During hibernation, animals cycle between periods of torpor, during which body temperature (Tb) and metabolic rate (MR) are suppressed for days, and interbout euthermia (IBE), during which Tb and MR return to resting levels for several hours. In this study, we measured respiration rates, membrane potentials, and reactive oxygen species (ROS) production of liver and skeletal muscle mitochondria isolated from ground squirrels (Ictidomys tridecemlineatus) during torpor and IBE to determine how mitochondrial metabolism is suppressed during torpor and how this suppression affects oxidative stress. In liver and skeletal muscle, state 3 respiration measured at 37°C with succinate was 70% and 30% lower, respectively, during torpor. In liver, this suppression was achieved largely via inhibition of substrate oxidation, likely at succinate dehydrogenase. In both tissues, respiration by torpid mitochondria further declined up to 88% when mitochondria were cooled to 10°C, close to torpid Tb. In liver, this passive thermal effect on respiration rate reflected reduced activity of all components of oxidative phosphorylation (substrate oxidation, phosphorylation, and proton leak). With glutamate + malate and succinate, mitochondrial free radical leak (FRL; proportion of electrons leading to ROS production) was higher in torpor than IBE, but only in liver. With succinate, higher FRL likely resulted from increased reduction state of complex III during torpor. With glutamate + malate, higher FRL resulted from active suppression of complex I ROS production during IBE, which may limit ROS production during arousal. In both tissues, ROS production and FRL declined with temperature, suggesting ROS production is also reduced during torpor by passive thermal effects.
- hydrogen peroxide
- top-down elasticity analysis
- oxidative phosphorylation
- oxidative stress
hibernation in small mammals involves repeated bouts of torpor, during which metabolic rate (MR) and body temperature (Tb) decline to low levels [typically 5% of basal metabolic rate (BMR) and 5°C, respectively] for several days (31). Torpor bouts are spontaneously interrupted by periods of interbout euthermia, during which MR rapidly returns to typical resting levels, and Tb returns to ∼37°C for several hours (31). The suppression of MR during torpor is not simply a consequence of lowered Tb but rather involves an active, regulated inhibition (38). Many studies have investigated whether active inhibition of mitochondrial respiration contributes to this MR suppression, since mitochondria are responsible for up to 90% of whole-animal oxygen consumption and have considerable control over cellular energy-demanding processes (63). Respiration rates of isolated mitochondria measured at 37°C are lower in ground squirrels during torpor in liver but not skeletal muscle compared with summer active controls (33, 50). However, there are clear indications of seasonal (i.e., summer vs. winter) metabolic cycles in hibernators that are distinct from the torpor-arousal cycle (16, 51). In addition, hibernators do not feed throughout the winter but do feed throughout the summer, which could have effects on mitochondrial properties (11, 67). Therefore, studies using summer active animals as controls cannot directly distinguish metabolic changes that occur in mitochondria during the transition from torpor to euthermia from unrelated seasonal and dietary effects.
Recent work from our laboratory has shown that ground squirrel liver mitochondrial respiration changes dramatically during the transition to/from torpor and interbout euthermia (3, 20). State 3 respiration rates measured at 37°C were nearly 85% lower during torpor compared with interbout euthermia. This suppression was also observed early during entrance into torpor (i.e., when Tb was still 30°C) but was only partially reversed by late arousal from torpor (i.e., even though Tb had risen to 30°C), a so-called “fast in, slow out” pattern of mitochondrial suppression. However, these studies examined only liver mitochondria, used only succinate (a substrate for complex II) to fuel respiration, and measured respiration at only 37°C. Therefore, the first objective of the present study was to compare mitochondrial respiration rates between torpor and interbout euthermia over a range of physiologically relevant temperatures (37°C, 25°C, and 10°C) in both liver and skeletal muscle of hibernating thirteen-lined ground squirrels (Ictidomys tridecemlineatus) using both glutamate + malate (a complex I-linked substrate) and succinate. For liver mitochondria, in which mitochondrial metabolic suppression during torpor has been previously well established, we further employed top-down elasticity analysis (8, 9) to determine which components of oxidative phosphorylation [substrate oxidation, ADP phosphorylation, and/or (IMM) proton leakiness] contribute to reduced mitochondrial respiration rates during torpor, and to what extent. Two previous studies of mitochondrial metabolism in hibernating animals examined the kinetics of some oxidative phosphorylation components (4, 33), but these studies used summer active or nonhibernating animals as controls and did not examine all three components of oxidative phosphorylation simultaneously, which precludes any comprehensive analysis of the regulation of mitochondrial respiration.
During torpor, blood flow to a number of tissues declines so dramatically (17, 29) that these tissues would be considered ischemic. During arousal from torpor, these tissues become reperfused (17, 40, 57). In nonhibernating mammals, ischemia-reperfusion (I-R) is associated with considerable tissue injury (24, 64), but hibernators appear to be resistant to the deleterious effect of repeated I-R insults, particularly during the hibernation season (23, 41, 74). One of the mechanisms by which I-R is thought to cause injury is via increased mitochondrial reactive oxygen species (ROS) production (6, 39, 62). ROS (e.g., superoxide, H2O2, and hydroxyl radical) cause oxidative injury to cellular macromolecules, particularly DNA, proteins, and fatty acids (27, 58). Therefore, it seems likely that hibernating animals employ mechanisms to mitigate oxidative damage during ischemia (i.e., torpor) and/or reperfusion (i.e., interbout euthermia).
