Effects of a gradual increase in temperature on the antioxidant defense system and plasma metabolic parameters in Antarctic fish Notothenia rossii


 Antarctica is considered a thermally stable ecosystem; however, climate studies point to increases in air and surface water temperatures in this region. These thermal changes may affect the biological processes of animals inhabiting such regions because they are stress factors and may promote metabolic changes, rendering the animals more vulnerable to oxidative damage. Plasma parameters are also good indicators of stress and allow analysis of the metabolic status of fish under temperature increases. The present study assessed the effect of acclimation temperature on the levels of plasma, osmoregulatory and oxidative metabolism parameters and antioxidant defenses in kidney, gill, liver and brain tissues of Notothenia rossii subjected to gradual temperature changes of 0.5°C/day until reaching temperatures of 2, 4, 6 and 8°C. Under the effect of the 0.5°C/day acclimation rate, gill tissue showed increased glutathione-S-transferase (GST) activity, and kidney tissue showed increased H⁺-ATPase at 9 days of the experiment (2°C). In the liver, consistent increases in the MDA concentration as an indicator of lipid peroxidation (9 (2°C),13 (4°C),17 (6°C) and 21 (6°C) days) were noted, as well as an increase in GSH at 9 days (2°C). In plasma, gradual decreases in the concentrations of total proteins and globulins were observed. These responses indicate the presence of thermal plasticity and an attempt at regulation to mitigate thermal stress. The changes showed that a gradual increase in temperature may cause opposite responses to the thermal shock model in N. rossii.

Every two days, the tanks were cleaned, and 50% of the water was exchanged with the temperature Enzymatic activity was expressed in nmol or µmol of substrate converted into product per minute (nmol.min -1 , mU; and µmol.min -1 , U; respectively). The speci c activities of the enzymes were expressed according to the protein concentration in the samples (mU.mg -1 or U.mg -1 ) after normalization to a protein concentration of 1 mg.ml -1 . The protein concentration was determined by the method described by Bradford (1976), with bovine serum albumin as the standard.
The activity levels of superoxide dismutase (SOD, EC 1. 15 Keen et al. (1976), where the reaction between GSH and 1-chloro-2,4-dinitrobenzene (CDNB) forms the thiolate anion (TBT), and the absorbance variation was monitored at 340 nm. The concentrations of GSH and other non-protein thiols were determined using the method described by Sedilak and Lindsay (1968) based on protein precipitation and the subsequent reaction of non-protein thiols with DTNB to generate a product that absorbs light at 415 nm.

Determination of oxidative damage marker levels
The lipid peroxidation (LPO) index was evaluated using the thiobarbituric acid reactive substances (TBARS) method adapted from Federici et al. (2007) in which the reaction of malondialdehyde (MDA) with thiobarbituric acid (TBA) produces a chromophore that can be measured at 535 nm. The protein carbonylation (PCO) index was determined using the method described by Levine et al. (1994), which is based on the reaction of carbonylated proteins with 2,4-dinitrophenylhydrazine (DNPH) generating dinitrophenylhydrazones, whose concentration was measured at 370 nm. The PCO analysis was not performed for the brain samples due to the limited volume of the samples obtained.  KÜLTZ;SOMERO, 1995). In the rst 4 replicates, only the reaction medium was added without inhibitors. In the next 4 replicates, a reaction medium containing 2 mM of ouabain (Na + /K + -ATPase inhibitor) was added. In the last 4 replicates, NEM (H + -ATPase inhibitor) was added. The readings were performed at 340-nm absorbance and a temperature of 20°C.

