Integrated Approach for Synthetic Cathinone Drug Prioritization and Risk Assessment: In Silico Approach and Sub-Chronic Studies in Daphnia magna and Tetrahymena thermophila

Synthetic cathinones (SC) are drugs of abuse that have been reported in wastewaters and rivers raising concern about potential hazards to non-target organisms. In this work, 44 SC were selected for in silico studies, and a group of five emerging SC was prioritized for further in vivo ecotoxicity studies: buphedrone (BPD), 3,4-dimethylmethcathinone (3,4-DMMC), butylone (BTL), 3-methylmethcathinone (3-MMC), and 3,4-methylenedioxypyrovalerone (MDPV). In vivo short-term exposures were performed with the protozoan Tetrahymena thermophila (28 h growth inhibition assay) and the microcrustacean Daphnia magna by checking different indicators of toxicity across life stage (8 days sublethal assay at 10.00 µg L−1). The in silico approaches predicted a higher toxic potential of MDPV and lower toxicity of BTL to the model organisms (green algae, protozoan, daphnia, and fish), regarding the selected SC for the in vivo experiments. The in vivo assays showed protozoan growth inhibition with MDPV > BPD > 3,4-DMMC, whereas no effects were observed for BTL and stimulation of growth was observed for 3-MMC. For daphnia, the responses were dependent on the substance and life stage. Briefly, all five SC interfered with the morphophysiological parameters of juveniles and/or adults. Changes in swimming behavior were observed for BPD and 3,4-DMMC, and reproductive parameters were affected by MDPV. Oxidative stress and changes in enzymatic activities were noted except for 3-MMC. Overall, the in silico data agreed with the in vivo protozoan experiments except for 3-MMC, whereas daphnia in vivo experiments showed that at sublethal concentrations, all selected SC interfered with different endpoints. This study shows the importance to assess SC ecotoxicity as it can distress aquatic species and interfere with food web ecology and ecosystem balance.


Introduction
The aquatic environment is the destination for diverse classes of contaminants, including psychoactive recreational drugs [1,2]. Nevertheless, little is known about their impacts

In Silico Study
Based on the recent reports by EMCDDA [4,7,22], Bade et al. (2022) [1] and Almeida et al. (2022) [29], 44 emergent SC were selected for the in silico approaches ( Figure 1). The results from the in silico studies obtained by the Estimation Program Interface (EPI) Suite TM program are shown in Table 1, as well as the predicted toxicity for the Tetrahymena pyriformis protozoan using the Toxicity Estimation Software Tool (TEST TM ) program. Chemical Abstracts Service (CAS) registry numbers and Simplified Molecular Input Line Entry System (SMILES) notations were required for both in silico programs (Table S1).
Both computational programs are quantitative structure-activity relationships (QSAR) models for toxicological predictions used to determine the potential adverse effects of chemical entities for environmental risk assessment. The toxicity of contaminants and their environmental fate are related to their chemical structure and intrinsic physical/chemical properties (e.g., polar surface area (PSA), water solubility (WSol), and log K OW ). These properties affect transport, permeability, bioavailability, bioconcentration, and bioaccumulation [30,31]. Information about data such as log K OW and bioconcentration factor (BCF) are required by international regulations such as the OECD guidelines [32], the United States Environmental Protection Agency (USEPA) criteria [33], and the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) criteria [34]. Indeed, log K OW is a very important parameter for predicting the distribution of a substance in environmental compartments (water, soil, air, and biota). If log K OW is lower than 3, the substance has no potential to bioconcentrate in living organisms.
In Table S2, a summary of the in silico data obtained for the 44 SC is available. For each parameter, the 44 SC are organized in increasing order of selected parameters (Table S2). The current study prioritized five SC for the in vivo ecotoxicity studies, namely 3,4-DMMC, BTL, 3-MMC, BPD, and MDPV, based on in silico approach data, consumption levels, EMCDDA reports [4,22], and recent reports in wastewaters [1].
Predicted BCF and BAF values for the 44 SC were very different according to the trophic level and the Arnot-Gobas method (no. 1 or 2). In general, considering the Arnot-Gobas method no. 1 and the three trophic levels of fish, the higher BCF and BAF predicted values were obtained for α-BHP (ranged from 793.200 to 1263.000 L Kg −1 wet wt −1 ), while the lower values were obtained for CATH (ranged between 2.257 and 2.947 L Kg −1 wet wt −1 ). On the other hand, in the Arnot-Gobas method no. 2 for the upper trophic level, BCF and BAF ranged from 15,020.000 to 568,100.000 L Kg  (Table 1).
Predicted BCF and BAF values assuming a biotransformation rate of zero ranged between 24.210 (BPD) and 973.500 L Kg −1 wet wt −1 (MDPV) and between 25.150 (BPD) and 2146.000 L Kg −1 wet wt −1 (MDPV), respectively (Table 1). For both Arnot-Gobas methods, BCF and BAF values order were the same for all trophic levels, namely MDPV > 3,4-DMMC > BTL > 3-MMC > BPD. Also, both BCF and BAF values were higher in the Arnot-Gobas method which assumes a biotransformation rate of zero. MDPV is the SC that presents higher BAF and BCF values. This result is expected since MDPV shows a log K OW greater than 3, and consequently, displays higher lipophilicity. Therefore, is expected to bioaccumulate in fish tissues.
For the protozoan T. pyriformis, NPP showed the highest predicted toxicity (1.11 mg L −1 ) whereas CATH showed the lowest predicted toxicity (148.21 mg L −1 ). Toxicity estimation order for the selected SC used for the in vivo experiments is MDPV > 3,4-DMMC > 3-MMC > BPD > BTL, ranging from 6.50 to 51.80 mg L −1 (Table 1). Similar to the other organisms, for the protozoan, MDPV showed the highest toxicity and BTL the lowest toxicity.

