Acute Adverse Effects of Metallic Nanomaterials on Cardiac and Behavioral Changes in Daphnia Magna

Nanomaterials are widely believed to induce toxic effects on organisms by evoking oxidative stress. In this study, we evaluated the toxic effects of nanomaterials on cardiac and behavioral changes in Daphnia magna under varying exposure conditions. Titanium dioxide nanoparticles (TiO 2 NPs), silver nanoparticles (AgNPs), and silver nitrate (AgNO 3 ) were selected for the acute toxicity tests. The adverse effects of the substances on the neonates including heart rate, swimming speed, and oxidative stress, were measured. The heart rate level decreased as the concentration of NPs and silver ions (Ag + ) increased. The average swimming speed was measured to be approximately 15 mm/min for the control group. The swimming speed generally increased for longer exposure to both NPs, although it reached a plateau at the lowest concentration of AgNPs. A similar but less clear trend was observed for Ag + . For all substances, the overall swimming speed exhibited no correlation or weak negative correlations with the exposure concentration. The oxidative stress levels increased after exposure compared to the control group. We conclude that aquatic nanotoxicity tests should consider multilevel physicochemical, physiological, and behavioral parameters for the ocial guidelines to quantify more robust adverse outcomes.


Introduction
Nanomaterials are used in a wide range of sectors due to their speci c physicochemical properties compared to their bulk materials which have the same chemical composition [1][2][3] . To ensure safe use, it is important to identify and manage the effects of exposure to nanomaterials on human and environmental health more effectively. In particular, the aquatic phase is considered a key starting point of potential entry and diffusion into the environment for understanding environmental fate and behavior. The aquatic phase creates connections with other environmental compartments such as soil, sediment, and air 4 .
Several studies indicate that nanomaterials are toxic to aquatic organisms such as sh and crustaceans [5][6][7] . Of these, shes have a direct impact on human health, but it is di cult to obtain information on the exact lethal concentration of nanomaterials from sh. Therefore, the impact of nanomaterials on non-target organisms will play an important role 8 . Aquatic invertebrates are a very large group of animals and some have over 1000 different types of species. Among them, arthropods are the largest invertebrate, consisting of the two largest groups: insects and crustaceans 9 . Crustaceans are the most important group of invertebrates, of which Daphnia magna is commonly used to assess the risk of nanomaterials 10 . Algae and crustaceans exposed to nanomaterials can poison and cause death to aquatic organisms such as sh through the food chain 5,11 . Nanomaterials are generally known to have higher toxicity than bulk materials. It has been reported from in vitro studies that nanomaterials can induce biological effects such as cell in ammation and apoptosis by oxidative stress [12][13][14][15] . The preliminary factors causing toxicity are size, shape, and surface area/charge 16 . Nanomaterials are more reactive due to their large surface area, and easy penetration into cells 16,17 . Depending on the surface properties, the nanomaterials may exhibit enhanced interactions with the cell membranes 18,19 .
Here, we focus on two representative types of metallic nanoparticles, titanium dioxide (TiO 2 ) and sliver (Ag) nanoparticles (NPs). TiO 2 NPs are generally known to be non-toxic and can be applied in various ways such as through cosmetics and catalysts 20,21 . AgNPs are also used in a wide range of consumer products as antibacterial agents due to their high cytotoxicity [22][23][24] . One of the most fundamental sources of AgNP toxicity has been attributed to the release of sliver ions (Ag + ) in tissue environments.
Daphnia magna is a standard organism for ecotoxicity test using toxic chemicals in the standardized protocols suggested by the Organization for Economic Cooperation and Development (OECD), International Organization for Standardization (ISO), and the United States Environmental Protection Agency (USEPA) 25,26 . As an organism in the bottom stage of the food chain, D. magna plays a key role in aquatic ecosystems since it is sensitive to environmental stresses, and population change can ultimately affect the population of upper predators 25,27 . Previous studies have reported the biological effects of hazardous chemicals on physiological factors such as cardiac and behavioral characteristics in D.
magna [28][29][30][31][32] . Fekete-Kertész et al. (2017) evaluated the change in heart rate using D. manga and found that heart rate was in uenced by several factors; chemical exposure level (triclosan), test medium, organism age, and exposure time 33 . Chung et al. (2016) also reported that the heart rate and swimming distance were linearly changed by TiO 2 nanoparticle concentration 34 . In particular, they adopted a video tracking method which has been recently used to evaluate the behavioral characteristics of organisms such as zebra sh and D. magna. Therefore, we aimed to evaluate the toxic effects on the physiological factors of D. manga using nanomaterials and to identify the relationships between the factors according to the exposure conditions in this study.