Statistical analyses
First, a univariate analysis of variance (ANOVA) was applied to the responses of the various biomarkers with a parametric distribution to determine differences between the groups subjected to temperature and time variations (ANDERSON, 2001). Accordingly, the sampling units were grouped into two factors (time factor: 9, 13, 17 and 21 days of the experiment; temperature factor: control (0°C) and experimental (2,4,6 and 8°C)). For the biomarkers with a non-parametric distribution, Kruskal-Wallis analysis was applied with similar groupings to those of the ANOVA. These analyses were performed in the vega package of the R platform (DEVELOPMENT CORE TEAM, 2013). Signi cant differences were considered when p ≤ 0.05.

Page 7/26
All animals survived for 21 days at all temperatures (0, 2, 4, 6 and 8°C). The analyses of variance showed that among the biomarkers tested, some showed signi cant differences and were in uenced by the factors temperature and time as well as the interaction between the two factors (Table 1).

ANOVA of oxidative metabolism biomarkers, non-protein thiol levels and oxidative damage markers
In  (Table 1).

Antioxidant enzyme activity levels
In N. rossii gills, no change in the activity levels of CAT (Fig. 1a) and SOD (Fig. 1e) were found compared to those in the respective control groups at each experimental temperature and between the different experimental times. The experimental and control groups at 13 days showed a decrease compared to those in the experimental and control groups at 21 days. The experimental and control groups at 13 and 17 days showed a decrease compared to those in the experimental and control groups at 21 days. The GPx activity level (Fig. 1b) increased in the control group at 9 days (0°C) compared to those in the other control groups at 13 and 21 days.
The gill GR activity level (Fig. 1c) increased in the control group at 21 days (0°C) compared to those in the other control groups (9, 13 and 17 days) and in the experimental groups. The GST activity level (Fig. 1d) increased in the experimental group at 9 days (2°C) compared to those in the respective control group (0°C) and the experimental groups at 13 and 17 days (4 and 6°C, respectively). The control group (0°C) showed increased GST activity levels at 21 days compared to those in the other control groups for the other experimental times (9, 13 and 17 days).
No changes in SOD (Fig. 2e), GR (Fig. 2c) and GST (Fig. 2d) activity levels were observed in the N. rossii kidneys compared with those in the respective control groups at all tested temperatures (2°C, 4°C, 6°C and 8°C) or between experimental times (9, 13, 17 and 21 days). The CAT activity level (Fig. 2a) increased in the control group at 17 days (6°C) compared to those in the other control groups at 9, 13 and 21 days (2°C, 4°C and 8°C). The GPx activity level (Fig. 2b) increased in the control group at 17 days (6°C) compared to that in the control group at 13 days (4°C).
In the liver of N. rossii, no change in the GR activity level (Fig. 3c) was observed compared to those in the respective control groups (2°C, 4°C, 6°C and 8°C) or between experimental times (9, 13, 17 and 21 days). The CAT activity level (Fig. 3a) was higher at 9 days (0°C) than those at the experimental times of 17 and 21 days.
The liver GST activity levels ( Fig. 3d) were lower in the control and experimental groups at 21 days than those in the control and experimental groups at the other experimental times (9, 13, 17 and 21 days). The SOD activity levels ( Fig. 3e) were higher in the control and experimental groups at 9 days than those in the control and experimental groups at 13 and 17 days.
In the brain of N. rossii, the GR activity level (Fig. 4c) decreased at 21 days compared to those in the control and experimental groups at the other experimental times (9, 13 and 17 days). The CAT activity level (Fig. 4a) was higher at 9 days (0°C) than those in the control groups at 13, 17 and 21 days.
The brain GPx activity level (Fig. 4b) decreased in the control and experimental groups at 21 days (0°C) compared to those in the control and experimental groups at 9 days. SOD activity levels (Fig. 4e) increased in the control and experimental groups at 9 days (0°C) compared to those in the other control and experimental groups at 13 and 17 days. At 21 days, in the control and experimental groups, the SOD activity levels decreased compared to those in the control and experimental groups at other experimental times (9, 13 and 17 days).
The brain GST activity levels (Fig. 4d) decreased in the control and experimental groups at 21 days compared to those in the other control and experimental groups at 9 and 17 days.