A number of studies have shown increased levels of some antioxidants during torpor and/or arousal in hibernators (18, 26, 45, 47, 52–54, 56, 59, 68), which could potentially be a mechanism to prevent oxidative damage during hibernation. In contrast, few studies have examined changes in mitochondrial ROS production during hibernation and have so far used only indirect methods, including plasma uric acid levels and markers of oxidative damage (54–56). These studies suggest that, at least in some tissues, ROS production may increase during arousal and/or interbout euthermia. Therefore, the second objective of the present study was to measure ROS production rates in vitro at physiologically relevant temperatures, using the same liver and skeletal muscle mitochondria used to assess respiration rates and the components of oxidative phosphorylation. ROS production was measured using both glutamate + malate and succinate (in the presence of rotenone), allowing assessment of ROS production rates at both complex I and III, the principle sites of ROS production in mitochondria (44, 70). This experimental design allows us to assess how changes to oxidative phosphorylation (via active inhibition and passive thermal effects) may affect ROS production in torpor and interbout euthermia.
MATERIALS AND METHODS
This project was approved by the local Animal Use Subcommittee (protocol 2008–055-06) and conformed to the guidelines of the Canadian Council on Animal Care. Thirteen-lined ground squirrels were either live-trapped in Carman, Manitoba, Canada (49°30′N, 98°01′W) or bred in captivity, according to established protocols (71). Both male and female individuals were used. Animals were housed individually in plastic cages (26.7 × 48.3 × 20.3 cm) and provided with corn-cob bedding, paper towel (for nest building), and a transparent red polycarbonate tube (for enrichment; 8 × 15 cm; BioServ, Frenchtown NJ). They were housed at 22°C ± 3°C with photoperiod adjusted weekly to match that of Carman, Manitoba. Food (Lab Diet 5P00), and tap water were provided ad libitum, with sunflower seeds (∼10) being provided three times per week. Summer active animals were acclimated to these conditions for at least 12 wk prior to being sampled in August. Body mass was measured weekly throughout the summer, as well as at the time of sampling. In October, animals began to hibernate and were moved to an environmental chamber and maintained at 4°C ± 2 on a 2:22-h light-dark photoperiod (lights on at 0800 EST). We continued to provide water ad libitum, but food was withheld after individual animals displayed 1 wk of uninterrupted torpor. Torpid and interbout euthermic animals were sampled throughout January and February. Summer active, interbout euthermic, and torpid animals were all sampled at the same time of day (0800–1000 EST).
Whole-animal metabolic rate and core body temperature.
Tb was measured in all animals via radiotelemetry, as in our previous studies (50). Torpid animals were sampled when Tb had been at or below 5°C for at least 72 h, and interbout euthermic animals were sampled following spontaneous arousal from torpor when Tb had been 37°C ± 1 for at least 3 h. Whole animal MR was measured simultaneously in some animals via flow-through respirometry, as in our previous studies (50).
As required by animal care protocols, torpid animals were euthanized by cervical dislocation, but summer active and interbout euthermic animals were euthanized by anesthetic overdose (Euthanyl, 270 mg/ml, 0.2 ml/100 g). Recent studies have justified the use of different euthanasia methods with respect to effects on mitochondrial properties (12, 13).
Purified liver mitochondria were isolated, as described previously (3). The liver was removed, immediately transferred to ice-cold liver homogenization buffer (LHB; in mM): 250 sucrose, 10 HEPES, 1 EGTA, pH 7.4 at 4°C, containing 1% BSA and cut into small (1 mm3) pieces using scissors. These liver pieces were homogenized using three passes of a loose-fitting Teflon pestle in a glass mortar, and filtered through a layer of cheesecloth. The filtered homogenate was centrifuged at 1,000 g for 10 min at 4°C, and the supernatant was centrifuged again in the same way. The resulting supernatant was centrifuged at 8,700 g for 10 min at 4°C, and the pellet was resuspended in 1 ml of LHB without BSA. This crude mitochondrial suspension was layered on top of a Percoll gradient containing 10 ml each of 10, 18, 35, and 70% Percoll (made in LHB) and centrifuged at 13,500 g for 35 min at 4°C. Mitochondria accumulated at the boundary between the 30% and 70% layers were removed, resuspended in LHB without BSA, and centrifuged at 8,700 g for 10 min at 4°C to remove residual Percoll. The mitochondrial pellet was resuspended again in LHB without BSA and centrifuged again in the same manner. This procedure results in a mitochondrial pellet of high purity (3) and was used for all measurements of respiration rate, oxidative phosphorylation kinetics, and ROS production. Liver mitochondrial yield was 8.3 ± 0.4 mg protein/ml and did not differ among metabolic states.
