IntroductionFunctional additives used in aquaculture usually are not tested for biosafety because it is believed that their properties can be determined from their chemical composition or by comparison with similar substances [1]. This formal evaluation cannot provide a guaran-teed safety, and therefore additional testing of possible toxic properties is required for feed additives.The Ames test is a standard technique for evaluat-ing genotoxicity, when genotoxic properties are de-termined based on the reverse mutation of Salmonella typhimurium. The disadvantage of this method is an early low sensitivity and a high percentage of false positive results [2]. An alternative way to assess genotoxicity is the сomet assay, which requires sophis-ticated analytical equipment [3]. In such conditions it is possible to use a micronu-cleus test based on the estimation of the frequency of nuclear anomalies. The micronucleus test can be performed in vivo, on living objects, and in vitro, on cell cultures. For example, the OECD chemical testing guidelines have methods for the micronucleus test on mammalian red blood cells [4] and in vitro on mam-malian cells [5], in which the recommended organism is a laboratory mouse. Other model systems are used for rapid toxicity and ecotoxicological studies, in par-ticular bony fish species [6].Danio rerio is a common model system for toxico-logical and ecotoxicological studies. This fish species has several advantages, in particular: small size, ease of maintenance, growth and maturation rate, estab-lished transgenetic methods, as well as the presence of orthologous genes with humans [7, 8]. Thus, the use of Danio rerio to evaluate the genotoxicity of feed addi-tives will allow the interpretation of the results on hu-mans.At the moment, there is a need for the inclusion of additional components in food and feed for animals that contribute to improving their functional properties [9]. Among the groups of such additives of the greatest interest are: antioxidants [10], pre/probiotic preparations [11] and micronutrient and vitamin complexes. In this paper, the possible genotoxicity of the following additives was considered: (i) butyric acid, which is an alternative to probiotics, has positive effects on the intestinal microflora and digestive health of fish [12]; (ii) organomineral chelated micronutrient compounds, which are a source of micronutrients with greater bioavailability than inorganic salts [13, 14]; (iii) lyco-pene, which is a promising antioxidant with proven multiple properties [15].Thus, the aim of this work is to evaluate the possi-ble genotoxic properties of three promising functional feed additives (butyric acid, organomineral chelated compounds of microelements and lycopene) by micronucleus test on Danio rerio.Materials and MethodsObject of Research. Aged 3 months Danio rerio were kept in 20 L aquariums at 20-24°C and pH 7.2-7.4 with natural light conditions (l/n 12 h), 25 in-dividuals in each aquarium, according to the standard housing protocols [16]. The fish were fed with Cop-pens vital (0.5-0.8 mm) commercial pelleted feed.Individuals without visible injuries 2.03 ± 0.17 cm in size and weighing 0.26 ± 0.03 g were selected for the experiment. Each experimental group included 25 individuals in three replicates (n = 25 ∙ 3). Identical conditions were maintained in the experimental tanks as well as in the holding tanks. Preparing experimental diets. Coppens vital 0.5-0.8 mm (Coppens, Netherlands) was used as a base feed. Dosages of the feed were determined on the basis of previous studies.Butyric acid (BA) in concentrations (mg/kg): 0.5, 1.0, and 2.0 [17, 18] was introduced into the base feed by microencapsulation with sodium alginate. For this purpose, sodium alginate was dissolved in distilled water (0.5 mg/100 ml) and sprayed onto the feed pel-lets. Then, the described dosages of BA were applied to the sprayed feed.Organomineral chelated micronutrients (Ch) (Jupi-ter LLC, Russia) with the following composition (g/l): Fe – 10; Mn – 15; Zn – 35; Se – 0.3; I – 1.1; Cu – 3. This composition was formed based on a number of publications describing the micronutrient requirements of vertebrate animals [19-21]. The chelating agent for all the elements was ethyl diamindianthic acid (EDDS). The chelated compounds were applied to the feed using the same method as the butyric acid at the following dosages (mg/kg): 0.5, 1.0, and 2.0.Lycopene extract (Lic), as a fat-soluble compound, was dissolved in 20 ml of corn oil. The following concentrations of lycopene (g/kg) were used in the