Concentrations of oxidative damage markers and non-protein thiols
In N. rossii gills, LPO (Fig. 5b) increased in the control and experimental groups at 21 days compared to that in the other experimental groups at the other experimental times (9, 13 and 17 days). No difference in LPO was found between the experimental and control groups. PCO (Fig. 5a) decreased in the control group at 13 days (0°C) compared to that in the control groups at 9 and 21 days. No signi cant difference in PCO was found between the experimental and control groups.
In the gills, the concentrations of glutathione and other non-protein thiols (Fig. 5c) increased in the control and experimental groups at 21 days compared to those in the control and experimental groups at 13 days. No signi cant variation in non-protein thiols was observed between the experimental and control groups.
In N. rossii kidneys, the levels of LPO (Fig. 6a) and PCO (Fig. 6b) did not change compared to those in the respective control groups (2°C, 4°C, 6°C and 8°C) or between the different experimental times (9, 13, 17 and 21 days).
In the kidneys, the concentrations of glutathione and other non-protein thiols (Fig. 6c) increased in the experimental group at 17 days (6°C) compared with those in the experimental groups at 13 and 21 days (4°C and 8°C). No signi cant changes in glutathione and other non-protein thiols were observed between the control and experimental groups.
In the liver of N. rossii, LPO levels ( Fig. 7a) increased at all temperatures (2, 4, 6 and 8°C) after 9, 13, 17 and 21 days compared to that in the control group (0°C). No change in PCO levels was found between the control and experimental groups (Fig. 7b) or between the different experimental times (9, 13, 17 and 21 days).
In the liver, the concentrations of glutathione and other non-protein thiols (Fig. 7c) increased in the experimental group at 9 days (2°C) compared to those in the control group. The experimental group at 21 days (8°C) showed lower concentrations compared with those in the experimental groups at 9 and 17 days (2°C and 6°C). The control group at 17 days (0°C) had a higher concentration of non-protein thiols compared to that in the control group at 9 days.
In the brain, no differences in LPO (Fig. 8a) or glutathione and other non-protein thiol levels ( Fig. 8b) were noted between the control and experimental groups or between the different experimental times (9, 13, 17 and 21 days).  Table 3). The biomarkers total protein (F = 9.17, p = 0.0025), globulin (F = 10.43, p = 0.0012) and LDH (F = 6.52, p = 0.0106) were in uenced by temperature (Table 3). Triglycerides and cortisol were not in uenced by time, temperature or their interaction ( Table 3). The interaction between time and temperature did not in uence any biomarker (Tables 2 and 3).

Concentration of plasma cortisol
In N. rossii plasma, the cortisol concentration (Fig. 9c) did not change in any experimental group compared to those in the respective control groups (2, 4, 6 and 8°C) or between experimental times (9, 13, 17 and 21 days).

Concentrations of plasma metabolic parameters
In N. rossii plasma, the albumin concentrations (Fig. 10d) decreased in the control and experimental groups at 13 days (4°C) compared with those in the control and experimental groups at 17 and 21 days (6°C and 8°C). No change in the albumin concentration was found in any experimental group compared with those in the respective control groups (2, 4, 6 and 8°C) or between experimental times (9, 13, 17 and 21 days). Cholesterol concentrations (Fig. 10a) increased in the control groups at 17 and 21 days (6°C and 8°C) compared with those in the control groups at 9 and 13 days (2°C and 4°C). No change in the cholesterol concentration was noted in any experimental group compared to those in the respective control groups (2, 4, 6 and 8°C) or between experimental times (9, 13, 17 and 21 days).
The plasma globulin (Fig. 10e) and total protein (Fig. 10c) concentrations decreased in all experimental groups (2, 4, 6 and 8°C) compared with those in the respective control groups. No differences in plasma globulin and total protein concentrations were noted when the groups were compared between experimental times (9, 13, 17 and 21 days).
Plasma LDH activity levels (Fig. 9b) increased in the experimental group at 9 days (2°C) compared to those in the other experimental groups at 13, 17 and 21 days (4, 6 and 8°C). No difference in LDH was observed between the experimental groups and their respective control groups (2, 4, 6 and 8°C).

