Lophiosilurus alexandri, a sedentary bottom �sh, adjusts its physiological parameters to survive to hypoxia condition.

We investigated blood gas, hematological and biochemical parameters and gill morphology and morphometry of Lophiosilurus alexandri juveniles submitted to hypoxia for 48 hours followed by recovery for 48 hours. A total of 48 juveniles (360.0 ± 141.6 g) were distributed among eight tanks (120 L) and subjected to hypoxia condition (water with dissolved oxygen at 2.12 ± 0.90 mg L − 1 ) or normoxia (at 5.60 ± 0.31 mg L − 1 ). Blood gas values (pH, PvCO 2 , PvO 2 , sO 2 , HCO 3− , stHCO 3− and base excess) in hypoxia were signi�cantly different from normoxia, while lactate and the electrolytes (K + , Na + , Cl − , Ca 2+ and HCO 3− ) there was no signi�cant change among treatments. The erythrocytes differed signi�cantly between hypoxia and normoxia at 24 h of recovery, while for hemoglobin and hematocrit there were no signi�cant differences. There was a signi�cant difference in glucose, triglycerides, and cholesterol for both normoxia and hypoxia, while plasma protein remained unchanged. All gill components (epithelial cells, erythrocytes, pillar cells, mucus cells, chloride cells, undifferentiated cells, and blood capillary lumen) differed signi�cantly between hypoxia and normoxia. A reduction in the length of the primary lamella was observed in the hypoxia and recovery treatments, when compared to normoxia. The secondary branchial lamella showed no signi�cant difference for both treatments. In general, juveniles of L. alexandri adapted well to hypoxia exposure for 48 h, as they were able to adjust most of their physiological variables to survive this stress condition and return to normoxia within 48 h.


Introduction
The stress response of sh occurs through neuroendocrine control in a way very similar to that of mammals.Fish have a cephalic kidney and chroma n and interrenal cells as structures homologous to the adrenal gland of mammals, and which perform the function of secreting circulating catecholamines (epinephrine and norepinephrine) and corticosteroids, such as cortisol (Barton 2000;Gallo and Civinini 2003).Stress presents three responses: the primary response, also known as the alarm response, is accompanied by high concentrations of catecholamines and corticosteroids in the plasma; the secondary response is related to increased blood hormonal levels, which promote increased heart rate and higher oxygen consumption; the tertiary response begins with the exhaustion of the organism, which causes a decrease in productive and reproductive performance and exposes the animal to various pathogens (Barton 2002;Lima et al. 2006;Wendelaar Bonga 1997).
Fish can experience several factors that are considered stressors in captivity, such as high con nement densities (Long et al. 2019;Refaey et al. 2018;Yarahmadi et al. 2016), handling and transportation (Jerez-Cepa et al. 2019;Sena et al. 2016) and changes in water quality (Lankford et al. 2003;Zhang et al. 2015).These stressors generate physiological disorders, which cause discomfort, impair well-being and potentially compromise survival.Thus, to maintain and restore body homeostasis, a stressed animal consumes energy due to breathing, locomotion and tissue repair, instead of using it for growth, reproduction, food intake and physiological status (Schreck and Tort 2016).
Pacamã, Lophiosilurus alexandri, is a Brazilian freshwater sh species endemic to the São Francisco River basin (Shibata 2003) that has excellent quality meat and is among regional species with high commercial value (Luz and Santos 2008).The species shows sedentary behavior and a preference for lentic environments in sandy or rocky regions (Travassos 1959).Due to its peculiar characteristics, however, it remains to be seen how this species responds physiologically to decreased water dissolved oxygen concentrations in the captive environment.Thus, the aim of this study was to evaluate the responses of blood gas, hematological and biochemical parameters, as well as gill morphology, of L. alexandri juveniles submitted to hypoxia and subsequent recovery.

