Mitochondrial hydrogen peroxide production alters oxygen consumption in an oxygen-concentration-dependent manner

Free Radical Biology & Medicine 38 (2005) 1594 – 1603 www.elsevier.com/locate/freeradbiomed

Original Contribution

Mitochondrial hydrogen peroxide production alters oxygen consumption in an oxygen-concentration-dependent manner Shane E. Munns, James K.C. Lui, Peter G. ArthurT School of Biomedical & Chemical Sciences, M310, University of Western Australia, Crawley, WA 6009, Australia Received 25 August 2004; revised 2 February 2005; accepted 22 February 2005 Available online 19 March 2005

Abstract Metabolic responses of mammalian cells toward declining oxygen concentration are generally thought to occur when oxygen limits mitochondrial ATP production. However, at oxygen concentrations markedly above those limiting to mitochondria, several mammalian cell types display reduced rates of oxygen consumption without energy stress or compensatory increases in glycolytic ATP production. We used mammalian Jurkat T cells as a model system to identify mechanisms responsible for these changes in metabolic rate. Oxygen consumption was 31% greater at high oxygen (150–200 AM) compared to low oxygen (5–10 AM). Hydrogen peroxide was implicated in the response as catalase prevented the increase in oxygen consumption normally associated with high oxygen. Cell-derived hydrogen peroxide, predominately from the mitochondria, was elevated with high oxygen. Oxygen consumption related to intracellular calcium turnover was shown, through EDTA chelation and dantrolene antagonism of the ryanodine receptor, to account for 70% of the response. Oligomycin inhibition of oxygen consumption indicated that mitochondrial proton leak was also sensitive to changes in oxygen concentration. Our results point toward a mechanism in which changes in oxygen concentration influence the rate of hydrogen peroxide production by mitochondria, which, in turn, alters cellular ATP use associated with intracellular calcium turnover and energy wastage through mitochondrial proton leak. D 2005 Elsevier Inc. All rights reserved. Keywords: Mitochondria; Hydrogen peroxide; Hypoxia; Metabolic rate; Proton leak; Oxygen sensing; Free radicals

ATP production from oxidative phosphorylation is critically dependent on an adequate oxygen supply. An inadequate oxygen supply results in a decline in pericellular oxygen concentration and elicits a multitude of responses in mammals aimed at maintaining ATP production. Systemic responses include increased cardiac output, red blood cell production, and the redistribution of blood flow (systemic arterial vasodilation and pulmonary arterial vasoconstriction) [1]. At the molecular level a declining oxygen concentration activates transcription of proteins such as

Abbreviations: FCCP, carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone; HBS, Hepes-buffered saline; CM-H2DCFDA, 5-(and 6-)chloromethyl-2V,7V-dichlorodihydrofluorescein diacetate; JC-1, 5,5V,6,6Vtetrachloro-1,1V,3,3V-tetraethylbenzimidazolylcarbocyanine iodide; EFA, ethoxyformic anhydride diethyl pyrocarbonate. T Corresponding author. Fax: (08) 6488 1148. E-mail address: [email protected] (P.G. Arthur). 0891-5849/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.freeradbiomed.2005.02.028

erythropoietin, nitric oxide synthase, and vascular endothelial growth factor [2]. These responses are designed to increase oxygen delivery to cells, minimizing the effects of the reduced oxygen supply. Metabolic responses of mammalian cells toward declining oxygen concentration are generally thought to occur when the concentration of oxygen limits cytochrome oxidase activity. It is not certain when oxygen becomes limiting, but it has been suggested to be less than 3 AM (0.2 kPa) for isolated mitochondria [3]. A lack of oxygen at the mitochondria typically causes a decline in oxygen consumption, cellular energy stress (reduced ATP synthesis leading to a decrease in ATP concentration), and a partial compensatory increase in the rate of glycolytic ATP production [4]. However, a novel metabolic response has been identified in which changes in the oxygen consumption of mammalian cells and tissue have been found to occur at oxygen concentrations markedly above (up to 90 AM)

