Cytotoxic Effects and Oxidative Stress Produced by a Cyanobacterial Cylindrospermopsin Producer Extract versus a Cylindrospermopsin Non-Producing Extract on the Neuroblastoma SH-SY5Y Cell Line

The incidence and interest of cyanobacteria are increasing nowadays because they are able to produce some toxic secondary metabolites known as cyanotoxins. Among them, the presence of cylindrospermopsin (CYN) is especially relevant, as it seems to cause damage at different levels in the organisms: the nervous system being the one most recently reported. Usually, the effects of the cyanotoxins are studied, but not those exerted by cyanobacterial biomass. The aim of the present study was to assess the cytotoxicity and oxidative stress generation of one cyanobacterial extract of R. raciborskii non-containing CYN (CYN−), and compare its effects with those exerted by a cyanobacterial extract of C. ovalisporum containing CYN (CYN+) in the human neuroblastoma SH-SY5Y cell line. Moreover, the analytical characterization of potential cyanotoxins and their metabolites that are present in both extracts of these cultures was also carried out using Ultrahigh Performance Liquid Chromatography-Mass Spectrometry, in tandem (UHPLC-MS/MS). The results show a reduction of cell viability concentration- and time-dependently after 24 and 48 h of exposure with CYN+ being five times more toxic than CYN−. Furthermore, the reactive oxygen species (ROS) increased with time (0–24 h) and CYN concentration (0–1.11 µg/mL). However, this rise was only obtained after the highest concentrations and times of exposure to CYN−, while this extract also caused a decrease in reduced glutathione (GSH) levels, which might be an indication of the compensation of the oxidative stress response. This study is the first one performed in vitro comparing the effects of CYN+ and CYN−, which highlights the importance of studying toxic features in their natural scenario.


Introduction
Cyanobacteria are some of the most abundant microorganisms on Earth, capable of producing oxygen and converting CO 2 into biomass using sunlight [1]. These photosynthetic prokaryotes grow to form blooms that are affected by temperature, sunlight, and the oxygen content of the environment where they proliferate, and thus, affect the biota in both aquatic and terrestrial ecosystems [2][3][4]. One of their most important features is the been studied yet. In this sense, different CYN-producing and non-producing cyanobacterial species can coexist in nature, and the presence of other substances different from CYN can influence their toxicity. Thus, in order to elucidate the possible effects on the nervous system of cyanobacterial extracts, the aim of the present work was to study, for the first time, the cytotoxic effects and oxidative stress produced by a CYN-non-producing extract (R. raciborskii) (CYN−) and a CYN-producing extract (C. ovalisporum) (CYN+) on the human neuroblastoma SH-SY5Y cell line, and to stablish the correlation between the toxic effects induced by the CYN or other bioactive compounds present in the extracts. For this purpose, both extracts were previously characterized for the first time using UHPLC-MS/MS and an appropriate data processing by the available software (Compound Discoverer and Traze Finder) to determinate other toxins, metabolites, or bioactive molecules.

Extracts Characterization
Once the bioactive compounds were extracted, they were analyzed by UHPLC-MS/MS and submitted to several databases related to metabolism and natural compounds (see Section 5.4). A total of 1596 compounds were found and identified in both extracts. The most outstanding compounds and whose presence have been corroborated in cyanobacteria are those present in Figure 1. The compounds found in the extract from R. raciborskii (CYN−) were: Anatabine, Annularin G, AS-I Toxin, Bisucaberin B, Leptosphaerina, Microcysbipterin B, and Microcysbipterin C., and from C. ovalisporum (CYN+) the following were found: Anabasine, Annularin G, Aphanorphine, and AS-I Toxin. Figure 2 shows the chromatographic profile of both extracts: CYN+ and CYN−.  In addition, numerous CYN metabolites have been identified in the extract from the C. ovalisporum culture, which are presented in Table 1. These metabolites could appear because of the following reactions: dehydration, reduction, oxidation, desaturation, nitro reduction, hydration, oxidative deamination to alcohol, sulfation, transformation of thiourea to urea, and acetylation. In addition, numerous CYN metabolites have been identified in the extract from the C. ovalisporum culture, which are presented in Table 1. These metabolites could appear because of the following reactions: dehydration, reduction, oxidation, desaturation, nitro reduction, hydration, oxidative deamination to alcohol, sulfation, transformation of thiourea to urea, and acetylation.