Results
Effects of nanomaterials on the heart rate Immobilization and heart rate were measured after acute exposure. The acute immobilization test was performed in accordance with OECD guidelines 202. A median lethal concentration (LC 50 ) value of TiO 2 in D. magna after 48 hours was determined to be more than 100 mg/ml and the LC 50 values of AgNPs and Ag + were 0.0134 and 0.0016 µg/ml, respectively (Fig. 1).
The heart rate decreased as the concentration of all substances (TiO 2 NPs, AgNPs, Ag + ) increased. The heart rate of the control group was observed to 364 ± 28 BPM (beats per minute). At the lowest concentration level of each substance after exposure for 3 and 48 hours (355-385 BPM), the heart rate was not signi cantly different from that of the control level, but it decreased by approximately 7.3% (TiO 2 , 345 ± 17 BPM), 4.3% (AgNP, 347 ± 26 BPM), and 15.0% (Ag + , 302 ± 15 BPM) at the highest concentration level after exposure for 3 hours, respectively ( Fig. 2A). After 48 hours of exposure, the heart rate recovered at the lowest and medium concentration levels compared to the control level. At the lowest concentration of each substance, the heart rate was 378 ± 25 BPM for TiO 2 NPs, 385 ± 14 BPM for AgNPs, and 375 ± 17 BPM for Ag + , but it gradually decreased with increasing concentration level for all substances (Fig. 2B). The reduction rate with a higher concentration was relatively lower than that after 3 hours of exposure (TiO 2 NPs, 5.8%; AgNPs, 2.4%; Ag + , 3.3%).
Effects on the behavioral performance Behavioral performance was measured as the swimming speed, which is an averaged moving distance per minute and the speed level of the control group was approximately 15 mm/min. Real-time tracking paths were recorded according to the instantaneous swimming speed as shown in Fig. 3 with examples of 10-minute exposure. The observed behavioral change upon substance exposure was limited to locomotive movements. The swimming speed generally increased for longer exposure to both NPs, although it reached a plateau at the lowest concentration of AgNPs. A similar but less clear trend was observed for Ag + . The swimming behavior appeared to be more activated at the lowest concentration level of each substance compared to higher concentration levels. For the neonates that were exposed to 0.1 µg/mL TiO 2 NPs, the swimming speed increased proportionally with the exposure duration (from 8.2 ± 0.9 mm/min for 3 hours to 21.6 ± 2.8 mm/min for 48 hours). In addition, the swimming speed increased with increasing exposure duration, although the swimming speed was lower at higher concentrations (1.0 and 10 µg/mL) than at the lowest level (0.1 µg/mL) (Fig. 4A). In the case of AgNPs, the swimming speed showed irregular patterns by the concentration level and exposure duration. It signi cantly increased up to the 2-hour exposure duration from 7.4 to 17.2 mm/min, and decreased slightly after 24-hour exposure at 0.0001 µg/mL. It continuously increased in proportion to the exposure duration at 0.001 µg/mL from 6.7 to 16.0 mm/min, but this trend was weakened at the highest concentration ( Fig. 4B). At the lowest concentration of Ag + (0.00001 µg/mL), the speed increased proportionally up to 24 hours of exposure (from 7.0 ± 0.8 mm/min for 3 hours to 21.7 ± 4.1 mm/min for 24 hours). There was no obvious pattern at higher concentration levels, but the speed peaked for the neonates exposed for 24 hours (Fig. 4C). Realtime temporal variations in swimming speed for each material are available in the supplementary information ( Fig. A1-A3).

