Cellular Respiration during Hypoxia
ROLE OF CYTOCHROME OXIDASE AS THE OXYGEN SENSOR IN HEPATOCYTES*

(Received for publication, December 23, 1996, and in revised form, April 21, 1997)

Navdeep S. Chandel , G. R. Scott Budinger , Sang H. Choe and Paul T. Schumacker Dagger

From the Department of Medicine, Section of Pulmonary and Critical Care, the University of Chicago, Chicago, Illinois 60637

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
REFERENCES


ABSTRACT

We previously reported that hepatocytes exhibit a reversible suppression of respiration during prolonged hypoxia (PO2 = 20 torr for 3-5 h). Also, isolated bovine heart cytochrome c oxidase undergoes a reversible decrease in apparent Vmax when incubated under similar conditions. This study sought to link the hypoxia-induced changes in cytochrome oxidase to the inhibition of respiration seen in intact cells. Hepatocytes incubated at PO2 = 20 torr exhibited decreases in respiration and increases in [NAD(P)H] after 2-3 h that were reversed upon reoxygenation (PO2 = 100 torr). Respiration during hypoxia was also inhibited when N,N,N',N'-tetramethyl-p-phenylenediamine (0.5 mM) and ascorbate (5 mM) were used to reduce cytochrome c, suggesting that cytochrome oxidase was partially inhibited. Similarly, liver submitochondrial particles revealed a 44% decrease in the apparent Vmax of cytochrome oxidase after hypoxic incubation. In hepatocytes loaded with tetramethylrhodamine ethyl ester (10 nM) to quantify mitochondrial membrane potential, acute hypoxia (<30 min) produced no change in fluorescence, consistent with the absence of an acute change in respiration. However, fluorescence increased during acute reoxygenation after prolonged hypoxia, suggesting an increase in potential. The control exhibited by NADH over mitochondrial respiration was not altered during hypoxia. Thus, changes in the Vmax of cytochrome oxidase during prolonged hypoxia correlate with the changes in respiration and mitochondrial potential. This suggests that the oxidase functions as an oxygen sensor in the intact hepatocyte.


INTRODUCTION

Cellular respiration is normally determined by metabolic activity and the corresponding rate of ATP utilization. Classical studies have shown that cellular oxygen uptake (VO2) remains essentially independent of oxygen tension (PO2) as long as the extracellular O2 tension exceeds a critical value ranging from 3-6 torr (1-4). Below this point, diffusion of O2 to the mitochondria begins to limit oxidative phosphorylation (5). However, in those studies the duration of hypoxia ranged from seconds to a few minutes at most, and this brief exposure may have precluded the development of an adaptive response. More recent studies have demonstrated that isolated cells undergo a significant suppression in their rate of O2 uptake when exposed to moderate hypoxia for longer periods (6, 7). This non-lethal suppression occurs at O2 tensions (PO2) between 10 and 40 torr and is rapidly reversed when the PO2 is restored to normoxic conditions. Because the decrease in respiration was not accompanied by a corresponding increase in lactate production (6, 7), the response would appear to reflect a decrease in ATP utilization rather than a mere shift in energy production from aerobic to anaerobic sources.

The mechanism responsible for the reversible decrease in respiration during moderate hypoxia has been the focus of a number of studies in our laboratory. In this regard, evidence points to a role for mitochondria as the site of cellular O2 sensing. For example, rat liver mitochondria incubated in buffer solutions at low PO2 (<2 torr) exhibited an inhibition in State 3 respiration after reoxygenation to PO2 = 20 torr, which was reversible when they were subsequently reoxygenated to PO2 = 100 torr (8). Inhibition of respiration during hypoxia also was observed when the mitochondria were given TMPD1 and ascorbate as substrates, which suggested that cytochrome oxidase function was reversibly inhibited. More recently, kinetic studies of isolated bovine heart cytochrome oxidase revealed that the apparent Vmax of the enzyme during hypoxia was reversibly inhibited by approximately 50% after incubation at PO2 = 20 torr (9). Those results suggested that O2 may interact with cytochrome oxidase, eliciting a reversible change in its catalytic function possibly through an allosteric interaction at a second binding site. Such an effect would produce a change in the reduction state of the enzyme in cells exposed to O2 tensions of approximately 20 torr for prolonged periods (i.e. >2-3 h), which in turn could allow cytochrome oxidase to function as a cellular O2 sensor in the physiological range of hypoxia.

The present study sought to determine whether the kinetic changes in cytochrome oxidase observed with the purified enzyme also contribute to the hypoxic responses in the intact cell. First, we assessed the function of cytochrome oxidase in intact cells and in submitochondrial particles to determine whether the changes in the oxidase were consistent with the changes in cellular respiration. Second, we studied the effects of hypoxia on mitochondrial transmembrane potential to assess the balance between electron flux and ATP consumption. Finally, we measured the respiratory control coefficients for NADH to determine whether processes controlling the supply and utilization of NADH were also affected by prolonged hypoxia. The results support the hypothesis that changes in the kinetic properties of cytochrome oxidase allow the enzyme to function as an O2 sensor in the intact cell, participating in the reversible inhibition of respiration during hypoxia.


EXPERIMENTAL PROCEDURES

Male Harlan Sprague Dawley rats weighing 225 ± 25 g were provided with food and water ad libitum. Hepatocytes were isolated by collagenase digestion of livers using methodology similar to that previously described (10), followed by Percoll centrifugation. This procedure yielded hepatocytes with >90% viability, as determined by trypan blue (0.4%) exclusion. Cells then were studied either in suspension or adhered to glass coverslips coated with 0.1% collagen (see below).

