4.1. Physico-Chemical Characterization and Stability Evaluation of Graphene Oxide
The fate and aggregation behavior of GO was characterized by DLS, electrophoretic mobility determination, and UV-Vis spectroscopy methods in the assembled test systems. About the stability of NPs, the zeta potential can give an idea. The zeta potential value represents the magnitude of the electrostatic or charge attraction between particles. Particles with a zeta potential below −30 and over +30 mV are considered stable (at pH 7), while values between −30 and +30 mV NPs are not considered stable. Instead, they are expected to aggregate over time [20
]. The zeta potential values measured in the assembled test systems indicate that GO suspensions are outside the stability limit and tend to form aggregates at all concentrations in which the phenomenon could be easily examined by bare eyes over time. Due to the strong instability of GO in the applied test medium, GO floccules were formed, which led to a slow phase separation-like behavior and daphnids tended to avoid the GO-enriched lower phase. However, the volume of this lower phase was different at each applied concentration and showed the systematic influence of the initial GO concentration and time as confirmed by Reference [21
] when diluting GO suspension with tap water. Due to GO flocculation and sedimentation representative sampling becoming particularly challenging in the inhomogeneous lower phase, no attempt was made to determine the GO concentration in the two phases. Nevertheless, a clear tendency of GO sedimentation over time was confirmed by the UV-Vis spectra as well. The above described phenomenon was also reported by Lv et al. [21
] when testing GO, highlighting that the presence of D. magna
resulted in enhanced aggregation of GO when compared to abiotic circumstances.
The magnitude of the average hydrodynamic diameter (HDD) did not correlate strongly with the concentration of GO after 48 h in the upper phase. It was similar (~1 µm) at 3.125 and 6.25 mg/L, indicating the presence of aggregated micron-size particles even at low GO concentrations. Aggregation and a change in GO characteristics such as HDD may be influenced by several parameters such as pH, ionic strength, and natural organic matter might affect its bioavailability and, hence, play an important role in determining toxicity [5
The interaction of nanoparticles with biological systems is widely studied by colloid chemical methods, i.e., the size change (aggregation) of particles by dynamic light scattering (DLS) and the change of surface charge, and, usually, as a consequence, the stability by zeta potential measurements. DLS can be used for the characterization of biomolecules, nanoparticles, and their aggregation [23
]. However, it is not selective. Thus, all components in a complex biological system might contribute to the signal. Considerable work was directed to improve the selectivity of the method, e.g., by combining fluorescent marking with DLS or using depolarization techniques [24
]. Similarly to DLS, it is difficult to get accurate values for zeta potential in complex systems since zeta potential is affected by various factors including the properties of particles (surface functionalization, diameter), particle concentration, particle-particle interactions, buffer components, the presence of proteins, or other biomolecules, etc. A very comprehensive review on the topic was published by Lowry et al. [25
]. By limiting the measurement time, the effects of particle aggregation and blackening (screening) the electrode by macromolecular or other coating of the electrode can be minimised. Using a low electric field can help avoid the change of the sample during the measurement. When keeping all limitations in mind, it is generally useful to compare the zeta potential with the rule-of-thumb stability limit (30 mV), but the possibly high standard error of zeta potential values must be taken into account.
4.2. Acute Ecotoxicity of Graphene Oxide to Daphnia Magna
The present study clearly displayed the potential of GO to affect Daphnia magna physiology, behaviour, and oxidative stress reactions, whereas the extent of the effects depend on the GO concentration, exposure period, and tested endpoints besides the physico-chemical parameters.
Similarly to the results of our conventional, standardized lethality, and immobilization assays, Lv et al. [21
] also described non-severe acute toxicities, including immobility (EC50
(72 h) = 44.3 mg/L) and mortality of D. magna
(72 h) = 45.4 mg/L) upon GO. These results are also in accordance with Zhang et al.’s [26
] findings reporting EC50
(48 h) = 84.2 mg/L and mitigation by 32.3% of the acute toxicity in the presence of humic acid (25 mg/L). The effective concentration values determined based on our results are not directly comparable with those determined by Lv. et al. [21
] since the maximum exposure period in the case of our experiments was 48 h. However, it could be stated that the EC20
(48 h) = 50 mg/L for mortality was close to their results.