Short-Term Exposure Assays with T. thermophila
Results from the growth inhibition assay with the protozoan are shown in Figure 2 and Table 2.
bioaccumulate in fish tissues.
For the protozoan T. pyriformis, NPP showed the highest predicted toxicity (1.11 mg L −1 ) whereas CATH showed the lowest predicted toxicity (148.21 mg L −1 ). Toxicity estimation order for the selected SC used for the in vivo experiments is MDPV > 3,4-DMMC > 3-MMC > BPD > BTL, ranging from 6.50 to 51.80 mg L −1 (Table 1). Similar to the other organisms, for the protozoan, MDPV showed the highest toxicity and BTL the lowest toxicity.

Short-Term Exposure Assays with T. thermophila
Results from the growth inhibition assay with the protozoan are shown in Figure 2 and Table 2.    Controls showed an OD decreases greater than 60% and the reference test with K 2 Cr 2 O 7 showed the reliability of the assay (Table S3). Determination of EC 50 or EC 20 was not possible as no relation between response and concentration was observed. Consequently, it was not feasible to determine the dose-response curves for T. thermophila. However, a growth inhibition effect was observed for MDPV, BPD, and 3,4-DMMC. No changes in growth inhibition were observed in organisms exposed to BTL whereas an increase in growth was noted for organisms exposed to 3-MMC ( Figure 2, Table 2).

Short-Term and Sublethal Exposure Assay with D. magna
No significant differences were found in mortality (<10%) for any of the SC. In fact, the results obtained for the daphnia mortality agree with the in silico data since all SC showed LC 50 values greater than 401.00 µg L −1 (MDPV) for daphnia, and the concentration studied in our in vivo study was considerably lower (10.00 µg L −1 ). Effects of SC on the morphophysiological parameters of D. magna are shown in Figure 3 and summarized in Table 3.
Controls showed an OD decreases greater than 60% and the reference test with K2Cr2O7 showed the reliability of the assay (Table S3). Determination of EC50 or EC20 was not possible as no relation between response and concentration was observed. Consequently, it was not feasible to determine the dose-response curves for T. thermophila. However, a growth inhibition effect was observed for MDPV, BPD, and 3,4-DMMC. No changes in growth inhibition were observed in organisms exposed to BTL whereas an increase in growth was noted for organisms exposed to 3-MMC ( Figure 2, Table 2).