Cytotoxicity Assays
A concentration dependent decrease in the SH-SY5Y cell viability was observed after their exposure to 1-10 µg CYN/mL of CYN+ during 24 and 48 h for all the parameters measured ( Figure 1). The EC 50 values obtained after 24 h of exposure for the different assays were 1.111 ± 0.325, 2.085 ± 0.204 and 4.423 ± 0.330 µg CYN/mL for MTS, NR and PC, respectively ( Figure 3A). Concerning the exposure for 48 h, CYN+ led to EC 50 values of 0.691 ± 0.165, 0.733 ± 0.165 and 2.350 ± 0.506 µg CYN/mL for MTS, NR and PC, respectively ( Figure 3B). At both exposure times, MTS resulted in the most sensitive biomarker, providing the lower EC 50 values.

Cytotoxicity Assays
A concentration dependent decrease in the SH-SY5Y cell viability was observed after their exposure to 1-10 µg CYN/mL of CYN+ during 24 and 48 h for all the parameters measured ( Figure 1). The EC50 values obtained after 24 h of exposure for the different assays were 1.111 ± 0.325, 2.085 ± 0.204 and 4.423 ± 0.330 µg CYN/mL for MTS, NR and PC, respectively ( Figure 3A). Concerning the exposure for 48 h, CYN+ led to EC50 values of 0.691 ± 0.165, 0.733 ± 0.165 and 2.350 ± 0.506 µg CYN/mL for MTS, NR and PC, respectively ( Figure 3B). At both exposure times, MTS resulted in the most sensitive biomarker, providing the lower EC50 values. In relation to the exposure to CYN− to the same amount of the extract than for CYN+, MTS provided EC50 values of 5.658 ± 1.180 and 5.164 ± 1.620 µg CYN+/mL equivalents after 24 and 48 h of exposure, respectively. The NR assay also led to a lower EC50 value after 48 h compared to the one after 24 h, being 9.669 ± 0.300 and >10 µg CYN+/mL equivalents, respectively. In addition, the PC assay led to EC50 values higher than 10 µg CYN+/mL equivalents at both exposure times ( Figure 4). In relation to the exposure to CYN− to the same amount of the extract than for CYN+, MTS provided EC 50 values of 5.658 ± 1.180 and 5.164 ± 1.620 µg CYN+/mL equivalents after 24 and 48 h of exposure, respectively. The NR assay also led to a lower EC 50 value after 48 h compared to the one after 24 h, being 9.669 ± 0.300 and >10 µg CYN+/mL equivalents, respectively. In addition, the PC assay led to EC 50 values higher than 10 µg CYN+/mL equivalents at both exposure times ( Figure 4).

Oxidative Stress Assays
The exposure to 0.275, 0.55 or 1.1 µg CYN/mL CYN+ (EC50/4, EC50/2 and EC50, respectively), led to significant changes in the ROS assay in SH-SY5Y cells after 8, 12 and 24 h ( Figure 5A). However, ROS changes were not detected in this cell line after 4 h of exposure for any of the concentrations tested. Furthermore, a decrease in the GSH levels after exposure to 0.55 µg CYN/mL CYN+ after 4 and 24 h of exposure was also observed ( Figure 5B).