S.E. Munns et al. / Free Radical Biology & Medicine 38 (2005) 1594–1603

those limiting to the mitochondria. Characteristically, a decline in oxygen concentration causes a decrease in oxygen consumption, but without the expected cellular energy stress (e.g., decline in ATP or phosphocreatine if present) or compensatory increase in glycolytic ATP production [5–7]. This response has been identified in a variety of mammalian models, including isolated primary hepatocytes [5], isolated primary cardiomyocytes [6], the C2C12 muscle cell line [7], whole tissue muscle hind-limb preparations [8,9], and in vivo hypoxic studies on rats [10]. There has also been a suggestion that low oxygen concentration affects cardiac basal metabolism [11]. The biological significance of this response toward low oxygen concentration is not known, but it may be related to prolonging function and/or viability of cells during times of compromised oxygen supply [12,13]. A reduced rate of oxygen consumption without a compensatory increase in glycolytic ATP production implies that the rate of ATP synthesis is decreased at low oxygen. A feature of this response, however, is the maintenance of intracellular ATP concentration at low oxygen. This implies that ATP-using processes decrease in conjunction with the rate of oxygen consumption, permitting the cell to maintain the concentration of intracellular ATP. The adjustment of cellular ATP use in response to declining oxygen concentrations suggests an intricate pathway of metabolic regulation via oxygen sensing. We have utilized mammalian Jurkat T cells as a model system, because they can be cultured and used in suspension, to identify the mechanisms underlying the oxygen-concentration-sensitive changes in cellular oxygen consumption. The main objectives of our study were to: (1) elucidate the mechanism through which oxygen consumption (cellular ATP/energy use) was altered by oxygen concentration and (2) identify ATP/energy-using processes sensitive to oxygen concentration. Our results point toward a mechanism in which changes in oxygen concentration influence the rate of hydrogen peroxide production by mitochondria, which, in turn, alters cellular ATP use associated with intracellular calcium turnover and energy wastage through mitochondrial proton leak.

Materials and methods Jurkat cell culture Jurkat cells were cultured in a spinner flask with continuous agitation (75 rpm) at 378C in a humidified incubator with 95% air and 5% CO2. Culture medium (RPMI 1640) was supplemented with 10% FCS, 2 mM glutamine, 50 IU/ml penicillin, and 50 Ag/ml streptomycin. All experiments and cell preparations were carried out in darkened work areas under red light to limit the production of reactive oxygen species from the exposure of culture/ experimental medium to light [14].

1595

Jurkat cell preparation for experiments The experimental medium was a Hepes-buffered saline (HBS) containing (in mM) Hepes, 20; Ca(NO3)2, 0.42; KCl, 5.33; MgSO4, 0.41; NaCl, 139; Na2HPO4, 5.63; glucose, 5; glutamine, 2, at pH 7.4, supplemented with FCS (10%). Unless otherwise stated, Jurkat cells were prepared for experiments by centrifugation (3 min
500g) and then resuspended at a cell concentration of 1
106 cells/ml in 378C HBS pregassed with a humidified low-oxygen (10 AM) gas mix. Aliquots of Jurkat cells were transferred to the experimental chamber and incubated for 30 min at low (10 AM) or high oxygen (150–200 AM) before experimental use. Oxygen consumption A Clark-type electrode was used to measure rates of oxygen consumption [15]. Briefly, Jurkat cells were suspended in HBS and stirred in a closed glass chamber maintained at 378C. Within the sealed closed chamber the oxygen electrode measured decreases in oxygen concentration as a function of cellular oxygen consumption. Oxygen consumption measurements taken between 5 and 10 AM were defined as low oxygen and between 150 and 200 AM as high oxygen. Fluorescence measurements Fluorescence measurements were carried out in a Shimadzu RF5000 spectrofluorophotometer modified to contain a glass closed-chamber system and oxygen electrode. Fluorescence measurements from cell suspensions were taken through the walls of the glass closed chamber, allowing the monitoring of oxygen concentration while taking simultaneous measurement of fluorescence and cellular oxygen consumption. Intracellular hydrogen peroxide The rate of intracellular hydrogen peroxide production in Jurkat cells was monitored by 5-(and 6-)chloromethyl-2V,7Vdichlorodihydrofluorescein diacetate (CM-H2DCFDA), a derivative of 2V,7V-dichlorodihydrofluorescein diacetate more efficiently retained by cells [16]. CM-H2DCFDA (1 AM) was added to cells in the spectrofluorophotometer closed chamber and incubated for 10 min, then fluorescence was measured at the excitation/emission wavelengths of 488 and 526 nm, respectively. Extracellular hydrogen peroxide Hydrogen peroxide in the extracellular medium was measured by the horseradish peroxidase substrate Amplex red (N-acetyl-3,7-dihydroxyphenoxazine) [17]. Amplex red (10 AM) and horseradish peroxidase (1 unit/ml) were added to HBS in the spectrofluorophotometer closed chamber and