Material And Methods
This study was conducted at Laboratório de Aquacultura (LAQUA) of the Universidade Federal de Minas Gerais (UFMG), Belo Horizonte, Minas Gerais, Brazil, and was approved by the Animal Ethics and Welfare Committee of UFMG (protocol number: 61/2019).

Fish and experimental conditions
A total of 48 juveniles (360.0 ± 141.6 g, 26.6 ± 3.0 cm) were distributed among eight tanks (120 L; six animals per tank) in a recirculating aquaculture system (RAS) with biological and mechanical lters, where they were acclimatized for 15 days.Fish were fed until apparent satiety twice a day (8 am and 4 pm) with a commercial extruded diet (38% of crude protein).The water quality during acclimatization remained at: dissolved oxygen (DO) 5.10 ± 0.56 mg L − 1 ; temperature 28.9 ± 0.28°C; pH 7.04 ± 0.11; and total ammonia 0.05 ± 0.01 mmol L − 1 .Dissolved oxygen concentration and temperature were measured using a digital oximeter (model EcoSense -YSI DO200A), pH was measured using a peagameter (model KASVI -K39) and total ammonia using a commercial kit (LabconTest®).
The water ow and arti cial aeration of four of the tanks were interrupted to decrease the DO concentration and promote a condition of hypoxia (hypoxia treatment).After reaching a DO concentration of about 2.0 mg L − 1 , the water ow was turned on again, but kept low, thus maintaining a DO concentration of 2.12 ± 0.90 mg L − 1 for 48 h.The other four other tanks were maintained at normoxia (normoxia treatment) with stable water ow and aeration and DO maintained at 5.60 ± 0.31 mg L − 1 .After 48 h, water ow and aeration were restored to the hypoxia tanks and the DO was maintained at 5.30 ± 0.47 mg L − 1 , similar to that of the normoxia tanks.The animals remained in these conditions for another 48 h, for a total experiment time of 96 h.During the experiment, the animals were fed twice a day until apparent satiety with the same commercial diet used during acclimatization.

Blood sampling and analysis
Six sh (from the same tank) of the hypoxia treatment were collected at 24 and 48 h of hypoxia and at 24 and 48 h of recovery, while six sh (from the same tank) of the normoxia treatment were collected at the same times.Two blood samples were collected from each sh, while contained in an appropriate damp cloth, by caudal venipuncture with ventral access using heparinized syringes.From the rst sample, 300 µl of blood was placed in microtubes on ice for blood gas analysis and subsequent determination of the following parameters: pH, PvCO 2 (partial pressure of carbon dioxide), PvO 2 (partial pressure of oxygen), SO 2 (oxygen saturation), cLac (lactate), K + , Na + , Cl − , Ca² + , HCO 3 − (bicarbonate), stHCO 3 − (standard bicarbonate) and BE (base excess).These analyses were performed using a blood gas analyzer (ABL800 BASIC-Radiometer®), with water temperature and oxygen concentration being corrected by the equipment, according to the experimental treatments.
From the second blood sample, approximately one milliliter of blood was collected to assess hematological and biochemical parameters.Hematocrit (%) was determined using the microhematocrit technique (Goldenfarb et al., 1971); plasma protein was measured by refractometry (portable refractometer, RHC 200-ATC, Huake Instrument Co., Ltd); and hemoglobin concentration was determined by spectrophotometry (Biochrom Libra S22) with the aid of a commercial kit (Ref.K023-1 QUIBASA Ltda.Bioclin).Erythrocyte counts were performed within 24 h after sampling using a hemocytometer and an optical microscope with 400 X magni cation.The remaining blood was centrifuged for 10 min at 4000 rpm for plasma separation and subsequent determination of glucose, triglycerides and cholesterol concentrations using an enzymatic-colorimetric method (Trinder reaction).All biochemical parameters were analyzed with commercial kits (Bioclin® -Belo Horizonte, Brazil -www.bioclin.com)and read with a spectrophotometer (Biochrom Libra S22).