Crude skeletal muscle mitochondria were isolated following a protocol modified from Bhattacharya et al. (7) and purified using the method of Yoshida et al. (73). Muscle tissue from both hind-limbs was excised and washed in ice-cold muscle homogenization buffer (MHB; 100 mM sucrose, 10 mM EDTA, 100 mM Tris-HCl, 46 mM KCl, pH 7.4 at 4°C). Fat, connective tissue, nerves, and hair were removed. The remaining muscle tissue was decanted and suspended in 9 volumes of MHB with protease (from Bacillus licheniformis, 5 mg/g wet muscle mass; Sigma, St. Louis, MO) and minced with fine scissors. After 5 min, muscle tissue was homogenized with three passes of a loose-fitting Teflon pestle in a glass mortar. The homogenate was then incubated on ice for 5 min, followed by further homogenization with three passes of a tight-fitting Teflon pestle in a glass mortar. The resulting homogenate was filtered through one layer of cheesecloth and centrifuged at 2,000 g for 10 min at 4°C. The supernatant was filtered through four layers of cheesecloth and centrifuged at 10,000 g for 10 min at 4°C. The pellet was suspended in 5 ml MHB with BSA (0.5%) and centrifuged again at 10,000 g for 10 min at 4°C. This pellet was then resuspended in 5 ml MHB, and this raw mitochondrial suspension was layered on top of 5 ml of 60% Percoll solution (made in MHB) and centrifuged at 21,000 g for 1 h at 4°C. Purified muscle mitochondria accumulated at the boundary between the MHB and 60% Percoll solution and were removed and suspended in MHB. To remove residual Percoll, this suspension was centrifuged at 21,000 g for 10 min at 4°C. This washing step was repeated three times. The final pellet was used for all measurements of respiration rate and ROS production, but low yields (2.8 ± 0.4 mg protein/ml for interbout euthermia; 1.4 ± 0.3 mg protein/ml for torpor) did not permit measurements of membrane potential.
Mitochondrial respiration rate and kinetics of oxidative phosphorylation.
Mitochondrial respiration rates and membrane potentials (ΔΨm) were determined using a high-resolution respirometer equipped with a tetraphenylphosphonium (TPP+)-selective electrode and an MI-401 micro-reference electrode (O2k-MiPNetAnalyzer, Oroboros, Innsbruck, Austria). Oxygen electrodes were calibrated to air-saturated buffer and oxygen-depleted buffer (obtained by addition of yeast suspension) using published oxygen solubilities (28), corrected for local atmospheric pressure.
Unless otherwise noted, all compounds were dissolved in MiR05 assay buffer (in mM): 110 sucrose, 0.5 EGTA, 3 MgCl2, 60 K-lactobionate, 20 taurine, 10 KH2PO4, 20 HEPES, pH 7.1 at 30°C, 1% BSA; 34. For glutamate + malate-fueled respiration, mitochondria (∼0.1 mg protein liver; ∼0.03 mg protein skeletal muscle) were added to 2 ml of MiR05 assay buffer equilibrated to 37°C, 25°C, or 10°C. Glutamate (10 mM) and malate (2 mM) were added to stimulate state 2 respiration, and state 3 respiration was achieved by adding ADP (0.2 mM). Conditions approximating state 4 respiration were subsequently achieved by adding oligomycin (2 μg/ml, dissolved in ethanol). Succinate-fueled respiration of skeletal muscle mitochondria was measured in the same way, except that succinate (10 mM; in the presence of rotenone, 2 μg/ml, dissolved in ethanol) was used instead of glutamate + malate.
State 3 and 4 respiration rates for succinate-fueled liver respiration were taken from the ADP phosphorylation and proton leak kinetic curves, respectively. To measure ADP phosphorylation kinetics, mitochondrial oxygen consumption and ΔΨm were measured simultaneously. ΔΨm was measured as an approximation of Δp. Nigericin (which can convert ΔpH into ΔΨm, such that Δp is expressed entirely as ΔΨm) was not used in this study, as ΔΨm is a dominant component of Δp and does not change appreciably, as we have discussed previously (13). The chamber of the respirometer was filled with 2 ml of MiR05 assay buffer equilibrated to 37°C, 25°C, or 10°C, and both the TPP+ electrode (filled with 10 mM TPP+) and reference electrode were introduced into the chamber. The TPP+ electrode was calibrated with five additions of TPP+: the first addition increased TPP+ concentration in the chamber to 1 μM, and each subsequent addition increased TPP+ concentration by 0.5 μM. Therefore, the total concentration of TPP+ present in the chambers during kinetic measurements was 3 μM. Once calibrated, liver mitochondria (0.1–0.3 mg protein) were added to the chamber. Subsequently, rotenone (2 μg/ml, dissolved in ethanol) and succinate (10 mM) were added to the chamber to stimulate state 2 respiration rate, and ADP (1 mM) was added to stimulate state 3 respiration. This amount of ADP was sufficient to support state 3 respiration for the duration of the measurement period. Substrate oxidation activity was titrated by up to five additions of malonate (0.5 mM each). Finally, carbonyl cyanide m-chlorophenylhydrazone (CCCP; 0.1 μM, dissolved in ethanol) was added to completely uncouple respiration and dissipate ΔΨm, allowing for later correction of any electrode drift. IMM proton leak kinetics was measured in a similar manner as described above. The chamber was filled with 2 ml of MiR05 assay buffer, and the TPP+ electrode was calibrated. Then, mitochondria (0.1–0.3 mg/ml) were added to the chamber, followed by rotenone and oligomycin (10 μg/ml, dissolved in ethanol). Subsequently, succinate was added to stimulate state 2 respiration (approximating state 4 respiration). Substrate oxidation was titrated by additions of malonate, and CCCP was added to allow for correction of electrode drift. ADP phosphorylation and proton leak curves were fit to a three-paramater exponential growth equation (using SigmaPlot 2001), as this model routinely fit the data considerably better than a straight line. Proton leak kinetics curves were also used to correct the ADP phosphorylation kinetics curve for the contribution made by proton leak-dependent respiration (13). This was accomplished by extrapolating rates of proton leak (using the proton leak kinetics curve) for all values of ΔΨm on each ADP phosphorylation kinetic curve and subtracting this value from the measured respiration rate. In some cases (particularly for torpid mitochondria), this led to negative values of respiration rate at low values of ΔΨm. This may reflect that this approach overestimates proton leak at low values of ΔΨm, or it may reflect a net hydrolysis of ATP (by reversal of ATP synthase; Ref. 30) in mitochondria in which the electron transport chain has been substantially suppressed, as ATP hydrolysis contributes to “proton leakiness” but would not be detected during measurements of proton leak because oligomycin inhibits ATP synthase. The kinetics of substrate oxidation were not measured directly but were rather approximated as the linear relationship between state 3 (from ADP phosphorylation curves, uncorrected for proton leak) and state 4 (from proton leak curves), as in our previous studies (12, 13). TPP+ measurements were corrected for the effects of compounds used on TPP+ concentration readings in the absence of mitochondria, as well as dilution effects caused by the addition of substrates and inhibitors. ΔΨm was calculated from TPP+ concentration using a modified Nernst equation described in Labajova et al. (42).
