**3. Discussion**

Several studies have discussed the cytotoxic and an anti-proliferative effect of ZEN mycotoxin and its metabolites in various cell lines, such as Caco-2 [11], HepG2 cells [13], CHO-K1 cells [32], and SH-SY5Y [6], and hose of BEA mycotoxin in Caco [14], CHO-K1 [19], and Hep G2 cells [17]. However, there are no reports on the effect of ZEN metabolites and BEA in neuronal cells. In the present study, we proved the toxicity of ZEN metabolites (α-ZEL and β-ZEL) and BEA in human neuroblastoma SH-SY5Y cells in relation to exposure time, mycotoxin concentration, and mixture of mycotoxins.

According to the IC50 values of single mycotoxins, β-ZEL was the most cytotoxic mycotoxin compared to the other mycotoxins assayed individually, which is in accordance with Marin et al. (2019) [33] who studied the cytotoxicity of ZEN and its metabolites in HepG2 cells, individually and in double combinations. On the contrary, Tatay et al. (2014) [32] demonstrated that α-ZEL was the most cytotoxic among three mycotoxins tested (α-ZEL, β-ZEL, and ZEN) in CHO-K1 cells. Regarding to double combinations, it was revealed that presence of two mycotoxins increased the cytotoxic potential in SH-SY5Y cells, as shown by the lower IC50 values. According to Figure 2a, IC50 for α-ZEL and BEA was not reached in individual treatment however, binary combination α-ZEL + BEA (5:1) inhibited cell proliferation from up to 50 to 90% for all times studied. For the β-ZEL + BEA (5:1) binary combination, as it can be observed in Figure 2b, the IC50 values at 48 and 72 h were lower than that of β-ZEL. This was also observed when β-ZEL was combined with α-ZEL, for which combination (α-ZEL + β-ZEL (1:1)), the IC50 value was the same as that found for β-ZEL alone. This result was not achieved by Tatay et al. (2014) [31] in CHO-K1 cells, although the mycotoxin concentrations studied in binary assays in that work were two times higher than the concentrations assayed in our study. The proliferation of CHO-K1 cells treated with the α-ZEL + β-ZEL mixture at the highest concentration

decreased only by 20% with respect to the values found when each mycotoxin was tested alone. In addition, in that study, the IC50 value was never reached for binary mixtures, whereas in our study in SH-SY5Y cells, after 48 and 72 h, the α-ZEL + β-ZEL combination inhibited cell proliferation up to 70% and 90%, respectively (Figure 2c). For the triple combination (α-ZEL + β-ZEL + BEA, (5:5:1)), cell proliferation inhibition was lower than when β-ZEL was assayed individually, and the same result was found for β-ZEL + BEA after 48 and 72 h and for α-ZEL + β-ZEL after 48 h in SH-SY5Y cells. This is in contrast with the results obtained for the tertiary combination of α-ZEL + β-ZEL + ZEN in CHO-K1 cells, as this combination was more cytotoxic than each mycotoxin tested alone [30].

As the co-occurrence of mycotoxins in food and feed is very common, some studies evaluated the toxicity and cytotoxicity of several mycotoxins, both individually and in combination, in different cell lines, using the isobologram model. In these experiments, HepG2 cells were exposed to ochratoxin A (OTA) and BEA [16], to double and triple combinations of alternariol, 3-acetyl-deoxynivalenol, and 15-acetyl-deoxynivalenol [28], and to combinations of ZEN and OTA or α-ZEL (tested also individually) [33], CHO-K1 cells in vitro were used to examine the interactions between the mycotoxins beauvericin, deoxynivalenol (DON), and T-2 toxin [26] as well as the combination of BEA, patulin, and ZEN [17], whereas Caco-2 cells were exposed to DON, ZEN, and Aflatoxin B1 [34]. It is important to understand whether the interaction between mycotoxins shows synergism, additive effects, and/or antagonism concerning cell viability.

In SH-SY5Y cells, almost all the combinations tested reduced cell viability more than the individual mycotoxins, except the β-ZEL + BEA (5:1), α-ZEL + β-ZEL (1:1), and α-ZEL + β-ZEL + BEA (5:5:1) combinations, for which the reduction in cell viability was not significantly different from that obtained when β-ZEL was assayed individually. According to Dong et al. (2010) [5], ZEN is degraded more efficiently to α-ZEL than to β-ZEL in almost all tissues, whereas it is converted more efficiently to β-ZEL than to α-ZEL in liver and lungs. Some studies demonstrated that β-ZEL is more cytotoxic than α-ZEL [31,35,36], whereas other studies found that α-ZEL is more cytotoxic [30,35]. Hence, there is a necessity to clarify the cytotoxicity of these two mycotoxins with studies of the toxicity mechanisms involved.

The IC50 values obtained by the MTT assay and the amount of mycotoxin detected in the media by LC–ESI–qTOF-MS were determined and translated into percentage values as an attempt to calculate the amount of each mycotoxin involved in the cytotoxic effect and in the type of interaction effect. Hence, the percentage of mycotoxin present in the media was considered in accordance to the IC50 value obtained from the MTT assay (Table 1). The results showed that among the individual mycotoxins assayed, the amount of α-ZEL that remained in the culture medium was above 50% of the administered quantity at all times assayed (Figure 5a). This can be related to the effect in Figure 1a, which shows that the viability was above 100% for the doses reported in Figure 5. This can be justified by the chemical structure of this compound, which might impede its access in the cell. Our results sugges<sup>t</sup> that the availability and capacity of the tested mycotoxins to ge<sup>t</sup> into cells were greater than those of α-ZEL, and as a consequence, the amounts of these mycotoxins detected in the media were lower than that of α-ZEL. To notice that the higher the amount of mycotoxin in the medium (at 24 h), the higher the cell viability, which might be related to the lower amount of mycotoxin affecting the live cells. On the contrary, BEA seemed to have easier access the cells, as its percentage in the medium was generally below 50%, but cell viability was maintained above 50% for the doses assayed, indicating the lower potential toxicity of BEA in SH-SY5Y cells compared to ZEN metabolites. In fact, among all three mycotoxins tested, BEA reached the IC50 values after long exposures times (72 h) (Table 1 and Figure 1c), highlighting again the mild toxic effect of BEA in SHY-SY5Y cells compared to ZEN metabolites.

According to this and when analyzing combinations, the amounts of ZEN metabolites found in the medium were in most cases below BEA's amounts, indicating easier access of these compounds in SH-SY5Y compared to BEA. In detail, for the α-ZEL + BEA combination (Figure 2a), it can be observed that the lower the amount of α-ZEL in the medium over time (Figure 5d), the lower the viability of SH-SY5Y cells, in particular at 72h. For triple mixtures, the cytotoxic effect was weaker at all times and for all mixtures compared with that of binary combinations; however, the amounts of each mycotoxin detected were all below 50%, and the cytotoxic e ffect seemed to be bearable for SH-SY5Y cells for doses administered in the first and second mixture but not for those of the third mixture (6.25 + 6.26 + 1.25) μM (α-ZEL + β-ZEL + BEA, 5:5:1), specifically at 48 and 72 h. We sugges<sup>t</sup> that cytotoxicity is due to the stimulation of di fferent biochemical mechanisms that, after a certain level of stimulation, cannot be controlled and cause cell death. Therefore, it is necessary to study in detail the mechanisms of action implicated in the cytotoxic e ffects that occur when several mycotoxins are present in the same food or diet.