Gill morphology and morphometry
The same animals used for blood sampling were euthanized (285 mg L − 1 eugenol) for gill collection.For histological analysis, the 2nd gill arch of each animal was xed in Bouin solution for 24 h, embedded in para n, sectioned at 5 µm thickness and stained with hematoxylin-eosin.The main structures in the primary and secondary gill lamellae were quanti ed using 40 histological elds from four individuals analyzed per treatment (24 and 48 h of hypoxia and 24 and 48 h of recovery for hypoxia treatment and same times for normoxia treatment).The relative proportion of the gill components (epithelial cell, erythrocytes, pillar cells, mucous cells, chloride cells, undifferentiated cells, and blood capillary lumen) were determined using a grid containing 560 intersection points overlaying each eld and ImageJ 1.52 software.Blank spaces and artefacts were excluded from the counts.To measure the diameter of the primary and secondary lamellae of the gills, 40 measurements extracted from four sh per group (10 measurements from each animal) were used, of which four treatments and one initial control.This measurement was performed using the AxioVision SE64 software coupled to an Axioplan 2 microscope (Zeiss).

Statistical analysis
All blood data were analyzed (SigmaPlot® Software, version 11.0) in a completely randomized design using a one-way ANOVA, to compare treatments over sampling time and the t-test to compare treatments (normoxia and hypoxia) at each time.Normality and homoscedasticity of the data were evaluated by the Shapiro-Wilk and Levene tests, respectively, followed by Tukey post-hoc test with 5% probability.Data are presented as mean and standard error (SE).Data for the relative proportion of gill structures were subjected to Kruskal-Wallis statistical analysis followed by Dunn post-hoc test with 5% probability (P < 0.05).

Blood gas parameters
Results for blood gas parameters are shown in Table 1.Values for pH did not differ signi cantly among sampling times for the normoxia (P = 0.153) and hypoxia treatments (P = 0.052).However, there was a signi cant difference between treatments at 24 h of recovery, with animals submitted to hypoxia showing higher blood pH (P = 0.010).
No signi cant differences were observed over time for either of the treatments nor between them for lactate (cLac) (P > 0.05) and to electrolytes K + (P > 0.05) Na + (P > 0,05) and Cl − (P > 0.05).There was a decrease in Ca 2+ levels in hypoxia treatment at 24 h of recovery (P = 0.004), when compared to 24 h of hypoxia and 48 h of recovery.For HCO 3 − there were signi cant differences between normoxia and hypoxia treatments at 24 h of hypoxia (P = 0.009) and 48 h of recovery (P = 0.042), in which the concentrations of HCO 3 − were higher in the hypoxia treatment.In relation to the sampling times, higher concentrations of HCO 3 − were observed at 48 h of hypoxia (P = 0.198) and at 24 h (P = 0.640) and 48 h (P = 0.042) of recovery.
Values for stHCO 3 − differed signi cantly among treatments at 24 h of hypoxia and 48 h of recovery, in which there was an increase in stHCO 3 − concentrations in hypoxia treatment at both sampling times, compared to the normoxia treatment (P = 0.051, at 24 h of hypoxia and P = 0.026, at 48 h of recovery).
Base excess (BE) remained constant over sampling times for normoxia (P = 0.075) and hypoxia (P = 0.156) treatments.Values for BE were signi cantly higher for the hypoxia treatment at 24 h of hypoxia (P = 0.033) and 48 h of recovery (P = 0.014), compared to the normoxia treatment.