Oxidative Stress Assays
The exposure to 0.275, 0.55 or 1.1 µg CYN/mL CYN+ (EC 50/4 , EC 50/2 and EC 50 , respectively), led to significant changes in the ROS assay in SH-SY5Y cells after 8, 12 and 24 h ( Figure 5A). However, ROS changes were not detected in this cell line after 4 h of exposure for any of the concentrations tested. Furthermore, a decrease in the GSH levels after exposure to 0.55 µg CYN/mL CYN+ after 4 and 24 h of exposure was also observed ( Figure 5B).
Concerning the non-producing extract, the same amount of CYN− caused an increase in the ROS levels after exposure to the highest concentrations only after the exposure for the longest periods (12 and 24 h) ( Figure 6A). With respect to GSH levels, there was a significant decrease after the exposure to all the concentrations assayed after 8 and 24 h of exposure ( Figure 6B). Concerning the non-producing extract, the same amount of CYN− caused an increase in the ROS levels after exposure to the highest concentrations only after the exposure for the longest periods (12 and 24 h) ( Figure 6A). With respect to GSH levels, there was a significant decrease after the exposure to all the concentrations assayed after 8 and 24 h of exposure ( Figure 6B).

Discussion
Cyanobacterial blooms have demonstrated to be able to cause toxicity to some extent in different experimental models as reviewed by Janssen [41]. Furthermore, it has also been reported that some cyanobacterial extracts containing cyanotoxins can cause different effects than the toxins themselves. In this sense, CYN is a cyanotoxin distributed