Concentrations of plasma osmo-ionic parameters
In N. rossii, the plasma calcium concentration (Fig. 11a) increased in the control and experimental groups at 13 days (4°C) compared with those in the control and experimental groups at 21 days (8°C). The concentrations in the control and experimental groups at 17 days decreased compared to those in the control and experimental groups at 9 (2°C) and 13 days (4°C). No difference in the calcium concentration was found between the experimental groups and their respective control groups (2, 4, 6 and 8°C). The plasma chloride concentration (Fig. 11b) decreased in the control and experimental groups at 21 days (8°C) compared to those in the control and experimental groups at 9 and 13 days (2°C and 4°C). No difference in the chloride concentration was observed between the experimental groups and their respective control groups (2, 4, 6 and 8°C).
The plasma phosphorus concentration (Fig. 11c) increased in the control and experimental groups at 21 days compared to those in the other control and experimental groups at 9, 13 and 17 days (2, 4 and 6°C). No difference in the phosphorus concentration was noted between the experimental groups and their respective control groups (2, 4, 6 and 8°C). The plasma magnesium concentration (Fig. 11d) increased in the control and experimental groups at 17 days (6°C) compared to those in the control and experimental groups at 9 and 13 days (2°C and 4°C). No difference in the magnesium concentration was found between the experimental groups and their respective control groups (2, 4, 6 and 8°C).

ANOVAs of the enzymes Na /K ATPase and H -ATPase in kidney and gill tissues
In N. rossii gills, the activity levels of Na /K ATPase (F = 7.9701, p = 2.0 − 4 ) and H -ATPase (F = 3.8672, p = 0.0148) were in uenced by the time factor (Table 4). In the kidney, Na /K ATPase (F = 13.4066, p = 0.0038) and H -ATPase (F = 14.4552, p = 0.0023) were in uenced by the interaction between the time and temperature factors ( Table 4). The time factor alone did not in uence the activity levels of Na /K ATPase and H -ATPase (Table 4) 3.9 Na / K ATPase and H -ATPase enzyme activity levels in the gills In the gills of N. rossii, Na / K ATPase activity levels at 17 days in the control (0°C) and experimental (6°C) groups were lower than those at the other experimental times and temperatures tested (Fig. 12a). H -ATPase activity levels at day 9 in the experimental (2°C) and control (0°C) groups were also lower than those at the other experimental times and temperatures tested (Fig. 12b).
3.10 Na / K ATPase and H -ATPase enzyme activity levels in the kidney In the kidney of N. rossii, Na / K ATPase activity levels were lower in the control group at 9 days than those in the control groups at 17 and 21 days (Fig. 13a). H -ATPase activity levels in the experimental group (2°C) at 9 days were higher than those in the respective control group and the control and experimental groups at 13 days (4 and 0°C, respectively) (Fig. 13b).