1596

S.E. Munns et al. / Free Radical Biology & Medicine 38 (2005) 1594–1603

measured at the excitation/emission wavelengths of 590 and 645 nm, respectively. Mitochondrial membrane potential Mitochondrial membrane potential was monitored using the lipophilic cationic probe 5,5V,6,6V-tetrachloro1,1V,3,3V-tetraethylbenzimidazolylcarbocyanine iodide (JC1; 200 nM) [18]. Jurkat cells were preincubated with JC-1 for 4 h at low oxygen in HBS at 378C to limit drift to a rate below 10% signal change per hour. Cells were then transferred to the closed chamber within the spectrofluorophotometer and fluorescence was monitored at the emission wavelengths of 535 nm for monomer and 590 nm for J aggregates after excitation at 514 nm. Intracellular calcium concentration Changes in intracellular calcium were monitored by the calcium-specific fluorescent probe indo-1. Jurkat cells were incubated with indo-1 (2 AM) for 25 min in 378C HBS FCS and then washed 2
in fresh 378C HBS FCS to remove extracellular indo-1. After loading, cells were transferred to the closed chamber within the spectrofluorophotometer and fluorescence was monitored, after excitation at 355 nm, at the emission wavelength of 400 nm. Fluorescent signal from indo-1 was converted to intracellular calcium concentration according to the equation 

Ca

2þ



 ¼ Kd

F  Fmin Fmax  F



where F represents fluorescent signal obtained from cells, F max represents maximal bound signal obtained after Triton X-100 (0.1%) treatment of cells, F min represents minimum bound signal obtained after EDTA (5 mM) treatment of cells, and K d is the dissociation constant for indo-1 and calcium [19]. Catalase inactivation To inactivate catalase a concentrated stock (25,000 units/ ml) was prepared in HBS. The solution was increased to pH 12 by the gradual addition of NaOH (10 M) and then returned to pH 7.4 by the gradual addition of HCl (1 M). Inactivation was verified by spectrophotometric assay of hydrogen peroxide stability in the presence of treated catalase (monitored at 240 nm). Lactate production and intracellular ATP concentration Rates of lactate production from Jurkat cells (3
106 cells/ml) were used to measure glycolytic ATP production at low or high oxygen. HBS was sampled at 15-min intervals and the concentration of lactate in the supernatant determined via spectrophotometric assay [20]. Intracellular ATP

concentration was monitored after 1 h incubation at low or high oxygen. Briefly, a sample of cells (250 Al) was taken via a Hamilton glass syringe and injected immediately into ice-cold perchloric acid (1.2 M) to deproteinize the cell suspension and inhibit further metabolic activity. Cell debris was pelleted by centrifugation (5 min
10,000g) and the supernatant was prepared for assay [21]. The concentration of ATP was measured spectrophotometrically [22]. Cell counts and viability Experimental cell number was determined by the mean of three separate hemacytometer counts from each experiment. Cell counts were carried out with the addition of eosin to determine cell viability by dye exclusion. Jurkat cell viability was maintained above 85% for all experimental protocols used in this study. Statistics All data points are expressed as means F SEM from measurements made on Jurkat cells taken from (n) separate experiments. Statistical difference was determined via Student’s t test. Data were considered significantly different when p b 0.05.