experiment: 25, 50, 75 [22, 23]. The resulting solution was evenly incorporated into the feed.The experimental feed was dried at 40°C to the original moisture content and stored at 4°C. The com-position of the experimental feed is shown in Table 1. Table 1Composition of experimental feeds IngredientsConCh05Ch1Ch2BA0.5BA1BA2Lic25Lic50Lic75Crude protein, %42424242424242424242Cride Fat, %13131313131313131313Crude Fibre, %2.392.392.392.392.392.392.392.392.392.39Crude Ash, %7.37.37.37.37.37.37.37.37.37.3Phosphorus, %0.850.850.850.850.850.850.850.850.850.85Calcium, %1.11.11.11.11.11.11.11.11.11.1Sodium, %0.20.20.20.20.20.20.20.20.20.2Vitamin A, IU/kg1 0001 0001 0001 0001 0001 0001 0001 0001 0001 000Vitamin D3, IU/kg2 2742 2742 2742 2742 2742 2742 2742 2742 2742 274Propulgallate, mg/kg53535353535353535353Butylated hydroxyanisole, mg/kg53535353535353535353Chelate trace elements, mg/kg–0.51.02.0––––––Butyric acid, g/kg––––0.51.02.0–––Lycopene, mg/kg–––––––255075 Micronucleus test. Before blood sampling, fish were sedated in MS-222 solution (1 mg/L), after which blood was drawn from the caudal vein. The micronu-cleus test was performed on days 5, 10, and 15 of the experiment according to the approved technique. On these dates, 3 individuals (n = 3 ∙ 10) of the same size and without visible lesions were randomly selected from each study group.Blood smears were prepared according to the stan-dard technique [24]. The blood sample was air-dried for 2-3 min, then fixed in Nikiforov mixture (1 : 1, methyl alcohol: diethyl ether) for 20 min and dried. Next, azure-eosin staining with Romanowsky-Giemsa was performed. For this purpose, 9 ml of dye was add-ed to 190 ml of buffer solution (pH 7.0), after which the preparations were stained for 10 minutes. The obtained preparations were washed three times in containers with distilled water and dried for 30 minutes.The frequency of micronuclei and nuclear anoma-lies in Danio rerio erythrocytes was determined by the methods [25, 26]. At least 2 500 blood cells were counted on each prepared blood slide and nuclear abnormalities were counted. ImageJ software (National Institutes of Health, USA) with a specially written script was used for automatic cell counting. The process of automatic cell counting is shown in Fig. 1. Fig. 1. Automatic blood cell count using the ImageJ program: a – initial photograph of erythrocytes at 400x magnification; b – image conversion to 16-bit format to highlight cell boundaries; c – setting color threshold with Image > Adjust > Threshold...; d – creating a two-color mask; e – cell counting with Analyze > Analyze Particles... (parameters: size 200-infinity; rounding 0.00-2.00); f – counting the final cell counts After making, the smears were viewed under an Olympus BX53 light microscope (Olympus Corpora-tion, Japan) with an ocular attachment Carl Zeiss ERc 5s (Zeiss, Germany) and ZEN lite software (Zeiss, Germany).In addition to micronuclei (MN), the following nu-clear anomalies were considered: notched nuclei (NN), lobbed nuclei (LN), and blebed nuclei (BN). Control-ling for multiple nuclear anomalies increases the valid-ity of testing. A number of distinguishing characteris-tics were used to reliably locate MN: same morphol-ogy of the nucleus and MN, small distance from the nucleus, size of 1/16 to 1/3 of the nucleus, minimal difference in color from the nucleus [27]. Examples of nuclear anomalies are shown in Fig. 2. Fig. 2. Nuclear anomalies detected in different experiments. The arrows indicate cells with nuclear anomalies: a, e – lobbed nuclei (LN); b, d – micronuclei (MN); c – notched nuclei (NN); f – blebed nuclei (BN) Positive control. Cr (VI) was obtained by dissolv-ing potassium dichromate (chemical pure 99.7%; CAS 7778-50-9) in pure distilled water for each experi-mental group. Concentrations of 2.5 and 4 mg/L, traditionally used as a positive control in the micronu-cleus test [28, 29], were used in the experiment. The conditions of the experiment with potassium bichromate were repeated and conducted in parallel with the feed additive test.In our work, the threshold value of MN occurrence, the excess of which is interpreted as a genotoxic effect for Danio rerio, was 0.5% per 1 000 blood cells. For other species of nuclear anomalies, no generally accepted threshold of genotoxicity was established, and their evaluation was performed by direct compared with control.Statistical processing. Numerical data distribution normality was determined using the Shapiro-Wilk test. Numerical data were compared between different groups using a two-way ANOVA with