Short-Term and Sublethal Exposure Assay with D. magna
No significant differences were found in mortality (< 10%) for any of the SC. In fact, the results obtained for the daphnia mortality agree with the in silico data since all SC showed LC50 values greater than 401.00 µg L −1 (MDPV) for daphnia, and the concentration studied in our in vivo study was considerably lower (10.00 µg L −1 ). Effects of SC on the morphophysiological parameters of D. magna are shown in Figure 3 and summarized in Table 3. BTL caused a significant increase in body size in both juveniles and adults (days 3 and 8, respectively). Different responses were found for juveniles and adults exposed to 3-MMC. Indeed, a significant increase in body size was observed for the juveniles at day 3, whereas a significant reduction of body size was observed in adults at day 8 ( Figure 3, Table 3). A significant decrease in body size was also noted for MDPV in adults. No changes in body size were observed on both days for BPD and 3,4-DMMC ( Figure 3, Table 3). SC also showed to interfere with heart area and size. Indeed, at day 3, a significant increase in heart area (except for BPD) and size of juveniles were observed for all SC. However, on day 8, only BTL, BPD, and 3,4-DMMC continued to stimulate heart area and size growth, whereas MDPV and 3-MMC caused a significant decrease in heart area and size ( Figure 3, Table 3).
Regarding heart rate, at days 3 and 8, a significant increase was observed for all SC except at day 3 in the organisms exposed to 3-MMC ( Figure 3, Table 3). No significant differences were observed in swimming speed and active time for all SC. However, total distance traveled was significantly increased in the organisms exposed to BPD and 3,4-DMMC ( Figure 4, Table 4).     Regarding reproductive parameters, although a tendency to the increase in the number of eggs per daphnia was observed for all SC (except for 3-MMC), only exposure to MDPV showed a significant increase ( Figure 5, Table 5).  A significant increase in reactive oxygen species (ROS) levels was ve and 3,4-DMMC exposures, but no changes were found for other 3-MMC, MD ( Figure 6, Table 6). A significant increase in thiobarbituric-acid-reactiv (TBARS) levels was observed for MDPV and 3,4-DMMC ( Figure 6, Table 6 in catalase (CAT) and acetylcholinesterase (AChE) enzymatic activity were cept for BPD, which caused a significant stimulation of AChE activity.  A significant increase in reactive oxygen species (ROS) levels was verified for BTL and 3,4-DMMC exposures, but no changes were found for other 3-MMC, MDPV, and BPD ( Figure 6, Table 6). A significant increase in thiobarbituric-acid-reactive substances (TBARS) levels was observed for MDPV and 3,4-DMMC ( Figure 6, Table 6). No changes in catalase (CAT) and acetylcholinesterase (AChE) enzymatic activity were observed except for BPD, which caused a significant stimulation of AChE activity.