For cells studied in suspension, approximately 1.0 × 107 cells were seeded into spinner flasks containing 250 ml of Dulbecco's modified Eagle's medium supplemented with Hepes (10 mM) and antibiotics. The media in the spinner flasks had been previously equilibrated with 21% O2, 5% CO2 or 5% O2, 5% CO2 passed through the head space in the flask. A polarographic electrode mounted in the flask was used to monitor the PO2 in the media. The spinner flasks were stirred at 60-70 rpm in an incubator regulated at 37 °C. Cellular O2 consumption rates were measured in aliquots of cells removed from the flask and studied in a magnetically stirred, water-jacketed (37 °C), anaerobic respirometer (2 ml volume) fitted with a polarographic O2 electrode. Prior to each measurement the respirometer was flushed with gas at the same PO2 used to gas the head space of the flask. Because the cell aliquots were transferred anaerobically from the spinner flask to the respirometer, contact with ambient air was prevented and the measurements of O2 uptake were obtained at approximately the same PO2 used for the incubation in the spinner flask. Mitochondrial respiration was determined as the difference between total O2 consumption and that obtained after addition of the electron transport inhibitor myxothiazol (150 ng/ml).

Changes in cellular NAD(P)H concentrations were assessed from changes in the autofluorescence of stirred cell suspensions studied in a spectrofluorimeter (Perkin-Elmer) at an excitation wavelength of 365 nm, an emission wavelength of 460 nm, and a slit width of 5 nm. Cells were transferred anaerobically from the spinner flask to a quartz cuvette (4 ml) containing a magnetic stir bar. The cuvette was filled or emptied via a 4-cm capillary chimney and was maintained at 37 °C in a water-jacketed holder within the spectrofluorimeter. The cuvette had been previously flushed with 3% O2 or 15% O2 prior to addition of cells. Different respiratory inhibitors were added to the cell suspensions after allowing 2-5 min for the cells to stabilize; the change in fluorescence was measured 2 min after addition. For each replicate experiment, the extent of maximum reduction of the mitochondrial NAD(P)H pool was determined by adding beta -hydroxybutyrate (20 mM) and rotenone (10 µM). Fluorescence changes are reported as percentage of fluorescence measured during maximum mitochondrial reduction.

Cytochrome c oxidase catalytic activity was measured in whole cells in some studies and in submitochondrial particles in others. Cells or submitochondrial particles were incubated in spinner flasks at different designated PO2. Oxygen consumption rates were measured in the presence of rotenone (10 µM), myxothiazol (100 ng/ml), TMPD (0.5 mM), and ascorbate (5 mM). The former two compounds were used to inhibit autologous mitochondrial electron transport. TMPD is a electron donor which reduces cytochrome c nonenzymatically (11). Therefore, when TMPD is used as a substrate, changes in O2 uptake rates reflect changes in cytochrome oxidase activity. Ascorbate was used to reduce TMPD, which would otherwise become progressively oxidized. Submitochondrial particles were prepared from rat liver mitochondria as described previously (12). Protein concentrations were determined by the biuret procedure with bovine serum albumin as the standard.

Mitochondrial membrane potential was measured in adherent hepatocytes on glass coverslips. Approximately 105 cells in 2 ml of media were pipetted onto coverslips (25 mm) and maintained in a humidified incubator for 12 h. Coverslips were then transferred to an anaerobic flow-through chamber (Sykes-Moore, Belco Glass Co.). A thermocouple within the chamber was used to maintain the perfusate temperature at 37 °C. The chamber was clamped to a temperature-controlled block on the stage of an inverted microscope and continuously perfused with Dulbecco's modified Eagle's medium supplemented with tetramethylrhodamine ethyl ester (TMRE, 10 nM). TMRE is a cationic membrane-permeant fluorescent probe that has been shown to be distributed between cytosol and mitochondria according to Nernstian equilibrium (13). As such, it has been shown to provide a qualitative measure of mitochondrial membrane potential. Because the dye is not significantly quenched in the cell, increases in fluorescence can be used to indicate relative hyperpolarization while decreases in fluorescence can be used to indicate depolarization. Cells were superfused with the dye for 1-2 h to effect loading prior to study.

Quantitative fluorescence images of cells were acquired using a digital imaging system. An inverted microscope was equipped with a 75-watt xenon lamp for epifluorescent illumination. Changes in fluorescence were determined using an excitation wavelength of 535 nm and an emission of 610 nm. The excitation wavelength was obtained using interference and neutral density filters mounted in the light path between the source and the cells. Images were acquired with a 12-bit chilled charge-coupled device (CCD) camera (Princeton Instruments) under computer control (Metamorph, Universal Imaging).

Differences among groups were analyzed using analysis of variance, followed by Student's t statistical comparisons. Statistical significance was determined at the 0.05 level. Values are reported as means ± S.D. Replicate experiments were carried out using cells from separate isolations.


RESULTS

Effects of Hypoxia on O2 Consumption and [NAD(P)H] in Intact Hepatocytes

This study sought to quantify the effects of hypoxic incubation on O2 consumption and NAD(P)H concentrations in rat hepatocytes. Cells were maintained in suspension in spinner flasks at PO2 = 100 or 20 torr for 5 h. Oxygen consumptions were measured in aliquots studied at approximately the same PO2 used for incubation, at time = 0, 1, 2, 3, and 5 h. As shown in Fig. 1, the O2 uptake at 20 torr decreased from 2.43 ± 0.82 initially to 1.39 ± 0.23 µmol/h/106 cells (p < 0.05) after 5 h at PO2 = 20 torr. By contrast, cells incubated at PO2 = 100 torr maintained their O2 uptake rates over the same interval. Cell viability was unchanged from the beginning to the end of the experiment. In separate experiments, cells were incubated at PO2 = 100 torr for 1 h, after which the oxygen tension was decreased to PO2 = 20 torr for 3 h. An aliquot of cells at 20 torr was then aspirated into a syringe containing 5% CO2 and 15% O2 and rotated axially for 5 min to restore the media PO2 to 100 torr (acute reoxygenation). Oxygen uptake rates and NAD(P)H concentrations during that study are summarized in Fig. 2. The NAD(P)H fluorescence increased by 25.8 ± 7.0% while O2 uptake decreased from 2.57 ± 0.05 to 1.16 ± 0.31 µmol/h/106 cells during hypoxia. During acute reoxygenation, O2 uptake and NAD(P)H fluorescence recovered to initial values. Thus, the changes in respiratory rate and NAD(P)H seen during prolonged hypoxia were reversible within 5 min after reoxygenation.


