Limited effects of exogenous glucose during severe hypoxia and a lack of hypoxia-stimulated glucose uptake in isolated rainbow trout cardiac muscle

The vertebrate heart must remain active to support whole-animal homeostasis and normal function. As a result, cardiomyocytes depend on an uninterrupted provision of energy substrates – either exogenous or endogenous – for production of ATP to sustain contractile function and the Ca²⁺ regulation governing it. The heart of aquatic ectotherms faces unique metabolic challenges because of the possibility of environmental and myocardial hypoxia. In particular, fishes possess a wide range of tolerances to reduced oxygen levels in water or by cardiac tissue. Similar to the mammalian heart, which fails within minutes during oxygen limitation (Allen et al., 1985; Matthews et al., 1986), athletic fishes, such as the rainbow trout (Oncorhynchus mykiss Walbaum), possess hearts that cannot maintain force development and cellular energy state during hypoxia (Gesser, 1977; Arthur et al., 1992; Driedzic and Gesser, 1994; Hartmund and Gesser, 1996). However, the heart of rainbow trout is capable of surviving severe hypoxia, without irreversible damage, and sustains contractile function at a reduced work rate (Arthur et al., 1992; Faust et al., 2004). At the same time, energetic needs of hypoxic cardiomyocytes are met by increasing anaerobic glycolysis (10-fold) and lactate output (>35-fold) for ATP production (Arthur et al., 1992).

Although the link between cardiac energy metabolism and function is of fundamental significance, the question of what metabolic fuel source is utilized or preferred under different conditions has not been defined in non-mammalian species, and it appears to be species and possibly sex specific. Similar to mammalian cardiac preparations, the use of exogenous glucose as a ‘potential’ energy substrate is a time-honored practice for perfused fish hearts and isolated cardiac tissue. The underlying assumption has been that glucose is beneficial for the support of myocardial ATP production and performance. When perfused hearts from sea raven (Hemitripterus americanus) are aerated (21% O₂) and deprived of exogenous fuel, they exhibit reduced contractile performance within 30 min (Driedzic and Hart, 1984). Inclusion of 10 mmol l⁻¹ glucose in the perfusion medium resulted in improved performance (15%) over preparations without exogenous substrate. In contrast, hearts from dogfish (Mustelus canis), little skate (Raja erinecea), American eel (Anguilla rostrata), Atlantic cod (Gadus morhua) and eelpout (Zoarces viviparus) do not require extracellular glucose to preserve contractile performance under aerobic conditions (Driedzic, 1978; Driedzic and Hart, 1984; Driedzic et al., 1984; Bailey et al., 2000a; Clow et al., 2004). Under aerobic conditions, exogenous glucose accounts for a very minor component (<10%) of energy metabolism of intact (West et al., 1993; Blasco et al., 1996) or isolated hearts of trout (Bailey and Driedzic, 1993). Conversely, physiological levels of exogenous glucose (5 mmol l⁻¹) maintained resting tension and increased developed tension in isolated ventricle strips from immature female and maturing males, respectively (Farrar et al., 2006).

Possibly related to this beneficial effect on myofilament function, glucose may influence the participation of the sarcoplasmic reticulum (SR) in cardiac excitation–contraction (EC) coupling. In American eel ventricle strips, extracellular glucose counteracts increases in resting tension in the face of elevated extracellular Ca²⁺ under fully oxygenated conditions. This effect is eliminated by blocking the release of Ca²⁺ from the SR with ryanodine (Bailey et al., 2000b). Experiments with cat heart muscle also suggest that the efficiency of EC coupling is locally controlled at the SR Ca²⁺ release sites by mechanisms utilizing ATP, produced by glycolytic enzymes closely associated with the release channel (Hüser et al., 2000). Thus, evidence from fishes and mammals demonstrates that an active glycolytic pathway is required to maintain cardiac EC coupling for intracellular Ca²⁺ homeostasis and mechanical performance under aerobic and oxygen-limiting conditions (Driedzic and Gesser, 1994; Gesser, 2002). Furthermore, under aerobic conditions, exogenous glucose appears to enhance SR Ca²⁺ uptake, storage and release in ventricle strips from sexually immature rainbow trout (males and females), but only helps to maintain resting tension in females (Farrar et al., 2006). To what extent exogenous glucose or tissue glycogen fuels glycolytic activity and ATP production in teleost hearts is unknown, although there is evidence for functional compartmentation of glycolysis and glycogenolysis in the isolated working rat heart (Anousis et al., 2004).

In the complete absence of oxygen, maintenance of cardiomyocyte ATP levels (and therefore cardiac performance) must rely solely on anaerobic glycolysis and either exogenous glucose or stored glycogen for metabolic fuel. Whether the ability to use exogenous glucose for anaerobic glycolysis and energy production defines hypoxia tolerance of the teleost myocardium remains unanswered and controversial. In the American eel (a hypoxia-tolerant species), incubation of ventricle strips under severe hypoxia activates glucose uptake (Rodnick et al., 1997) and the presence of exogenous glucose allows ventricle strips to maintain force production during hypoxia (Bailey et al., 2000a). Conversely, exposure of hypoxia-tolerant short-horned sculpin (Myoxocephalus scorpius) to 2.0 mg O₂ l⁻¹ (62.5 μmol l⁻¹) for 240 min did not increase cardiac glucose uptake, and this may be related to the corresponding bradycardia and reduced energy demands of the myocardium in vivo (MacCormack and Driedzic, 2007). It is also unclear whether substrate preference and utilization changes with metabolic state, such as reoxygenation immediately following a bout of severe hypoxia.