Discussion
Some studies reported the occurrence of SC in wastewaters as influents and effluents, but information about surface waters is scarce [13,14,37,38]. For instance, 4-MMC (up to 106.00 ng L −1 ), bk-MAP (up to 12.00 ng L −1 ), and MDPV (up to 6.00 ng L −1 ) were measured in influent wastewater samples from eight European countries [15]. Another study reported the presence of BTL, MDPV, BPD, and 3,4-DMMC in a low range (1.00 to 20.00 ng L −1 ) in urban wastewaters from WWTP from different European cities [12]. The occurrence of BPD and 3,4-DMMC in Portuguese surface waters and effluent samples was reported though below their limit of quantification of 125.00 and 250.00 ng L −1 , respectively [16]. Therefore, in addition to the significant human health risks, assessment of their toxic effects on aquatic organisms has become important to provide scientific evidence about their ecotoxicity. In this study, 44 SC were selected for in silico studies, and a group of five emerging SC was prioritized for further in vivo ecotoxicity studies.
In silico tests were applied to predict the physicochemical properties (WSol and log K OW ), the BCF, BAF and potential toxicity for diverse organisms from different trophic levels. The 3,4-DMMC, BTL, 3-MMC, BPD, and MDPV were selected for the further in vivo experiments based on in silico approaches data, consumption levels, EMCDDA reports [4,22], and recent reports in wastewaters [1].
The log K OW is important to identify potential contaminants of concern according to OECD guidelines and USEPA criteria [32,33]. Predicted log K OW values were very different among SC ranging from 1.38 for CATH to 5.47 for 4-MeO-α-POP. The order of SC with log K OW values ≥ 3.00 was 4- The 44 SC order of toxicity was similar for the three model organisms (fish, daphnid, and green algae). In general, the higher acute and chronic toxicity were obtained for 4-MeO-α-POP > NPP > 4-MeO-α-PHPP ≈ 5-PPDI, whereas the lower toxicity values were found for EPH > bk-MAP > MTP > CATH for the three organisms (green algae, daphnia, and fish). For the SC selected for the in vivo assays, the acute and chronic toxicity showed the highest toxicity of MDPV towards fish, daphnia, and green algae, whereas BTL showed the lowest toxicity. The susceptibility of organisms was different with green algae showing higher sensitivity to SC in comparison to fish and daphnia. Regarding chronic toxicity, daphnia showed higher vulnerability. Overall, T. pyriformis ICG 50 predicted values showed higher toxicity for NPP (1.11 mg L −1 ) and the lower toxicity for CATH (148.21 mg L −1 ), whereas for the prioritized five SC, the predicted toxicity values ranged between 6.50 and 51.80 mg L −1 in the following order: MDPV > 3,4-DMMC > 3-MMC > BPD > BTL.
BCF and BAF values for the 44 SC varied according to the trophic level and the Arnot-Gobas method. Considering the Arnot-Gobas method no. 1 (three trophic levels of fish), the higher BCF and BAF predicted values were obtained for α-BHP, whereas the lowest were predicted for CATH. Regarding the Arnot-Gobas method no. 2 for the upper trophic level, the highest BCF and BAF predicted values were obtained for 4-MeO-α-POP, while the lower values were found for CATH. In the five SC selected for the in vivo ecotoxicity tests, and the predicted BCF and BAF values estimated the same order of potential for bioaccumulation for three trophic levels (MDPV > 3,4-DMMC > BTL > 3-MMC > BPD). According to the REACH criteria, substances can be categorized as very bioaccumulative (>5000 L Kg −1 wet wt −1 ), bioaccumulative (5000 ≥ 2000 L Kg −1 wet wt −1 ), or not bioaccumulative (<2000 L Kg −1 wet wt −1 ) [34]. Considering these criteria, SC are not considered bioaccumulative except MDPV (2146 L Kg −1 wet wt −1 ), which is categorized as a bioaccumulative substance. These results are in concordance with predicted log K OW values and the chemical structures.
For a comprehensive study regarding the prioritization of the five SC' short-term exposure, tests were performed with protozoan and daphnia and compared with in silico SC toxicity prediction.
Protozoan growth inhibition assay was accomplished from 1.25 to 40.00 mg L −1 to estimate EC 50 values. Within this range of concentrations, no EC 50 values were possible to be determined. Higher concentrations were not tested as they are not expected to be measured in environmental aquatic ecosystems. Nevertheless, protozoan growth response depended on the SC showing different susceptibilities to the substances. The highest growth inhibition was observed for MDPV followed by BPD and 3,4-DMMC. No toxicity was found for BTL and 3-MMC caused an increase in protozoan growth. The protozoan is not included in the ECOSAR TM model organisms; however, toxicity was possible to obtain using TEST TM program. Data showed higher toxicity of MDPV and lower toxicity of BTL. These models have been pointing out that toxicity increases with the increase in the number of atoms and degree of methylation per compound and that toxicity decreases with an increase in nitrogen substitution [39]. In vivo experiments with protozoan are in accordance with the in silico data that showed higher toxicity of MDPV and lower toxicity of BTL. However, 3-MMC showed to stimulate protozoan growth, whereas in silico predictive data indicated toxicity of this substance after 48 h of exposure. Similar results, i.e., growth increase instead of growth inhibition, have been reported for other organisms as bacteria exposed to environmental contaminants. These organisms may use the contaminants as sources of carbon that causes stimulation of growth. Mennillo et al. (2018) reported a growth increase at the highest concentration of ketoprofen on the bacteria Vibrio fischeri [40]. Differences between the in silico data and in vivo experiments can also be related to the time of exposure (28 h in the in vivo studies whereas 48 h for the in silico study).
Regarding daphnia, in silico data showed toxicity at high concentrations (mg L −1 ), with MDPV showing the higher toxicity and BTL the lower toxicity. As these concentrations are not expected to occur in the environment, a sublethal concentration, 10.00 µg L −1 , was selected and different parameters as checkpoints of toxicity were evaluated. Indeed, in silico tests are based on acute immobilization and daphnia reproductive assays that can be insufficient to evaluate toxicity as other endpoints can be affected at lower concentrations and affect the survival of the organisms. Thus, morphophysiological, behavioral, reproductive, and biochemical parameters were evaluated as biomarkers of toxicity [41][42][43].
Morphophysiological parameters showed to be affected in a substance and daphnia life cycle dependent manner. BTL affected all morphophysiological parameters causing a significant increase in body size, heart area, heart size, and heart rate in both juveniles and adults. All other SC also stimulated heart rates in both juveniles (except 3-MMC) and adults. Regarding heart area and size, different responses were observed among SC. All SC affected heart size and area in juveniles, causing an increase except BPD in juveniles, but at day 8 a decrease in heart area and size was observed for both MDPV and 3-MMC. Studies of the effects of SC on the morphophysiological parameters of this microcrustacean are non-existent so far; however, interferences in morphophysiological parameters such as heart rate, thoracic limb activity, and mandible movements have been reported for other chemicals including psychoactive substances [43,44].
Behavioral parameters, such as swimming speed, distance traveled and active time, gained special attention since SC are NPS acting at the level of the central nervous system that directly affects the locomotive abilities of daphnia [41][42][43]45]. Although a tendency to the increase in swimming speed was observed for some SC (BTL, BPD, and 3,4-DMMC) no significant effects were noted. Additionally, no changes in active time were observed for any of the SC. Regarding total distance traveled, an increase was observed for BPD and 3,4-DMMC. Changes in swimming activity (distance moved and swimming speed) were observed in D. magna after 21 days of exposure to COC at 0.05 and 0.50 µg L −1 [45].
Only MDPV showed to interfere with first reproductive events causing a significant increase in the number of eggs per daphnid. Changes in reproduction events have been reported for other psychoactive drugs (namely, methamphetamine and COC) after 21 days of exposure [18,45], but the 21-day reproductive assay was not performed in our current work.
SC also exhibited distinct effects on biochemical parameters. Although BTL and 3,4-DMMC caused an increase in ROS levels, no changes were observed in CAT activity. Additionally, no changes were noted in TBARS levels for BTL, but 3,4-DMMC caused an increase in its levels corroborating the increase in ROS levels. Parolini et al. (2018) reported that benzoylecgonine induces oxidative stress on D. magna after 48 h of exposure at environmental concentrations (0.50 and 1.00 µg L −1 ) [46]. A tendency to increase in TBARS levels was also observed for all SC; however, significant changes were noticed for 3,4-DMMC and for MDPV. CAT is a relevant antioxidant enzyme, which acts in the protection of cells from ROS species, transforming hydrogen peroxide in oxygen and water. Although an increase was observed for ROS and TBARS, no changes in CAT activity were noted. Changes in CAT activity have been reported in daphnia exposed to psychoactive substances [18,47]. AChE is an enzyme that plays an important role in the normal regulation of the central nervous system, being an important biomarker [46]. Only organisms exposed to BPD showed significant alterations (increase) in AChE levels.
In vivo experiments with daphnia, at sublethal concentrations, showed that all SC can interfere with different endpoints, and therefore, it was not possible to observe a distinct SC toxicity. BTL interfered with all morphophysiological parameters in contrast to the lower in silico data toxicity potential, and thus toxicity can be underestimated. Therefore, care should be taken when using in silico data as contaminants can interfere with endpoints not considered in these programs.

