Etazene induces developmental toxicity in vivo Danio rerio and in silico studies of new synthetic opioid derivative

Synthetic opioids are gaining more and more popularity among recreational users as well as regular abusers. One of such novel psychoactive substance, is etazene, which is the most popular opioid drug in the darknet market nowadays. Due to limited information available concerning its activity, we aimed to characterize its developmental toxicity, including cardiotoxicity with the use of in vivo Danio rerio and in silico tools. Moreover, we aimed, for the first time, to characterize the metabolite of etazene, which could become a potential marker of its use for future forensic analysis. The results of our study proved severe dose-dependent developmental toxicity of etazene (applied concentrations 10–300 µM), including an increase in mortality, developmental malformations, and serious cardiotoxic effects, as compared with well-known and used opioid—morphine (applied concentrations 1–50 mM). In silico findings indicate the high toxic potential of etazene which may lead to drug-drug interactions and accumulation of substances. Furthermore, phase I metabolite of etazene resulting from N-dealkylation reaction was identified, and therefore it should be considered as a target for toxicological screening. Nonetheless, the exact mechanism of observed effects in response to etazene should be further examined.


Results
In vivo developmental toxicity of ETZ and MORPH in FET. Figure 2 shows the effect of ETZ and MORPH in the FET test. Exposure to both compounds induced concentration-and time-dependent mortality effect. The LC 50 at 96 hpf was determined to be 182.8 μM (95% confidence interval: 178.5-187.4 μM, R = 0.9966) and 25.21 mM (95% confidence interval: 23.89-26.53, R = 0.9927) for ETZ and MORPH, respectively. LC 50 of ETZ is approximately 135 × fold higher than MORPH in the D. rerio FET test. Survival at negative control was higher than 95% at the end of the test. Table 1 represents the D. rerio hatching rate during ETZ exposure at 72 hpf (one-way ANOVA F(3,11) = 95.86, p < 0.0001) and 96 hpf (one-way ANOVA F(3,11) = 81.94, p < 0.0001). Post hoc Dunnett test showed that, following exposure to 25 μM of ETZ, the hatching rate decreased to 36% (p < 0.0001) and reached 0% (p < 0.0001) in 50 μM treated D. rerio at 72 hpf. At 96 hpf, hatching rate decreased from 47 (p < 0.001) to 3% (p < 0.0001) with increasing ETZ exposure concentration. In the negative control group, 100% of D. rerio embryos were hatched within 72 hpf.
Acute exposure of ETZ induces arrhythmia. Initial studies of acute time-and concentration-dependent ETZ administration revealed cardiac arrhythmias induced by concentrations of 200 and 300 µM (Fig. S1). The main study revealed that after 40 min of incubation of ETZ at various concentrations, a dose-dependent decrease in the number of atrial and ventricular contractions was observed (two-way ANOVA: concentration [F(5,202) = 292.5, p < 0.0001], contraction atrium/ventricle [F(1,202) = 19.82, p < 0.0001], and interactions effect [F(5,202) = 4.337, p = 0.0009) (Fig. 6a). The post hoc Tukey's test revealed that ETZ 50 µM significantly decreased contraction of atrium (p = 0.0033), ventricle (p = 0.0048); ETZ 100-300 µM significantly decreased contraction of atrium and ventricle (p < 0.0001) as compared with negative control group, while number of contractions of ventricle relative to atrium were significantly lower after exposure of ETZ 200 µM (p = 0.0005) and ETZ 300 µM (p = 0.0007). The beat-to-beat interval analysis of the atrium and ventricle were in a 1:1 ratio up The same post hoc test indicated that time interval of atrium was significantly decreased compared to negative control after ETZ 300 µM treatment (p = 0.011), whereas time interval of ventricle was lower compared to negative control after ETZ 200 and 300 µM treatment (p < 0.0001). The majority were disorders characterized by   www.nature.com/scientificreports/ prolongation of ventricular systole phase relative to the atrium (Fig. 6d). Sample videos of the described rhythm disturbances shown in Fig. 5c, d are provided in supplementary materials (Video S1,S2). ADME profiling. All predicted properties absorption, distribution, metabolism, excretion (ADME) are presented in Table 3. Lipophilicity is represented by the logP, which helps predict in vivo permeability through biological membranes. ETZ has been shown to have almost 3 times greater lipophilicity value than MORPH. The further analysis predicts that both compounds show high gastrointestinal absorption and the possibility to penetrate BBB. Only MORPH is P-gp substrate. Metabolite profiling revealed that MORPH is an inhibitor against CYP2D6, and ETZ for CYP: 1A2, 2C19, 2D6, 3A4. Druglikeness analysis did not show any violation and meet the range criteria described in used models, both compounds have a bioavailability score of 0.55 and are likely per os active in humans.
