P < 0.05 was assumed as statistically significant. Six protocols were considered to establish the hepatocytes dissociation method for P. lineatus. The non-enzymatic dissociation with EDTA (2 mM in PBS) was unsuccessful due to lyses of most cells during the procedure.

The same drawback happened after using the trypsin (0.05%) and pancreatin (0.25%) enzymes. However, collagenase IV (0.25 U ml−1), collagenase IV (0.15 U ml−1) associated with dispase (0.5 U ml−1), and only dispase (1 U ml−1) presented satisfactory results considering cell yield and viability ( Table 1). The protocol using collagenase IV resulted in 88% of cell viability and 1.01 × 107 hepatocytes per gram of liver, whereas collagenase IV and dispase enzymes resulted in about 3-fold increase in hepatocytes yield ( Table 1) maintaining similar cell viability. However, 97% of cell viability Z-VAD-FMK and 6.36 × 107 hepatocytes per gram of liver were obtained using dispase ( Table 1). On this way, the latter protocol was selected for further tests of cell attachment. Hepatocytes adhered properly on two of the three culture flask brands tested, TTP® and Biofil®. However, there was no improvement in hepatocyte attachment Enzalutamide with coat pretreatments and four days were necessary for cell attachment after seeding, as evidenced by the clustering

of hepatocytes in large groups ( Fig. 1 and Fig. 2). On this way, we adopted the following protocol for investigation of cylindrospermopsin effects. Hepatocytes were dissociated with dispase, seeded on TTP® flasks, cultured during four days for cell recovery and attachment, and then exposed to cylindrospermopsin through replacement of culture medium. Cell viability decreased 8% in hepatocytes exposed to the two lowest concentrations of purified cylindrospermopsin (0.1 and 1 μg l−1), but not at the highest concentration of 10 μg l−1 (Fig. 3A). Cells exposed to the three concentrations of cylindrospermopsin have similar GST and G6PDH activities in comparison to the control

group (Fig. 3B and C), although Tukey post test indicated that the GST activity of PRKD3 the hepatocytes exposed to 10 μg l−1 was 12% lower than of those exposed to 1 μg l−1 (Fig. 3B). Similarly, this post test showed that the G6PDH activity of the hepatocytes exposed to cylindrospermopsin at 10 μg l−1 was 19% lower than of those exposed to 0.1 and 1 μg l−1 (Fig. 3C). No significant alterations were observed for GSH concentration (53.6 ± 15.8 μmoles of non protein thiols per mg of protein) and also for the 2GSH/GSSG ratio (p > 0.7188) after exposure ( Fig. 3D), despite of the 25% increase of reactive oxygen/nitrogen species levels (mainly hydrogen peroxide) in all cylindrospermopsin-exposed groups in comparison to the control group ( Fig. 3E). Likewise, MXR activity decreased in about 22% in exposed groups, but without a concentration–response relation ( Fig. 3F), demonstrating that cylindrospermopsin may be able to make hepatocytes of P.