The primary objective of the present study was to test the hypothesis that exogenous glucose increases contractile performance of isolated cardiac tissue from rainbow trout subjected to conditions that challenge energy homeostasis and EC coupling. Based on limited evidence for a mechanical benefit of exogenous glucose, a secondary objective was to determine mechanisms and kinetics of glucose uptake in cardiac tissue. Contrary to the proposed hypothesis, our data demonstrate a lack of dependence of contractile performance on exogenous glucose in this species and, for the first time, an ability to modulate cardiac glucose uptake via substrate concentration and reduction in tissue temperature.

Female rainbow trout (~250–550 g) were obtained from a local breeder (Funder-Holme, Jutland, Denmark), transported to the University of Aarhus, Department of Zoophysiology, and maintained in aerated, flow-through freshwater tanks (900 l) at 15±2°C and under a controlled photoperiod (12 h:12 h light:dark cycle). Fish were fed an extruded pellet diet consisting of 44% protein, 29% fat, 12% carbohydrate and 7% ash (available energy 22.1 MJ kg⁻¹, Efico Enviro 920, BioMar A/S, Brande, Denmark) ad libitum. All animals were sexually immature at the time of sampling. Fish care and the sampling protocol were approved by the Institutional Animal Care and Use Committee at the University of Aarhus.

Sexually immature male and female rainbow trout (~300–500 g) were obtained from Clear Springs Foods (Buhl, ID, USA) and transported to the Aquatic Research Facility at Idaho State University. Fish were maintained in 1000 l circular tanks receiving recirculating, filtered, UV-sterilized freshwater at 14±1°C. These fish were also exposed to a 12 h:12 h light:dark cycle and fed extruded pellets (1% of body mass, 46% protein, 21% fat, 18% carbohydrate and 8% ash, available energy 22.5 MJ kg⁻¹, produced by Clear Springs Foods) every other day. All experiments performed at Idaho State University were approved by the university's Institutional Animal Care and Use Committee.

Individual animals in Aarhus were netted, euthanized by a blow to the head, decapitated, and the heart was quickly excised and transferred to ice-cold oxygenated physiological saline in a Petri dish, in which one to four longitudinal myocardial strips were prepared from each ventricle. Each strip represented a complete cross-section of the ventricle and consisted of both compact (epicardial) and spongy (endocardial) tissue. The physiological saline consisted of (mmol l⁻¹): 115 NaCl, 2.5 KCl, 1 MgSO₄, 1 NaH₂PO₄, 1.5 CaCl₂ and 15 NaHCO₃. The saline was equilibrated with 0.5% CO₂ and 99.5% O₂, and pH was 7.6 at 15°C. Ventricular strips [N=56, thickness: 1.4±0.03 mm, assuming a square cross-section, mass: 11.9±0.6 mg (means ± s.e.m.)] were mounted vertically using 3-0 surgical silk; one end was attached to a thin glass rod and the other end to one of the two platinum stimulation electrodes. The second stimulation electrode was positioned close to the upper end of the preparation. Both stimulation electrodes were connected to a Grass SD 9 stimulator (Quincy, MA, USA). The glass rod was connected to a force transducer (Fort 10, World Precision Instruments, Sarasota, FL, USA). Glass, water-jacketed muscle baths contained 50 ml of physiological saline and were maintained at 15 or 23°C with a refrigerated recirculating water bath (Lauda K2 RD, Köningshofen, Germany). Unless stated otherwise, ventricle strips were electrically paced at 0.5 Hz, with pulses of 5 ms and a voltage 1.5 times that eliciting maximal response.

Each preparation was allowed to stabilize for 20 min before tissue length was adjusted with a micrometer screw to the peak of the contractile force–length relationship. Preparations were then left to stabilize for another 30 min before experimentation. Signals from the force transducers were recorded by an MP100 data-acquisition system (Biopac, Santa Barbara, CA, USA) at 200 Hz and analyzed by AcqKnowledge software (Version 3.7.1, Biopac). At the end of each experiment, length and wet mass of each cardiac strip were measured. Force (mN) relative to cross-sectional area (mm²) was estimated assuming a tissue density of 1.06 mg mm⁻³ (Layland et al., 1995) and uniform thickness of the strips.

After stabilization of twitch force in the absence of exogenous substrate, two of four ventricle strips from each of six hearts (run in parallel) received 5 mmol l⁻¹ glucose (d-glucose, unless stated otherwise) and the other two received 5 mmol l⁻¹ mannitol (membrane impermeable control). To determine SR participation in EC coupling, one of each of the two ventricle strips in glucose and mannitol was exposed to 10 μmol l⁻¹ ryanodine. Ryanodine at this concentration provides maximal effect on twitch force development of trout heart preparations (Hove-Madsen, 1992), probably by locking the SR Ca²⁺ release channel in a closed state and rendering the SR ineffective in Ca²⁺ cycling (Rousseau et al., 1987). After 40 min, electrical stimulation was turned off and was then resumed after 5 min to assess post-rest potentiation (Bers, 1985). When force was stable, stimulation rate was elevated from 0.5 to 1.1 Hz for 15 min and then returned to 0.5 Hz. These rates are comparable to heart rates for quiescent and active rainbow trout held at similar temperatures (Wood et al., 1979; Gamperl et al., 2011), and we noted in preliminary trials that some preparations become arrhythmic at frequencies >1.5 Hz (data not shown). After an additional 15 min at 0.5 Hz, a maximally effective dose of adrenaline (10 μmol l⁻¹) was added to all four preparations. Ten minutes later, stimulation rate was elevated again to 1.1 Hz for 15 min and finally returned to 0.5 Hz for 15 min at 0.5 Hz.

To address potential effects of elevated temperature on myocardial substrate preference and cardiac performance, experiments were repeated with ventricle strips from six hearts, except that 30 min after the addition of glucose or mannitol temperature was elevated from 15 to 23°C over 10–15 min and maintained at 23°C for 30 min before the rest of the protocol was applied.