Tukey's post hoc test. The level of significance was chosen as P ≤ 0.05 and results are presented as mean ± SD (standard deviation). Data were statistically processed using GraphPad Prism version 9.0 software (GraphPad, San Diego, CA, USA).ResultsPositive control (potassium bichromate). When potassium bichromate (Dic) was used as a positive control, a genotoxic effect was observed on the fifth day of the experiment (Fig. 3, a). Fig. 3. Frequency of MN and various nuclear anomalies when using potassium bichromate in different concentrations: a – percentage of MN (red line – genotoxicity threshold); b – percentage of NN; c – percentage of BN; d – percentage of LN There was also a reliable difference in nuclear anomalies of the “notched nucleus” type, on days 5, 10, and 15 of the experiment (Table 2). Table 2Numerical values of the occurrence of MN and various nuclear anomalies using potassium bichromate at different concentrations*DaysVariantsMNNNLNBN5Control0.07 ± 0.01300.025 ± 0.0180Dic2.50.432 ± 0.016**0.217 ± 0.0380.167 ± 0.0590.184 ± 0.024Dic40.519 ± 0.023***0.288 ± 0.019**0.164 ± 0.0270.203 ± 0.0510Control0.064 ± 0.0310.015 ± 0.0180.027 ± 0.140Dic2.50.35 ± 0.032**0.142 ± 0.030.051 ± 0.0280.093 ± 0.03Dic40.365 ± 0.033**0.117 ± 0.019**0.082 ± 0.0190.133 ± 0.03515Control0.032 ± 0.02300.014 ± 0.010.021 ± 0.002Dic2.50.169 ± 0.012**0.046 ± 0.0310.024 ± 0.0130.034 ± 0.015Dic40.142 ± 0.014**0.047 ± 0.006**0.021 ± 0.0090.062 ± 0.021*Values are presented as mean ± standard deviation. Significance values from controls were obtained using two-direction ANOVA test with Tukey's post hoc test: **p < 0.01; ***p < 0.05. When evaluating other nuclear anomalies, no reli-able differences were found.Butyric acid. Record of MN and other nuclear anomalies showed that on day 5 of the experiment, the incidence of MN in all groups studied was no more than 0.23% (Fig. 4). The highest value was recorded in the group with butyric acid concentration of 1 mg/L and made 0.28%. This is the maximum recorded value throughout the experiment, nevertheless, it is well below the threshold of detectable genotoxicity. Fig. 4. Frequency of MN and various nuclear anomalies when using butyric acid in different concentrations: a – percentage of MN (red line – genotoxicity threshold); b – percentage of NN; c – percentage of BN; d – percentage of LN On day 10 of the experiment, the number of MN at concentrations of 0.5 and 1 mg/l significantly de-creased to values of 0.08 and 0.06%, respectively (Table 3). Table 3Numerical values of the occurrence of MN and various nuclear anomalies when using butyric acid in different concentration*DaysVariantsMNNNLNBN5Control0.07 ± 0.01300.025 ± 0.0180BA0.50.232 ± 0.0120.033 ± 0.0050.057 ± 0.020.091 ± 0.091BA10.28 ± 0.063**0.107 ± 0.053**0.035 ± 0.0250.072 ± 0.033BA20.23 ± 0.028**0.075 ± 0.0360.055 ± 0.0420.055 ± 0.4210Control0.064 ± 0.0310.015 ± 0.0180.027 ± 0.140BA0.50.088 ± 0.0370.015 ± 0.0210.013 ± 0.0180.062 ± 0.062BA10.067 ± 0.0140.027 ± 0.020.044 ± 0.040.033 ± 0.036BA20.14 ± 0.0590.027 ± 0.03900.096 ± 0.03915Control0.032 ± 0.02300.014 ± 0.010.021 ± 0.002BA0.50.042 ± 0.0020.029 ± 0.0200.027 ± 0.019BA10.02 ± 0.0280.02 ± 0.02800.019 ± 0.028BA20.015 ± 0.0450.011 ± 0.0160.011 ± 0.0160.033 ± 0.027*Values are presented as mean ± standard deviation. Significance values from controls were obtained using two-direction ANOVA test with Tukey's post hoc test: **p < 0.05. At the same time, in the group of fish fed with bu-tyric acid added at a concentration of 2 mg/l, the number of MN was 0.14%.On the last day of the experiment, there was also a negative dynamics in the number of MN; in all stud-ied groups their values did not exceed 0.4%. The ob-tained data on the number of MN did not differ significantly from the reference values of the control, which suggests the absence of genotoxicity of the studied substance in the given concentrations.Over the entire period of the experiment, the num-ber of other nuclear anomalies did not exceed the con-trol values (Fig. 4, b, c, d)). Just like the number of MN, the percentage of nuclear anomalies decreased with the course of the experiment. Thus, on the last day of the experiments, their number did not exceed 0.02%.Organomineral chelate compounds of trace ele-ments. Record of MN and nuclear anomalies showed the following results (Fig. 5, Table 4). Fig. 5. Frequency of MNlei and various nuclear anomalies when using an organomineral chelated trace element compound at different concentrations:a – percentage of MN (red line – genotoxicity threshold); b – percentage of NN; c – percentage