Fig. 1. Relationship between oxygen tension and oxygen consumption in hepatocytes. Cells were maintained in suspension in experimental and control spinner flasks for 5 h. Aliquots were acquired anaerobically at t = 0, 1, 2, 3, and 5 h for oxygen consumption measurements at approximately the same PO2 used for incubation. *, p < 0.05 between groups, n = 5. Values are means ± S.D.
[View Larger Version of this Image (19K GIF file)]


Fig. 2. O2 consumption rate (top) and NAD(P)H fluorescence (bottom) during hepatocyte exposure to PO2 = 100 torr (1 h), PO2 = 20 torr (3 h), and PO2 = 100 torr (min, Acute Reoxygenation). NAD(P)H fluorescence changes are reported as percentage of maximum mitochondrial NAD(P)H reduction. *, p < 0.05 between groups, n = 5. Values are means ± S.D.
[View Larger Version of this Image (16K GIF file)]

Effects of Hypoxia on Cytochrome c Oxidase Catalytic Activity in Hepatocytes

This study sought to determine whether the changes in kinetics observed for purified cytochrome oxidase (9) also occur within intact hepatocytes during prolonged hypoxia. Hepatocytes were maintained in suspension in spinner flasks at PO2 = 100 or 20 torr for 5 h. Cellular O2 uptake rates were measured at time = 0, 1, 2, 3, and 5 h at the same PO2 used for incubation, immediately after addition of TMPD (500 µM), ascorbate (5 mM), and rotenone (10 µM) (Fig. 3). Because TMPD supplies electrons directly to cytochrome c (1), respiration with TMPD can provide an index of maximal cytochrome oxidase activity if cytochrome c is maximally reduced. After exposure to 20 torr for 5 h, TMPD respiration decreased from 7.76 ± 0.57 to 4.65 ± 0.2 µmol/h/106 cells (p < 0.05). By contrast, TMPD respiration remained unchanged in cells maintained at 100 torr for the same duration. After incubation under hypoxia for 5 h, cells were acutely reoxygenated to PO2 = 100 torr, and TMPD respiration increased to 7.39 ± 0.27, which was not different from the initial value at PO2 = 100 torr (p = not significant).


Fig. 3. Relationship between oxygen tension and cytochrome c oxidase activity as reflected by TMPD respiration in intact hepatocytes. Cells were maintained in suspension in experimental and control spinner flasks for 5 h. Aliquots were acquired anaerobically at t = 0, 1, 2, 3, and 5 h for oxygen consumption measurements in the presence of rotenone (10 µM), myxothiazol (100 ng/ml), TMPD (0.5 mM), and ascorbate (5 mM). After the 5-h measurement, an aliquot was reoxygenated to PO2 = 100 torr, and O2 uptake rate was measured. *, p < 0.05 between groups, n = 5. Values are means ± S.D.
[View Larger Version of this Image (20K GIF file)]

The O2 dependence of cytochrome oxidase activity was also examined in submitochondrial particles from rat liver mitochondria. Submitochondrial particles were incubated without substrates in stirred suspensions at different PO2 for 4 h. Oxygen consumption rates were then assessed in aliquots studied at the same PO2 used for incubation, after addition of TMPD (500 µM) and ascorbate (5 mM) as substrates. As shown in Fig. 4 (top), submitochondrial particles incubated at PO2 < 50 torr exhibited a 43.8% decrease in cytochrome oxidase activity, compared with those studied at higher O2 tensions. The TMPD concentration dependence of the O2 consumption rate was also measured in submitochondrial particles incubated at PO2 = 100 or 20 torr for 4 h (Fig. 4, bottom). As the concentration of TMPD was increased from 0 to approximately 100 µM there was a proportional increase in the rate of O2 uptake, consistent with a progressive increase in cytochrome c reduction at these low concentrations of TMPD. At higher concentrations of TMPD the increase in O2 uptake appeared to plateau, suggesting that cytochrome c reduction was complete. Submitochondrial particles incubated under hypoxia also showed a TMPD concentration dependence of O2 uptake, but the rate was decreased by approximately 50% at each TMPD concentration compared with the normoxic particles (Fig. 4, bottom). Thus, cytochrome oxidase activity in submitochondrial particles was significantly and reversibly inhibited over a wide range of TMPD concentrations, after exposure to prolonged hypoxia.


Fig. 4. Relationship between cytochrome c oxidase activity and oxygen concentration in submitochondrial particles isolated from hepatocytes. Top, cytochrome c oxidase activity was measured as a function of oxygen concentration in SMP given TMPD (0.1 mM) and ascorbate (2 mM) as substrates. Bottom, cytochrome c oxidase activity as a function of TMPD concentration in SMP. As [TMPD] was increased, the ratio of [ascorbate] to [TMPD] was maintained at 20.
[View Larger Version of this Image (20K GIF file)]

Effects of Prolonged Hypoxia on Mitochondrial Membrane Potential in Hepatocytes

Mitochondrial transmembrane potential reflects a balance between processes contributing to the electrochemical gradient and those tending to dissipate it. Measurement of the potential can therefore provide information regarding the relative rates of these opposing processes and insight into the processes inhibited during hypoxia. For example, if the decrease in cellular respiration during hypoxia were due solely to an inhibition of cytochrome oxidase activity without any change in ATP utilization, then a marked decrease in mitochondrial potential should occur. By contrast, if the decreases in cellular respiration were caused solely by a decrease in ATP utilization without any change in mitochondrial function then membrane potential should increase during hypoxia. To distinguish between these, we measured mitochondrial potential in intact hepatocytes superfused with normoxic or hypoxic media.