Four strips from each of 10 ventricles were run in parallel. After stabilization of twitch force at a stimulation rate of 0.5 Hz, two ventricle strips received 5 mmol l⁻¹ glucose and the other two 5 mmol l⁻¹ mannitol (controls). Fifteen minutes thereafter, severe hypoxia was imposed for 100 min by replacing O₂ with N₂ in the gas mixture. After 10 min of severe hypoxia, one of the two ventricle strips in glucose and mannitol was exposed to 10 μmol l⁻¹ adrenaline. To assess temporal changes in ventricle strip performance under aerobic conditions, additional preparations were oxygenated continuously and subjected to the same time schedule and treatments as the above experiments (N=4).

To assess the potential importance of glucose and increasing concentrations of adrenaline to tissue mechanical performance during aerobic recovery, severe hypoxia was imposed on paced (0.5 Hz) ventricle strips not receiving exogenous substrate. Five minutes before reoxygenation, three of the four ventricle strips were exposed to 5 mmol l⁻¹ glucose, 10 mmol l⁻¹ glucose, or 5 mmol l⁻¹ mannitol. The fourth ventricle strip received 5 mmol l⁻¹ l-glucose (N=4), 5 mmol l⁻¹ sucrose (N=1) or nothing (N=1) for a total of 6 experiments. After 60 min of reoxygenation, adrenaline was added stepwise to attain final concentrations of 0.1, 0.5, 2 and 20 μmol l⁻¹ with stabilization of force at each concentration. To control for oxygenation state, additional experiments were performed with identical treatments but under aerobic conditions throughout (N=10).

Four ventricle strips from each of six hearts received 2 mmol l⁻¹ of Na-butyrate, Na-acetate or Na-octanoate, or 5 mmol l⁻¹ of glucose after 55 min of severe hypoxia. Five minutes later, N₂ was replaced by O₂ for aerobic recovery and ventricle strips remained oxygenated for 60 min.

Similar to mechanical performance studies performed in Aarhus, ventricle strips were prepared from rainbow trout in Pocatello. Glucose uptake was measured for individual ventricle strips using the glucose analog 2-deoxy d-glucose (2-DG) as described previously for American eel (Rodnick et al., 1997) and Atlantic cod (Clow et al., 2004). 2-DG is transported into cardiomyocytes, phosphorylated by hexokinase and trapped inside the myocyte as 2-deoxyglucose 6-phosphate (2-DG-6-P), allowing for the determination of rates of sugar uptake over extended periods (minutes versus seconds). Uniform ventricle strips (1.6–2.0 mm thick, weighing 14–24 mg) were rinsed in an ice-cold, filtered (0.22 μm) teleost Ringer's solution consisting of (in mmol l⁻¹) 111 NaCl, 5.0 KCl, 1.5 CaCl₂, 1.0 MgSO₄, 0.5 NaH₂PO₄, 10 NaHCO₃ and 5 glucose, pH 7.6 at 14°C. The solution was equilibrated with 0.5% CO₂ and 99.5% O₂ and osmolality was kept constant (~300 mOsmol kg⁻¹ water) by varying the concentration of mannitol, such that the sum of glucose, pyruvate, mannitol and 2-DG was 40 mmol l⁻¹. Select measurements of 2-DG uptake were conducted on individual strips from the same animal at multiple temperatures (4, 14 and 24°C, N=6 for each temperature) and over a range of 2-DG concentrations (1–40 mmol l⁻¹, N=2–5 per concentration) to define thermal sensitivity and kinetics. All solutions were equilibrated in stoppered, 25 ml Erlenmeyer flasks at the appropriate temperature before experiments and serial incubations occurred in a reciprocating water bath (Haake-Fisons SWB20, Karlsuhe, Germany, set at 60 cycles min⁻¹) and two custom flask incubators connected to refrigerated recirculating baths (Neslab RTE 7, Portsmouth, NH, USA).

Ventricle strips were (1) preincubated for 1 h in 2 ml of Ringer's solution containing 0.1% (w/v) bovine serum albumin (BSA; fraction V, fatty acid free) pregassed with either 0.5% CO₂:99.5% O₂ (aerobic, ~640 mmHg P_O2) or 0.5% CO₂:99.5% N₂ (severe hypoxia, 6–10 mmHg P_O2), (2) rinsed for 10 min in Ringer's solution containing 40 mmol l⁻¹ mannitol plus 0.1% BSA, minus glucose to remove glucose from the extracellular space, and (3) incubated for 20 min in 1.5 ml of glucose free Ringer's solution containing 0.1% BSA, 2 mmol l⁻¹ sodium pyruvate (extracellular substrate to support oxidative metabolism in the absence of glucose), 1–40 mmol l⁻¹ 2-DG including 1,2-[³H(N)] deoxy d-glucose (40 Ci mmol⁻¹, specific activity 1.0 μCi ml⁻¹) and d-mannitol-1 [¹⁴C] (53 mCi mmol l⁻¹, specific activity 0.1 μCi ml⁻¹). Incubations were terminated and extracellular space and intracellular 2-DG and 2-DG-6-P were determined as described previously (Rodnick et al., 1997; Clow et al., 2004). Preliminary studies (N=4) demonstrated that 2-DG uptake (at 5 mmol l⁻¹ and 14°C) remains linear for at least 60 min, which is consistent with measurements on ventricle strips from American eel (Rodnick et al., 1997). For comparative purposes, data are reported as μmol 2-DG ml⁻¹ intracellular water 20 min⁻¹ or nmol 2-DG g⁻¹ tissue min⁻¹. Cytochalasin B (25 μmol l⁻¹, final concentration), a nonselective inhibitor of facilitative glucose transport proteins (GLUTs) in rainbow trout (Teerijoki et al., 2001), or excess d-glucose (25–100 mmol l⁻¹, final concentrations) were included in some incubations to determine the relative contribution of GLUTs versus other transmembrane mechanisms (e.g. simple diffusion). Cytochalasin B was dissolved in absolute ethanol [final concentration 0.025% (v/v)] and added to incubation flasks just prior to experiments. Separate controls demonstrated that the presence of ethanol did not affect 2-DG uptake measurements, and incubations with increased cytochalasin B (100 μmol l⁻¹) did not significantly reduce 2-DG uptake compared with 25 μmol l⁻¹ cytochalasin B (data not shown).

