Oxygen is critical for both terrestrial and aquatic aerobic organisms. Most of the organisms need molecular oxygen to support metabolic processes (Chandel and Schumacker, 2000). Compared with terrestrial environments, aquatic environments exhibit much wider temporal and spatial variations in oxygen concentration, and water contains extremely lower oxygen in the same volume of air at the same atmosphere pressure (Rytkönen et al., 2007). As a result, aquatic organisms are frequently exposed to variations in oxygen levels. In most cases, they will usually respond by escaping to other environments when encountering hypoxia condition. However, if hypoxia is unavoidable, aquatic organisms including fish have to make responses either by increasing oxygen transfer to tissues or reducing oxygen consumption through metabolic depression to survive (Moraes et al., 2002; Douxfils et al., 2012).

The earlier studies of fish under hypoxia stress are focused on tissue changes including blood, heart, gill, and brain. In rainbow trout, hypoxia markedly decreases the oxygen uptake effectiveness in blood (Randall et al., 1967), whereas during progressive hypoxia, the rate of oxygen uptake does not alter (Holeton and Randall, 1967b). During hypoxia the percentage saturation of both arterial and venous blood are decreased, while blood pressure in the dorsal and ventral aortae increases, all these changes are associated with a marked bradycardia in fish after hypoxia stress (Holeton and Randall, 1967a). Bradycardia does not affect fish cardiac output but stroke volume of cardiac enhanced, and vascular resistance to blood flow correspondingly increased (Holeton and Randall, 1967b; Butler and Taylor, 1975). Hypoxia also increases fish Hb-O₂ affinity and oxygen capacity, and directly provides higher and more stable effectiveness of oxygen uptake in blood; however the total P_O2 is till decreased (Holeton and Randall, 1967b; Wood and Johansen, 1972). During hypoxia, the ATP concentration and ATP:Hb ratio in the red cells decreased (Wood et al., 1975), further demonstrate that the erythrocytic ATP:Hb4 molar ratios declined with increasing hypoxic stress as did the pH gradient between the erythrocyte and plasma (Boutilier et al., 1988). Hypoxia leads to the fish bradycardia, while the cardiac output is regulated by changes in the stroke volume not heart rate (Hemmingsen et al., 1972). Long period of hypoxia has no effect on fish heart functions but leads to the decrease of myocardial glycogen concentration and the occurrence of mitochondrial necrosis (Leknes, 1985; Lennard and Huddart, 1992). Primarily, fish moves higher in the water column, increases gill ventilation and exhibits aquatic surface respiration during aquatic hypoxia (Urbina et al., 2011). Gills display protruding lamellae after 1 day of hypoxia and reach its greatest extent after 7 days, and the changes are completely reversible after 7 days reoxygenation (Sollid et al., 2003). The capillary diameter of fish brain increases after hypoxia, and 90 min hypoxia leads to the disappearance of brain pyruvate kinase (Scheich et al., 1972; Lushchak, 1993). Brain sensitivity to hypoxia in fish is reflected by changes in extracellular K⁺ activity, and in the hypoxia-tolerant fish crucian carp the Na⁺/K⁺-ATPase level in hypoxia crucian carp brain is maintained (Hylland et al., 1997).

The recent studies show that some hematological parameters including red and white blood cell count, hematocrit and hemoglobin concentrations are significantly increased after hypoxia treatment in fish (Petersen and Gamperl, 2011; Richards, 2011; Ni et al., 2014). Meanwhile, the biochemical parameters, such as blood glucose, lactate, corticosteroids, ATP, and GTP are also altered under hypoxia (Rees et al., 2009; Omlin and Weber, 2010; Urbina and Glover, 2012; Ni et al., 2014). During hypoxia, metabolic activity decreases and most of the energy will be used for the maintenance of primary physiological functions, probably leading to the changes of the immune defense. It is reported that in Eurasian perch and Catla, the lysozyme activity is significantly decreased under hypoxia conditions (Douxfils et al., 2012; Singh et al., 2016).