Mitochondrial reactive oxygen species production.
Mitochondrial ROS production was measured using a protocol that detects H2O2 released from the mitochondria into the surrounding medium. While ROS that are produced toward the cytosol/surrounding medium would be completely detected by our assay, ROS that are produced toward the mitochondrial matrix would be only partially detected because they would be first subjected to degradation via endogenous mitochondrial antioxidants (e.g., glutathione peroxidase). Therefore, our assay underestimates total mitochondrial ROS production and has been elsewhere described as mitochondrial ROS “release” rate (22, 65, 66). Nevertheless, ROS “release” rate is a physiologically important parameter, as it is likely that ROS that are not degraded by mitochondrial antioxidants are most likely to contribute to cellular oxidative stress.
Mitochondrial ROS production was determined using peroxidase and homovanillic acid, following established protocols (5, 14), except that the concentration of peroxidase was increased to 20 U/ml, so that peroxidase activity did not limit ROS detection at low temperatures. The reaction was carried out in 1.7-ml centrifuge tubes, immersed in a temperature-controlled water bath. ROS production was measured under nonphosphorylating conditions (i.e., without the addition of ADP, approximating state 4) at 37°C, 25°C, and 10°C using either glutamate + malate (concentrations as above for mitochondrial respiration) or succinate (with rotenone). Rates of ROS production were linear over the time course examined and were corrected for background rates measured without substrates. Basal ROS production rates were measured in the absence of ETC inhibitors acting downstream of the ROS-producing sites, whereas maximal ROS production rates were measured in the presence of antimycin A (5 μg/ml, dissolved in ethanol) and rotenone, for succinate and glutamate + malate, respectively. Free radical leak (FRL) was calculated using the equation from Barja et al. (5).
In all cases, values presented are means ± SE; n = 8 for summer active and torpid animals, and n = 5 for interbout euthermic animals, except for MR values, where n = 2 for summer active animals and n = 4 for torpid and interbout euthermic animals. For each tissue, separately, we used a repeated-measures mixed-model (PROC MIXED in SAS 9.2) to test the effects of metabolic state and temperature on mitochondrial respiration rates, ROS production rates, and FRL. FRL values, which are expressed as percentage values, were arcsine transformed prior to analysis to ensure normal distribution for statistical analysis. We used a two-way ANOVA (PROC GLM in SAS 9.2) to test differences in Q10 values between mitochondrial respiration and ROS production rates within a particular metabolic state, as well as differences in either Q10 value between metabolic states. In all cases, nonsignificant interactions were dropped from the models. When significant interactions were detected, we used the slicing procedure in SAS 9.2 to examine simple effects (i.e., differences in one factor at fixed levels of the second factor). Tukey's test was used for post hoc analyses where necessary.
Differences in kinetics between metabolic states were evaluated using a Monte Carlo-based approach, as described in Brown and Staples (12). This approach evaluates differences in kinetic curves only at state 3 and 4 values (i.e., it does not determine whether the two curves differ over their entire length), which permits us to determine how each component of oxidative phosphorylation contributes to changes in mitochondrial respiration among metabolic states. Partial integrated responses, which indicate the quantitative change in respiration rate brought about by changes in activity in a particular component of oxidative phosphorylation, were calculated on the basis of equations provided in Hafner et al. (37) and Ainscow and Brand (1).
P values for all analyses are indicated on the appropriate figures and tables, and P values have been interpreted as suggested by Curran-Everett and Benos (21), with P < 0.05 suggesting a real effect and P < 0.10 suggesting a possible effect.
Whole-animal characteristics of ground squirrels during summer and hibernation.
By the time of sampling, body mass of hibernating animals (both torpid and interbout euthermic; 146.0 g ± 6.7) had dropped to levels 45% lower than observed in summer active animals (264.8 g ± 18.4). Hibernating animals exhibited a typical pattern of changes in Tb and MR over the hibernation season (Fig. 1). During torpor, Tb and MR were typically 3.8°C ± 0.1 and 0.21 ml O2·g−1·h−1 ± 0.03, respectively. During interbout euthermia, steady-state Tb was 37.1°C ± 0.9 and MR was 4.43 ml O2·g−1·h−1 ± 1.5. Although Tb of summer active animals (36.0°C ± 1.1) did not differ from that of interbout euthermic animals, MR of summer active animals (1.4 mlO2·g−1·h−1) was considerably lower.
Liver and skeletal muscle mitochondrial respiration rate.