To confirm that incubation of ventricle strips under severe hypoxia stimulated anaerobic glycolysis and examine the importance of exogenous glucose uptake for glycolytic metabolism, we included 5 mmol l⁻¹ glucose in the aerobic and hypoxic Ringer's solution instead of 2-DG, and the final incubation was extended from 20 to 60 min. Cytochalasin B (25 μmol l⁻¹) was added to some preparations to inhibit facilitated glucose uptake. Net lactate efflux from ventricle strips (N=9–10 per treatment) was determined by sampling the incubation solution and measuring this metabolite according to the method of Fleischer (Fleischer, 1970) as modified by Battiprolu et al. (Battiprolu et al., 2007).

Unless specified otherwise, chemicals were purchased from Sigma-Aldrich (St Louis, MO, USA). We used several indices of mechanical performance to assess cardiac contractility and relaxation (Hartmund and Gesser, 1996; Gesser, 2002). Twitch force development was normalized (%) to values measured after the initial stabilization period. Although twitch force was calculated in absolute terms and standardized for ventricle strip cross-sectional area (values for 56 Aarhus preparations were 11.1±0.5 mN mm⁻²), relative values were used in assessing the effects of experimental interventions. Rates of rise of contraction and 50% relaxation were also measured from waveforms.

The adrenaline-dependent change in twitch force (ΔF) was calculated as the force just before adrenaline additions subtracted from the twitch force in the presence of adrenaline. The adrenaline concentration (ED₅₀) eliciting half of the maximal adrenaline effect on force was assessed with the Michaelis–Menten equation written as ΔF=ΔF_m–ED₅₀(ΔF/[Adr]), whereby ED₅₀ is provided by the slope of ΔF plotted against ΔF/[Adr]. The change in 2-DG uptake with incubation temperature was quantified using Q₁₀ and the formula Q₁₀=(R₂/R₁)^{[10/(T₂–T₁)]}, where R₁=2-DG uptake at the lower temperature (T₁), and R₂=2-DG uptake at the higher temperature (T₂).

Prior to statistical analysis, a variety of tests were conducted to determine whether the data were normally distributed and justify the use of parametric tests. For normalized measurements of contractile performance (values ranging from above 100% to well below 100%), graphical assessments were performed according to Zar (Zar, 1999). Additional tests for normality included Kolmogorov–Smirnov and Shapiro–Wilk tests (2-DG uptake and dose response data), and Mauchly's test of sphericity (for repeated measurements of lactate efflux and temperature effects on 2-DG uptake). In each case the data were normally distributed. However, it is also worth pointing out that the parametric tests used in the present study are robust enough to be applied to data not severely deviating from normality (Zar, 1999). Comparisons of ventricle strip performance were assessed by either one-way ANOVA with Tukey's post hoc test or Student's t-test. Analyses of 2-DG uptake measurements and net lactate efflux from ventricle strips were initially completed using a mixed model protocol (SPSS software, Version 16.0, Chicago, IL, USA) or repeated measures two-way ANOVA. If no differences were found between male and female fish in Pocatello, the data for both sexes were pooled and analyzed by repeated-measures, one-way ANOVA. Data are expressed as means ± s.e.m. and statistical significance was established at P<0.05.

Preliminary experiments demonstrated that twitch force in the presence of 5 mmol l⁻¹ glucose or 2 mmol l⁻¹ lactate or pyruvate (N=5) did not differ significantly under aerobic conditions at 0.5 and 1.1 Hz stimulation rates (data not shown). Therefore, subsequent aerobic studies focused on possible effects of exogenous glucose on cardiac performance. In the presence of ryanodine and adrenaline, twitch force was similar in ventricle strips receiving glucose and isomolar mannitol, at both 0.5 and 1.1 Hz (Fig. 1A). With exogenous glucose or mannitol, ryanodine significantly depressed twitch force, not only after 5 min of rest, but also in the presence of adrenaline and at the higher stimulation frequency (Fig. 1A,B). The addition of adrenaline resulted in a twofold increase in force development (positive inotropy) at 15°C.

Consistent with studies performed at 15°C, twitch force for ventricle strips at 23°C was comparable with glucose and mannitol under all conditions applied (Fig. 1C). At the warmer temperature, negative effects of ryanodine on twitch force also occurred (Fig. 1C), although ryanodine did not diminish force at the elevated stimulation rate (Fig. 1D). The results suggest that exogenous glucose did not affect contractile performance under aerobic conditions with elevated metabolic demands, in the presence of intact or reduced SR function.