During the aerobic metabolism, the negative tetravalent atom of O₂ produces water at the end of the mitochondrial electron transport chain, while negative monovalent atom of O₂ generates several reactive oxygen species (ROS) such as superoxide anions (O2-•), hydroxyl anions (OH^−•), and hydrogen peroxide (H₂O₂) (Turrens, 2003). These ROS have damage effect on organisms including induction of oxidation of proteins, DNA and steroid components, and peroxidation of unsaturated lipids in cell membranes (Chance et al., 1979; Southorn and Powis, 1988). To minimize the detrimental influence of ROS, aerobic organisms have evolved enzymatic antioxidant mechanism accompanying the formation of antioxidant defenses such as superoxide dismutases (SOD), catalases (CAT), and glutathione (GSH), to scavenge the superoxide radicals and hydrogen peroxide for maintaining the homeostasis (Scandalios, 2005; Kobayashi and Yamamoto, 2006). SOD catalyzes the dismutation of O2-• to H₂O₂, and CAT and glutathione peroxidases (GSH-Px) reduce H₂O₂ to H₂O (Scandalios, 2005). When the ROS generation exceeds their removal, oxidative stress occurs. It has been reported that hyperoxia or even hypoxia alters cellular ROS levels in fish (Ross et al., 2001; Cooper et al., 2002; Garcia Sampaio et al., 2008). In some fish species, hypoxia alone causes elevation of the activities of enzymes involved in antioxidant defense (Lushchak et al., 2001, 2005; Lushchak and Bagnyukova, 2007), while in other fish species the antioxidant enzyme activities are decreased or/and unchanged in hypoxia stress (Leveelahti et al., 2014; Huang et al., 2015; Ransberry et al., 2016). Moreover, ROS-detoxifying enzyme activities rather than direct reactive intermediate levels are investigated in most experiments to indicate the altered status of ROS in hypoxia.

Megalobrama amblycephala (Wuchang bream) as an important freshwater fish has been cultured for half a century, whereas intensive aquaculture has brought lots of problems, such as crowding and hypoxia. More seriously, the absence of oxygen in water usually causes the large-scale suffocation death of M. amblycephala, thus resulting in great economic losses. In the present study, M. amblycephala were treated by different dissolved oxygen (DO) concentration and hypoxia time to systematically study their responses to hypoxia in physiological and immune levels by evaluating blood parameters. In addition, oxidative stress status in hypoxia was studied by analyzing ROS and related antioxidant enzymes in liver. Finally, the distributions of the stress response proteins in different tissues were also investigated.

In order to provide insight into mechanisms of the hypoxia adaptation in physiology, immune response and oxidative stress in M. amblycephala, some well-known hypoxia indicators, Hb, MetHb, glucose, lactate, glucose, and hepatic glycogen concentrations were investigated after hypoxia stress, and showed the consistency with the previous studies, for instance, Hb and MetHb, functioning in oxygen transport and utilization are induced when the pressure of oxygen decreases in blood (Grek et al., 2011). Our results show that Hb and MetHb are more significantly induced at 6 h for moderate hypoxia treatment, suggesting that Hb and MetHb may be more sensitive to hypoxia time rather than oxygen levels in fish. Na⁺ influx in fish erythrocytes can indicate the activation of Na⁺/H⁺ exchanger and the increase of the pH value, inducing a decrease of hemoglobin affinity to oxygen in fish erythrocytes, as a result the oxygen is delivered into tissues (Nikinmaa, 1992; Reid and Perry, 1994; Nielsen, 1997). In this study, Na⁺ stayed at high levels with the decrease of DO, but decreased at 6 h for hypoxia, indicating that Na⁺ probably flows into M. amblycephala erythrocytes (Figure 1). SDH is involved in both the citric acid cycle and respiratory electron transfer chain (Rutter et al., 2010), and its activity under hypoxia stress was obviously increased in M. amblycephala.