In liver, when measured at 37°C, state 3 respiration rate of mitochondria respiring on glutamate + malate did not differ between interbout euthermia and torpor; however, at both 10°C and 25°C, state 3 was ∼60% lower in torpid animals (Fig. 2A). By contrast, state 4 respiration rate with glutamate + malate did not differ between metabolic states at any assay temperature (Fig. 2B). When succinate fueled respiration, state 3 respiration rate was up to 77% lower in torpid animals at 37°C and 25°C, but no difference between metabolic states was observed at 10°C (Fig. 2C). On the other hand, state 4 respiration rate was up to 37% lower in torpid animals at all temperatures (Fig. 2D). In all cases, respiration rate declined significantly with temperature, but this effect did not differ between torpor and interbout euthermia.
In skeletal muscle, when glutamate+malate fueled respiration, neither state 3 (Fig. 3A) nor state 4 (Fig. 3B) respiration rate differed between torpid and interbout euthermic animals at any assay temperature. With succinate, state 3 respiration was 30% lower in torpid animals than interbout euthermic animals at 37°C, but not at 25°C or 10°C (Fig. 3C), and state 4 respiration rate did not differ among metabolic states at any temperature (Fig. 3D). As with liver, respiration rates declined significantly with temperature in skeletal muscle mitochondria, but this effect did not differ between torpor and interbout euthermia.
Q10 values for state 3 and state 4 respiration did not differ among metabolic states in either tissue, except that, in liver, the Q10 of state 3 glutamate + malate-fueled respiration was significantly higher in torpor compared with both interbout euthermia and summer (Table 1).
Oxidative phosphorylation kinetics of liver mitochondria.
At all assay temperatures, substrate oxidation activity was considerably lower in mitochondria from torpid animals compared with interbout euthermic animals under state 3 conditions, whereas under state 4 conditions, substrate oxidation activity was lower in torpor at 37°C and 25°C, but not 10°C (Fig. 4, A–C). ADP phosphorylation activity was also lower in torpid animals compared with interbout euthermic animals under state 3 conditions at all assay temperatures (Fig. 4, D–F). By contrast, IMM proton leakiness was significantly higher under state 4 conditions in mitochondria from torpid animals when measured at 37°C but was significantly lower in mitochondria from torpid animals when measured at 25°C and 10°C (Fig. 4, G–I).
Figure 5 summarizes the results of our top-down elasticity analysis. Partial integrated response values were used to show the quantitative contribution of each component of oxidative phosphorylation to changes in mitochondrial respiration rate between states. Changes in the activity of oxidative phosphorylation components between interbout euthermic animals measured at 37°C and torpid animals measured at 10°C best approximates the in vivo changes that likely occur during torpor. Active suppression of state 3 mitochondrial respiration rate during torpor was largely achieved by suppression of substrate oxidation, which accounted for nearly 95% of the total reduction in respiration rate between states. As torpid mitochondria were cooled from 37°C to 10°C, respiration rate dropped further, with substrate oxidation and ADP phosphorylation accounting for 75% and 25% of the decline in respiration rate, respectively. By comparison, when interbout euthermic mitochondria were similarly cooled, the decline in mitochondrial respiration rate was entirely achieved by substrate oxidation, with absolutely no contribution from ADP phosphorylation. Active suppression of state 4 mitochondrial respiration rate during torpor was entirely achieved by inhibition of substrate oxidation activity, as IMM proton leakiness was higher during torpor, which partially opposed the suppression of substrate oxidation. As torpid mitochondria were cooled from 37°C to 10°C, the decline in state 4 respiration rate was driven by nearly parallel declines in substrate oxidation and proton leak activity. When mitochondria from interbout euthermic animals were similarly cooled, however, the decline in state 4 respiration was driven entirely by a decline in substrate oxidation, which was partially opposed by an increase in proton leakiness with declining temperature.
Mitochondrial reactive oxygen species production in liver and skeletal muscle.
In both liver and skeletal muscle, basal and maximal ROS production rates were measured at 37°C, 25°C, and 10°C. Basal and maximal ROS production rates were measured in the absence or presence, respectively, of ETC inhibitors acting downstream of the ROS-producing complexes (i.e., rotenone for glutamate + malate-fueled respiration and antimycin A for succinate-fueled respiration).
In liver mitochondria, with glutamate + malate, basal ROS production was 87% lower in interbout euthermic animals, but 43% higher in summer active animals, compared with torpid animals, when measured at 37°C. At lower temperatures (25°C and 10°C), however, no differences among metabolic states were observed. Furthermore, while basal ROS production declined with temperature in both summer active and torpid animals, it did not fall with temperature in interbout euthermic animals (Fig. 6A). FRL was up to 89% lower in interbout euthermic animals compared with torpid animals at all assay temperatures and did not change with temperature (Fig. 6C). Our estimated FRL calculations for summer active animals were more similar to torpid animals at 37°C and 25°C, but more similar to interbout euthermic animals at 10°C. Maximal ROS production with glutamate + malate was 32% higher in torpid animals compared with both summer active and interbout euthermic animals, which did not differ from each other, when measured at 37°C, but no difference was observed among metabolic states at either 25°C or 10°C. Moreover, as with basal ROS production, maximal ROS production declined significantly with temperature only in summer active and torpid animals (Fig. 6E).