Initial studies were carried out to confirm that the fluorescent probe TMRE behaved in the manner consistent with that expected for a mitochondrial potentiometric dye. Hepatocytes on coverslips maintained at 37 °C were imaged every 30 s for 12 min while the chamber was perfused with media at PO2 = 100 torr. Ouabain (100 µM), an inhibitor of the membrane Na+-K+-ATPase, was added to the perfusate at t = 5 min. Fig. 5 (top) illustrates the increase in fluorescence observed after the addition of ouabain, consistent with the expected hyperpolarization of mitochondrial membrane potential. Fig. 5 (bottom) illustrates the response to gramicidin D, which should decrease the membrane potential by increasing Na+-K+-ATPase activity and thereby increasing the supply of ADP to mitochondria. Cells given gramicidin D (1 µg/ml) at PO2 = 100 torr exhibited significant decreases in fluorescence, consistent with mitochondrial depolarization.


Fig. 5. Effect of ouabain (top) and gramicidin D (bottom) on mitochondrial membrane potential as determined with TMRE in hepatocytes on coverslips. Top, cells were imaged every 30 s prior to and after addition of ouabain (100 µM) to inhibit ATP utilization by Na+-/K+-ATPase at PO2 = 100 torr. Bottom, cells were imaged every 30 s prior to and after addition of gramicidin D (1 µg/ml) to stimulate ATP utilization by Na+-/K+-ATPase, at PO2 = 100 torr.
[View Larger Version of this Image (20K GIF file)]

A decrease in the activity of the mitochondrial ATP synthase should result in hyperpolarization of the membrane potential. Fig. 6 (top) shows the effect of oligomycin (1 µg/ml) on TMRE fluorescence in hepatocytes during normoxia. The increase in fluorescence that was observed is consistent with membrane hyperpolarization. Similarly, the increase in fluorescence elicited using carboxyatractyloside (50 µM) to block the mitochondrial adenine transporter (Fig. 6, bottom) was consistent with membrane hyperpolarization.


Fig. 6. Effect of mitochondrial ATP synthase activity and adenine nucleotide transporter activity on mitochondrial membrane potential as determined with TMRE in hepatocytes on coverslips. Top, cells were imaged every 60 s prior to and after addition of oligomycin (1 µg/ml) to inhibit ATP synthase activity, at PO2 = 100 torr. Bottom, cells were imaged every 15 s prior to and after addition of atractyloside (50 µM) to inhibit the mitochondrial adenine nucleotide transporter, at PO2 = 100 torr.
[View Larger Version of this Image (17K GIF file)]

Because the mitochondrial potential is generated by electron flux, an inhibition of electron transport would limit proton pumping and cause a depolarization. Fig. 7 (top) shows the effect of myxothiazol (25 ng/ml) administered to cells at PO2 = 100 torr on TMRE fluorescence. The decrease in fluorescence is consistent with the expected decrease in mitochondrial potential. Likewise, administration of the protonophore carbonyl cyanide p-trifluoromethoxyphenylhydrazone (5 µM) (Fig. 7, bottom) caused a large decrease in fluorescence, consistent with the loss of membrane potential.


Fig. 7. Effect of electron transport inhibition and mitochondrial uncoupling on mitochondrial membrane potential as determined with TMRE in hepatocytes on coverslips. Top, cells were imaged every 30 s prior to and after addition of myxothiazol (25 ng/ml) to inhibit electron transport, at PO2 = 100 torr. Bottom, cells were imaged every 30 s prior to and after addition of carbonyl cyanide p-trifluoromethoxyphenylhydrazone (5 µM), a mitochondrial membrane protonophore, at PO2 = 100 torr.
[View Larger Version of this Image (18K GIF file)]

Effects of Hypoxia and Reoxygenation on Mitochondrial Membrane Potential

Collectively, the above studies indicated that TMRE is capable of providing a sensitive measure of mitochondrial potential in hepatocytes. Accordingly, this probe was used to assess mitochondrial potential during hypoxia in two different studies. First, hepatocytes loaded with TMRE were equilibrated at PO2 = 100 torr for 1 h. The cells were then imaged every 30 s for 10 min, after which the PO2 of the perfusate was abruptly decreased to 20 torr and the cells were imaged for an additional 10 min (acute hypoxia). This produced no detectable change in fluorescence (Fig. 8), suggesting that mitochondrial potential remained constant. This was consistent with the absence of change in O2 uptake or TMPD respiration by intact hepatocytes during acute hypoxia (see Fig. 1).


Fig. 8. Effect of acute hypoxia on mitochondrial membrane potential as determined with TMRE in hepatocytes on coverslips. Cells were imaged every 60 s for 10 min at PO2 = 100 torr. Subsequently, the PO2 in the perfusion system was changed to 20 torr, and cells were imaged every 60 s for 10 min. Values are means ± S.D., n = 4.
[View Larger Version of this Image (26K GIF file)]

By contrast, exposure to prolonged hypoxia was associated with a decrease in O2 uptake by 3-4 h (see Fig. 1). Unfortunately, measurement of TMRE fluorescence over the same time frame is imprecise, due to drifts in the signal caused by additional uptake of dye, photobleaching, or other artifacts. As an alternative, we examined the response to acute reoxygenation to PO2 = 100 torr in hepatocytes that had been incubated for 3 h at PO2 = 20 torr. As shown in Fig. 2, acute reoxygenation had been associated with a rapid restoration of O2 consumption, indicating a reversal of the effects of prolonged hypoxia. As shown in Fig. 9, cells that had been perfused with media at PO2 = 20 torr for 3 h were imaged for 10 min to establish a fluorescence baseline. Subsequently, the media PO2 was abruptly increased to 100 torr as the cells continued to be imaged. The TMRE fluorescence increased significantly during reoxygenation, which suggested that the mitochondrial potential had increased. Such a change suggests that the rate of electron transport was acutely increased, consistent with an increase in the activity of cytochrome oxidase.