Metabolite identification. ChemSketch 2018 2.5 (ACD/Labs, Toronto, Canada) was used to draw structures of hypothetical metabolites and to calculate their exact masses. Mass spectral interpretations were based on general fragmentation rules and metabolites were identified based on their precursor mass (PM), the calculated molecular formulae, and the fragmentation patterns compared to those of the parent compound and already known metabolites of ETZ homologs. Figure  Discussion. In the era of rapid development of psychoactive substances, it is crucial to understand the risks posed by NPS. The present study has been designed to evaluate for the first time toxicity of a newly introduced non-fentanyl synthetic opioid derivative-ETZ. A previous study revealed that, ETZ was assessed to be 70 times more potent than morphine in an animal model of analgesia 4 . The toxicological data for ETZ obtained by clinicians are not available, but β-arrestin/mini-Gi recruitment assay revealed high potency of ETZ (MOR-βarr2, EC 50 = 54.9 nM; MOR-mini-Gi, EC 50 = 164 nM) against MORPH (MOR-βarr2, EC 50 = 338 nM; MOR-mini-Gi, EC 50 = 385 nM) 13 . Current knowledge regarding its activity is available on the internet forums, where users describe the feeling of euphoria. Concerning prescribed opioids intake in humans, analgesic, euphoric effects, reduction of the body temperature, slow heart rate, a decrease of blood pressure, as well as respiration depression was observed 14 . Thus, it was necessary to evaluate the toxicological and pharmacological profile of ETZ compared to the known opioid MORPH. Our study revealed strong developmental toxicity of ETZ at the micromolar concentration (LC 50 = 182.8 µM) in D. rerio embryos, in particular, morphological abnormalities, cardiotoxicity and delayed hatching. ETZ treatment at a concentration of 10 µM did not produce noticeable signs of sublethal endpoints during the FET test, whereas the concentration of 25 µM highly reduced the hatching rate at 72 hpf larvae. Hatching is a critical period of D. rerio embryo development and occurs at approximately 72 hpf. The hatching process is described as the combined effect of the activity of hatching enzyme released by the embryo's hatching gland cells to weaken the chorion and then the mechanical tearing of egg envelopes. It occurs when the embryo reaches a size, at which its energy requirements for oxygen exceed the diffusion capabilities of this gas through the egg envelopes and perivitelline fluid 15 . Factors such as the inhibition of hatching enzymes, strong mechanical properties of chorions, environmental hypoxia, and weakened spontaneous movement of embryos may generally lead to delayed or failed embryo hatching 16 . Embryo development is delayed, and thus results in disturbed hatching, and/or chemical exposure induced deformations, which could impede movement 17 . Regardless of the reason, a failure in hatching usually leads to a lethal outcome.