With succinate, basal rates of liver mitochondrial ROS production measured at 37°C were 20% higher in summer active animals compared with both torpid and interbout euthermic animals, which did not differ from each other; however, no differences among metabolic states were observed at lower temperatures (Fig. 6B). By contrast, FRL was 27% higher in torpid animals compared with interbout euthermic animals (and our estimated FRL levels in summer active animals) at all assay temperatures (Fig. 6D). Maximal ROS production rates differed among metabolic states only at 37°C, being 35% lower in torpid animals compared with summer active and interbout euthermic animals, which did not differ from each other (Fig. 6F). Temperature significantly reduced basal and maximal ROS production, as well as FRL, in summer active, interbout euthermic, and torpid animals.
In skeletal muscle, basal ROS production rates with glutamate + malate were nearly 50% higher in torpor than in interbout euthermia at all temperatures (Fig. 7A), but FRL did not differ among metabolic states at any temperature (Fig. 7C), which reflects that state 4 respiration rate with glutamate + malate tended to be higher during torpor as well. Maximal ROS production rates also did not differ between torpid and interbout euthermic animals at any assay temperature (Fig. 7E). In both metabolic states, basal and maximal ROS production, as well as FRL, declined with temperature. When succinate fueled respiration, there were no differences between metabolic states in terms of basal ROS production (Fig. 7B), FRL (Fig. 7D), or maximal ROS production (Fig. 7F) at any assay temperature. In addition, while basal ROS production and FRL declined with temperature, maximal ROS production rate actually increased with decreasing temperature.
For the most part, Q10 values for ROS production were higher than Q10 values for state 4 respiration in both tissues, except in liver of interbout euthermic and torpid animals respiring on glutamate + malate (Table 2).
The present study represents one of the most comprehensive investigations of mitochondrial metabolism during hibernation to date. It examined mitochondria from both liver and skeletal muscle, two tissues that contribute significantly to whole animal metabolism. It is the first study to compare mitochondrial respiration rates between torpor and interbout euthermia over a physiologically relevant range of temperatures using substrates for both complex I and II. It is also the first study to use top-down elasticity analysis to identify potential mechanisms of mitochondrial metabolic suppression and assess the quantitative contributions of each component of oxidative phosphorylation to mitochondrial suppression observed during torpor. Moreover, this study is the first to measure mitochondrial ROS production in hibernation over a range of relevant temperatures using substrates and inhibitors that permitted evaluation of ROS production from forward-electron flow at both complex I and complex III. Therefore, the present study greatly advances our understanding of the regulation of mitochondrial respiration and ROS production during hibernation.
Active, regulated suppression of mitochondrial respiration during torpor.
The present study and previous work (3, 4, 20, 33, 50) suggest that succinate-fueled liver mitochondrial respiration is considerably suppressed during torpor when measured at 37°C. Glutamate + malate-fueled respiration was not suppressed during torpor in the present study when measured at this same temperature, consistent with previous work in thirteen-lined ground squirrel (50) but not long-tailed ground squirrels (15). Top-down elasticity analysis (conducted at 37°C with succinate) showed that substrate oxidation activity was suppressed during torpor, suggesting that the suppression of succinate oxidation during torpor results, at least in part, from inhibition of the substrate transport and/or the electron transport chain. Given that liver mitochondrial respiration was suppressed with succinate but not glutamate + malate, the likely site of inhibition is succinate dehydrogenase (SDH) or the dicarboxylate transporter, as these sites are unique to succinate oxidation. Competitive inhibition of SDH by oxaloacetate has previously been shown to account for up to 25% of the suppression of liver mitochondrial respiration during torpor (3), but we are unaware of any studies examining dicarboxylate transport.
ADP phosphorylation activity was also suppressed in torpid animals compared with interbout euthermic animals. No previous studies have explicitly examined ADP phosphorylation kinetics during hibernation, although a few studies have suggested that the adenine nucleotide translocator (ANT) may be suppressed (2) and/or have a reduced affinity for adenine nucleotides (10) during torpor. ADP phosphorylation was suppressed in two Mus musculus strains during fasting-induced daily torpor (12) but not during spontaneous daily torpor in Phodopus sungorus, where ANT content did not change during torpor (13). Although ADP phosphorylation was suppressed in torpor in this study, it made only a minor contribution to the suppression of mitochondrial metabolism during torpor as it had little control over respiration rate.
Previous studies in hibernators (4, 33) showed no change in proton leakiness during torpor, except when fed diets containing levels of polyunsaturated fats above or below those found in standard rodent diets (33); however, summer and/or nonhibernating animals were used as controls. In the present study, liver IMM proton conductance was higher at 37°C in mitochondria from torpid animals compared with interbout euthermic animals. This is consistent with previous studies of spontaneous daily torpor in Phodopus sungorus (13) and suggests that restricting proton leakiness is not a primary mechanism for reducing mitochondrial respiration during torpor.
To our knowledge, this study is the first demonstration of skeletal muscle mitochondrial metabolic suppression in hibernation. Succinate-fueled (but not glutamate + malate-fueled) skeletal muscle mitochondrial respiration was suppressed during torpor compared with interbout euthermia, although the extent of this suppression was less than that observed in liver. This suggests that similar mechanisms may suppress mitochondrial respiration during torpor in both liver and skeletal muscle but to different extents. Given that mitochondrial recovery from skeletal muscle tissue was 50% lower during torpor than IBE, it remains possible that a reduction in muscle mitochondrial density during torpor accounts for a considerable amount of mitochondrial metabolic suppression, though this warrants further investigation. Previous studies of skeletal muscle respiration have shown no suppression during torpor (4, 50). The most likely reason that our findings differ from previous work is that we compared torpid animals with interbout euthermic animals rather than nonhibernating animals, thus controlling for both seasonal and dietary effects.