The GO-mediated feeding behaviour showed a strong relation with the polydispersity index (PDI) and the hydrodynamic diameter (HDD) determined in the test medium at different GO concentrations. There was no significant inhibition at the lowest GO concentrations (3.125 and 6.25 mg/L), while the average HDD was much lower (0.97–0.98 µm) than at the next, higher concentrations tested at which the PDI and the HDD increased exponentially. At these higher PDI and HDD values, the feeding inhibition became significantly elevated, indicating that the toxicity was closely related to these parameters. The elevated toxicity may be due to the higher particle size of GO at 12.5–50 mg/L (4.86–9.49 µm), which could be retained by the filtering apparatus of daphnids, while the accumulation level of the smaller particles was found to be very low [19
Based on previously reported depuration studies at lower GO (<1 mg/L) concentrations, achieving almost complete depuration of GO [27
], our aim was to study the effect of the maximum GO concentration in the test medium, which would still induce a satisfactorily restorable adverse effect on feeding. Our results revealed altered feeding behaviours after 24 and 48 h of GO exposure, which could be recovered only to a certain extent even after 24 h of a recovery period in a clean test medium. Guo et al. [27
] measured the rate of GO depuration after the exposure of daphnids to GO at 50, 100, and 250 µg/L for 24 h, and then replaced them into clean artificial freshwater for 1, 4, 10, and 24 h. GO uptake was expressed in body burden units (µg graphene/mg Daphnia dry mass) before and after depuration. The results showed that Daphnia excreted 46% and 64% of the accumulated graphene from their guts after being exposed to a graphene concentration of 100 and 250 μg/L, respectively, in clean artificial freshwater (AF), resulting in roughly constant body burden after elimination for 24 h in clean AF for Daphnia exposed to different concentrations during the uptake experiments. In addition, if Daphnia were fed algae during the depuration period, the body burdens decreased more than 90% during the first 4 h.
Furthermore, the presence of 10 mg/L humic acid (HA) in the AF also increased the graphene excretion rate compared to the clean AF condition. It was supposed that this enhancement may be attributed to the interactions between HA and graphene or to the similarities between the molecule size of HA and algae, but further studies are necessary to understand the mechanism [27
Assuming a direct link between the GO depuration rate and restored feeding activity, it could be stated that, in our experiments, roughly a complete GO depuration could be reached even after 4 h in clean growth medium in the case of exposure to a lower GO concentration of up to 12.5 mg/L for 48 h, according to the restored feeding activity, compared to the control, without addition of algae or of any kind of natural organic matter, which proved to facilitate GO depuration from the gastric tract, according to Guo et al. [27
]. Based on these results, acute exposure to GO at a relatively high concentration (12.5 mg/L) did not result in irreversible damage in the normal food uptake rate. However, the effect of chronic exposure to GO at a low and a very low (ng–µg/L) level is still an open question to be investigated as well as the potential risks of other contaminants, such as organic pollutants and heavy metals present in the aquatic ecosystems, which may adsorb onto GO nanoparticles [28
However, Loureiro et al. [30
] investigated the effect of pegylated GO (nGO-PEG) nanoparticle suspension at 2–400 µg/L in 21 d chronic toxicity test, following reproduction rate and body length, which were strongly connected to the normal rate of feeding, revealing that body length was not affected at any of the applied nGO-PEG concentrations. However, all tested concentrations of nGO-PEG resulted in a significant increase in the reproductive output. These findings support the importance of the ecotoxicity testing of different GO particles with and without further modifications, altering bioavailability and resulting in different toxicity patterns, even eco-friendly forms of functionalized GO.
Based on our knowledge, the effect of GO on the heartbeat rate of D. magna
has not been investigated yet. However, it could be hypothesized that this physiological endpoint may be altered by previously reported GO induced oxidative stress [12
]. According to our results, the heartbeat rate was found to be lowered due to GO exposure for 48 h at 12.5, 25, and 50 mg/L concentrations. However, it was also found that, after replacement of the daphnids into clean growth medium for 4 h, the heartbeat rate was normalized and no significant difference compared to the control group was determined. In the current literature, there is only one example of the testing of CNM with a D. magna
heartbeat rate assay. Lovern et al. [31
] investigated the effect of nano-C60
and a fullerene derivative, exposing D. magna
to 2 ppm C60
for 1 h. They found that the nano-C60
was the only suspension to cause a significant change in the heartbeat rate when compared to the control, increasing the average rate from 43.6 bpm to 317.09 ± 7.21 bpm after exposure. This suggests that the functional groups may have a drastic effect on the sublethal effects (such as behavior and physiology) of nanoparticles.
The key potential mechanism of nanoparticles’ toxicity is supposed to be via oxidative stress with reactive oxygen species [32
]. Although the oxidative stress inducing property of GO at certain concentrations in different aquatic organisms [33
] and in D. magna
] was confirmed, our aim was to determine the extent of GO-induced oxidative stress within the concentration range used in our experiments as well as to complement our ecotoxicity toolkit with a method having the ability to reveal the potential underlying molecular mechanisms potentially influencing the physiological and behavioral endpoints. We determined a concentration-dependent significant increase of the oxidative stress level within the 6.25–50 g/L GO concentration range. Zhang et al. [26
] found oxidative damage caused by GO in D. magna,
while resulting approximately 135%, 130%, and 250% increase in ROS, superoxide dismutase (SOD), and catalase (CAT) activity, respectively, at 10 mg/L GO concentration in the acute test. A significant increase in ROS generation was confirmed by Cano et al. [12
] when exposing D. magna
to GO even at a very low concentration (0.1 mg/L) for 72 h.