Results The oxygen-concentration-sensitive changes in cellular oxygen consumption are not associated with energy stress and occur through changes in mitochondrial oxygen consumption For experiments, all cell preparations were initially carried out at low oxygen because the transition from low to high oxygen was quicker and more reproducible than gassing cells from high to low oxygen. Jurkat cells displayed a 31% greater rate of oxygen consumption at high oxygen (150–200 AM) compared to low oxygen (5–10 AM) (Fig. 1), with no indications of energetic stress (low oxygen 6.8 F 0.3 nmol ATP/106cells vs high oxygen 7.0 F 0.2 nmol ATP/106cells, n = 3) or increases in glycolytic ATP production (low oxygen 13.7 F 1.4 nmol lactate/min/ 106cells vs high oxygen 14.9 F 1.4 nmol lactate/min/ 106cells, n = 3). The response was reversible with rates of oxygen consumption measured under low oxygen, after 40 min of high-oxygen incubation, not significantly different from the low-oxygen control (Fig. 1). Oxygen consumption in Jurkat cells was predominantly mitochondrial, with combined inhibition by ethoxyformic anhydride diethyl pyrocarbonate (EFA; complex I inhibitor, 11.5 mM) and myxothiazol (complex III inhibitor, 4 AM) reducing rates of oxygen consumption by 97% at low oxygen (to 0.6 F 0.1 pmol oxygen/s/106cells, n = 3) and 95% at high oxygen (to 1.2 F 0.1 pmol oxygen/s/106cells,

S.E. Munns et al. / Free Radical Biology & Medicine 38 (2005) 1594–1603

Fig. 1. Jurkat cells increase oxygen consumption after an increase from low oxygen (column 1, 5–10 AM) to high oxygen (column 2, 150–200 AM). Jurkat cells maintain normal rates of oxygen consumption when returned to low oxygen (column 3) after 40 min of high-oxygen incubation. *Significant difference ( p b 0.05) from low oxygen (n = 3).

n = 3). This predominance of mitochondrial oxygen consumption is consistent with previous studies on primary rat lymphocytes [23]. Hydrogen peroxide affects the rate of oxygen consumption in Jurkat cells

1597

from low oxygen to high oxygen for the same set of cells (Fig. 3a). Inhibition of mitochondrial function with the electron transport chain inhibitor myxothiazol (4 AM) combined with the mitochondrial proton ionophore carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone (FCCP; 3.5 AM) abolished the significant difference in hydrogen peroxide production rates (Fig. 3a). Hydrogen peroxide can cross cell membranes [24] so Amplex red, a cell-membrane-impermeable fluorescent dye, was used to measure hydrogen peroxide accumulation in the extracellular medium at low and high oxygen. HBS in the absence of cells produced hydrogen peroxide; however, this background production was not significantly different at low or high oxygen (Fig. 3b). For the same set of cells, hydrogen peroxide production increased after the transition from low oxygen to high oxygen (Fig. 3b). It was not possible to quantify hydrogen peroxide production before and after the transition in oxygen concentration for the same set of cells. However, for separate sets of cells the rate of hydrogen peroxide production measured at low oxygen was 1.1 F 0.3 pmol/min/106 cells (n = 3) and 5.2 F 0.8 pmol/min/106 cells (n = 3) at high oxygen.

We enzymatically altered extracellular concentrations of hydrogen peroxide at low and high oxygen to establish whether these changes could affect the rate of oxygen consumption. Hydrogen peroxide crosses cell membranes so the intracellular concentration of hydrogen peroxide will change in response to alterations in the extracellular concentration [24]. Catalase (2500 units/ml), an enzyme specific for hydrogen peroxide degradation, reduced rates of oxygen consumption at high oxygen to a level that was not significantly different from rates of oxygen consumption with catalase at low oxygen (Fig. 2a). Inactivated catalase had no significant effect on the rate of oxygen consumption compared to noncatalase controls, excluding nonspecific actions of the catalase protein (Fig. 2a). Glucose oxidase (150 munits/ml), an enzyme specific for hydrogen peroxide generation, was used to establish a steady-state gradient of hydrogen peroxide between the extracellular and the intracellular environment [24]. Under these conditions, rates of oxygen consumption at low oxygen increased by 36% (Fig. 2b). Therefore manipulation of hydrogen peroxide concentration influenced cellular oxygen consumption in Jurkat cells independent of oxygen concentration. Hydrogen peroxide from Jurkat cells is produced in an oxygen-concentration-dependent manner and is predominantly mitochondrial in origin Hydrogen peroxide production from cells has been shown to be dependent upon oxygen concentration, so we measured hydrogen peroxide from Jurkat cells to establish whether changes in oxygen concentration altered the production of hydrogen peroxide [25–30]. Rates of intracellular hydrogen peroxide production, measured via CMH2DCFDA, were significantly increased after transition