Fig. 9. Effect of acute reoxygenation on mitochondrial membrane potential as determined with TMRE in hepatocytes on coverslips. Cells were superfused with media at PO2 = 20 torr for 3 h to induce hypoxic conditioning. Cells were imaged every 60 s for 10 min at that PO2. The PO2 in the perfusion system was then increased acutely to 100 torr while TMRE fluorescence images continued to be recorded. Values are means ± S.D., n = 6.
[View Larger Version of this Image (27K GIF file)]

Effects of Hypoxia on the Control of Mitochondrial Respiration by [NADH]

Cell processes that alter mitochondrial [NADH] produce a corresponding change in cellular O2 consumption (14). In steady state, mitochondrial [NADH] reflects a balance between "upstream" processes controlling the supply of reducing equivalents from the matrix dehydrogenases to NADH and the "downstream" steps responsible for electron transport to cytochrome oxidase and O2. The control exerted by [NADH] on cellular respiration has been described by Brown et al. (15) in terms of the relative sensitivities of the upstream and downstream arms of this metabolic pathway to changes in the [NADH]. The resulting respiratory control coefficient for NADH contains information related to the extent that [NADH] controls cellular respiration. In hepatocytes, Brown et al. (15) reported control coefficients for NADH of 0.15 to 0.30. We carried out a similar analysis to determine whether the control coefficients for NADH were altered in hepatocytes incubated at PO2 = 20 torr. Such an effect would indicate that hypoxia must influence metabolic control at multiple sites within the cell. An absence of change in the control coefficient for NADH during hypoxia would support the notion that cytochrome oxidase acts alone in this response.

To determine the sensitivity of downstream processes to [NADH], mitochondrial [NADH] was experimentally altered by adding beta -hydroxybutyrate or acetoacetate, which influence mitochondrial redox states via their interconversion by the NAD+-linked beta -hydroxybutyrate dehydrogenase located in the mitochondrial matrix. Ketone bodies were added to suspended hepatocytes after incubation at either PO2 = 100 torr or 20 torr for 4 h, and the change in mitochondrial respiratory rate was measured. Addition of beta -hydroxybutyrate to normoxic cells produced an increase in NAD(P)H fluorescence and an increase in mitochondrial respiration, whereas addition of acetoacetate produced a decrease in fluorescence and a decrease in respiration (Table I). Although cells incubated under prolonged hypoxia exhibited increased levels of NAD(P)H fluorescence compared with normoxic cells (see Fig. 2), the responses to beta -hydroxybutyrate and acetoacetate were not different between the two groups (Table II). Thus, the suppression of cellular respiration elicited by exposure to prolonged hypoxia did not alter the extent to which mitochondrial [NADH]/[NAD+] ratios control respiration at sites downstream from NADH (Fig. 10).

Table I. Respiratory control ratios in hepatocytes incubated at PO2 = 20 torr for 4 h

Cells were exposed to hypoxia (PO2 = 20 torr) for 3 h. Cells were administered various substrates for 2-5 min, and NAD(P)H changes and mitochondrial respiration were measured. For each cell preparation, the extent of maximum reduction of the mitochondrial NAD(P)H pool was determined by adding 20 mM beta -hydroxybutyrate and 10 µM rotenone. Fluorescence changes are reported as percentage of maximum mitochondrial reduction. Mitochondrial respiration was calculated by subtracting the myxothiazol-insensitive respiration from the cellular respiration rate. Values are means ± SD, n = 5. 

Interventions Mitochondrial respiration NAD(P)H

  % change
 beta -Hydroxybutyrate (20 mM) 14.1  ± 2.4 40.9  ± 10.1
Acetoacetate (20 mM)  -12.8  ± 4.7  -37.4  ± 14.3
Gramidicin D (10 µg/ml) 25.5  ± 7.1  -17.1  ± 5.9
Myxothiazol (50 ng/ml)  -66.2  ± 19.9 57.5  ± 18.0
Oligomycin (5 µg/ml)  -70.3  ± 12.3 63.4  ± 7.8
Atractyloside (100 µM)  -74.0  ± 3.9 71.5  ± 4.3

Table II. Respiratory control ratios in hepatocytes incubated at PO2 = 100 torr for 4 h

Cells were exposed to normoxia (PO2 = 100 torr) for 3 h. Cells were administered various substrates for 2-5 min, and NAD(P)H changes and mitochondrial respiration were measured. For each cell preparation, the extent of maximum reduction of the mitochondrial NAD(P)H pool was determined by adding 20 mM beta -hydroxybutyrate and 10 µM rotenone. Fluorescence changes are reported as percentage of maximum mitochondrial reduction. Mitochondrial respiration was calculated by subtracting the myxothiazol-insensitive respiration from the cellular respiration rate. Values are means ± SD, n = 5. 