Quanti cation of oxidative stress
Reactive oxygen stress (ROS) levels were measured by 2,7-dichloro uorescein diacetate (DCFDA) assay. Figures 5 & 6 show the results of ROS measurements in D. magna. All three materials exhibited stronger intensities than the control group after 3 hours of exposure. The error bars in Fig. 6 represent the arithmetic mean and standard deviation (n = 3, mean ± SD). In particular, the ROS intensity of TiO 2 signi cantly increased in proportion to the concentration, but the AgNPs and their ions showed concentration-dependent tendencies and then decreased at the highest concentration (0.01 µg/mL for AgNPs and 0.001 µg/mL for Ag + ). The results indicate that exposure to all substances for 3 hours signi cantly affected ROS levels, while exposure for 48 hours decreased ROS levels at a higher concentration, which is likely due to the alteration of the effective concentration after 48 hours of exposure.

Discussion
The toxicity test using D. magna is well documented in the OECD guidelines 202 and it is determined by immobilization at 24 and 48 hours after exposure to nanoparticles. In most cases, it is di cult to determine accurate toxic effects based on the method because the data varies depending on the different groups. Considering this limitation, we tried to test the acute toxicity on D. magna using in uential indices including immobilization, heart rate, swimming performance and ROS level. for AgNO 3 , which were lower than our results 36 . The results occur depending on the strain and water physicochemical parameters, which means that even if we measured the LC 50 considering experimental conditions, the outcomes might be different between laboratories.
The change in heart rate as an index of toxic effects on D. magna is widely used in acute toxicity tests. The heart rate was measured repeatedly using more than 10 neonates exposed to the materials and a cross check for all measurements was also performed to minimize bias by observers. The results showed that the heart rate decreased with increasing concentration for all substances. It was possible to obtain a similar pattern even for 3-hour exposure with the OECD guidelines (48-hour exposure) 37 . Thus, we con rmed that an exposure duration of 3 hours was su cient to test acute toxicity. In addition, TiO 2 NPs considered to be non-toxic, also reduced the heart rate by increasing the exposure level. This showed that TiO 2 NPs affect the heart rate of D. magna, similar to other toxic NPs (e.g., AgNPs), but it is limited to conclude that TiO 2 NPs in uence the immobilization (behavioral performances).
Locomotion-based behavior is a highly sensitive index for identifying toxic effects of chemicals 28,38 . Thus, immobilization was characterized by measuring swimming speed (video tracking) in addition to heart rate counting. In particular, the behavioral change was measured at each exposure duration (3, 6, 12, 24, 48 hours), thus we could identify the temporal effects in detail compared to most previous studies that examined tests after 48 hours of exposure according to the OECD guidelines 202 (Fig. A1-A3 in the Supplementary Information).
To date, information on not only the temporal variations in immobilization but also the relationship between heart rate and behavioral responses to the toxic effects of nanoparticles remains limited. NPs, fullerenes and fullerene derivatives 31 . The authors found that an increased exposure level in uenced the low heart rate and movement. Additionally, the effect was not statistically signi cant for D. manga exposed to TiO 2 NPs, but the temporal variation was not characterized. In our study, the heart rate increased after 48 hours of exposure compared with that after 3 hours of exposure, while it decreased in proportion to the exposure level (Fig. 2). The swimming speed at an earlier exposure duration (3 hours) was low for all materials and increased as time elapsed (Fig. 4). Therefore, it can be inferred that the exposure duration has more in uences on the cardiac and behavioral effects at earlier periods, and D. magna are more active after a certain stabilization period.
The ROS level was additionally measured to support the outcomes in this study. The ROS levels of the target materials increased with the exposure level, but they gradually decreased from a certain level after 48 hours of exposure. The relative intensities of ROS were observed with inverted U-shaped patterns after 48 hours of exposure. This can be interpreted as the ROS levels recovered by enhanced antioxidative responses at high concentrations 39,40 and the overall uorescence is enhanced by ROS production from the nanoparticles. The excessive ROS generation can cause oxidative stress 41 . Previous studies have also reported that oxidative stress can cause toxic effects on heart rate, swimming speed, and reproduction in D. magna [42][43][44] . In this study, it was observed that the nanoparticles enhanced the ROS levels compared to the control group, which is also consistent with previous studies 13, 45 .
There have been signi cant differences between the results on immobilization, cardiac effects and behavioral changes in previous studies. Therefore, a comprehensive test including these indices would be necessary to improve the quality of the results on acute toxicity. Furthermore, the effects of nanomaterial mixtures on aquatic toxicity need to be explored in terms of invertebrate heart physiology and swimming patterns 46,47 .