Discussion
Cyanobacterial blooms have demonstrated to be able to cause toxicity to some extent in different experimental models as reviewed by Janssen [41]. Furthermore, it has also been reported that some cyanobacterial extracts containing cyanotoxins can cause different effects than the toxins themselves. In this sense, CYN is a cyanotoxin distributed worldwide due to the number of species able to produce this secondary metabolite. However, studies considering the effects of CYN− versus CYN+ are very scarce, while no studies in this regard have been reported in neuronal cells so far.
Taking all the above into account, cytotoxicity and oxidative stress assays were performed using a R. raciborskii CYN− strain and a C. ovalisporum CYN+ strain in the human neuroblastoma SH-SY5Y cell line. Concerning CYN−, the MTS assay led to EC 50 values of 5.658 ± 1.180 and 5.164 ± 1.620 µg of CYN+/mL equivalents after 24 and 48 h of exposure, respectively, indicating cytotoxicity to some extent. This toxicity could be mainly produced by Anatabine: a major nicotinic receptor agonist [42], AS-I Toxin [43], Bisucaberin B: known cytostatic [44], and by Microcysbipterin B and C: protease inhibitors [45]. However, the same effects were observed after exposure to CYN+ at almost 5 times less concentration (EC 50 values of 1.111 ± 0.325, and 0.691 ± 0.165 after 24 and 48 h, respectively), meaning that the effects observed after exposure to CYN+ could be mainly due to CYN and its different detected variants and not to the rest of the compounds present in the cyanobacterial extract, such as Anabasine: a major nicotinic receptor agonist [46], and AS-I Toxin [43], and both were detected in this extract. In addition, other potentially bioactive or toxic compounds could be present in the extracts and have been not identified in this work due to the conditions used in the CYN extraction. Moreover, the present study is limited by the inability to quantify minority metabolites detected in both extracts, which is due to the lack of commercially available standards. This obstacle is particularly prevalent when dealing with metabolites and compounds that have not yet been extensively described in scientific literature, as is the case in this particular research. However, the identification of such compounds can only be achieved through the use of reference spectra libraries.
In regard to the CYN concentrations found in the environment, they can vary considerably based on multiple factors [4]. Up to 1050 µg/L CYN has been detected in an Australian water supply [11]. Taking into account that up to 90% of the total CYN in surface water is in the dissolved fraction, rather than intracellular [22], the concentrations used in this study (0-10 µg/mL) reflect actual exposure scenarios in the environment.
In order to compare the effects observed after exposure to CYN+, Hinojosa et al., [35] in the same experimental model, obtained an EC 50 value of 0.87 ± 0.13 and 0.32 ± 0.08 µg/mL after the same exposure times, but used pure CYN [35,47]. Thus, CYN seems to be a bit less toxic in the extract than if it were isolated, probably due to some interaction between the rest of the compounds of the extract with the toxin, highlighting the importance of considering other substances that may be present in natural conditions. The only study performed in vitro with neuronal models exposed to cyanobacterial extracts, reported no alterations in neurons of Helix pomatia that were acutely exposed to crude extracts of R. raciborskii containing CYN [48]. However, the CYN-concentration was not indicated. Concerning the effects of pure CYN in cytotoxicity in neuronal cell models, Takser et al. [49] also obtained a concentration-dependent effect after exposure to 0.042 and 4.2 µg/mL in N2a murine neuroblastoma derived cells and BV-2 microglia murine cells with N2a cells being more sensitive to CYN than BV-2 cells, which highlights the importance of the experimental model. Furthermore, the effects of pure CYN have been investigated in terms of synapsis effects on murine primary neurons [36], and pure CYN led to a decrease on cell viability at the same concentrations than those tested in the present study, together with a decrease on the number of synapsis at similar concentrations than EC 50 and EC 50/2 in our study [36].
Regarding the comparison between cells exposed to CYN+ and pure CYN in nonnervous cellular models, different results have been described in scientific literature. Contrary to the results obtained in our study, Gutierrez-Praena et al. [50] performed the comparison between CYN+ and pure CYN in the human liver cancer cell line HepG2, obtaining higher toxicity exerted by CYN+. These differences obtained in the CYN-extracts from R. raciiborski could be mainly due to the experimental model used (SH-SY5Y vs. HepG2 cell lines). In addition, even though the CYN-species is the same, the cultivation conditions (growth time, biomass concentration, etc.) and extraction process could cause variations in the composition of bioactive substances other than cyanotoxins in cyanobacterial CYNextracts. However, in the same experimental model, several authors obtained different responses. In this sense, Bain et al. [51] obtained significant changes from 1 µg/mL after exposure for 24, 48, and 72 h to pure CYN, while Neumann et al. [52] demonstrated again that the extract produced a lower response than the one obtained by the pure toxin, in agreement with our data.
In a different cell line, Bain et al. [51] performed the MTS assay in Caco-2 cells using pure CYN, detecting a significant reduction of cell viability after 24 and 48 h of exposure from 2.5 µg/mL, while after 72 h, the changes were significant from 1 µg/mL. However, when Gutierrez-Praena et al. [30] performed the same experiment, also using pure CYN, they reported a significant decrease in cell viability at lower concentrations. In this regard, according to Neumann et al. [52], CYN+ from R. raciborskii was demonstrated to be less toxic after 24 h compared to the results obtained by pure CYN, which agrees with our results by leading to significant changes from 2.5 µg CYN/mL CYN+, while after 48 h these changes started to be significant from 0.25 µg CYN/mL. Furthermore, in the present study, a significant increase in ROS levels was detected after all the concentrations of exposure assayed for 8, 12, and 24 h in the cells exposed to the CYN+ extract. However, almost no significant changes were observed in GSH levels, which could indicate an ineffective response against ROS levels. Nonetheless, CYN− extracts led to the opposite by increasing ROS levels only after the highest concentrations and times of exposure, but caused a significant reduction of GSH levels after all the concentrations of exposure at different time points. This fact could indicate that the oxidative stress produced by the extracts might not be due just to the presence of CYN but also the presence of some other compounds, such as Anatabine, AS-I Toxin, Bisucaberin B, Microcysbipterin B and C, even though the oxidizing activity of these substances has not been described yet in the literature.
In a similar way, other authors have also found increased ROS levels in HepG2 cells with CYN+ and pure CYN [53,54]. Concerning CYN−, there are no previous in vitro studies performed with these extracts in this regard, in any experimental model, to be able to compare our results.
Specifically, in terms of neurotoxicity, in vivo studies after exposure to CYN cyanobacterial extracts have been performed in aquatic animals. In this regard, Kinnear et al. and White et al., [37,38] reported histopathological changes in the encephalon and behavioral changes, respectively, after seven-days exposure to whole cell extracts and live cell cultures of R. raciborskii in Bufo marinus tadpoles. Furthermore, Guzmán-Guillén et al. [39] exposed tilapia fish to a culture of C. ovalisporum containing CYN for 14 days, which led to the inhibition of acetylcholinesterase activity, an increase in lipid peroxidation, and serious histopathological damage. In addition, these authors detected CYN in all brain samples, which agrees with the results obtained in Da Silva et al. [40], who exposed Hoplias malabaricus to both purified CYN and an extract of R. raciborskii containing CYN for 7 and 14 days by intraperitoneal injection. In this study, the authors also reported an increase on the acetylcholinesterase activity after 7 days of exposure to CYN+, and a decrease after 14 days of exposure to CYN. Furthermore, these authors also detected an increase in oxidative stress by an increase in glutathione-S-transferase activity (GST), and lipid peroxidation after both 7 and 14 days of exposure to both forms of CYN (CYN+ and pure CYN). Thus, these authors reported differences between the effects produced by CYN+ and CYN itself.
In general, these facts highlight that the importance of studying the cyanobacterial extracts, as well as their own toxicity, cannot be ignored. In addition, the presence of these bioactive compounds may play an important role that should be taken into account in order to perform a more rigorous evaluation of the toxic implications of cyanobacterial blooms in different organisms, and in this case, the possible neurotoxic properties of several compounds that may be present in the cyanobacterial extracts.