Fig. 2. Manipulating extracellular hydrogen peroxide concentration alters rates of oxygen consumption. (a) Jurkat cells were prepared in HBS containing catalase (2500 units/ml) or inactivated catalase and then incubated for 30 min at low (n) or high oxygen (5) before oxygen consumption measurements. (b) Jurkat cells were preincubated for 40 min in the absence (n) or presence (5) of glucose oxidase (150 munits/ml) at low oxygen or low oxygen with EDTA (1 mM). Glucose oxidase generated 0.24 F 0.02 AM hydrogen peroxide/min (n = 3) at low oxygen concentration. Mitochondrial inhibitors (myxothiazol, 4 AM, and EFA, 11.5 mM) were used to account for oxygen consumption by glucose oxidase (data not shown). *Significant difference ( p b 0.05) from control (n = 3).

1598

S.E. Munns et al. / Free Radical Biology & Medicine 38 (2005) 1594–1603

84% (Fig. 4). The endoplasmic reticulum ATPase inhibitor thapsigargin (1 AM) was used acutely (measurement taken approximately 10 min after addition) to increase intracellular calcium by releasing endoplasmic reticulum stores [36]. Thapsigargin increased the rate of oxygen consumption at low oxygen thereby reducing the difference in oxygen consumption between low and high oxygen by 61% (Fig. 4). The ryanodine receptor antagonist dantrolene (20 AM) abolished the difference in oxygen consumption between low and high oxygen, indicating that calcium release from the endoplasmic reticulum through the ryanodine receptor-associated calcium channel contributed to the response (Fig. 4). There was no detectable change in intracellular calcium concentration during the transition from moderate low oxygen to high oxygen

Fig. 3. Hydrogen peroxide production is associated with increased oxygen concentration. (a) Intracellular hydrogen peroxide measured by CMH2DCFDA (1 AM) in Jurkat cells was produced in an oxygen-concentration-dependent manner (n). The mitochondrial inhibitor myxothiazol (4 AM) and the proton ionophore FCCP (3.5 AM) blocked the increase in fluorescence at high oxygen (5). (b) Extracellular hydrogen peroxide, measured by the membrane-impermeant fluorescent dye Amplex red (10 AM), was produced in an oxygen-concentration-dependent manner in the presence of Jurkat cells (n). HBS medium without cells produced hydrogen peroxide but this background fluorescence was not significantly different between low and high oxygen (5). *Significant difference ( p b 0.05) from low oxygen (n = 3).

Intracellular calcium turnover influences oxygen consumption in an oxygen-concentration-dependent manner Hydrogen peroxide has been shown to influence intracellular calcium turnover by releasing endoplasmic reticulum calcium stores and facilitating extracellular calcium influx [31–35]. We examined whether the oxygen-concentration-sensitive mitochondrial hydrogen peroxide production influenced cellular oxygen consumption, through increasing cellular ATP use associated with intracellular calcium turnover. Jurkat cells were exposed to glucose oxidase (hydrogen peroxide generator) in the presence of the chelator EDTA. Pretreatment of Jurkat cells with EDTA (1 mM), before glucose oxidase exposure, abolished the effect of the enzyme on the rate of oxygen consumption (Fig. 2b), indicating that a divalent ion, possibly calcium, could be involved in the oxygen-concentration-sensitive changes in cellular oxygen consumption. We manipulated aspects of intracellular calcium turnover in Jurkat cells to examine the effects on cellular oxygen consumption. Chelation of extracellular divalent ions with EDTA (1 mM) decreased the rate of oxygen consumption at high oxygen, reducing the difference in oxygen consumption between low and high oxygen by

The fluorescent dye indo-1 was used to determine whether mitochondrial hydrogen peroxide production was altering cellular oxygen consumption through changes in intracellular calcium concentration. No increase in intracellular calcium was detectable during the transition from low to high oxygen compared to the low-oxygen control (Fig. 5). Exogenous hydrogen peroxide was added to Jurkat cells to establish whether elevated hydrogen peroxide concentration could increase intracellular calcium concentration in Jurkat cells (Fig. 6). Indo-1 fluorescent signal increased significantly with the addition of 50 AM hydrogen peroxide. Therefore intracellular calcium concentration in Jurkat cells can be influenced by hydrogen peroxide concentration. However, during the transition from low to high oxygen changes in mitochondrial hydrogen peroxide production did not cause a detectable change in intracellular calcium by this method.

