment used as control (Fig. 2B).
The weight loss (%) of all the analyzed samples were negligible at 900 °C due to the inorganic nature of the sediments. DTGA analysis of the obtained TGA curves was carried out in order to better understand the steps of degradation (Fig. 2C).
The preliminary DTGA investigation performed on the sediments exposed to MWCNT dispersions did not clearly highlight the presence of negative peaks in correspondence with that of native MWCNTs (600 °C, cf. Fig. 2A). This could be most probably due to the experimental conditions adopted to individuate trace amount of MWCNTs particulate. Further investigation will be devoted to fine tune and optimize the experimental condition in order to detect even traces of MWCNTs in the sediment. However, the use of TGA analysis could represent a promising tool for the detection of MWCNTs in inorganic sediments due to the different degradation behaviour Physiological parameter. Mortality. At control (28 + 0.00 mg/L MWCNTs) and 21 + 0.00 mg/L MWCNTs, 100% of survival was recorded after 28 days' exposure.
Exposures to 0.10 mg/L under salinity 28, mortality was 2.22% while exposures to the same MWCNTs concentration under salinity 21 resulted in 6.66% of mortality.
Biochemical parameters. Energy reserves and metabolic activity. At both salinities (28 and 21) the lowest GLY content was observed in polycheates exposed to the highest MWCNTs concentration (1.00 mg/L), with significant differences compared to the remaining conditions (control and 0.10 mg/L). No significant differences in GLY content were observed between polychaetes exposed to control and 0.10 mg/L MWCNTs. No significant differences were observed in terms of GLY content when comparing the organisms under both salinities for each of the tested concentrations (control-0.00, 0.10, 1.00 mg/L) (Fig. 3A). Under both salinities, polychaetes showed significantly lower PROT content when exposed to MWCNTs compared to control individuals, with significant differences among all tested concentrations (Fig. 3B). When comparing organisms exposed to different salinities at each of the tested concentrations (control-0.00, 0.10, 1.00 mg/L), lower PROT content was observed in organisms exposed to the lowest salinity, but no significant differences were observed between salinity levels.
With the increase of MWCNTs exposure concentration, the ETS activity increased in polychaetes under both salinities (control-28 and 21), with significant differences among all exposure concentrations (Fig. 3C). Organisms under the lowest salinity presented significantly lower ETS activity in polychaetes exposed to 0.10 mg/L (Fig. 3C).
Antioxidant and biotransformation enzymes activity. The activity of SOD increased in polychaetes with the increase of exposure concentrations, regardless of the salinity level (control-28 and 21), with significant differences among exposure concentrations (Fig. 4A). Comparing SOD activity in polychaetes at each of the exposure concentrations (control-0.00, 0.10, 1.00 mg/L) no significant differences were observed between individuals exposed to different salinities (28 and 21) (Fig. 4A).
The activity of CAT showed no significant differences among exposure concentrations, both in polychaetes exposed to salinity 28 and salinity 21 (Fig. 4B). Furthermore, at each exposure concentration no significant differences were observed between organisms exposed to different salinities (control-28 and 21) (Fig. 4B). At both tested salinities, the activity of GPx was significantly higher in polychaetes exposed to MWCNTs (0.10 and 1.00 mg/L) relative to control organisms, with the highest values in polychaetes exposed to 0.10 mg/L (Fig. 3C). In general organisms exposed to salinity 21 presented higher GPx activity, however significant differences between organisms exposed to salinities 21 and 28 were only observed at 0.10 mg/L (Fig. 4C).
Comparing to control, at both salinities 28 and 21, the activity of GSTs significantly increased in polychaetes exposed to 0.10 mg/L while in polychaetes exposed to 1.00 mg/L, the activity of these enzymes significantly decreased (Fig. 5). At each of the exposure concentrations (control-0.00, 0.10, 1.00 mg/L), a similar trend was found, with no significant differences between polychaetes exposed to both salinities (control-28 and 21) (Fig. 5).
Indicators of cellular damage. Regardless of salinity (control-28 and 21), polychaetes exposed to MWCNT concentrations showed increased LPO levels along the increasing exposure gradient, with significant differences among all conditions (Fig. 6A). Comparing LPO levels in polychaetes under different salinities at each of the exposure concentrations (control-0.00, 0.10, 1.00 mg/L), significant differences were only found at the highest concentration, with organisms under salinity 21 presenting higher LPO values (Fig. 6A).
GSH/GSSG significantly decreased in polychaetes with the increase of exposure concentrations under both salinities, showing significant differences among all concentrations (Fig. 6B). Comparing GSH/GSSG values in organisms at each of the tested concentrations (control-0.00, 0.10, 1.00 mg/L), significantly different values were only observed in organisms at the control conditions, with higher values at salinity 28 (Fig. 6B).