Furthermore, our studies have demonstrated that ETZ induced a series of cardiac-related deformities in the larva, suggesting an influence on cardiac development and function. A statistically significant decrease of heartbeat and pericardial edema at concentrations of 25 µM and 50 µM, respectively observed at 96 hpf was noted. Additionally, AO staining revealed death cells of myocardial tissue after ETZ 50 µM exposure. As the highest dose (200 µM) caused 87% decrease in a heartbeat compared to the control, we may suggest that the Table 3. Results of in silico ADME profiling of ETZ and MORPH. cLogP: logarithm of compound partition coefficient between n-octanol and water, GIA: gastrointestinal absorption, P-gp: P-glycoprotein substrate. ( +) positive, ( −) negative. www.nature.com/scientificreports/ recorded bradycardia indicates a strong cardiotoxic potential of ETZ. As the heart is the first functional organ that is formed during D. rerio development, ETZ-cardiotoxicity can be attributed to disruptions in one or more developmental stages. However, we may also consider it as a part of general toxicity, as it is well known that one of the side effects of opioids is induction of the cardiovascular ailments. In both, opioid-consuming patients and abusers, prolongation of the QT interval, which can lead to torsades de pointes, is the potential of causing sudden death 18 . Although D. rerio presents a prototypic vertebrate heart with a single atrium and ventricle, the mechanisms of its activity appear to be similar to higher vertebrates, especially that more than 95% of drugsinducing QT prolongation in humans have the same outcomes in ZF model 5 . In the present study after 40 min. of acute incubation with ETZ, we noticed bradycardia of both chambers with regularly beating heart (1:1) at lower concentrations (50 and 100 µM), while bradycardia and 2:1 atrioventricular block occurred after exposure on the highest concentrations (200 and 300 µM). Extended exposure time or higher concentration can lead to a ventricular beat arrest or cardiac arrest. So far, many substances mainly zERG (D. rerio ether-a-go-go-related gene) channel inhibitors and calcium channel activators have been identified as inducing ventricular arrhythmias in the 2:1 and higher degree from in the ZF model 19,20 . As regards to ETZ, the cardiotoxicity mechanisms require further investigation, but we may speculate that these abnormalities may develop also in humans. Therefore, the obtained data should be taken into account by physicians treating ETZ intoxication. By comparing toxicological outcomes (Fig. S2), we noticed that MORPH is ~ 135 times less toxic than ETZ in D. rerio FET assay. Determined LC 50 = 25.21 mM for MORPH stays in line with previous studies (LC 50 = 23.39 mM) carried out in FET under static exposure 21 . Contrary to the findings of Sanchez 12 , we did not note any malformations, nor consequences on the heart in nanomolar range treated larvae. However, it is difficult to discuss obtained results in the context of the abovementioned studies as the information on used MORPH salt and purity was not provided. In our studies MORPH-treated larvae in higher concentration-20 mM exhibit heart edema and show slightly less impact a heartbeat (14%, highest concentration vs negative control) in comparison to the strong cardiotoxic effect of ETZ. No influence on the hatching process was observed after incubation in MORPH solution. The evaluation of toxicological endpoints at 96 hpf of ETZ vs MORPH (Fig. S2) exposed a different pattern of action. It should be noted that the appearance of cardiac edema is preceded by a significant hatching inhibition in ETZ incubated larvae. At the highest concentration of the drug, mortality is a consequence of emerging cardiac dysfunctions in combination with general malformations of the body and inhibition of development. The toxicological profile of MORPH observed in our studies significantly differs www.nature.com/scientificreports/ from ETZ. The hatching process does not affect developmental toxicity, while the appearance of malformations coincides with an increase in mortality. Based on the chemical structure and similarity to known opioids, we suggest that observed ETZ-induced developmental abnormalities are related to activation of opioid system. During embryogenesis, the D. rerio opioid system plays an important role, particularly in neurodevelopment, proliferation, and differentiation 22 . Studies suggest that different opioids receptors are involved in a specific process during development. It was revealed that the opioid receptor (zfMOR) mRNA was detected at higher levels than the other opioid receptors during the segmentation period (up to 24 hpf), when the CNS is beginning its subdivision and differentiation into the complex structure 23 . Microarray analyses after chronic exposure to 10 nM MORPH from 5 to 24 hpf identified alteration of several genes associated with µ opioid receptor expression and genes involved in neuronal development, CNS patterning processes, neuronal differentiation, dopaminergic neurotransmission, the serotoninergic signaling pathway, and glutamatergic neurotransmission 24 .
Since the structure of the studied substance can affect its physicochemical properties, lipophilicity, and pharmacokinetic, it is also crucial to take its impact into consideration. Thus, to better understand the observed toxicological differences between tested substances, we performed an ADME in silico analysis. As compound uptake and its toxicity are likely to be influenced by lipophilicity (logP), chemicals with higher logP, and thus more lipophilic, tended to be more often toxic to the developing embryos than chemicals with lower logP 25 . Accumulation of highly lipophilic compounds in D. rerio embryos has been previously described 26 . The predicted clogP of ETZ is significantly higher than MORPH and thus makes the drug able to easily penetrate biological membranes including chorion, and accumulate in the larva's body. The results imply that the greater toxicity of ETZ towards MORPH is associated with its lipophilic character, which is in line with the aforementioned literature. Following further, in a toxicological context, the effect on Pgp should be also considered. This protein is a member of the ATP-binding transmembrane glycoprotein family [ATP-binding cassette (ABC)], which can excrete drugs from inside the cells to outside the surrounding 27 . In accordance with our in silico profiling, according to ADME prediction, ETZ does not meet the criteria of the P-gp substrate, whereas MORPH has been already identified as P-gp substrate in cell culture system 28 , through radiolabeling and in P-gp KO mice 29 .