Hematological and biochemical parameters
Erythrocytes (Fig. 1A) did not differ signi cantly among sampling times for the normoxia treatment (P > 0.05), but there was a signi cantly lower count for the hypoxia treatment at 48 h of hypoxia (P = 0.015).
The hypoxia treatment had a signi cantly higher erythrocytes count at 24 h of recovery than normoxia treatment (P = 0.036).
Hemoglobin concentrations (Fig. 1B) did not differ signi cantly among sampling times for both treatments (P > 0.05) or between treatments at the different sampling times (P > 0.05).Hematocrit (Fig. 1C) differed signi cantly among sampling times for the hypoxia treatment, with it decreasing over time (P = 0.015), while it did not differ signi cantly among sampling times for the normoxia treatment (P > 0.05).Hematocrit did not differ among treatments in each sampling time (P > 0.05).
Plasma glucose (Fig. 2A) for normoxia treatment differed signi cantly among sampling times (P = 0.007), with it being lowest at the time corresponding to 24 h hypoxia.The hypoxia treatment also differed signi cantly among times, with it being highest at 48 h of recovery (P = 0.004).Glucose levels for hypoxia treatment at 48 h of hypoxia and 24 h of recovery were signi cantly lower than that for normoxia (P = 0.006), however, at 24 h of hypoxia and 48 h of recovery the treatments did not differ signi cantly (P > 0.05).
The levels of triglycerides (Fig. 2B) differed between the sampling times for normoxia treatment, with the lowest values corresponding to 24 h of hypoxia (P < 0.001), and 24 (P = 0,160), and 48 h of recovery (P = 0.281).Hypoxia treatment did not show a signi cant difference in triglycerides between sampling times but had lower levels in 48 h of hypoxia (P = 0.003) compared to normoxia treatment.
Cholesterol (Fig. 2C) differed completely over the sampling times for both treatments.For hypoxia treatment, the lowest level of cholesterol was found at 24 h of recovery (P = 0.020) and the highest value at 24 h of hypoxia (P = 0.022).The normoxia treatment also differed between sampling times, with the lowest value at 24 h of hypoxia (P = 0.016).Plasma protein (Fig. 2D) did not show difference between treatments at the same sampling times, however for the normoxia treatment there was a decrease in plasma protein at 24 h of recovery (P = 0.818), compared to normoxia in the other sampling times.
The proportion of undifferentiated cells was lower at 24 and 48 h of recovery, when compared to hypoxia times and normoxia treatment.The percentage of blood capillary lumen for hypoxia treatment was lower at 24 and 48 h of hypoxia and at 24 h of recovery compared to normoxia treatment; however, at 48 h of recovery, no signi cant difference was observed in relation to normoxia (Table 2).A reduction in the length of the primary lamella was observed in the hypoxia and recovery treatments, when compared to the control group.The secondary branchial lamella showed no signi cant difference (P > 0.05) for both treatments and over the sampling times (Table 2).