Ventricle strips did not maintain contractile performance during oxygen limitation (Fig. 2A,B). Force development decreased rapidly or gradually in the presence of exogenous mannitol and glucose, respectively. In the absence of added adrenaline, there was a biphasic response in ventricle strips receiving glucose, whereby twitch force was greater compared with controls for approximately 30 min after the onset of hypoxia, but not during the remainder of the hypoxic period. In fact, twitch force was significantly lower with exogenous glucose at the end of the hypoxic exposure (Fig. 2A). Ten minutes after the onset of hypoxia, the addition of adrenaline increased twitch force to that of baseline values under aerobic conditions after 20 min (Fig. 2B). However, independent of the presence of exogenous glucose, and similar to studies without added adrenaline, twitch force decreased progressively during severe hypoxia. Ventricle strips receiving glucose did maintain twitch force at a higher value compared with mannitol-treated ventricle strips at the end of the exposure to N₂. The possibility that degradation of adrenaline limited twitch force production was discounted by the finding that supplemental adrenaline added after 45 min of hypoxia was without effects on force development (data not shown). Furthermore, the adrenaline response did not diminish in similar experiments run with continuous oxygenation instead of hypoxia. In contrast to hypoxic experiments, aerobic controls with added adrenaline displayed limited force decay during the experiment (Fig. 2C).

Following severe hypoxia, the recovery of twitch force during 60 min of reoxygenation was not affected by the addition of exogenous glucose (5 or 10 mmol l⁻¹) compared with controls (Fig. 3A). Increasing concentrations of adrenaline promoted contractile performance in a staircase manner; however, the presence of glucose did not alter the maximal or submaximal responses. Similarly, when ventricle strips were oxygenated continuously, twitch force was identical in the presence or absence of glucose, and in response to incremental additions of adrenaline (Fig. 3B). The presence of glucose did not alter the results under any conditions. In addition, the concentration of adrenaline eliciting half of the maximal effect on twitch force (ED₅₀) ranged from 0.3 to 0.5 μmol l⁻¹, and was not influenced by variations in substrate availability (Table 1). Results obtained with the ventricle preparations exposed to either 5 mmol l⁻¹ l-glucose (N=4), 5 mmol l⁻¹ sucrose (N=1) or no addition (N=1) were similar to the results for the other three groups (data not shown) and support the lack of dependence of adrenaline-induced intropism on exogenous glucose.

When recovering after oxygen deprivation, force development by cardiac muscle from rainbow trout did not increase when exposed to non-carbohydrate compared with carbohydrate substrates. During severe hypoxia, force development by ventricle strips decreased by 80–90% over 60 min (Fig. 4). The recovery of twitch force during reoxygenation was not different between strips receiving glucose, butyrate and acetate, and was decreased with octanoate relative to the recovery with added glucose and acetate after 40 to 60 min of incubation.

Under aerobic conditions, uptake rates for 5 mmol l⁻¹ 2-DG at 14°C were similar for male (1.16±0.08 μmol ml⁻¹ intracellular water 20 min⁻¹, N=6) and female (1.09±0.03 μmol ml⁻¹ intracellular water 20 min⁻¹, N=6; Table 2) ventricle strips from Pocatello rainbow trout, so data for both sexes were combined. Exposure of non-contracting ventricle strips to severe hypoxia for 60 min did not stimulate 2-DG uptake; however, hypoxic exposure increased net lactate efflux by nearly sixfold above aerobic values for male and females (sexes combined, N=9–10; Table 2).

Cytochalasin B inhibited uptake of 2-DG (5 mmol l⁻¹) by 59±3 and 59±2% under aerobic and hypoxic conditions, respectively (Table 2), suggesting that GLUTs are responsible for a significant portion of extracellular glucose uptake by quiescent cardiomyocytes, independent of oxygenation state. It also appears that another uptake mechanism (i.e. passive or simple diffusion) is responsible for the balance (~41%). Similar to cytochalasin B, when 50 or 100 mmol l⁻¹ glucose were added to the Ringer's solution (N=8), uptake of 2-DG was inhibited by 25±8 and 43±6%, respectively (data not shown). These data provide additional evidence for a common, competitive uptake mechanism for both sugars.

Consistent with an incomplete inhibition of 2-DG uptake with cytochalasin B and d-glucose, we observed a linear increase for 2-DG uptake and lack of saturation between 1 and 40 mmol l⁻¹ (Fig. 5). In addition, cytochalasin B inhibited proportionately less of the total 2-DG uptake as 2-DG concentration increased. At 40 mmol l⁻¹ 2-DG, cytochalasin B inhibited just 25% of total 2-DG uptake, suggesting a greater enhancement of simple versus facilitated diffusion at increasing sugar concentrations. At each concentration of 2-DG, 92–95% of the 2-DG taken up by cardiac tissue was phosphorylated to 2-DG-6-P and this was unaffected by the presence of cytochalasin B (N=4, data not shown). Together, these results provide evidence that both facilitated and passive transport mechanisms exist for glucose uptake and cardiac tissue has a high capacity to phosphorylate intracellular 2-DG.

As shown in Fig. 6, when ventricle strips were warmed acutely from the acclimation temperature (14°C) to 24°C, there were no significant changes in total 2-DG uptake (Q₁₀=1.24±0.21). In contrast, reducing temperature from 14 to 4°C decreased overall 2-DG uptake by 45±5% (P<0.01, Q₁₀=1.91±0.18). There were no significant differences between ventricle strips from male and female rainbow trout and data for the sexes were combined. The addition of cytochalasin B decreased 2-DG uptake via facilitated glucose transporters to a similar extent at every temperature tested (51±6% at 24°C, 68±6% at 14°C, and 61±9% at 4°C), and yet the absolute contribution of non-transporter mediated 2-DG uptake was decreased at 4°C compared with 14 and 24°C (Fig. 6).

The initial objective of this study was to examine potential effects of exogenous glucose on contractile performance of the rainbow trout heart under conditions known to alter performance, challenge ion regulation and increase energy demands. The addition of glucose has been a mainstay for heart preparations from fish because of (1) historical use of glucose in buffers for similar mammalian preparations, (2) the presence of glucose in extracellular fluid of fishes, and (3) the assumption that exogenous glucose is required to support energy production, especially during hypoxia. However, the use and benefits of exogenous versus endogenous energy substrates in the teleost heart appear to be species and sex specific with diverse links to contractility and tolerance to varied extracellular conditions (Driedzic, 1992; Battiprolu et al., 2007).