Discussion
Although Antarctic sh can live in extremely cold and stable environments, due to higher tissue When ROS production increases, the body must respond with compensatory mechanisms such that the resulting metabolic processes occur in a balanced manner to preserve a good physiological state among cells (FRANKLIN, 2010).
A response of the gills to heat stress is expected because the gills are in close contact with water and are directly exposed to changes in the surrounding aquatic environment, such as temperature changes (PERRY; LAURENT, 1993; SABER, 2011). In their normal physiological state, the gills already have a high metabolic rate mainly due to the process of ionic regulation (JOHANSE; PETTERSON, 1981;MOMMSEN, 1984).
In the gills at 9 days of the experiment, when the temperature reached 2°C, an increase in GST activity was observed (Fig. 14). This enzyme is responsible for the conjugation of xenobiotics to GSH, facilitating their elimination. Other GST isoforms with diverse functions have been identi ed, such as intracellular protein transport and regulation of apoptosis, a process that may be increased under stressful circumstances (HALLIWELL; GUTTERIDGE A similar response was observed in a study of acute stress in N. rossii maintained at 8°C for 6 days, where GSH levels increased signi cantly over time (MACHADO et al., 2014).
Increased availability of GSH can increase the availability of substrate, allowing higher turnover of enzymes that require GSH and thus conferring cells with a greater ability to neutralize ROS. In this study, however, no changes in the activity levels of GSH-dependent enzymes were observed in the liver, possibly because the constitutive concentrations of these enzymes were su cient, which was also observed by Machado et al., 2014 in N. rossii.
During the acclimation period, the liver MDA levels remained high, re ecting a continuous lipid peroxidation response (Fig. 14). A high LPO index is expected for Antarctic sh under oxidative stress given the high levels of mono-and polyunsaturated fatty acids in these animals (SIDELL, 1998).
The products generated by LPO are detoxi ed through their conjugation with GSH by GST, which may justify the observed increase in thiols at 9 days (2°C) as an attempt to maintain redox balance. However, the permanent increase in LPO concentrations indicates that equilibrium was not achieved.
In N. rossii, increases in liver LPO did not occur in acute stress models (MACHADO et al., 2014;KLEIN et al., 2017). Increased levels of liver LPO were also observed in other Antarctic sh, such as Pagothenia borchgrevinki (ALMROTH et al., 2015) and Pachycara brachycephalum (HEISE et al., 2007), exposed to high temperatures in the long term, which may suggest that for LPO, the exposure time to a stressor agent may be more relevant than its magnitude.
The In N. rossii plasma, a decrease in the concentration of total proteins (albumin and globulins) (Fig. 14) accompanied by a decrease in the globulin concentration was observed, indicating that the former decrease occurred at the expense of globulin consumption (Fig. 14). This result is similar to that found in a long-term stress model (90 days The increase in H + -ATPase in the kidneys at 9 days and 2°C may indicate that plasma acidi cation was already occurring (Fig. 15). Renal intercalated cells are one of the types of specialized cells that contain In hypertensive and heat shock models, a hyperglycemic effect was observed in response to heat stress in N. The concentrations of the osmo-ionic parameters calcium, chloride, phosphorus and magnesium also did not change with the gradual increase in temperature. However, the time of the inter-renal response to adrenocorticotrophic hormone (ACTH) for cortisol production, as well as changes in the plasma levels of ions, has a latency from minutes to hours (PANKHURST, 2011).
Due to the experimental design of our study, where the rst samples were collected 9 days after exposure to the stressor agent, we may have missed plasma variations for some markers. When subjected to heat shock, N. rossii presented changes only in the Mg ion until up to 12 hours of exposure to a temperature of 8°C (KANDALSKI et al., 2018).
In Ramdhia quelen, a subtropical sh, a cortisol peak occurred rapidly in juveniles between 5

Conclusion
The gradual increase in temperature triggered plasma and oxidative metabolic responses unlike those reported in studies of thermal shock. Such responses include increased gill GST activity, indicating an increase in ROS production and a persistent increase in liver LPO, which suggests that for LPO, the exposure time to a stressor agent seems to be more relevant than its magnitude.
The data obtained for liver LPO may be related to increased GSH, and liver impairment seems to have in uenced the plasma protein composition because the synthesis of globulins may have been altered by the gradual increase in temperature.
The increased H + -ATPase activity in kidneys may aggravate the imbalance in the energy demand in this organ leading to activation of the gluconeogenic pathway.
In cases of thermal acclimation, ionic parameters, cortisol and glucose are not considered good markers possibly due to their rapid release in the plasma.