Discussion
Our study aimed to evaluate the effects of hypoxia during 24 and 48 h, on blood gas parameters, hematological, biochemical and possible changes in gill morphology of Lophiosilurus alexandri.Mortality was not observed at 24 and 48 h of hypoxia or at 24 and 48 h of recovery, similar to the study by Mahfouz et al. (2015) for Oreochromis niloticus exposed to short (24 h) and prolonged (30 days) hypoxia, and Mattioli et al. (2019) for L. alexandri juveniles submitted to air exposure for 30 min.In a more recent study of the same species, Baldissera et al. (2020) observed only one case of mortality after exposure to hypoxia for 72 h.According to Richards (2009), conditions of hypoxia allow animals to reveal their adaptive responses, such as enhanced survival due to changes in the expression of some genes by sh exposed to long-term hypoxia (more than a few hours).
In general, the present study found that the pacamã juveniles to reveal adaptive responses to hypoxia, with few changes in blood gas, hematological and biochemical parameters, and with just slight changes in gill morphology without remarkable histopathological alterations.Lophiosilurus alexandri is a species of sedentary behavior and low locomotor activity that lives at the bottom of rivers in its natural environment.Like what Lays et al. (2009) found with the sedentary species Anarhichas minor, conditions of hypoxia did not alter the behavior of L. alexandri juveniles in our study and the sh remained calm at the bottom of the tanks throughout the experiment.
As observed in blood gas parameters, there was an increase in blood pH values in the treatment submitted to hypoxia at period of 24 h of recovery, compared to normoxia.In addition, an increase in bicarbonate (HCO 3 − and stHCO 3 − ) and BE values were observed at 24 h of hypoxia and 48 h of recovery.
According to Joyce et al. (2015), these changes may represent strategies for the prevention of blood acidemia, which could lead to respiratory and metabolic acidosis, as was observed by Mattioli et al. (2019) for pacamã juveniles submitted to air exposure.Base excess (BE) alterations may be related to the maintenance of the acid-base balance of the internal environment and are of great physiological and biochemical importance, since the activities of cellular enzymes and electrolytic exchanges and maintenance of the structural state of the proteins in organisms are deeply in uenced by small changes in blood pH (Macari 1994).Blood PvO 2 for the hypoxia treatment decreased at 24 h of hypoxia, compared to the normoxia treatment, however, this variable quickly returned to normal levels at 24 h of recovery.
The concentrations of K + , Na + , Cl − and Ca 2+ ions in hypoxia treatment did not show signi cant differences during hypoxia nor during recovery, showing that there were no electrolytic changes in L. alexandri juveniles.This nding differs from that of Mattioli et al. (2019) with pacamã juveniles subjected to air exposure, where responses such as acidosis and osmotic stress were found.However, juveniles of O. niloticus submitted to transport did not show changes in Ca 2+ ion levels (Moreira et al. 2015).Ionic homeostasis is essential to ensure proper cell function (Hwang et al. 2011), and consequences of routine aquaculture management can generate stress in different species and, thereby, affect ionic regulation in sh (Ashley 2007).
According to Fazio et al. (2013), hematological parameters can be used to determine the physiological status of sh, water quality and other important variables.Hypoxia did not in uence erythrocyte counts, as well as hemoglobin and hematocrit, of the L. alexandri juveniles of the present study, which may re ect an adaptation of this species to the stress conditions imposed.In contrast, increases in erythrocyte counts, hemoglobin concentrations and hematocrit values soon after exposure to hypoxia are common for many sh species as a way to improve the oxygen transport capacity of blood circulation and its distribution to tissues, as found by Abdel-Tawwab et al. (2015) for Oreochromis niloticus, Aboagye and Allen (2018) for Polyodon spathula and Val et al. (2015) for Prochilodus nigricans.
Among biochemical variables, we observed a decrease in glucose concentrations at 24 h of hypoxia for normoxia and hypoxia treatments, compared to 48 h of hypoxia.This decrease may have been the result of the suspension of feeding on the day prior to the 24 h hypoxia sampling, which was then reestablished the following day and continued until the end of the trial.However, was observed that sh in the hypoxia treatment only returned to feeding after 24 h of recovery, when their glucose concentrations were similar to those of sh of the normoxia treatment at 48 h of recovery.This same response was also observed by Kupittayanant et al. (2011) with owerhorn sh (Amphilophus trimaculatus x Amphilophus citrinellus x Vieja synspilum), for which there was a reduction in feed intake after 12 h of exposure to hypoxia.However, unlike our study, these authors found an increase in glucose levels with hypoxia, which is explained by the occurrence of hepatic glycogenolysis.In the present study L. alexandri juveniles submitted to hypoxia may have experienced a decrease in metabolic rate, leading to decreased feed intake and, consequently, to decreased plasma glucose concentrations.It was also observed that the restoration of food re ected an increase in triglyceride concentrations (at 48 h of hypoxia) in the normoxia treatment, while the hypoxia treatment did not show any signi cant changes.