Taken together with other predicted properties, that is high GIA, and the possibility to cross BBB, ETZ is a potential neurotoxicant. This assumption is supported by the fact that analysis of post-mortem samples revealed a high accumulation of isotonitazene in the brain and pericardial fluids 30 . This compound is another benzimidazole-derived opioid analgesic drug related to ETZ, and was characterized as stronger than ETZ (MOR-βarr2, EC 50 = 1.63 nM; MOR-mini-Gi, EC 50 = 3.72 nM). Interestingly, its metabolite N-desethylisotonitazene presents higher potency (MOR-βarr2, EC 50 = 0.614 nM; MOR-mini-Gi, EC 50 = 1.16 nM) 13 , which may intensify or be the major reason the toxicity effect. In the context of present studies identification of ETZ metabolites seems to be crucial. Therefore, since the D. rerio model is successfully used to identify metabolites, our investigation based on this model revealed the creation of one metabolite by N-dealkylation process. For isotonitazene, four metabolites in human blood and urine specimens were found, and authors suggest that the N-dealkylation and/or O-dealkylation biotransformation may occur in the other benzimidazole-based opioid analogs 31 . Thus, our results lead to a similar conclusion N-desalkyl species are appropriate metabolite biomarkers for routine screening of biological specimens. Thus, confirmation of the metabolite presence and its structure elucidation is important for further research as the findings presented by us, and discoveries of strong pharmacological potential of 2-benzylbenzimidazole family metabolites by Vandeputte 13 , highlight a possible major contribution of metabolite or synergism with the parent compound in toxicological, and especially cardiotoxic outcomes.
Another crucial factor for assessing metabolic toxicity is the interaction with cytochrome P450s enzymes, which are produced in the liver and are involved in phase I of metabolism. In silico profiling revealed that ETZ is a possible inhibitor of 1A2, 2C19, 2D6, and 3A4 subfamilies. Inhibition of CYPs impairs the biotransformation or clearance of many substances resulting in different half-life, higher plasma levels of drugs, and possible drugdrug interactions within multidrug users. Many NPS were already identified as substrates of mentioned isozymes. Methylone is metabolized primarily by CYP2D6 with less contribution of CYP1A2 and 2C19 32 , mephedrone by CYP2D6 33 , 4F-MDMB-BINACA by CYP: 1A2, 2C19, 3A4 11 . Recently CYP3A4 and CYP2D6 were found as enzymes involved in two fentanyl derivates, 4F-Cy-BAP and Fu-BAP, pathways of metabolism 9 .
In conclusion, our studies showed severe dose-dependent developmental toxicity of ETZ-cardiotoxicity, increase in mortality, developmental malformations and a delayed hatching process are more pronounced than in well-known opioid-MORPH. Further, phase I metabolite of ETZ resulting from N-dealkylation reaction was identified, and therefore it should be considered as a target for toxicological screening. In silico findings indicate the high toxic potential of ETZ, as it has been identified as an inhibitor of several CYPs (1A2, 2C19, 2D6, 3A4) which may lead to drug-drug interactions and accumulation of substances. Nonetheless, the exact mechanism of observed effects in response to ETZ should be examined further. In the next step evaluation of metabolites' properties of as well as their interactions with parent compound are necessary. www.nature.com/scientificreports/ D. rerio culture and embryo toxicity test (FET). D. rerio of the AB strain (Experimental Medicine Centre, Medical University of Lublin, Poland) were maintained at 28.5 °C, on a 14/10 h light/dark cycle, under standard aquaculture conditions (pH of water 6.9-7.5, conductivity 500-800 µS, average daily water change cycle 10-15%). Fertilized eggs were collected via natural spawning. Embryos were reared in E3 embryo medium (pH 7. Quantification of ETZ and MORPH in post-exposed solution. Sample was diluted with methanol and 10 μL was transferred to a 150 μL vial with a cap containing 80 μL of methanol and 10 μL of internal standard (IS) (4-MMC-d 3 , 1 μg/mL). Solution was transferred to silanized glass insert and analyzed by UPLC-QQQ-MS with already developed method 34 . In vivo cell death assay. The mechanism of cellular toxicity was determined in live D. rerio using acridine orange (AO) staining, which is a nucleic acid selective metachromatic dye. Three randomly selected embryos (4 dpf) after FET experiment were incubated with 5 µg/ml of AO for 1 h at 28 °C. Then, embryos were washed in E3 buffer and mounted in methylcellulose. Apoptotic cells were captured using a fluorescence microscope and the fluorescence intensity of apoptotic cells in the heart region was quantified using ImageJ.