of BN; d – percentage of LNTable 4Numerical values of the occurrence of MN and various nuclear anomalies when using organomineral chelate compounds of trace elements in different concentrations*DaysVariantsMNNNLNBN5Control0.07 ± 0.01300.025 ± 0.0180CH0.50.133 ± 0.0230.013 ± 0.0180.045 ± 0.00040.017 ± 0.017CH10.177 ± 0.0470.034 ± 0.0240.043 ± 0.0330.066 ± 0.059CH20.195 ± 0.070.062 ± 0.0280.02 ± 0.0280.082 ± 0.0410Control0.064 ± 0.0310.015 ± 0.0180.027 ± 0.140CH0.50.127 ± 0.0640.017 ± 0.02400.027 ± 0.038CH10.111 ± 0.0430.017 ± 0.0250.031 ± 0.0220.017 ± 0.025CH20.091 ± 0.0360.047 ± 0.006**00.018 ± 0.02515Control0.032 ± 0.02300.014 ± 0.010.021 ± 0.002CH0.50.096 ± 0.017**0.02 ± 0.02800.009 ± 0.013CH10.061 ± 0.03700.034 ± 0.0280.023 ± 0.016CH20.048 ± 0.02900.007 ± 0.010.009 ± 0.014*Values are presented as mean ± standard deviation. Significance values from controls were obtained using two-direction ANOVA test with Tukey's post hoc test; **p < 0.05. The maximum number of MN was observed with chelated trace element compounds on day 5 of the experiment, at a concentration of 2 mg/kg and reached 0.19% (1 000 cells). At a concentration of 1 mg/kg and 0.5 mg/kg, the number of MN was 0.17% and 0.13%, respectively. Significant differences were observed in the number of MN at day 15 at a concentration of 0.5 mg/kg; attached nuclei at day 10 at a concentration of 2.0 mg/kg (p < 0.05). These values are lower than the genotoxicity threshold determined as 0.5% (or 5 MN per 1,000 cells).All types of nuclear anomalies occurred in all ex-perimental groups. Their frequency did not exceed 0.06%, which is also a sign of the absence of genotox-icity. Numerical values of anomalies and MN are shown in Fig. 5 (b, c, d).On day 10 of the experiment, the number of MN decreased significantly compared to the beginning of the experiment. Thus, in all groups under study, the number of MN did not rise above 0.12%. The distribu-tion of nuclear anomalies had a random character and was not observed in all replications, remaining at a low level, not significantly different from the control.On day 15 of the experiment, there was a further negative dynamic in the number of MN. At the con-centration of 0.5 mg/kg the number of MN was 0.09%, which was the highest value among all concentrations for this period of time. The dynamics of the number of MN and nuclear anomalies in the CH study is shown in Fig. 5 and Table 4.The studies suggest that the investigated concentrations of organomineral chelate compounds of trace elements are not genotoxic. Lycopene. The study of the genotoxic properties of lycopene showed that this substance in the investi-gated concentrations presumably possesses I anti-genotoxic effect. If chelated compounds and fatty acids on day 5 of the experiment showed an increased level of MN compared to the control, when lycopene was used, its level did not exceed 0.14%. On days 10 and 15 of the experiment, the number of MN in Danio rerio blood erythrocytes decreased to values not different from those of the control in all concentrations studied (Fig. 6 and Table 5). Fig. 6. Frequency of MN and various nuclear anomalies when using lycopene at different concentrations: a – percentage of MN (red line – genotoxicity threshold); b – percentage of NN; c – percentage of BN; d – percentage of LNTable 5Numerical values of the occurrence of MN and various nuclear anomalies using potassium bichromate at different concentrations*DaysVariantsMNNNLNBN25Control0.07 ± 0.01300.025 ± 0.0180LIC250.122 ± 0.008**00.055 ± 0.0390.041 ± 0.012LIC500.078 ± 0.03100**0.026 ± 0.018LIC750.133 ± 0.0350.023 ± 0.0180.014 ± 0.020.05 ± 0.05450Control0.064 ± 0.0310.015 ± 0.0180.027 ± 0.140LIC250.053 ± 0.008000.039 ± 0.013LIC500.048 ± 0.0390.032 ± 0.0450.015 ± 0.0220.015 ± 0.022LIC750.021 ± 0.0150.024 ± 0.0180.01 ± 0.014075Control0.032 ± 0.02300.014 ± 0.010.021 ± 0.002LIC250.057 ± 0.0450.013 ± 0.0180.001 ± 0.0010.001 ± 0.001LIC500.0270 ± 0.0190.026 ± 0.0180.057 ± 0.0390.013 ± 0.019LIC750.055 ± 0.0390.038 ± 0.0290.038 ± 0.0290*Values are presented as mean ± standard deviation. Significance values from controls were obtained using two-direction ANOVA test with Tukey's post hoc test: **p < 0.05. The measured number of nuclear anomalies was also lower than the values obtained with the other feed additives studied. Despite this, nuclear anomalies were present throughout the experiment and their number and distribution by type was random. Anomalies of the BN type were noted most frequently on blood preparations, the number of which on average reached 0.057% at the concentration of 