In Silico Study
For in silico studies, the EPI Suite TM program (version 4.11, November 2012) [48] developed by the USEPA with KOWWIN TM , WSKOWWIN TM , ECOSAR TM , and BCFBAF TM programs were used. The EPI Suite TM program uses a single input to run diverse validated estimation programs allowing to predict log K OW , WSol, bioaccumulation, and estimate toxicity for fish (96 h and 14 days) [31,49,50], daphnia (48 h and 21 days) [31], and green algae (48 h) [51], and ChV for fish, daphnia, and green algae (Chlorophyta). ChV is defined as the geometric mean by the following equation: where the NOEC is the no-observed-effect concentration and the LOEC is the lowestobserved-effect concentration.
For estimation of these parameters, the chemical names, CAS registry numbers, and SMILES notations were introduced on in silico computational programs. These parameters as well as PSA values and IUPAC (International Union of Pure and Applied Chemistry) names were obtained from PubChem website searches (https://pubchem.ncbi.nlm.nih.gov, accessed on 21 December 2022 and on 8 March 2023). For more detailed information, please see the Supplementary Materials (Table S1). Table S1 provides the acronyms, IUPAC names, CAS numbers, chemical structures, SMILES notations, and PSA values of the 44 SC selected for in silico studies.
The KOWWIN TM program (version 1.68, September 2010) was used to estimate the log K OW (log octanol-water partition coefficient) of chemicals using an atom/fragment contribution method [33,52]. The WSKOWWIN TM program (version 1.42, September 2010) estimates the WSol of an organic compound using the log K OW previously estimated by KOWWIN™ program and then applicable correction factors if needed [53][54][55]. The ECOSAR TM (version 1.11, July 2012) estimates the aquatic toxicity of chemicals, namely acute toxicity and chronic toxicity to aquatic organisms such as fish, aquatic invertebrates, and green algae. BCFBAF™ program (version 3.01, September 2012) provides screening levels of BCF and BAF prediction by regression model based on log K OW values and includes correction factors for biotransformation and ionization. The Arnot-Gobas method [56] was considered (including biotransformation rate estimates and assuming a biotransformation rate of zero) to calculate BCF and BAF for three trophic positions of fish. As no information for protozoan is possible to obtain using EPI Suite TM program, the TEST TM program (version 5.1.2, October 2022) was used [57]. This application was developed by USEPA to estimate toxicity values for several endpoints including the 48 h assay on the protozoan T. pyriformis by accessing the 50% of the growth inhibition concentration (IGC 50 ).