Materials and methods
Cardiac rhythm analysis. Five, new D. rerio embryos at 4 dpf stage of development (total n = 75; from three independent experiments) were pooled at 6 well plate per well in E3 solution and habituated for 30 min to miscroscope light and room temperature. Then solution was changed to 1 ml of ETZ at different concentrations (10, 50, 100, 200, 300 µM) and incubated up to 40 min. Concentrations and time were determined through previous experiments. The plate with the embryos was kept on the microscope during the incubation time. Every 10 min. heart region of each embryo was recorded at 30 fps, 1080p for 12 s. by a camera installed onto microscope ocular. Videos processing were done according to the already published procedure 36 . VirtualDub 1.10.4 was used to extract 10 s (300 frames) and convert it into uncompressed AVI format. Obtained files were opened by ImageJ 1.53, then the atrium and ventricle region was selected, subsequently using "plot z-axis profile" with "live" option, the time profile of pixel intensity change (as beat) was generated. These values were plotted using OriginPro 9.0, adjacent averaging (points of window 5, without weighted average) smoothing method with automatic positive peak detection was applied. The total number of positive peaks is equal to the number of atrium or ventricle beats per 10 s. Beat-to-beat interval was calculated as the difference between the next and previous peak values.
In silico ADME profiling. The ADME-related physicochemical properties of ETZ and MORPH were predicted by the SwissADME online Web server 37 . Consensus logP (logarithm of compound partition coefficient between n-octanol and water), as an average of different model predictions (iLOGP, XLOGP3, WLOGP, MLOGP, SILICOS-IT), was calculated. In addition, the BOILED-Egg model of tested compounds was predicted to show the capability of gastrointestinal (GI) absorption and permeability of the blood-brain barrier (BBB). Bioavailability score was evaluated according to certain rules, Lipinski's rule of five, Ghose, Veber, Egan, Muegge. Furthermore, possible interactions with major human cytochrome P450 (CYP) isoforms involved in drug metabolism, 1A2, 2C19, 2C9, 2D6, 3A4 were generated.

Ex vivo identification of metabolites.
In compliance with an already described procedure 8 for metabolism studies, ten D. rerio larvae at 4 dpf (total n = 40) were pooled in a 6-well plate containing 3 ml of 75 μM ETZ. Concentration was chosen based on maximum tolerated concentration (MTC) results performed according to Richter et al. 2019 8 . After 24 h of exposure at 28 °C, larvae, and medium were collected. Three independent repetitions were performed (total n = 30). After the experiment, larvae were immediately anesthetized with 15% www.nature.com/scientificreports/ tricaine and sacrificed by decapitation. To confirm the absence of interfering compounds, a blank group of D. rerio larvae without substance were prepared. Furthermore, a control medium sample containing only the drug was prepared, for the detection of compound degradations during the incubation step. For precipitation to 50 µl of homogenate, 200 µl of cold ACN with 0.1% (v/v) HCOOH was slowly added and vortexed. Subsequently, samples were sonicated for 15 min., at + 4 °C. After that tubes with solution were mounted in a rack, and the solvent was evaporated at room temperature using nitrogen stream. The content was dissolved in 30 µl of MeOH and transferred to a vial with insert mounted. 5 µl was injected onto the HPLC-HRMS/MS system as described below. Statistical analyses. The median lethal concentration (LC 50 ) was determined by nonlinear, four parameters regression analysis. Obtained data were tested with the use of Shapiro-Wilk and Brown-Forsythe test. The results were processed by the one-way ANOVA analysis with Dunnett's post-test or two-way ANOVA analysis with Tukey's test. A p-value < 0.05 was considered statistically significant. All statistical analyses were performed using GraphPad Prism 8.