During the hypoxia treatment, glucose increased while the hepatic glycogen content decreased, which has also been reported in fish and other species (Routley et al., 2002; Douxfils et al., 2012; Polak et al., 2013). It is suspected that hepatic glycogen is the primary source of glucose, when animals are subjected to acute hypoxia, glycogenolysis is enhanced and hepatic glucose is released to blood to provide sufficient blood glucose and energy (through anaerobic glycolysis), thus ensure adequate basal metabolism for organism survival (Chen et al., 2007). Therefore, glucose homeostasis during the hypoxia treatment depends primarily on hepatic glucose output (Wahren and Ekberg, 2007).

It is well-known that the ROS increase naturally stimulates the activities of antioxidant enzymes, SOD, CAT, GSH-Px, GST, etc. to remove ROS (Niki, 2010). However, the increase of antioxidant enzyme activities does not mean the ROS generation, or vice versa, especially under combined-factor conditions. Additionally, the antioxidant enzyme levels could only reflect the antioxidant capacity rather than the status of oxidative stress. Thus, the levels of the reactive intermediates, H₂O₂ and ROS are analyzed and the results showed that hypoxia decline H₂O₂ and ROS levels integrally. However, ROS reaches the peak after severe acute hypoxia (DO: 1.0 ± 0.2 mg/L) treatment for 0.5 h, suggesting that anaerobic metabolism has probably taken the place of aerobic metabolism during this DO point since M. amblycephala suffocation point ranges from 0.64 to 0.35 mg/L (Chen et al., 2012). Meanwhile, the lactate level representing anaerobic metabolism also reached the peak, indicating the anaerobic metabolism is occurred (Segura et al., 1996). The activities of the antioxidant enzymes (SOD, CAT), and the H₂O₂ and ROS contents retained at lower levels under hypoxia conditions, which are consistent with some published studies in fish, but contrary with others (Table S1). After careful and scrupulous analysis and comparison, we suggest that the biomarkers of oxidative stress caused by hypoxia and especially co-hypoxia factors probably be primarily focused on ROS itself. Moreover, moderate hypoxia treatment may positively affect M. amblycephala by driving liver into lower oxidative stress status. It is reported that chronic hypoxia leads to a marked improvement in mouse survival, body weight, body temperature, behavior, and enhancement against mitochondrial toxicity (Jain et al., 2016).

It has been known that some eukaryotic mRNAs can be translated via internal initiation by specific mRNA regions termed internal ribosome entry site (IRES) (Holcik et al., 1999; Cornelis et al., 2000). Remarkably, many IRES-containing mRNAs encode proteins that have important roles in development, differentiation, cell cycle progression, cell growth, cell apoptosis, and stress response (Holcik et al., 1999; Henis-Korenblit et al., 2000; Stoneley et al., 2000; Prats and Prats, 2002). The presence of IRES elements allows crucial survival factors to be transiently translated under stress conditions that require immediate changes in protein levels (Hellen and Sarnow, 2001; Holcik and Sonenberg, 2005; Komar and Hatzoglou, 2005). The IRES on 5′-UTR of Hsp70 has been studied to drive cap-independent Hsp70 protein translation in apoptosis, hypertonic and thermal stress (Rubtsova et al., 2003; Hernández et al., 2004; Rocchi et al., 2013; Chen et al., 2014). Our results firstly indicate that Hsp70 could respond to hypoxia stress through IRES-mediated manner, moreover, the Hif-1α and junior Hsp70 expressions in limited tissues indicate that hypoxia stress induce tissue-specific response.

Taken together, we have performed hypoxia/reoxygenation treatment in M. amblycephala and systematically studied their responses to hypoxia in aspect of physiology, oxidative stress, and immune. The results demonstrate that hypoxia significantly induces M. amblycephala hematological and biochemical parameters, and moderate hypoxia drives the liver into a lower oxidative stress status to protect fish from oxidative damage. Furthermore, we have firstly observed that hypoxia-reoxygenation treatment stimulates M. amblycephala innate immune activity.

NC and HLW conceived and designed the experiments. MW, HJW, CXH, XW, YH, BZ, and CHH carried out the experiments and analyzed the data. NC, WW, HL, and HLW supervised the project. NC wrote the manuscript and all authors reviewed the manuscript.