Combined effects of aflatoxin B1 and deoxynivalenol on the expression of glutathione redox system regulatory genes in common carp.

The purpose of the present study was to evaluate the short-term effects of aflatoxin B1 (AFB1 ) and deoxynivalenol (DON) exposure on the expression of the genes encoding the glutathione redox system glutathione peroxidase 4a (gpx4a), glutathione peroxidase 4b (gpx4b), glutathione synthetase (gss) and glutathione reductase (gsr) and the oxidative stress response-related transcription factors Kelch-like ECH-associated protein 1 (keap1) and nuclear factor-erythroid 2 p45-related factor 2 (nrf2) in liver, kidney and spleen of common carp. During the 24-hr long experiment, three different doses (5 µg AFB1 and 110 µg DON; 7.5 µg AFB1 and 165 µg DON or 10 µg AFB1 and 220 µg DON/kg bw) were used. The results indicated that the co-exposure of AFB1 and DON initiated free radical formation in liver, kidney and spleen, which was suggested by the increase in Nrf2 dependent genes, namely gpx4a, gpx4b, gss and gsr. Expression of keap1 gene showed upregulation after 8 hr of mycotoxin exposure, and also upregulation of nrf2 gene was found in kidney after 8 hr of exposure, while in the liver, only slight differences were observed. The changes in the expression of the analysed genes suggest that level of reactive oxygen species reached a critical level where other signalling pathway was activated as described by the hierarchical model of oxidative stress.

the more than 20 different aflatoxins, aflatoxin B 1 (AFB 1 ) is the most toxic one and frequently occurs in nuts, cereals, dried fruits and spices (Taniwaki, Pitt, & Magan, 2018). DON, on the other side, is less toxic than AFB 1 but is the most widely distributed mycotoxin and can be usually found in cereals such as wheat, maize, barley, oat and rice (Wu et al., 2017).
A maximum of 20 µg aflatoxin B 1 kg −1 complete feed is the regulatory limit for farm animals according to the Commission Regulation 574/2011, while for DON, the maximum proposed limit is set at 5 mg/kg complete feed (2006/576/EC).
There are still only limited data about the detrimental effects of AFB 1 or DON on fish health, and the experiments showed marked differences among fish species with different endpoints. However, there is a detailed risk assessment for mycotoxin contamination in fish feeds (Pietsch, 2020).
Both AFB 1 and DON inhibit the protein synthesis at the levels of initiation, transcription and translation, and also DNA synthesis in eukaryotic cells (Holladay, Smith, & Luster, 1995;McLean & Dutton, 1995), and have negative effects on growth rates and immune response, thus both mycotoxins are immune-suppressive or immune-depressive compounds, even in carp (Pietsch, 2015;Sahoo & Mukherjee, 2001). Reduction in feed intake, poor growth rates and feed efficiency in rainbow trout (Oncorhynchus mykiss) were reported by Woodward, Young, and Lun (1983) and Hooft, Elmor, and Encarnação (2011) as effect of DON. Döll et al. (2010) also described reduction in feed intake, increase in feed conversion and reduction of growth rate in Atlantic salmon (Salmo salar) as effect of 3.7 mg DON/kg feed. Moreover, histopathological changes were observed in the hepatopancreas and intestine in trout (Hooft et al., 2011), and hepatopancreas in carp (Pietsch, Schulz, Pere Rovira, & Kloas, 2014). AFB 1 also significantly reduced growth parameters of carp fingerlings (Akter, Rahman, & Hasan, 2010) and has negative effect on growth of other fish species, such as channel catfish (Ictalurus punctatus) and Nile tilapia (Oreochromis niloticus) at concentrations ranging from 1.88 to 100 mg AFB 1 kg −1 feed (Ahn Tuan, Grizzle, Lovell, Manning, & Rottinghaus, 2002;Chavez-Sanches, Martinez, & Moreno, 1994;Encarnacao, Srikhum, Rodrigues, & Hofstetter, 2009;Jantrarotai & Lovell, 1990). Histopathological alterations, like dystrophy of liver, were also described as effect of AFB 1 in carp at high doses, 20 or 200 µg/kg feed (Svobodova & Piskac, 1980). Toxic effects on primary hepatocytes in vitro have also been identified both as effect of AFB 1 and DON (He et al., 2010). Moreover, Zhou et al. (2017) revealed synergistically enhanced toxic effects of AFB1+DON on fish cell line (BF-2) and zebrafish larvae. Šišperová et al. (2015) observed that DON causes oxidative stress in rainbow trout and it has also been reported that AFB 1 induces the formation of reactive oxygen species, which may cause oxidative stress (Shen, Shi, Shen, & Ong, 1996). Both AFB 1 -and DON-induced oxidative stress affect the enzymatic and non-enzymatic antioxidant defences and alter the transcription factor nuclear factor-erythroid 2 p45-related factor 2 (Nrf2), which is the key regulator of the oxidative stress response, and its main negative regulator Kelch-like ECH-associated protein 1 (KEAP1) gene expression (Kövesi, Pelyhe, Zándoki, Mézes, & Balogh, 2018;Pelyhe et al., 2016aPelyhe et al., , 2016b. Under basal conditions, Keap1 binds Nrf2 and recruits it into the Cul3containing E3 ubiquitin ligase complex for ubiquitin conjugation and proteasomal degradation (Motohashi & Yamamoto, 2004). However, as effect of oxidative stress, the activation of Nrf2 increases due to conformational changes in Keap1 cysteine side chains; therefore, newly synthesised Nrf2 proteins bypass Keap1 and translocate into the nucleus, bind to the antioxidant response element (ARE) and drive the expression of Nrf2 target, antioxidant system encoding, genes (Taguchi, Motohashi, & Yamamoto, 2011). Glutathione peroxidase enzymes have a central role in antioxidant defence, and GPx4 is the most important in fishes, because it gives one third of the total GPx activity (Grim, Hyndman, Kriska, Girotti, & Crockett, 2011) Two gpx4 genes, gpx4a and gpx4b, were identified in carp (Hermesz & Ferencz, 2009). GPx4 catalyses the reduction in reactive oxygen species, mainly phospholipid hydroperoxides, using reduced glutathione as co-substrate. Oxidised form of glutathione, glutathione disulphide, can be reduced by the enzyme glutathione reductase, but glutathione homeostasis also depends on its synthesis, catalysed by glutathione synthetase (Halliwell & Gutteridge, 1989).
The purpose of the present study was to evaluate the toxic effects of AFB 1 and DON dual exposure on the oxidative stress response in liver, kidney and spleen of common carp juveniles.
Therefore, a short-term (24 hr) in vivo toxicological experiment was carried out where the expression of several genes responsible for the regulation of the glutathione redox system was analysed.