Experimental scheme
This study consists of three steps to examine the toxicity of nanomaterials using D. magna. The procedures include sample preparation (test organism, test substances), treatment (exposure), and analytical tests using instruments. One-day-old neonates that were newly born within 24 hours were prepared and certain exposure levels of nanomaterials were treated for different exposure times. After exposure duration at each concentration level, physiological effects on the neonates including the cardiac effect and behavioral changes were characterized. The test procedures for each process are described in the following sections in detail.
Daphnia magna culture D. magna cultured in the laboratory was used in a series of experiments. The test organism was cultured in 5 L beakers in a 21 ± 1ºC thermostatic incubator under a constant light cycle for 24 hours. The light conditions were maintained in the light for 16 hours, and then the light was turned off for 8 hours (12 W, 356 mm lamp, 672 Lumen, 4200 K). To maintain the breeding condition, the D. magna adults were cultivated in OECD M4 medium, and daily feeding for the organisms was conducted using an algae containing a mixture of Chlorella vulgaris and additional nutrition (yeast, cerophyll and trout chow; YCT) 48 . One-day-old neonates of D. magna were used for the exposure test. To check the sensitivity of the D. magna culture, an acute immobilization toxicity test was performed with potassium dichromate (K2Cr2O7) as a reference. The sensitivity of D. magna culture to K2Cr2O7 ranged within the limits (halfmaximal effective dose, EC 50 : 0.6-2.1 mg/L for 24 hours) suggested by OECD guideline 202. For each test concentration and control, ve neonates each (< 15 hours-old) were placed in 6-well plates (VWR Tissue Culture Plates, VWR, Darmstadt, Germany) containing 10 mL of either test solution or suspension. The neonates (< 24 hours-old) obtained from the fth generation were used in the toxicity tests to minimize variability. After incubation for 3 and 48 hours, D. magna was used to measure the heart rate and swimming speed.

Preparation of nanomaterials
A total of three commercially available materials including two metallic nanoparticles (TiO 2 NPs, AgNPs) and AgNO 3 were used for tests in this study. The metallic nanomaterials were purchased from manufacturers (TiO 2 : Aeroxide® TiO 2 P25, Evonik Industries, Germany; AgNP: NanoXactTM, NanoComposix Inc., USA) and the particle size of each nanomaterial was measured using a transmission electron microscope (TEM). Individual TiO 2 particles are normally 21 nm in size, and their aggregates are distributed hundreds of nanometers in size. Spherical AgNPs were provided at 0.02 mg/mL in a 2 mM sodium citrate solution, and the primary particle size was 31 ± 3 nm. Silver nitrate (ACS reagent, purity ≥ 99.0%; Sigma-Aldrich, St. Louis, MO, United States) was used as a source of Ag + . All the chemicals were used without further puri cation.
After each material was dispersed in the medium, the size distribution of each material was measured using a TEM (JEM-2100 LaB 6 ; JEOL, Tokyo, Japan) to identify whether the particle size was changed by medium type (distilled water vs. ISO medium) (Fig. 7). Characterization of the NP suspensions was conducted under a 200 kV accelerating voltage. An aliquot of each suspension was rinsed over a holey carbon TEM grid (type S147-4, Plano, Wetzlar, Germany) and then dried at room temperature. More than 100 particles were taken at three magni cations due to the different particle sizes. The images were analyzed for average length (diameter) using the pixel ruler via Image J software. The primary particle size of TiO 2 was 24 ± 5 nm and the AgNP was 31 ± 3 nm (Fig. 7A & 7C). However, we found that the primary particles aggregated in the medium over 48 hours (Fig. 7B & 7D). Thus, we con rmed that D. magna can be exposed to primary and aggregated particles simultaneously.