Temperature affects oxidative phosphorylation in different ways during torpor and euthermia.
While active, regulated inhibition of mitochondrial respiration was observed at 37°C in both liver and skeletal muscle when succinate was used to fuel respiration, it was not observed at 10°C in liver, nor at 10°C or 25°C in skeletal muscle, consistent with previous work in hibernators (33, 50) and daily heterotherms (12, 13). Therefore, it is likely that low temperature is sufficient to reduce succinate-fueled mitochondrial respiration to torpid levels and that active inhibition of succinate oxidation serves largely as a mechanism to initiate metabolic suppression during entrance into torpor, when Tb is still high. By contrast, at least in liver, active, regulated inhibition of mitochondrial respiration with glutamate + malate was not observed at high assay temperature (37°C) but was observed at 25°C and 10°C. This suggests that active suppression of glutamate + malate oxidation may not be important for initiating metabolic suppression during entrance into torpor but may be more temperature-sensititve in torpid animals (as suggested by higher Q10 values) to maximize the energy-saving impact of passive thermal effects as Tb falls during torpor.
Some other components of oxidative phosphorylation were also more temperature-sensitive in torpid animals than interbout euthermic animals. ADP phosphorylation was more temperature-sensitive during torpor than interbout euthermia. Figure 5 shows that ADP phosphorylation activity decreased in torpid animals as temperature declined, but actually increased slightly in interbout euthermic animals with a similar change in assay temperature. This pattern has been observed previously in two daily heterotherms, dwarf Siberian hamsters (13) and mice (12). While ADP phosphorylation had little control over respiration rate, particularly at low temperatures, it did exert considerable control over mitochondrial membrane potential (data not shown), since it is the principle consumer of ΔΨm under phosphorylating conditions. In ground squirrels (this study), hamsters (13), and mice (12), state 3 mitochondrial membrane potential fell by 17 mV when euthermic mitochondria were cooled from 37°C to 10°C or 15°C, but did not change under equivalent conditions with torpid mitochondria, largely because ADP phosphorylation and substrate oxidation declined to similar degrees in response to temperature changes in torpid mitochondria. Therefore, mitochondria from torpid animals may be distinguished from mitochondria from interbout euthermic animals in that the former can maintain homeostasis of mitochondrial membrane potential at low temperatures via a more temperature-responsive ADP phosphorylation component, which may be an important adaptation to prevent, for example, stimulation of apoptotic pathways (35).
IMM proton leakiness also showed a differential response to temperature between torpid and interbout euthermic animals. Previous studies in rats (25), insects (19), hamsters (13), and mice (12) have shown that mitochondrial membrane proton conductivity declines with temperature. Overall, this same pattern was observed for torpid (and summer active; not shown) mitochondria in the present study, but not for interbout euthermic mitochondria, which showed an opposite pattern, whereby proton leakiness increased as temperature was reduced. Parallel declines in substrate oxidation and proton leak with temperature (as typically observed) may be important to help minimize changes in membrane potential in the face of changing temperatures. For example, in the present study, when torpid mitochondria were cooled from 37°C to 10°C, membrane potential increased by only 3 mV, whereas, in interbout euthermic mitochondria, this same temperature change induced a decrease in membrane potential of nearly 20 mV. However, perhaps the unusual lack of temperature sensitivity of proton leakiness in interbout euthermic animals serves to prevent large, temperature-dependent fluctuations in mitochondrial proton leak during periodic arousals. Proton leakiness was lower in mitochondria from interbout euthermic animals when measured at 37°C, but it was lower in torpid animals when measured at 10°C, close to torpid Tb, suggesting that ground squirrels may be able to keep proton leakiness low throughout the hibernation season despite dramatic changes in Tb.
Basal ROS production rates are higher in liver, but not skeletal muscle mitochondria, in torpor.
In liver mitochondria isolated from torpid animals, electrons were more likely to leak prematurely from ETC complex III when compared with mitochondria from interbout euthermic or summer active animals at all temperatures, as witnessed by higher FRL with succinate (in the presence of rotenone) in torpid animals. This is consistent with our previous observations in both mice and dwarf Siberian hamsters during daily torpor (11). Given that maximal ROS production with succinate was actually lower in torpid animals than either interbout euthermic or summer animals, particularly at 37°C, increased FRL likely reflects that ETC complex III is in a considerably more reduced state during torpor than during euthermia. In contrast to liver, no difference in FRL between metabolic states was observed in skeletal muscle.Therefore, the increased susceptibility to ROS production at complex III in torpid liver mitochondria does not appear to be related to SDH inhibition, as SDH appears to be inhibited in both liver and skeletal muscle mitochondria.
Changes in ΔΨm are not likely responsible for increased liver mitochondrial FRL in torpor because neither we (12, 13; this study) nor Barger et al. (4) have shown that state 4 ΔΨm is consistently higher in torpid liver mitochondria than euthermic controls. Therefore, some other mechanism(s) must be responsible. Petrosillo et al. (61) showed that loss of cardiolipin increases ROS production from complex III, but Chung et al. (20) showed no change in cardiolipin content of liver mitochondria between interbout euthermic and torpid animals. A recent study by Pasdois et al. (60) suggested that a decline in oxidized cytochrome c during periods of ischemia—via loss of cytochrome c and/or inhibition of cytochrome-c oxidase—triggers mitochondrial ROS production without compromising mitochondrial respiration rate. Indeed, the capacity of liver mitochondria to oxidize cytochrome c is significantly lower during torpor and late arousal, compared with summer activity, in I. tridecemlineatus (50) and could be responsible for increased ROS production rates in torpor.