Interventions Mitochondrial respiration NAD(P)H

  % change
 beta -Hydroxybutyrate (20 mM) 16.9  ± 5.1 40.7  ± 9.0
Acetoacetate (20 mM)  -16.1  ± 4.5  -43.2  ± 9.8
Gramidicin D (10 µg/ml) 22.0  ± 5.2  -13.5  ± 2.9
Myxothiazol (50 ng/ml)  -66.9  ± 17.8 60.6  ± 13.4
Oligomycin (5 µg/ml)  -70.4  ± 6.1 62.7  ± 7.0
Atractyloside (100 µM)  -75.4  ± 5.8 71.4  ± 3.5


Fig. 10. Effect of hypoxic incubation on the control of respiration by NADH. beta -Hydroxybutyrate (20 mM) increased [NAD(P)H] and increased respiration. Acetoacetate (20 mM) decreased [NAD(P)H] and decreased respiration. Gramicidin D (10 µg/ml) decreased [NAD(P)H] and increased respiration. Myxothiazol (50 ng/ml) increased [NAD(P)H] and decreased respiration. Both oligomycin (5 mg/ml) and atractyloside (100 µM) increased [NAD(P)H] and decreased respiration. The changes in respiration in response to these changes in [NAD(P)H] were similar in cells incubated at PO2 = 100 torr and 20 torr.
[View Larger Version of this Image (27K GIF file)]

An experimental intervention that alters the rate of electron transfer from NADH to O2 will elicit a change in [NADH] reflective of how the supply of reducing equivalents from the Krebs cycle to NADH is regulated. We studied this regulation in hepatocytes by measuring the changes in NADH when electron transport was partially inhibited by administering myxothiazol (50 ng/ml), or when electron transport was accelerated by adding gramicidin D (10 µg/ml) to stimulate Na+-K+-ATPase activity. The myxothiazol caused a 66% decrease in mitochondrial respiration, which elicited a 58% increase in NAD(P)H fluorescence. The gramicidin D caused a 26% increase in respiration and a 17% decrease in NAD(P)H fluorescence (Table I). The slope of the line relating change in [NAD(P)H] and the change in respiration was similar between hypoxic (0.83) and normoxic cells (0.82), suggesting that the control of electron supply to NADH was not influenced by hypoxic incubation (Fig. 10).

The above relationships were used to calculate the respiratory control coefficient for NADH, which is the ratio of the slope relating Delta [NAD(P)H]/Delta VO2 (obtained from gramicidin D and myxothiazol studies) to the Delta VO2/Delta [NAD(P)H] (generated by the titration of ketone bodies). The overall control coefficient was 0.24 ± 0.03 during normoxia compared with 0.23 ± 0.03 in hypoxia. When comparing the slope generated by myxothiazol, oligomycin, and carboxyatractyloside to that obtained with addition of ketone bodies, a value of 0.35 ± 0.04 was obtained during normoxia, and a value of 0.32 ± 0.04 was found during hypoxia. Thus the overall control coefficient of NADH over mitochondrial respiration ranged from 0.2-0.38 in both normoxic and hypoxic groups. The similar control coefficients seen in the normoxic and hypoxic cells suggest that the control exerted by NADH over mitochondrial respiration was not influenced by hypoxia.


DISCUSSION

Mitochondrial respiration involves the transfer of reducing equivalents from metabolic substrates to O2. As reducing equivalents are shuttled through the electron transport chain, some protons are pumped out of the mitochondrial matrix while others are chemically consumed, thereby generating an electrochemical gradient across the inner mitochondrial membrane consisting of a membrane potential (Psi ) and pH gradient. The ATP synthase obtains the free energy needed for ATP synthesis from the return of protons down this gradient. Many factors are known to influence the control of mitochondrial respiration under different conditions. For example, models have been proposed to characterize the regulation of respiration by [ADP] + [Pi] (16), the phosphorylation potential (17), the supply of reducing equivalents to NADH (14), the ATP synthase, the adenine nucleotide carrier (18), and mitochondrial [Ca2+] (19), and the interdependence of these factors in a multisite regulatory model (20). However, our understanding of the regulation of respiration in intact cells during prolonged hypoxia is still incomplete.

In the present study and in a previous report (6) we found that hepatocytes decrease their rate of O2 uptake when kept under physiological levels of hypoxia for several hours. These changes were reversed when the PO2 was restored to 100 torr and were not associated with a loss of cell viability. Moreover, lactate production by the cells did not increase significantly during prolonged hypoxia, indicating that anaerobic ATP production was not increased. Collectively, these observations suggest that hepatocytes may be capable of suppressing cellular ATP utilization at O2 tensions well above the critically low levels necessary to limit respiration. The present study sought to clarify the narrower question of whether the changes in cytochrome oxidase observed in vitro mediate the detection of O2 within the intact cell during prolonged hypoxia.

Effects of Hypoxia on the Regulation of Respiration by [NADH]

To determine whether processes upstream or downstream of NADH contributed to the suppression of respiration during hypoxia, we examined the effects of hypoxia on the control exerted by NADH on mitochondrial respiration. An inhibition of the matrix dehydrogenases could decrease the supply of reducing equivalents to NADH, which could reduce O2 consumption and decrease mitochondrial membrane potential. Such a response could conceivably decrease the phosphorylation potential and thereby affect ATP synthesis. However, the data showed that NADH levels were increased in hepatocytes incubated under hypoxia, which indicated that prolonged hypoxia was associated with a change in the regulation of electron transport downstream from the NADH pool. Yet the control exerted by NADH supply over mitochondrial respiration was similar in the two groups. In both groups, the changes in [NAD(P)H] were large compared with the changes in O2 consumption, indicating that the control of respiration by [NADH] is small compared with that exerted by ATP. We therefore conclude that inhibition of matrix dehydrogenases is unlikely to contribute importantly to the hypoxic suppression of respiration. Rather, the responsible mechanism appears to reside exclusively within the mitochondrial electron transport chain downstream of NADH supply.