For rainbow trout in Aarhus, glucose did not alter contractile performance when energy demands of oxygenated cardiac muscle were increased due to chronotropic (0.5 to 1.1 Hz), inotropic (adrenaline) or temperature (15 to 23°C) stimuli. As a result, the current findings extend previous results that question the value and use of exogenous substrates as effective fuel for aerobic energy metabolism in the fish heart (Driedzic and Hart, 1984). It appears that isolated cardiac tissue from rainbow trout can be supported principally – if not completely – by endogenous energy stores (i.e. glycogen and or triglyceride) under most conditions. Conversely, in the presence of substrate-free buffer, endogenous substrates are not able to support the higher energy needs of the mammalian heart, even under aerobic conditions (Neely et al., 1970) and increased heart work stimulates the uptake of exogenous glucose (Depre et al., 1999). While it appears that hearts from American eel (Rodnick et al., 1997), Atlantic cod (Clow et al., 2004) and rainbow trout at Aarhus (present study) do not require extracellular glucose to maintain mechanical function in vitro, over extended periods (several hours), low levels of exogenous glucose uptake may be important in rainbow trout to replenish and maintain glycogen reserves and support the oxidative pentose phosphate pathway, and could help explain intraspecific differences in cardiac energy metabolism (Battiprolu et al., 2007) and hypoxia tolerance (Faust et al., 2004).

The EC coupling of the vertebrate heart depends upon trans-sarcolemmal movement of extracellular Ca²⁺ and to a degree that varies among species on an intracellular Ca²⁺ regulation by SR (Tibbits et al., 1990). The majority (60–70%) of energy needs of the mammalian myocardium fuels the myosin ATPase for cross-bridge cycling and the creation of muscle tension. However, ~15% of metabolic energy needs of contracting cardiomyocytes is devoted to pumping Ca²⁺ (two Ca²⁺ ions per ATP consumed) from the cytoplasm into the SR by its Ca²⁺ATPase for mechanical relaxation (Schramm et al., 1994). Several lines of evidence support an active role of the SR in cycling Ca²⁺ during EC coupling in the rainbow trout (reviewed in Shiels and White, 2005). In the present study we used ryanodine to confirm SR participation in cardiomyocyte EC coupling. Ryanodine reduced twitch force development at the first stimulus after a prolonged rest period [post-rest potentiation (Bers, 1985)] and the force developed at 1.1 Hz in the presence of adrenaline at 15°C and in the absence of adrenaline at 23°C – a temperature close to the upper thermal limits of this species (Rodnick et al., 2004). Overall, the current results for Aarhus rainbow trout confirm the importance of the SR as a source of activator Ca²⁺ and, in contrast to a previous study (Farrar et al., 2006), do not support an essential role of exogenous glucose for Ca²⁺ loading/release in the SR of female rainbow trout.

Even under aerobic conditions, glycolysis appears to be crucial for normal EC coupling in the heart of rainbow trout (Gesser, 2002) and mammals (Kockskämper et al., 2005). For example, there is close association between glycolytic enzymes and the SR, and ATP generated from these enzymes may selectively support SR Ca²⁺ uptake in the mammalian heart (Xu and Becker, 1998; Zima et al., 2006). Intermediates and products of glycolysis also modulate the open probability of the RyR and SR Ca²⁺ release (Kockskämper et al., 2005; Zima et al., 2006). However, there is also evidence for glycogenolytic enzymes associated with the cardiac SR (Entman et al., 1977), suggesting that endogenous glycogen might serve as a preferential substrate over exogenous glucose for glycolysis and Ca²⁺ homeostasis.

It is well established that adrenergic stimulation increases contractile performance and energy demands of the vertebrate heart. Conversely, very little is known about how adrenaline might affect cardiac fuel use and energy metabolism in the teleost heart. In the rainbow trout, inotropic effects of adrenaline are due to activation of the β-adrenergic signal-transduction pathway, leading to greater trans-sarcolemmal Ca²⁺ influx (Tibbits et al., 1992; Shiels and Farrell, 1997) and possibly increased Ca²⁺-induced release of Ca²⁺ from the SR. β-Adrenergic stimulation is also thought to accelerate the Ca²⁺ pumping rate of the SR to shorten diastole and accelerate relaxation (Tada et al., 1975). Consistent with previous studies (Gesser et al., 1982; Gesser, 1996), adrenaline resulted in a rapid (<5 min) doubling of twitch contractile force of ventricle strips from Aarhus rainbow trout under aerobic conditions. The expectation is that increased energy consumption with adrenaline will help support the myosin ATPase and Ca²⁺ transport mechanisms to enhance inotropic and lusitropic responses, respectively. However, previous (Gesser et al., 1982; Farrar et al., 2006; Battiprolu et al., 2007) and current data demonstrate that positive inotropic effects of a maximally effective dose of adrenaline occur during oxygen deprivation and recovery after oxygen deprivation, and these effects are independent of the presence of exogenous glucose. An additional study using iodoacetate to block glycolysis showed that inotropic responses to adrenaline under aerobic conditions do not require an active glycolytic pathway (Farrar et al., 2006). Together, these results suggest that aerobic and anaerobic metabolic pathways are capable of supporting adrenaline-induced changes in mechanical performance of the rainbow trout heart without exogenous glucose.