| Production of mycotoxins and analyses
For experimental contamination of the feed, AFB 1 was produced in ground corn which was artificially infected with an aflatoxin pro- was produced by Fusarium graminearum (NRRL 5883) strain on corn substrate according to Fodor et al. (2006). DON and 15-acetyl DON content of the feed were determined by HPLC method after immune affinity purification according to Pussemier et al. (2006).

| Animals, experimental design, sample preparations
A total of 96 one-year-old common carp (Cyprinus carpio) juveniles (body weight: 22.68 ± 6.22 g) was used for the experiment. Animals were randomly divided into four treatment groups (control, MIX1, MIX2 and MIX3) into four aquaria (150 L each) after a seven day long acclimatisation period. The aquaria were used in a semi-static system with dechlorinated, continuously aerated tap water. A 12 hr light:12 hr dark light regimen was used. The effect of the mycotoxin mixture (AFB 1 and DON) was investigated in three different dosage groups: control; low mix (MIX1: 5 µg AFB 1 and 110 µg DON/ kg b.w); medium mix (MIX2: 7.5 µg AFB 1 and 165 µg DON/kg b.w) and high mix (MIX3: 10 µg AFB 1 and 220 µg DON/kg b.w). The dose range was selected based on our previous short-term studies with individual aflatoxin B 1 (Kövesi et al., 2018) or DON (Pelyhe et al., 2016b) exposure in carp. An appropriate amount of mycotoxin-containing fungal culture was mixed with ground growth feed for carp (GARANT Aqua Classic™, Garant-Tiernährung) and diluted to 5 ml of water immediately before use. The nutrient content of the diet (on dry matter basis) was 30% crude protein, 7% crude fat, 5% crude fibre, 7.5% crude ash and 50.5% nitrogen-free extract.
At the start of the experiment, two animals were taken out from each experimental group and six of them served as absolute control (0 hr). The AFB 1 and DON contaminated diet were applied by gavage directly into the gut once. The transit time of feed particles was also investigated; six fish from the control group received methyl orange dyed (1% w/w) control feed by gavage with an amount of 1% body weight. Water temperature during the experiment was 19 ± 1°C which in case of the common carp means a moderate metabolic rate.
At samplings, namely at 0, 8, 16 and 24 hr after exposure, six carps from each experimental group were investigated. Fish were over-anaesthetised with clove oil and decapitated, and then, liver, kidney and spleen samples were removed and taken into 1.5 ml collection tubes and immediately frozen in liquid nitrogen and stored at −80°C until analysis to prevent RNA degradation.