Neurotoxicity markers. No significant differences in ATChI-ChE activity were observed between polychaetes exposed to different concentrations for both salinities tested (Fig. 7A). Moreover, no significant differences were also observed when comparing the ATChI-ChE activity in polychaetes under different salinities at each of the exposure concentrations (Fig. 7A).

Discussion
In the present study, H. diversicolor was exposed to MWCNTs and biochemical alterations induced in the organisms after 28 days of exposure to two salinity levels (28 and 21) at two concentrations (0.10 and 1.00 mg/L) were investigated. Our finding demonstrated that the salinity decrease, may not change the toxicity of MWCNTs neither the sensitivity of organisms to these NMs. Indeed, our results showed that under salinity 21 the toxicity of MWCNTs was similar to the impacts measured under control (28) in terms of oxidative responses, organism's metabolism and  69 demonstrated salinity-dependent toxicity of NMs in fish. These authors used different concentrations of embryo-rearing medium (ERM) (1x, 5x, 10x, 15x, 20x and 30x) (1 × ERM consisted of 1.0 g NaCl, 0.03 g KCl, 0.04 g CaCl 2 2H 2 O and 0.163 g MgSO 4 7H 2 O in 1 L of ultrapure water). The results showed that from freshwater (1 × ERM) to seawater (30 × ERM), the toxicity of silver nanocolloids (SNC) to medaka embryos increased. However, the same authors found that at different ERMs, the zeta potentials of a CNTs such as SWCNTs were stable, without differences of particle sizes from freshwater to seawater 54 . In fact, raw CNTs, such as MWCNTs or SWCNTs, are nearly insoluble in any solvent owing to their bundle-shaped configurations 70 and these properties could partially explain why salinity didn't increase the toxicity of MWCNTs.
In the present study H. diversicolor showed increased ETS activity with the increase of exposure concentrations, at both salinity conditions, which might indicate that the increase of metabolic capacity of H. diversicolor was necessary to activate defence mechanisms to mitigate oxidative stress induced by MWCNTs, leading to the consumption of energy reserves. Our findings are in agreement with the present hypothesis, showing a decrease of energy reserves (measured by GLY and PROT content) when H. diversicolor specimens were exposed to MWCNTs, a response that was similar under control (28) and low (21) salinity conditions, indicating that the impacts induced by MWCNTs were not altered by salinity shifts. When organisms are exposed to anthropogenic environmental disturbances, they can increase their energy expenditure as a mechanism of cellular protection. Sokolova et al. 71 already demonstrated that marine organisms are able to increase their energy reserves to protect the cell against environmental disturbances. Gagné et al. 72 pointed out that energy expenditure in invertebrates is enhanced when exposed to different pollutants, restricting the amount of energy reserves necessary for survival, homeostasis and reproduction. Also, De Marchi et al. 32 showed that H. diversicolor and Diopatra neapolitana polychaetes presented higher metabolic activity with the increase of NMs exposure concentration, although energetic expenditure was not observed.
The impacts of NMs may also be observed by oxidative stress responses developed by organisms under stressful conditions, which can be assessed by measuring the activity of antioxidant and biotransformation enzymes (such as SOD; CAT; GPx; GSTs), and cellular damage (assessed by LPO) induced due to overproduction of reactive oxygen species (ROS) 24,25 .
In the present study, the activity of SOD, which is the enzyme responsible for the removal of superoxide anion (O 2 − ) forming hydrogen peroxide (H 2 O 2 ) 73 , increased in organisms exposed to MWCNTs, especially at the highest concentration (1.00 mg/L) showing a possible adaptive response to increased ROS production as a  32 , using the same carbon NMs, showed that this antioxidant enzyme in D. neapolitana and H. diversicolor polychaetes tended to maintain the activity among the exposure conditions. These results may be explained by the fact that the excess of H 2 O 2 produced by SOD could be eliminated by another antioxidant enzyme with similar capacity (i.e. degradation of H 2 O 2 ) and, in the present study, this hypothesis was confirmed by an increase of GPx activity in polychaetes contaminated with 0.10 mg/L MWCNTs compared to non-contaminated. However, in the present study the activity of GPx decreased at the highest exposure concentration (1.00 mg/L) in comparison to the activity observed at 0.10 mg/L. These results could be due to the oxidative stress persisting, or with extreme high ROS levels, which can cause proteins damage and a consequently decrease of the activity of this antioxidant enzyme, as already demostrated by Chatziargyriou et al. 74 , where the inhibition of enzyme activity (i.e. GPx) under extreme stressful conditions effects was observed.