Conclusions
For the first time, the present study showed the cytotoxic effects which are caused by the exposure to cyanobacterial extracts in the neuronal SH-SY5Y cell line, resulting in more toxicity in CYN+ compared to CYN−. In addition, the characterization of both extracts (UHPLC-MS/MS) revealed the presence of different substances and products derived from CYN with neurotoxic potential in these culture extracts, such as Anatabine, AS-I Toxin, Bisucaberin B, Anabasine, Microcysbipterin B and C. Furthermore, our results demonstrated that the CYN+ extract caused an increase in ROS levels without significant alterations of the GSH levels. On the other hand, the CYN− extract induced an increment in ROS levels only at the highest concentrations and times of exposure assayed, but a significant decrease in GSH levels at all the concentrations assayed, which could be explained by the presence of other compounds able to induce oxidative stress in the neuronal cells. In general, comparing the effects produced in both culture extracts (CYN+ vs. CYN−), the greatest toxicity seems to be due to the presence of CYN when compared to the other detected compounds. Thus, these results highlight the importance of studying not only the isolated cyanotoxin but also in their natural environment, as well as no toxins producing cyanobacterial cultures to test more realistic exposure scenarios.

Model System
Human neuroblastoma SH-SY5Y cells were obtained from the American Type Culture Collection (CRL-2266). This cell line was maintained in an MEM and F-12 (1:1) medium supplemented with 10% FBS, 1% L-glutamine 200 mM, 1% sodium pyruvate, 1% non-essential amino acids, and 1% penicillin/streptomycin solution, in an atmosphere containing 5% CO 2 at 95% relative humidity at 37 • C (CO 2 incubator, NuAire ® , Madrid, Spain). Cells were grown to 80% of confluence in 75 cm 2 plastic flasks. Cells were harvested with 0.25% trypsin-EDTA (1X) twice per week. The quantification of the cells was performed in a Neubauer chamber. SH-SY5Y cells were plated at a density of 1 × 10 5 cells/mL to perform all the experiments.
The SH-SY5Y cell line has been widely used in preliminary neurotoxicity studies because it has important advantages [55]. There are also several publications that use this cellular model to measure GSH content and show it to be a suitable model for this purpose [56,57].