| RNA Isolation and quantitative Real-Time PCR
Nucleozol reagent (Macherey-Nagel) was used for total RNA isolation from 5-10 mg liver, kidney and spleen homogenates as described by the manufacturer's instructions, and DNase I treatment was also performed according to the supplier's protocol (Thermo Fisher Scientific) to avoid genomic DNA contamination. The integrity and quality of the RNA samples were verified by agarose gel electrophoresis and NanoPhotometer (Implen GmbH) measurements. Only those RNA samples were accepted for further investigation which had the ratios of absorption 260:280 nm higher than 2.0. Then, pools were formed from equal (1 µg) amounts of DNase treated RNA per 6 individual carp samples for each sampling point per treatment. cDNA production was performed using a standard protocol with RevertAid Reverse Transcriptase and random nanomer primer from 1 μg of total RNA pool.
The real-time PCR measurements were performed in five technical replicates. According to the results of previous experiments, no measurable differences were found if the determination was made from pooled and not individual samples.
The primers for the quantification of the mRNA transcriptional levels of the gpx4a, gpx4b, nrf2, keap1, gss and gsr and endogenous control, β-actin gene were chosen based on the literature (Hermesz & Ferencz, 2009;Jiang et al., 2015;Safari, Hoseinifar, Nejadmoghadam, & Jafar, 2016) and are shown in Table 1. β-actin has no known interaction with oxidative stress or mycotoxins and served as an endogenous control gene in other studies screening the effects of mycotoxins in fish species (El-Barbary, 2016).
The applied real-time PCR procedure was described previously by Pelyhe et al. (2016b) and Kövesi et al. (2018). Shortly, Maxima SYBR Green method was used, where no template controls were also performed for each primer pair. SYBR Green signal was detected at the end of the extension period, and the amplified products were verified by melting curve analysis and gel electrophoresis.

| Statistical methods
All data are presented as mean ± standard deviation (SD). Firstly, the data were tested by Shapiro-Wilk normality test, and to confirm homogeneity of variance, both Bartlett and Browne-Forsythe tests were performed. All data passing both conditions were analysed using one-way ANOVA. Significance of differences between groups was evaluated using post hoc Tukey's test (p < .05). In case of those data that did not pass Bartlett and Browne-Forsythe test, a non-parametric Kruskal-Wallis test with pairwise comparisons was used (p < .05). Analyses were performed with GraphPad Prism 7.0 (GraphPad Software).