Organisms may also present increased activity of GSTs in the presence of contaminantes, a biochemical marker widely used to evaluate the detoxification capacity of organisms catalysing the conjugation of the reduced form of glutathione (GSH) to xenobiotic substrates 73 . Studies, using polychaetes as sentinel species of different quantum dots and metal-based NMs exposures, already showed an increase of these biotransformation enzymes as sensitive biomarkers of defence responses to NMs 24,25,[75][76][77][78][79] . However, in the present study the activity of GSTs increased at the lowest tested concentration (0.10 mg/L MWCNTs) but decreased in organisms exposed at the highest concentration (1.00 mg/L MWCNTs), at both salinities. These results may indicate that the activity of this group of enzymes can be activated up to a certain level of stress (i.e. up to a certain exposure concentration) while the activity decrease may be associated to a possible inhibition due to the high MWCNTs concentration. This behaviour it is in line with the results of De Marchi et al. 80 , which using carboxylated form of MWCNTs showed that the GSTs activity was decreased compared to control at the highest exposure concentration in the clams Ruditapes philippinarum. This findings indicated that this group of enzymes was not able to biotransform CNTs into a less toxic excreted substance beyond certain concentration levels, being inhibited at high stress conditions.
In the present study, although the activation of antioxidant enzymes activities in H. diversicolor exposed to MWCNTs, the LPO, which is a marker of cellular injury and is an indicator of oxidative damage in cell  32 already observed that the activation of antioxidant enzymes activities in H. diversicolor and D. neapolitana exposed to MWCNTs was not enough to prevent cellular damage and LPO level increased along the increasing exposure gradient. Also, Anisimova et al. 82 , exposed the bivalve Modiolus modiolus to 12-14 nm diameter MWCNTs (100 mg/L) for 48 h and demonstrated that CNTs were responsible for the increase of LPO levels and increase of reduced glutathione (GSH), one of the most important antioxidant involved as cellular protector from oxidative stress, acting as a free radical scavenger preventing LPO. Furthermore, our results showed an increase of LPO when organisms were exposed at the highest MWCNTs concentration under low salinity (21). This result may be explained by the fact that higher salinity (in this case 28) decreased the availability of NMs due to their aggregation 54 . Thus, higher availability of NMs to polychaetes exposed to low salinity water media (21) may lead to higher NMs uptake by the organisms causing higher cellular damage (e.g. LPO generation) at this condition.
Stress-induced ROS accumulation is counteracted also by non-enzymatic antioxidant systems that include a variety of scavengers. The role of scavengers is to neutralize ROS by direct reaction with them, and cytosolic reduced glutathione (GSH) is one of the crucial nonprotein thiols that plays an important role in intracellular defence against ROS-induced oxidative damage 83 . GSH is able to neutralize several reactive species through its oxidation, generating oxidized glutathione (GSSG). Thus, the ratio between GSH and GSSG is often used as a marker of oxidative stress 73,84,85 . In the present study, the GSH/GSSG values decreased with the increase of exposure concentrations, at both salinities, indicating that, regardless of different salinity conditions, MWCNTs induced oxidative stress in H. diversicolor. Also, in the study performed by De Marchi et al. 32 the concentration of this scavenger, in two polychaete species (D. neapolitana and H. diversicolor) exposed to different MWCNT concentrations, strongly decreased among conditions, a pattern that confirms the development of oxidative damage in organisms exposed to CNTs.
Cholinesterase (ChE) activities are a core biomarker of neurotoxicity which interferes with behavioural responses 24 . Physiologically, this class of serine hydrolases are able to remove acetylcholine from the synaptic cleft 86 . Behavioural biomarkers are sensitive tools to assess the impact of contaminants and they are particularly relevant from an ecological point of view 87 . Our results revealed that MWCNTs impaired the hydrolytic activity of ChEs. In fact, a significant inhibition of PTChI-ChE activity in H. diversicolor under MWCNTs exposure compared to control condition was proved. Previous studies have already used these neurotoxicity markers as a useful tool to evaluate the impact of different metal-based nanomaterials on polychaetes (Ag NPs, Au NPs, CuO NPs, CdS NPs, ZnO NMs) 25,77,88 and recently also carbon nanomaterials (MWCNTs NMs) 32,89 showing, in most of the cases, a inhibition of ChEs activities in polychaetes exposed to NMs demonstrating a possible neurotoxicity induction when organisms are exposed to these emerging pollutants.
Given the ecological relevance of polychaetes, any significant injury on these key organisms may bring consequent impacts at the ecosystem level. Thus, our findings provide valuable information regarding the potential risk of CNTs to H. diversicolor considering the projected increase of nanotechnology and industrial uses of NMs and consequent release to aquatic ecosystems.