50 mg/kg on day 5 of the experiment, which is also a very low value.Discussion The studies showed the absence of genotoxic effect of the applied feed additives on Danio rerio blood erythrocytes in the whole concentration range. As a result of long-term use of the supplements under consideration, a slight increase in the number of MN and nuclei anomalies on day 5 of the experiment and their further decrease on days 10 and 15 were re-corded. The mechanism of MN reduction is probably based on the following reactions: natural cycles of erythropoiesis in Danio rerio; adaptation to the new composition of the feed mixture, which can take up to 10 days.In the course of work to establish the genotoxicity of butyric acid there was shown the absence of a genotoxic effect on all the studied indicators, including MN and nuclear anomalies. The negative dynamics of the number of MN, also recorded in the study of or-ganomineral chelate compounds of trace elements, is probably a consequence of the adaptation of the organism to the new composition of feed. It is worth noting that butyric acid is an energy substrate and can be included in energy metabolism without additional catabolic reactions. Earlier studies [30] on the inclusion of butyric acid in the diet of animals, including fish, showed a significant change in blood biochemical pa-rameters, as well as increased glycogen deposition in the liver. This suggests that butyric acid can cause some acceleration of erythropoiesis processes by direct incorporation into metabolic pathways. However, this hypothesis requires additional verification by molecu-lar biology methods. Comparing the results obtained on the number of MN when adding butter acid and organomineral chelated micronutrient compounds to the feed, we can conclude that butter acid has a greater effect on the formation of nuclear abnormalities. The use of these feed additives, based on the results of the studies ob-tained, can help to reduce oxidative stress, which may have a positive effect on the incidence of genetic ab-normalities.Since low concentrations of chelate compounds were used in the study, the absence of their genotoxic effect during the entire experiment was shown. The maximum values of the number of MN and nuclear abnormalities recorded on day 5 of the experiment are probably the result of adaptation of fish to the condi-tions of maintenance and food ration. It is also likely that as a result of micronutrient deficiency, more red blood cells are released into the bloodstream because the chelate supplement includes iron compounds in a convenient form for assimilation. After adaptation and habituation to the feed, on the 10th day of the experiment, this effect decreases, and the negative dynamics of the number of MN continues until the end of the experiment. It should be noted that the number of MN as well as other nuclear anomalies did not ap-proach the values recorded in individuals when genotoxic compounds were used. The antigenotoxic effect demonstrated by lycopene in all concentrations studied is probably related to its antioxidant properties. Since lycopene has the property of accumulating in tissues and being released into the blood [15], the negative dynamics of MN throughout the chronic experiment is evident. The presence of other abnormalities of the erythrocyte nucleus shape may be related to other DNA repair mechanisms unrelated to oxidative stress. For these reasons, the decrease in the number of NA may not occur as intensively in comparison with MN. In general, the results of this experiment demon-strate significant prospects for the use of lycopene in aquaculture feeds, since it not only does not show genotoxic properties, but, on the contrary, has an anti-genotoxic effect.ConclusionsTesting the genotoxic properties of feed additives is an important step in biosafety evaluation, because without such measures their use carries significant risks. Danio rerio is a suitable model for the evalua-tion of genotoxic effects. Butyric acid has no genotox-icity in the entire range of concentrations studied, even under the conditions of a long experiment (15 days). The maximum recorded value of micronuclei (0.28%) was well below the genotoxicity thresholds. Or-ganomineral chelated metal compounds based on EDDS have no genotoxic effect at concentrations from 0.5 to 2 mg/kg. Lycopene showed pronounced antigenotoxic properties, which makes it the most promising component of aquaculture. Its properties are probably based on its high antioxidant activity preventing the formation of reactive oxygen species.