Heart rate counting
The most commonly used endpoint of toxicity is to measure the death rates of D. magna. The immobilization of D. magna immobilization was observed at each concentration before counting the heart rate. LC 50 is de ned as the concentration of toxic substances that kill 50% of the test organism within a certain exposure period. Survival data were plotted and LC 50 values were calculated using logistic 3-parameter curve tting with Sigmaplot 13.0 software (Systat Software Inc., San Jose, CA, USA).
The heart rate was measured to evaluate the in uences of exposure to each nanomaterial in triplicate.
Neonates hatched within 24 hours from the fth generation of D. manga adults were newly prepared on the day of the experiment. Every ve neonates were exposed to each concentration level for 3 and 48 hours. The control group was also prepared for comparison with exposure groups under the same conditions without chemical treatment. After exposure, an aliquot of the methyl cellulose solution (4% V/W, Lot No. SLCC9072, Sigma-Aldrich Corp., St. Louis, MO, United States) was used to x individual neonates on glass plate. We observed heart rate conditions for one minute using an optical microscope at 4X magni cation (Model CKX41, Olympus Inc., Tokyo, Japan), and recorded them with video les. The heart rate was nally counted manually in the play condition of low speed (× 0.3).

Swimming performance monitoring
As a behavioral index, the swimming performance of D. magna after exposure to each concentration level of nanomaterials was measured using a direct-reading instrument (Model Zebrabox, View Point Life Science Inc., Lyon, France). Individual neonates exposed to each concentration were transferred to each well of a 96-well plate separately. Each well plate batch was placed into the Zebrabox and all exposure groups were kept stable in entirely dark conditions for 30 minutes before measurement to minimize the effects of sudden environmental changes. After stabilization for 30 min, the swimming speed (distance per minute) and moving route were measured under tracking mode every minute for 80 minutes. The realtime data were analyzed using automated observation software (Zebralab-2, View Point Life Science Inc., Lyon, France).

Measurement of oxidative stress
The measurement of ROS in D. magna was measured using a 2, 7-dichloro uorescein diacetate (DCFDA) cellular ROS detection assay kit (ab113851, Abcam, Berlin, Germany). After nanoparticle exposure for 3 and 48 hours, 10 for each concentration of D. magna were washed in a beaker with pure water and transferred to Eppendorf tubes with 200 µL of phosphate buffered saline (PBS). D. magna was homogenized by VWR® Disposable Pellet Mixers and Cordless Motor (VWR, Darmstadt, Germany). Then, the homogenates were centrifuged at 13,000 x g for 20 minutes, and the supernatant of samples was collected. Samples were kept at -80 ºC until the assay was performed. Twenty microliters of each collected supernatant and 80 µL of assay buffer were placed on a black 96-well microplate (Thermo-Scienti c, Karlsruhe, Germany) and treated with 100 µL of 10 µM DCFDA. Then, uorescence measurements were conducted immediately after incubation for 30 minutes in the dark using Spark® Multimode Microplate Reader (Tecan Trading AG, Männedorf, Switzerland). The excitation state was maintained at a wavelength of 485 nm, and the emission state was maintained at a wavelength of 535 nm. Each concentration was measured ve times and the mean value was obtained. The total protein content of each sample was quanti ed by a BCA protein assay for the normalization of samples. The DCFDA levels in the D. magna were also visualized using a Zeiss SteREO Discovery V8 microscope with a Plan S 1.0×FWD 81 mm objective (Carl Zeiss NTS, Ltd., Jena, Germany). After that, the captured images were analyzed using ZEN imaging software (Carl Zeiss NTS Ltd., Jena, Germany).