Hypoxia conditions did not alter the plasma cholesterol concentrations of juvenile L. alexandri, compared to the normoxia treatment.This nding differs from other studies that found an increase or decrease in this metabolite in blood of sh also submitted to hypoxia (Bera et al. 2017;Kuppitayanant et al. 2011), feed restriction (Assis et al. 2020;Favero et al. 2019) or low water temperature (Costa et al. 2016;Favero et al. 2019;He et al. 2015) and may be related to increased energy consumption (He et al. 2015) and decreased synthesis of endogenous cholesterol by sh during stress situations.According to Costa et al. (2019), exposure to stressors and the intensity of stress can lead to increased plasma protein concentrations in sh, as the energy need of these animals increases under such conditions.However, hypoxia for 48 h was not su cient to promote an increase in juvenile L. alexandri, since plasma protein concentrations did not differ signi cantly from sh in normoxia.
The branchial epithelium has key functions in gas exchange, such as oxygen uptake and carbon dioxide release, in addition to other important functions, such as osmoregulation, acid-base regulation, excretion of nitrogen compounds, and detoxi cation (Evans et al. 2005;Wilson and Laurent 2002).The present study found a reduction in primary lamella length in hypoxia and recovery periods when compared to normoxia.This reduction can be seen as a mechanism to reduce metabolic costs in ion transport (Evans et al. 2005).Some species reduce gill surface area in response to hypoxia exposure, with the intention of decreasing ion uncontrol (Matey et al. 2011;Boeck et al. 2013;Borowiec et al. 2015;Chasiotis et al. 2012).
In our study, changes in the relative proportions of gill components were also detected, such as an increase in the proportion of epithelial cells at 24 h of hypoxia.Epithelial cells play a key role in gas exchange and ion balance and their increase during hypoxia can be explained as a strategy to increase gas exchange and restore osmoregulatory balance (Wong and Wong 2000).
The lamellae of the secondary branches are covered by epithelial cells and pillar cells (Laurent and Dunel 1985;Takashima and Hibiya 1995).The pillar cells have contractile properties and perform blood transport for hematosis, in addition to being the rst site of excretion and playing a role in ion regulation (Baldisserotto 2002).In the present study, hypoxia caused a decrease in the percentage of pillar cells, which remained so during recovery.In a study with Crucian carp (Carassius carassius), Sollid et al. (2003) found a decrease in the total amount of cells in the gills as well as an increase in the respiratory surface area, which was not observed in the present study.Also, in contrast to the present study, hypoxia was found to cause a disorganization of pillar cells in the gills of the species Hippocampus reidi (Negreiros et al. 2011).
Erythrocytes, which are responsible for transporting the oxygen molecule, had higher percentages in the rst hours of recovery from hypoxia, showing that the animal was still in the process of reestablishing homeostasis and needing a rapid recruitment of erythrocytes to assist in the transport of oxygen to the gills (Inoue et al. 2011).The mucus cells present in the branchial epithelium are responsible for secreting a layer of glycoproteins and glycolipids that acts as a barrier, provides mechanical and biological protection of the epithelium (Breseghelo et. al. 2004) and facilitates ionic regulation (Diaz et al. 2000).
This was evidenced in the present study by the lowest concentration being found when the animals were recovering from hypoxia, demonstrating that ionic regulation was already in the recovery phase from the stress suffered.
An increase in the percentage of chloride cells was observed at 48 h of hypoxia, which may indicate a compensatory response of the animals to improve ion absorption and maintain ionic and osmotic homeostasis in the face of the challenge of low oxygenation (Fernandes and Mazon 2003).

Conclusion
Juveniles of Lophiosilurus alexandri adapted well to hypoxia for 48 h, as they were able to adjust most of their physiological variables to survive this stress condition and return to normoxia within 48 h.

-Con icts of interest
The authors declare that they have no known competing nancial interests or personal relationships that could have appeared to in uence the work reported in this paper.
-Ethics approval This study was approved by the Animal Ethics and Welfare Committee of UFMG (protocol number: 61/2019).-Consent to participate (include appropriate statements) Not applicable -Consent for publication (include appropriate statements) Not applicable -Availability of data and material/ Data availability Not applicable -Code availability (software application or custom code) applicable -Authors' contributions (include all authors):