We also considered the possibility that exogenous energy sources may be important for adrenergic sensitivity of the increased contractile response. The concentration of adrenaline eliciting half of the maximal inotropic effect on force production in ventricle strips from Aarhus rainbow trout (ED₅₀, 0.33–0.55 μmol l⁻¹) is similar to values reported for (1) adrenaline (0.28 μmol l⁻¹) stimulation of CO₂ production from glucose in isolated cardiomyocytes from rainbow trout (Milligan, 1991) and (2) adrenaline (0.46 μmol l⁻¹)-induced glucose release from hepatocytes of fed rainbow trout at similar acclimation temperatures (McKinley and Hazel, 1993), and is within the range of resting and stress levels of adrenaline (1 nmol l⁻¹ to 1 μmol l⁻¹) for this species (Tetens et al., 1988). Our results showing that exogenous glucose did not alter sensitivity of ventricle strips to adrenaline do not discount the possibility that overall cardiac substrate availability (exogenous and endogenous) modulates adrenergic control of inotropy and lusitropy, and non-mechanical functions in vivo.

Assuming that oxygen is not limiting, the majority of ATP production to support heart performance in fish comes from aerobic metabolism (Driedzic et al., 1983). Salmonid fishes have been labelled as ‘hypoxia sensitive’ and respond to environmental hypoxia with a rapid reflex bradycardia and yet maintain cardiac output because of an increase in stroke volume (Wood and Shelton, 1980). The heart must rely on anaerobic glycolysis for energy production as oxygen supply becomes limiting. Importantly, environmental hypoxia increases plasma levels of glucose (Van Raaij et al., 1996) and adrenaline (Tetens and Christensen, 1987) in rainbow trout. This combination could potentially increase cardiac glucose uptake and utilization via concentration- or hormonal-dependent mechanisms. In the present study, the transition from aerobic to severe hypoxic conditions resulted in a gradual decline of twitch force development to just 10–20% of original values, which agrees with previous studies on this species (Gesser et al., 1982; Hartmund and Gesser, 1992; Overgaard and Gesser, 2004; Battiprolu et al., 2007). Also consistent with previous findings for various teleost species [American eel (Bailey et al., 2000a) and Atlantic cod (Clow et al., 2004)] was the observation that exogenous glucose reduced the decline in twitch force production in ventricle strips during exposure to severe hypoxia (Fig. 2). However, in the absence of added adrenaline we observed a biphasic response, with early contractile performance in the presence of glucose above that of the controls, whereas it fell below controls by the end of experiments. This response may reflect enhanced ATP production from extracellular glucose and glycolysis (after 15–45 min) followed by inhibition of glycolytic activity and/or reduced glycogenolysis (later, after 100 min).

In the presence of adrenaline, twitch force development was enhanced, but this gain was not significantly influenced by extracellular glucose. A likely explanation for the reduced influence of extracellular glucose on contractile function during adrenergic stimulation involves the increased provision of glycolytic substrate (glucose and G6P) via intracellular glycogenolysis. Furthermore, regardless of oxygenation state, lactate efflux from non-contracting ventricle strips did not change when cytochalasin B was present (Table 2), indicating that facilitative uptake of exogenous glucose is not vital for increasing glycolytic flux under these conditions.

While several studies demonstrate that exogenous glucose improves contractile function in the teleost heart during hypoxia (Driedzic and Gesser, 1994; Bailey et al., 2000a; Gesser, 2002; Clow et al., 2004), it was not known whether various exogenous substrates could help support recovery of mechanical function during reoxygenation following severe hypoxia. A previous study demonstrated that glucose was beneficial for rainbow trout ventricle strips during reoxygenation following an extended period (120 min) of severe hypoxia and adrenaline exposure in the absence of exogenous glucose in order to deplete glycogen (Gesser, 2002). Early studies on hearts from frogs (Clark et al., 1938) and turtles (Reeves, 1963) indicate that the use of exogenous glucose may require a high rate of glycogenolysis and/or significant glycogen depletion. Results from the present study indicate that ventricle strips deprived of exogenous substrate recover contractile function following 60 min of severe hypoxia, and the addition of various substrates during reoxygenation had no (acetate, butyrate or glucose) or even negative (octanoate) inotropic effects. Whether more prolonged periods of hypoxia – and significant glycogen depletion – could shift fuel selection in favor of exogenous substrates during reoxygenation will require further investigation. In contrast to the present findings, exogenous glucose helps the recovery of hypoxia-induced contractile depression and sarcolemma integrity in isolated feline cardiac muscle (Burton et al., 1980).

Given the limited effects of exogenous glucose on mechanical performance or post-hypoxic recovery of rainbow trout cardiac tissue, a second objective of this study was to define the mechanism(s) and kinetics of glucose uptake in non-contracting ventricle strips. Under aerobic conditions, the uptake of 2-DG (~30 nmol g⁻¹ min⁻¹; Table 2) in quiescent ventricle strips from Pocatello rainbow trout was comparable to values reported for American eel (Rodnick et al., 1997), Atlantic cod (Clow et al., 2004) and rates of glucose utilization in perfused, normoxic rainbow trout hearts (West et al., 1993), but greater than in vivo cardiac glucose utilization [5 nmol g⁻¹ min⁻¹ (West et al., 1993)]. In the present study, the majority of intracellular 2-DG was phosphorylated to 2-DG-6-P, suggesting that hexokinase activity was greater than transmembrane transport and glucose transport was a rate-limiting step for glucose utilization. These data agree with previous measurements of maximal hexokinase activity (12 μmol g⁻¹ min⁻¹) in ventricles from Pocatello rainbow trout (Battiprolu et al., 2007), which exceed 2-DG uptake rates by 400-fold. Conversely, in vivo studies of brown trout, Salmo trutta (Blasco et al., 1996), and short-horned sculpin, Myoxocephalus scorpius (MacCormack and Driedzic, 2007), report the accumulation of much more cardiac 2-DG than 2-DG-6-P and a greater importance of hexokinase over glucose transport in controlling glucose uptake. Similar to in mammals, the control of cardiac glucose utilization in rainbow trout and other teleosts may vary between enzymatic steps depending on experimental conditions (Kashiwaya et al., 1994).