Mitochondrial Membrane Potential

The mitochondrial potential provides the driving force for ATP synthesis, and its measurement can provide insight into the factors contributing to the suppression of respiration in hypoxia. Factors that increase the flux of reducing equivalents in the electron transport chain tend to increase membrane potential by augmenting proton pumping. Conversely, factors that increase the entry of protons across the inner mitochondrial membrane such as the ATP synthase or proton leaks tend to decrease mitochondrial potential. During steady state the mitochondrial potential therefore reflects a balance between the rate of electron transport and the rate of ATP utilization by the cell. If the suppression of respiration during prolonged hypoxia occurred via an inhibition of the matrix dehydrogenases or the electron transport chain without a change in the rate of ATP utilization, then mitochondrial potential would decrease. Such a depolarization could conceivably limit the rate of ATP synthesis and consequently the rate of ATP utilization. On the other hand, if hypoxia suppressed respiration by inhibiting ATP utilization at the ATPases directly, then hyperpolarization of the membrane would be predicted in response to the decrease in ADP availability.

We found that reoxygenation of cells that had been incubated under hypoxia caused an abrupt restoration in O2 uptake and a significant increase in mitochondrial potential. This suggests that the membrane potential was depolarized during hypoxia, as a consequence of a decrease in mitochondrial electron transport. The results also suggest that electron transport must have increased acutely at reoxygenation without any change in the rate of ATP utilization. Interestingly, acute hypoxia elicited no decrease in potential. However, neither basal nor TMPD respiration decreased until cells had been kept hypoxic for 2-3 h, suggesting that the changes in respiratory control required several hours of hypoxia to become apparent. These results were consistent with the changes in cytochrome oxidase function observed in submitochondrial particles, where a decrease in the apparent Vmax required 3-4 h of hypoxic incubation. These observations implicate cytochrome oxidase as a possible oxygen sensor in intact hepatocytes, through an effect of hypoxia on the kinetic activity of the enzyme. However, the significance of the changes in membrane potential as a component in the signal transduction of hypoxia is not yet known.

Cytochrome Oxidase Function

Experiments using submitochondrial particles supplied with TMPD and ascorbate as substrates confirmed that changes in the kinetic function of the electron transport system paralleled the changes in cellular respiration and membrane potential that were observed during hypoxia and reoxygenation. In this regard, incubation of submitochondrial particles under prolonged hypoxia induced a reversible decrease in the rate of enzymatic turnover at each concentration of TMPD. Because TMPD reduces cytochrome c, measurements of respiration with TMPD reflect the activity of the cytochrome c-cytochrome oxidase complex directly. Assuming that cytochrome c reduction increases as the concentration of TMPD was increased (see Fig. 4, bottom), these results suggest that prolonged hypoxia acts on cytochrome oxidase to cause a decrease in its turnover rate at any given level of cytochrome c reduction. This observation is consistent with our previous findings demonstrating that the purified oxidase undergoes a reversible decrease in its apparent Vmax when incubated in buffer solutions at similar levels of hypoxia (9). Collectively, these observations implicate a physiological role for cytochrome oxidase in the cellular response to hypoxia and suggest a mechanism by which this oxidase can function as a cellular O2 sensor.

These results generate two important questions. First, how could cytochrome oxidase function as an oxygen sensor at O2 concentrations much higher than its apparent Km for oxygen? Second, why should a 50% decrease in the apparent Vmax of the enzyme cause a decrease in respiration if the enzyme normally functions at only a fraction of its maximal velocity? A conceptual model that addresses these questions is shown in Fig. 11, which is based on the experimental data shown in Fig. 4, bottom. Note that cytochrome c and the oxidase are normally present in roughly equal concentrations in mitochondria (11). Under normoxic conditions (point A), cytochrome c exists in an incompletely reduced state, and cytochrome oxidase operates well below its maximal capacity (Vmax). Prolonged hypoxia appears to decrease the apparent Vmax of the oxidase, as evidenced by (a) the decrease in TMPD respiration observed in intact cells (Fig. 3), (b) the decrease in enzymatic turnover seen in submitochondrial particles incubated under hypoxia (Fig. 4), or (c) the decrease in Vmax measured for isolated cytochrome oxidase (9). If a shift to this "hypoxic" state could occur instantaneously, O2 consumption would fall acutely (Fig. 11, top, point B). However, continued electron flux would cause an increase in cytochrome c reduction and tend to restore O2 uptake. Indeed, the normoxic level of O2 uptake could be maintained if cytochrome c were sufficiently reduced. However, a simultaneous decrease in the rate of ATP utilization by the cell would produce the inhibited state (Fig. 11, top, point C) seen after 3-4 h of hypoxia (see Fig. 1, t = 3-5 h). We did not measure cytochrome c reduction directly, but we did observe an increase in [NAD(P)H] in cells incubated under prolonged hypoxia, which suggests that cytochrome c reduction increased.


Fig. 11. Hypothetical relationship between cytochrome oxidase function and cellular respiration during hypoxia. Top, during prolonged hypoxia, cytochrome oxidase undergoes a change in functional state from its "normoxic" to "hypoxic" state (A right-arrow B). However, respiration can be maintained by increases in cytochrome c reduction (B right-arrow C). Bottom, upon reoxygenation, cytochrome oxidase undergoes a rapid return to its "normoxic" state. This is associated with acute increases in electron transport, respiration, and mitochondrial potential (C right-arrow D). In steady state normoxia, the basal level of respiration is restored (D right-arrow A).
[View Larger Version of this Image (16K GIF file)]

Reoxygenation of cytochrome oxidase appears to cause an immediate restoration of its normal Vmax, as evidenced by (a) the rapid increase in electron transport and O2 uptake observed in cells and submitochondrial particles during reoxygenation, (b) the rapid increase in cellular TMPD respiration seen at reoxygenation, and (c) the decrease in [NAD(P)H] and the increase in mitochondrial potential seen upon reoxygenation. These changes can be interpreted according to the model shown in Fig. 11, bottom. Cells adapted to hypoxia (point C) undergo a rapid increase in respiration when reoxygenated, consistent with a rapid return of the oxidase to its native state (Fig. 11, bottom, point D). The increase in O2 uptake would cause an acute decrease in cytochrome c reduction. In steady state, ATP utilization would be restored and the cell would return to its normoxic state (point A).