Sub-Chronic Assay with T. thermophila
Protozoan growth inhibition toxicity assay was performed based on the procedures described in the Standard Operating Procedures for Toxkit tests (Protoxkit F TM , MicroBioTests Inc., Gent, Belgium) and in accordance with OECD Guideline no. 244 [28] [27]. Daphnia were maintained in groups of 30 individuals per 800 mL of MHRW, supplemented with a vitamin mixture (biotin, thiamine, and cyanocobalamin), algae extract (Ascophyllum nodosum, Extract Sol-Plex Sierra acquired from Alltech Naturally (Sintra, Portugal)) and Saccharomyces cerevisiae yeast acquired from Pura Vida (Lisbon, Portugal). Organisms were fed with a microalgae suspension of Raphidocelis subcapitata at 3.0 × 10 5 cells mL −1 day −1 (neonates/juveniles) or 6.0 × 10 5 cells mL −1 day −1 (adults). The culture media was renewed every 2 days. Microalgae were cultured in Woods Hole MBL medium, in a semicontinuous 4 L batch culture at a 16:8 h light: dark cycle (20 ± 2 • C) [58]. Culture maintenance details can be found in Supplementary Material (Section S2.1.2.1). Individuals born between the 3rd and 5th brood with less than 24 h were used to start new cultures and to perform the experimental assays.

Experimental Design
Five replicates were used for the control and for each compound experiment. Each experimental unit contained 15 neonates per 250 mL of culture medium. From the individual SC stock solutions prepared at 1.00 mg mL −1 in UPW, individual intermediate solutions were prepared at 1.00 mg L −1 (in 100 mL of UPW) and stored at 4 • C, and used to prepare the concentration of 10.00 µg L −1 with MHRW for the exposure experiments. Every 2 days, the culture medium was renewed, and organisms were fed with a microalgae suspension.
Mortality was monitored over the exposure time. On days 3 and 8, the morphophysiological parameters were determined using a Zeiss Axiostar plus optical microscope (Carl Zeiss, Jena, Germany) coupled to a PowerShot G9 digital video camera (Canon, Tokyo, Japan). Three organisms were randomly collected from each replicate, photographed and video recorded for 1 min (min), and later analyzed to determine the body size, heart size and area, and heart rate. To assess morphometric parameters free image measurement software was used and DaVinci Resolve software (version 17.2 Build 11) was used to change clip speed (to 25% frame reduction) for heart rate determination. From day 8, photographs were taken and used for the determination of the number of eggs per daphnia.
On day 5, five individuals were randomly collected from each experimental replicate and placed into a 6-well plate with~5 mL of the respective exposure medium, and video was recorded for 1 min with a digital video camcorder (Canon Legria HF R506, Japan). Each well was previously filled with 5 mL of melted 1% agarose and once solidified a circular swimming area of 27 mm was created using a plastic cylinder. These holes provide a swimming area with excellent optics and visibility for video recording. The clips were analyzed with TheRealFishTracker program to obtain swimming speed, active time, and total distance travelled [59,60].