CYN Determination and Characterization of Cultures
CYN extraction from the lyophilized cultures (C. ovalisporum CYN+ and R. raciborskii CYN−) was performed according to the validated method published in Guzmán-Guillén et al. [59].
CYN was only detected in the CYN+ extract with a retention time of 1.31 min, obtaining 91.50 µg CYN/mL concentration in the CYN-producing extract. No presence of CYN was found in the non-CYN-producing extract (Figure 7).
Moreover, both extracts were analyzed for the presence of other cyanotoxins or metabolites using analytical techniques previously used in our laboratory [60]. For this purpose, standards of the following cyanotoxins were used: Anatoxin A, Microcystin-LR, -RR and -YR. Finally, a metabolomic study was carried out to search for molecules derived from the toxins and bioactive compounds using the Compound discoverer 3.2 software (Thermo Fisher Scientific, Madrid, Spain) and the databases: BioCyc, Food and agricultural organization, FooDB, Human Metabolome database, KEGG, mass bank, Nature chemistry, Phenol-explorer and Plant-Cyc.

Cytotoxicity Assays
SH-SY5Y cells were seeded out for basal toxicity tests in 96 well tissue culture plates, and after being incubated at 37 °C for 24 h, they were exposed to the extracts. To select the amount of extract for exposure, the producer extract was used as a reference. Thus, once CYN was quantified in this extract (CYN+), the required amounts of extract containing CYN concentrations in the range of 0-10 µg CYN/mL of this toxin were calculated. The same amount of CYN− extract was used to perform the cytotoxicity assays. After this time, the medium was replaced with the exposure solutions in the plates and then they were incubated for 24 and 48 h at 37 °C at 5% CO2. The basal cytotoxicity endpoints assayed were tetrazolium salt reduction (MTS), supravital dye neutral red cellular uptake (NR) and protein content (PC). All the assays present in this work were performed by triplicate.

Cytotoxicity Assays
SH-SY5Y cells were seeded out for basal toxicity tests in 96 well tissue culture plates, and after being incubated at 37 • C for 24 h, they were exposed to the extracts. To select the amount of extract for exposure, the producer extract was used as a reference. Thus, once CYN was quantified in this extract (CYN+), the required amounts of extract containing CYN concentrations in the range of 0-10 µg CYN/mL of this toxin were calculated. The same amount of CYN− extract was used to perform the cytotoxicity assays. After this time, the medium was replaced with the exposure solutions in the plates and then they were incubated for 24 and 48 h at 37 • C at 5% CO 2 . The basal cytotoxicity endpoints assayed were tetrazolium salt reduction (MTS), supravital dye neutral red cellular uptake (NR) and protein content (PC). All the assays present in this work were performed by triplicate.
MTS reduction was measured according to the procedure described in Barltrop et al. [61]. After the addition of the MTS compound to the wells and the incubation during 3 h in darkness, the absorbance was immediately measured at 490 nm.
Neutral red uptake was performed according to the procedure indicated in Borenfreund and Puerner [62]. The NR absorbed by the cells was quantified at 540 nm.
PC was quantified following the method reported in Bradford [63] with modifications [64] and read the absorbance at 620 nm. 5.6. Oxidative Stress Assays 5.6.1. Reactive Oxygen Species (ROS) Generation The ROS production was assessed by using the dichlorofluorescein (DCF) assay in 96 well plates according to Puerto et al. [65] with some modifications [35,48,66,67].

Glutathione Content (GSH)
Glutathione (GSH) content was evaluated by reaction with the fluorescent probe monochlorobimane (mBCL). Cells were exposed to the extracts at the same concentrations as the ones used in the ROS generation assay and then incubated for 4, 8, 12, or 24 h. The GSH synthesis inhibitor, buthionine sulfoximine (BSO, 1 mM), was used as the positive control. After the exposure, cells were incubated for 20 min at 37 • C in the presence of 40 µM mBCL. Later, cells were washed with PBS and the fluorescence measured at 380 nm (excitation) and 460 nm (emission).

Calculations and Statistical Analysis
The data were presented as mean ± standard deviation (SD) in relation to control. Statistical analysis was carried out using analysis of variance (ANOVA), followed by Dunnett's multiple comparison tests using GraphPad InStat software (GraphPad Software Inc., La Jolla, CA, USA). Differences were considered significant from p < 0.05. EC 50