Complex I ROS production is suppressed during interbout euthermia in liver.
When glutamate + malate fueled respiration, FRL was considerably higher in torpid animals compared with interbout euthermic animals. Unlike what we saw with succinate, however, our estimated FRL for summer active animals was similar to that observed in interbout euthermic animals only at 10°C; at both 25°C and 37°C, FRL was considerably lower in interbout euthermic animals compared with summer animals, even though respiration rate does not differ between these two states (3). This is not simply a seasonal effect, as mitochondria from torpid animals had FRL levels comparable to summer animals, suggesting that the observed differences in FRL between torpor and interbout euthermia with glutamate + malate may reflect suppression of ROS production during interbout euthermia. In support of this idea, when mitochondria from interbout euthermic animals were warmed from 10°C to 37°C, neither basal nor maximal ROS production increased, even though mitochondrial respiration rate nearly doubled over this same temperature range, which explains why FRL remains so low at the higher assay temperatures. Given that no such difference was observed with succinate (with rotenone), where ROS production is derived from complex III only, we believe that the lower rate of ROS production during interbout euthermia reflects suppression of ROS production from complex I. Either complex I is modified, such that its capacity for ROS production is actively reduced during interbout euthermia (and quite possibly during the preceding arousal), or levels of endogenous mitochondrial antioxidants are higher during interbout euthermia than summer, leading to higher rates of endogenous ROS degradation and, therefore, lower rates of H2O2 release from complex I in interbout euthermia. Indeed, several studies have shown increased levels of some mitochondrial antioxidants in liver during the hibernation season (e.g., 18, 59), but the extent of this difference (∼50% higher in hibernation) is unlikely to fully account for the difference seen in basal ROS production rates (∼90% lower during interbout euthermia vs. summer). Therefore, complex I ROS production capacity is likely reduced during interbout euthermia. Lower release rates of ROS from complex I during interbout euthermia may be important because tissue reperfusion—as occurs during arousal and interbout euthermia—is thought to promote ROS production (39, 62), especially from complex I (69). Therefore, modification of complex I itself and/or increased degradation of complex I-derived ROS production in liver during interbout euthermia, may help to alleviate any burst in ROS production from this site, thereby limiting oxidative damage to liver during periodic arousals.
Declining temperature reduces mitochondrial ROS production.
There appear to be two different mechanisms by which basal ROS production declined with temperature. First, as predicted by thermodynamics, the capacity of ETC complexes I and III to produce ROS typically declined with temperature. Second, ROS production was more temperature-sensitive than mitochondrial respiration rate. Therefore, at lower temperatures, a smaller proportion of electrons leaked from the ETC prematurely, which decreased ROS production. We previously observed this same effect in liver mitochondria from hamsters and mice (11), which suggests that it may broadly characterize mammalian mitochondria. To our knowledge, however, no studies have measured how mitochondrial ROS production changes with temperature in strictly homeothermic mammals.
Perspectives and Significance
Both liver and skeletal muscle mitochondrial oxidative capacity is clearly suppressed during torpor in hibernating ground squirrels, and similar mechanisms may be involved in both tissues (e.g., SDH inhibition). Given that liver and skeletal muscle, together, account for as much as 38% of BMR in mammals (48), active suppression of mitochondrial respiration in these two tissues, to the extent seen in the present study, could reduce MR by as much as 20%, although measurements of mitochondrial respiration under more physiological conditions are needed to confirm this assertion. Together with active suppression of heart rates and ventilation rates (49), hibernators may be able to reduce MR quite considerably even before Tb begins to decline. We must now examine mitochondrial metabolism during hibernation in other tissues, especially brain, heart, and kidney, to determine whether mitochondrial metabolic suppression during torpor occurs throughout the animal, and we should begin to examine changes in mitochondrial metabolism during torpor in other hibernating species, especially nonrodents, to determine whether mitochondrial metabolic suppression during torpor characterizes all mammalian orders in which hibernation occurs and involves common mechanisms. This could shed light on the evolutionary origins of mammalian hibernation, which are presently unclear (32, 36). Moreover, in addition to reducing energy expenditure, hibernation likely also reduces oxidative stress. We have demonstrated that reduced Tb likely reduces mitochondrial ROS production, particularly by keeping the ETC relatively oxidized, and we have further shown that bursts in ROS production during arousal and interbout euthermia, especially at complex I, are likely avoided in hibernators. This may help to explain previous observations that hibernation seems to contribute to longevity in several species, including hamsters (46), bats (72), and lemurs (43), and could suggest that the alleviation of oxidative damage accumulation was a potential selective pressure driving the evolution of mammalian hibernation.
Funding for this research was provided in the form of Discovery and Research Tools and Infrastructure Grants from the Natural Sciences and Engineering Council of Canada (to J. F. Staples) and a Queen Elizabeth II Scholarship in Science and Technology (to J. C. L. Brown).
No conflicts of interest, financial or otherwise, are declared by the authors.
Alvin Iverson and his staff at the Carman and Area Research Center (University of Manitoba) were very helpful in trapping animals. We thank Manitoba Conservation for providing permission to trap animals and Lynda McCaig for her assistance with animal dissection.
- Copyright © 2012 the American Physiological Society