| RE SULTS
Clinical signs of toxicity and mortality were not observed during the 24 hr long trial in the experimental groups. According to our previous studies (Kövesi et al., 2018), we confirmed that the transit time of feed particles that were coloured with methyl orange dye was 16 hr at the applied water temperature.
The relative expression of keap1 gene was significantly lower in the liver at 8th hour in the lowest (MIX1) but significantly higher in the highest (MIX3) dose group (Figure 1a), and significantly higher values were observed also in kidney in the moderate and high dose group (MIX2 and MIX3) than the control (Figure 1b), while no significant differences were found in spleen (Figure 1c). After 16 hr of exposure, the keap1 gene expression was significantly higher in the liver in the lowest (MIX1) and moderate (MIX2), but not in the highest dose groups than the control (Figure 1a), while in kidney, no significant differences were found (Figure 1b). In the spleen, significantly higher values were observed in as effect of the highest dose group (MIX3) at the same sampling (Figure 1c). After 24 hr of exposure, significantly higher keap1 gene expression values were found in the liver as effect of all applied dose groups (MIX1, MIX2 and MIX3) than the control (Figure 1a), while in the kidney and spleen, there were no significant differences (Figure 1b,c).
There were no significant changes in the nrf2 gene expression in liver at 8th hour, but at 16th hour, significantly lower values were observed in the lowest dose group (MIX1) than the control (Figure 2a).
In the kidney, upregulation of nrf2 gene was found as effect of the In case of gpx4a expression, opposite tendencies were found after 8 hr of exposure between the liver and kidney. It was downregulated in the liver (Figure 3a) but upregulated in the kidney (Figure 3b) in all treatment groups, and also in spleen in the lowest (MIX1) dose group, (Figure 3c) as compared to the control. After 16 hr of exposure, significantly higher values were observed in all treatment groups than the control in both liver and kidney (Figure 3b,c), and as effect of moderate dose group (MIX2) in spleen (Figure 3c). Later, at 24th hour, opposite tendencies were found between the liver and TA B L E 1 Primers of target (gpx4a, gpx4b, nrf2, keap1, gss and gsr) and endogenous control (β-actin) genes