Statistical Analysis
Protozoan growth inhibition assay data were analyzed in GraphPad (Version 8.0.1.244) using nonlinear regression with variable slope (four parameters) and least-squares fit method. To assess which concentrations caused significant growth inhibition, one-sample unidirectional t-tests (% growth inhibition > 0) were performed for each concentration and all substances (a total of 35 one-sample t-tests were performed); to overcome the multiplicity problem (testing related multiple hypotheses; [70]), p-values were adjusted by controlling the false discovery rate [71] using the p.adjust() function in R [72].
Data from sublethal assay on daphnia were obtained using Jamovi program (version 2.2.5), a free statistical software application [73]. To evaluate the five SC data effects, both general linear models (one-way ANOVA) and generalized linear models were applied [74,75]. General linear model fit by OLS was employed for survival, morphophysiological, behavioral, and biochemical parameters, while negative binomial generalized linear models (model for count data) were applied for reproductive events (number of eggs per daphnia). In both cases, the occurrence of a significant effect of SC was additionally examined with Dunnett contrasts to evaluate significant differences between treatment and control. The differences were considered statistically significant if the p < 0.05.

Conclusions
The presence of SC at a low ng L −1 range has been reported in WWTP effluents. However, the information available concerning the ecotoxicological effects on aquatic model species is still poorly explored.
For that, 44 SC were selected for the in silico studies and a group of five SC prioritized based on consumption and recent reports in wastewaters for an integrative in vivo study using two ecologically relevant organisms belonging to different trophic levels (protozoan and microcrustacean), namely BTL, BPD, MDPV, 3-MMC, and 3,4-DMMC. The in silico data revealed that MDPV is the SC with the most potential for fish bioaccumulation and also the most toxic (at acute and chronic levels) for the three organisms (fish, daphnia, and green algae). Additionally, different organism susceptibilities were obtained regarding both acute (green algae) and chronic (daphnia) evaluation. For T. pyriformis, MDPV showed the highest toxicity, while BTL showed the lowest toxicity.
Similar results were found for the in vivo experiments using the T. thermophila, namely increasing growth inhibition (MDPV > BPD > 3,4-DMMC), whereas no effects were observed for BTL. However, a stimulatory effect was observed for 3-MMC. Considering the sublethal ecotoxicity assays in the microcrustacean, different susceptibilities were noted depending on the type of SC and endpoints study.
Our work shows that SC do not affect mortality at the concentrations studied; however, they interfered at sublethal levels with several endpoints in D. magna (precisely, morphophysiology (BTL significantly increases the body size, heart size, heart area and heart rate in juveniles and adults daphnia), swimming behavior (BPD and 3,4-DMMC significantly raise the total distance traveled), reproduction (MDPV significantly increase the number of eggs per daphnia), and oxidative stress and biochemical activity (3,4-DMMC and MDPV significantly increase the TBARS levels)).
These results emphasize that, although in silico programs provide important data about the toxicity potential of diverse substances, in vivo experiments at sublethal or environmental reported levels focusing on other endpoints demonstrate that contaminant toxicity can be underestimated.
In this sense, further in vivo experimental studies should be conducted to extend the current knowledge about SC ecotoxicity effects on non-target aquatic organisms and to expand to other organisms belonging to different trophic levels, such as fish, improving ecotoxicological testing in risk assessment procedures. Additionally, the potential of mixtures should be considered as additive or synergistic effects can occur.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules28072899/s1, Table S1: Acronyms, International Union of Pure and Applied Chemistry (IUPAC) name, Chemical Abstracts Service (CAS) number, chemical structure, Simplified Molecular Input Line Entry System (SMILES) notation, and polar surface area (PSA) value of 44 SC for in silico studies; Table S2: Summary of predicted data (physical/chemical properties (i.e., WSol and log K OW ) and toxicity) for 44 SC using EPI Suite TM program (green algae, daphnia, and fish) and TEST TM program (protozoan, Tetrahymena pyriformis); Table S3: Percentage of growth inhibition vs. log concentration of Tetrahymena thermophila after 28 h of exposure to K 2 Cr 2 O 7 at five concentrations tested.