The importance of facilitated glucose transport across the sarcolemma for rainbow trout cardiomyocytes was suggested by a major cytochalasin B-sensitive component of 2-DG uptake (Fig. 5) and is consistent with similar studies of American eel (Rodnick et al., 1997) and Atlantic cod (Clow et al., 2004). The current finding of non-carrier-mediated glucose uptake in rainbow trout ventricle strips also agrees with studies of American eel and Atlantic cod hearts, and other tissues of rainbow trout [erythrocytes (Tse and Young, 1990) and hepatocytes (Mannerström et al., 2003)]. The existence of multiple mechanisms for trans-sarcolemmal permeability of glucose raises important new questions about the regulation of cardiac glucose utilization and the possibility that glucose uptake is determined primarily by extracellular conditions and not cardiac performance in rainbow trout.

In contrast to previous studies on hearts from American eel (Rodnick et al., 1997) and Atlantic cod (Clow et al., 2004), our results are the first to demonstrate an inability of severe hypoxia and elevated glycolytic activity to stimulate 2-DG uptake in rainbow trout cardiac muscle in vitro. It has been known for over 50 years that oxygen limitation stimulates membrane transport and intracellular phosphorylation of glucose in isolated rat hearts (Morgan et al., 1959). Constraints involving the stimulation of trans-sarcolemmal transport of extracellular glucose could define a fundamental difference between the rainbow trout, some hypoxia-tolerant teleosts, and mammals. Although GLUTs have been identified in striated muscle from trout (Planas et al., 2000; Teerijoki et al., 2000), altered intracellular trafficking may limit sarcolemmal permeability of glucose in rainbow trout compared with mammals (Díaz et al., 2007). It is noteworthy that limiting the increase of glucose transport and glycolytic energy production by cardiomyocytes during hypoxia could be advantageous by preventing the accumulation of metabolic end products (lactate and protons) in the cytosol, which is consistent with reduced cardiac function in rainbow trout. It is also possible that our in vitro preparations lack essential extracellular factors to stimulate or inhibit the utilization of exogenous substrates for energy production and improved contractile performance of rainbow trout cardiac muscle.

The present study provides new and compelling information about the concentration and temperature dependence of glucose uptake in the rainbow trout ventricle. The importance of sugar concentration is not surprising given the presence of a cytochalasin B-insensitive component to glucose uptake and may highlight an opportunistic strategy to increase the utilization of extracellular glucose. While severe hypoxia, per se, did not stimulate 2-DG uptake in rainbow trout ventricle strips, hypoxic hearts will be exposed to elevated extracellular glucose levels in vivo (Van Raaij et al., 1996), and a more favorable concentration gradient should increase membrane transfer of glucose proportionately. In contrast to varying exogenous substrate levels, an acute increase in temperature (14 to 24°C) did not affect overall 2-DG uptake (at 5 mmol l⁻¹) in ventricle strips. This indicates a lack of thermal sensitivity as fish approach their physiological limits and may necessitate the use of other energy substrates (e.g. fatty acids or glycogen) at proportionately higher rates to increase cardiac ATP production. However, our data show temperature dependence of 2-DG uptake between 14 and 4°C, which may parallel tissue oxygen consumption requirements over this range and reflect a phase transition of membrane lipids, reduced membrane fluidity and/or reduced intrinsic activity of membrane-bound glucose transporters (Whitesell et al., 1989).

In the absence of increasing rates of extracellular glucose uptake, glycogen is the primary substrate for elevating glycolytic flux (Newsholme and Leech, 1983), especially during severe hypoxia (Reeves, 1963). A possible explanation for differences in cardiac use of exogenous glucose in rainbow trout (Battiprolu et al., 2007) and a lack of stimulation of 2-DG uptake in the present study may be that high glycogen levels in ventricle strips promote glycogenolysis, which in turn elevates intracellular glucose-6-phosphate, inhibits hexokinase and limits glucose uptake. The importance of glycogen in regulating glucose uptake in mammalian skeletal muscle stimulated by contractions and hypoxia has been documented (Hespel and Richter, 1990; Reynolds et al., 1998) and there may be a threshold for glycogen below which cardiomyocyte uptake and utilization of exogenous glucose is increased in rainbow trout. During hypoxia, the preferential use of glycogen could also be related to a sizable energetic advantage (50%, three versus two ATP per glucose unit) over exogenous glucose for anaerobic glycolysis (Newsholme and Leech, 1983).

In summary, this study demonstrates that despite greater energetic needs (increased frequency of contraction, temperature and inotropism) and glycolytic requirements (severe hypoxia), isolated cardiac tissue from rainbow trout can rely predominantly on intracellular energy stores and maintain mechanical performance independent of exogenous energy substrates. Depending on the timing of measurements, exogenous glucose can have beneficial and negative effects on contractility during severe hypoxia, and the addition of various exogenous substrates did not impact mechanical recovery following oxygen deprivation or the sensitivity to adrenaline. Rainbow trout use a combination of facilitated glucose transport and simple diffusion for 2-DG uptake in cardiac tissue, and yet severe hypoxia or elevated temperature does not increase 2-DG uptake under basal conditions. Ultimately, the extracellular concentration of glucose may be a key determinant of myocardial glucose metabolism in this species, and glucose uptake is compromised at cold temperatures.

The authors extend a special thanks to Dr William Driedzic at Memorial University for valuable comments and suggestions, and to technician Dorte Olsson at the University of Aarhus for excellent assistance.