| D ISCUSS I ON
The present study revealed that AFB 1 -DON co-exposure probably induced oxidative stress in the liver, kidney and spleen of one-yearold common carp juveniles.
In this experiment, it was confirmed that the transit time of feed particles in the gastrointestinal tract was 16 hr in the control group that was fed with methyl orange dyed feed. As fishes are poikilothermic animals, their metabolic rate is defined by the actual water temperature. As there was no further feed intake, this transit time defines the time for the absorption of mycotoxins from the intestine; therefore, it can be used as an exposure time window. Hence, the changes in the values of the target genes' expression induced by the co-exposure of AFB 1 and DON presumably correlate with that transit time.
The expression of keap1, the negative regulator of Nrf2, was upregulated both in the liver and kidney, but it was slightly affected in spleen; however, in case of kidney, this upregulation was observed only after 8 hr of exposure as the relative gene expression returned to the control levels later. This probably means, that Nrf2, the transcription activator of numerous antioxidant genes, including glutathione peroxidases, glutathione reductase and glutathione synthetase, might have not been dissociated from the binding of Keap1, probably because the level of reactive oxygen species might have reached a high level, where other, for instance NF-κB signalling pathway (Lingappan, 2018), is activated as described by the hierarchical model of oxidative stress (Gloire, Legrand-Poels, & Piette, 2006). The NF-κB signalling is different from the Keap1-Nrf2-ARE pathway, and NF-κB regulates the pro-inflammatory, but not antioxidant genes (Li et al., 2008). This hypothesis is confirmed by the expression of nrf2 which showed the same changes as keap1 in kidney; a dose-dependent induction at 8th hour which later turned to downregulation even in case of spleen. Also, expression of nrf2 in liver showed only slight differences comparing to the control group.
At lower level of ROS formation, cysteine side chains of Keap1 go under conformational changes; therefore, the level of unbound newly synthesised Nrf2 increasing, which then activates the ARE containing genes encoding glutathione metabolism enzymes (gss and gsr) and glutathione peroxidases, namely gpx4a and gpx4b (Taguchi et al., 2011). The expression of gss (glutathione synthetase), which links glycine to the dipeptide of glutamate and cysteine (Espinosa-Diez et al., 2015), showed similar results in kidney and spleen as nrf2, which also suggested that the Keap1-Nrf2-ARE pathway was not activated. In case of the liver, a delayed effect was observed, as gss gene expression was upregulated only after 24 hr of mycotoxin exposure. This suggests that higher oxidative stress was induced earlier in kidney and spleen, which activated other pathway. In case of liver, which has a higher steady-state level of antioxidants, showed delayed effect and also lower oxidative stress, which was activated by the Keap1-Nrf2-ARE pathway. The same response was observed F I G U R E 6 Effect of AFB1 and DON treatment on the relative expression of gsr gene in the liver (a), kidney (b) and spleen (c) of common carp (mean ± SD; n = 6 in a pool, equal amounts of cDNA per individual). a,b Different superscripts mean significant difference at p < .05 level. MIX1: 5 µg AFB 1 and 110 µg DON/kg b.w; MIX2: 7.5 µg AFB1 and 165 µg DON/kg b.w; MIX3: 10 µg AFB 1 and 220 µg DON/kg b.w in our previous study with the same dose of AFB 1 alone (Kövesi et al., 2018). The expression of gsr, which plays important role in the reduction of glutathione disulphide to reduced glutathione (Espinosa-Diez et al., 2015), showed nearly the same changes in kidney and spleen as gss. This means that the reduction in GSSG to GSH was not effective in the last 16 hr of the trial. This was possibly caused by the lack of continuous nutrient supply, because this reaction requires NADPH as hydrogen donor for the reduction, and NADPH synthesis requires optimal carbohydrate supply. In the case of liver, an induction was observed after 8 and 24 hr of mycotoxin exposure.
These findings are contradictory with our previous findings with the lower dose of AFB 1 alone, where the induction was found after 16 hr of exposure (Kövesi et al., 2018). These results suggested that there are some synergistic effects between AFB 1 and DON, which modify not only the level of oxidative stress, but also the expression of genes encoding the regulatory elements of oxidative stress response. Phospholipid hydroperoxide glutathione peroxidase (GPx4) plays an important role in the elimination of lipid hydroperoxides in membranes and protection of their integrity from the detrimental effects of lipid peroxidation, in particular in fishes, where it has the highest activity among GPx enzymes (Antunes, Salvador, & Pinto, 1995). Grim et al. (2011) demonstrated that the higher amount of polyunsaturated fatty acids of fish tissue and elevated temperature-dependent oxidative activity are accompanied with higher GPx4 activity. The expression of gpx4a and gpx4b genes in the liver showed a dual response; downregulation at 8th hour, while upregulation in the last 16 hr of the trial. The same changes were found in our previous short-term studies with AFB 1 (Kövesi et al., 2018) or DON alone (Pelyhe et al., 2016b). These results suggest that the applied doses had delayed effects on gene expression, which may be related to the effect of the higher absorbed amounts of mycotoxins as a function of time. In case of kidney, opposite changes were observed as the relative expression was higher as compared to the control in the first 16 hr, and later dropped to the control level. Spleen showed similar changes than liver as there was an induction at the 16th and 24th hour.
In conclusion, the results of the present study suggested that the co-exposure of AFB 1 and DON induce oxidative stress in liver, kidney and spleen in carp. The results showed that reactive oxygen species formation and consequently gene expression of regulatory and synthesis encoding genes of the antioxidant system were provoked earlier in the kidney than liver and spleen, as the expression of the glutathione redox system-related genes and even the genes of Keap1-Nrf2 pathway was upregulated even after 8 hr of exposure.
Lack of Nrf2 activation was probably due to the enhanced formation of reactive oxygen species; therefore, it reached a critical level where another signalling pathway was activated as described the hierarchical model of oxidative stress. The results also suggested a synergistic interaction between AFB 1 and DON, because the response was different than the two mycotoxins used at the similar dose range individually. Moreover, there are no publications indicating the gene expression levels of different antioxidant enzymes in different tissues (liver, kidney and spleen) of common carp intoxicated with AFB1 and DON at the same time, and to our best knowledge, this study is the first to address the expression of the above-mentioned genes in three different tissues.

ACK N OWLED G EM ENT
The present study was supported by the ÚNKP-18-3 New National Excellence Program of the Ministry of Human Capacities and EFOP-3.6.3-VEKOP-16-2017-00008 project, co-financed by the European Union and the European Social Fund to BK, KB and MM.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

A N I M A L WE LFA R E S TATE M E NT
The author confirms that the ethical policies of the journal, as noted on the journal's author guidelines page, have been adhered to, and the appropriate ethical review committee approval has been re-