An assumption of the above analysis is that the decrease in respiration during prolonged hypoxia reflects a decrease in the rate of ATP utilization. In cultured cardimyocytes, prolonged moderate hypoxia caused a reversible decrease in respiration and contractile motion without a decrease in [ATP] or phosphocreatine levels (7), which was indicative of a decrease in ATP utilization. We speculate that a similar inhibition of ATP-dependent reactions occurs in hepatocytes during hypoxia. The results of the present study are supportive of that hypothesis, but additional studies will be required to test that question directly.

Under steady state conditions, ATP synthesis must match ATP utilization in the intact cell. It is unlikely that the decrease in respiration observed during prolonged hypoxia is solely a consequence of an inhibition of the oxidase, because such a limitation would limit ATP synthesis and deplete cellular ATP stores, unless ATP utilization were also inhibited. Data from the present study are consistent with the hypothesis that cytochrome oxidase acts as the O2 sensor during prolonged hypoxia. Through a signaling sequence that is not yet known, we speculate that mitochondrial responses to hypoxia initiate a reversible suppression of ATP utilization elsewhere in the cell. Reoxygenation rapidly reverses the inhibition of cytochrome oxidase, which increases mitochondrial potential by restoring electron transport while ATP utilization remains inhibited. It is conceivable that the changes in mitochondrial potential or cytochrome c reduction shown in Fig. 11 participate in this signaling sequence. However, further understanding of the details of this signaling cascade and its significance for O2 sensing in other cell types will require additional study.


FOOTNOTES

*   This work was supported by NHLBI, National Institutes of Health Grants HL32646 and HL35440.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dagger    To whom correspondence should be addressed: Pulmonary and Critical Care Medicine, the University of Chicago, MC6026, 5841 South Maryland Ave., Chicago, IL 60637. Tel.: 773-702-9363; Fax: 773-702-4736; E-mail: pschumac{at}medicine.bsd.uchicago.edu.
1   The abbreviation used is: TMPD, N,N,N',N'-tetramethyl-p-phenylenediamine.

REFERENCES

  1. Wilson, D. F., Rumsey, W. L., Green, T. J., and Vanderkooi, J. M. (1988) J. Biol. Chem. 263, 2712-2718 [Abstract/Free Full Text]
  2. Jones, D. P. (1986) Am. J. Physiol. 250, C663-C675 [Abstract/Free Full Text]
  3. Kennedy, F. G., and Jones, D. P. (1986) Am. J. Physiol. 250, C374-C383 [Abstract/Free Full Text]
  4. Warburg, O., and Kubowitz, F. (1929) Biochem. Z. 214, 5-18
  5. Jones, D. P., and Mason, H. S. (1978) J. Biol. Chem. 253, 4874-4880 [Abstract]
  6. Schumacker, P. T., Chandel, N., and Agusti, A. G. N. (1993) Am. J. Physiol. 265, L395-L402 [Abstract/Free Full Text]
  7. Budinger, G. R. S., Chandel, N., Shao, Z. H., Li, C. Q., Melmed, A., Becker, L. B., and Schumacker, P. T. (1996) Am. J. Physiol. 14, L37-L53
  8. Chandel, N., Budinger, G. R. S., Kemp, R. A., and Schumacker, P. T. (1995) Am. J. Physiol. 268, L918-L925 [Abstract/Free Full Text]
  9. Chandel, N. S., Budinger, G. R. S., and Schumacker, P. T. (1996) J. Biol. Chem. 271, 18672-18677 [Abstract/Free Full Text]
  10. Seglen, P. (1972) Exp. Cell. Res. 74, 450-454 [Medline] [Order article via Infotrieve]
  11. Cooper, C. E. (1990) Biochim. Biophys. Acta 1017, 187-203 [Medline] [Order article via Infotrieve]
  12. Ardail, D., Gasnier, F., Lermé, F., Simonot, C., Louisot, P., and Gateau-Roesch, O. (1993) J. Biol. Chem. 268, 25985-25992 [Abstract/Free Full Text]
  13. Farkas, D. L., Wei, M.-D., Febbroiello, P., Carson, J. H., and Loew, L. M. (1989) Biophys. J. 56, 1053-1069 [Abstract]
  14. Hafner, R. P., Brown, G. C., and Brand, M. D. (1990) Eur. J. Biochem. 188, 313-319 [Abstract]
  15. Brown, G. C., Lakin-Thomas, P. L., and Brand, M. D. (1990) Eur. J. Biochem. 192, 355-362 [Abstract]
  16. Chance, B., and Williams, G. R. (1955) J. Biol. Chem. 217, 383-393 [Free Full Text]
  17. Erecinska, M., and Wilson, D. F. (1982) J. Membr. Biol. 70, 1-14 [Medline] [Order article via Infotrieve]
  18. Klingenberg, M. (1980) J. Membr. Biol. 56, 97-105 [Medline] [Order article via Infotrieve]
  19. Kadenbach, B., Barth, J., Akgun, R., Freund, R., Linder, D., and Possekel, S. (1995) Biochim. Biophys. Acta 1271, 103-109 [Medline] [Order article via Infotrieve]
  20. Jones, D. P., Shan, X., and Park, Y. (1992) Annu. Rev. Nutr. 12, 327-343 [CrossRef][Medline] [Order article via Infotrieve]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.