Fig. 4. Manipulation of intracellular calcium turnover alters oxygen consumption in Jurkat cells at low oxygen (n) and high oxygen (5). EDTA (1 mM), a chelator of divalent ions, and dantrolene (Dant.; 20 AM), an antagonist of the endoplasmic reticulum ryanodine receptor, decreased the oxygen-concentration-sensitive changes in cellular oxygen consumption. Thapsigargin (Thaps.; 1 AM), a stimulator of endoplasmic reticulum calcium release, acutely increased the oxygen consumption of cells at low oxygen. Jurkat cells were exposed to thapsigargin or dantrolene for approximately 10 min before oxygen consumption measurements. *Significant difference ( p b 0.05) from equivalent control (n = 3).

S.E. Munns et al. / Free Radical Biology & Medicine 38 (2005) 1594–1603

Fig. 5. There was no detectable increase in intracellular calcium concentration measured with indo-1 (2 AM) during the transition from low to high oxygen. Arrow indicates beginning of reoxygenation period (n = 3).

The increased rate of oxygen consumption at high oxygen reflects elevated cellular ATP use and increased energy wastage through mitochondrial proton leak The involvement of intracellular calcium release from the endoplasmic reticulum suggested that changes in ATP ion pump activity were involved in the oxygen-concentration-sensitive changes in cellular oxygen consumption. To test this, ATP use by the endoplasmic reticulum calcium ATPase and the plasma membrane calcium ATPase was inhibited, as these enzymes represent the major ATPases regulating calcium turnover in Jurkat cells [37]. We also inhibited the plasma membrane sodium potassium ATPase, as this enzyme is indirectly involved in the transport of intracellular calcium via the sodium calcium exchanger [37]. Combined inhibition with thapsigargin (endoplasmic reticulum ATPase, 1 AM), vanadate [38] (plasma membrane calcium ATPase, 1 mM), and ouabain (sodium potassium ATPase, 2 mM) accounted for approximately 70% of the oxygen-concentration-sensitive changes in cellular oxygen consumption (Fig. 7). Thapsigargin exposure for this experiment was long term (approximately 30 min) to allow the transient increase in intracellular calcium to pass before oxygen consumption measurements were taken.

Fig. 6. Sensitivity of indo-1 toward changes in intracellular calcium concentration induced by exogenous hydrogen peroxide addition to Jurkat cells. Indo-1 (2 AM)-loaded Jurkat cells were exposed to increasing concentrations of hydrogen peroxide at 5-min intervals. Intracellular calcium concentration was measured 2 min after peroxide addition. *Significant difference ( p b 0.05) from 0 AM hydrogen peroxide (n = 3).

1599

Fig. 7. The increase in oxygen consumption at high oxygen was related to an increase in intracellular calcium turnover. Jurkat cells were exposed to thapsigargin (1 AM) to deplete endoplasmic reticulum stores of calcium and then incubated for 30 min to allow excess intracellular calcium to be removed. Ouabain (2 mM) and vanadate (1 mM) were then added simultaneously at high (n) or low oxygen (5). Rates of oxygen consumption were measured immediately after the addition of ouabain and vanadate (labeled inhibitors). *Significant difference ( p b 0.05) from equivalent control (n = 3).

To further verify that cellular ATP use was changing during the oxygen-concentration-sensitive changes in cellular oxygen consumption we inhibited oxygen consumption related to ATP use with oligomycin (1 AM), an inhibitor of the mitochondrial ATP synthase. Oligomycin reduced the change in oxygen consumption caused by the change in oxygen concentration from 5.4 to 1.3 pmol oxygen/s/106 cells, confirming that oxygen consumption related to cellular ATP use was sensitive to oxygen concentration. The residual oxygen consumption after oligomycin treatment (low oxygen 4.6 F 0.2 pmol oxygen/s/106cells vs high oxygen 5.9 F 0.4 pmol oxygen/s/106cells, n = 4, p b 0.05) was also sensitive to changes in oxygen concentration, indicating that energy wastage through mitochondrial proton leak is additionally involved in the oxygen-concentrationsensitive changes in cellular oxygen consumption. Mitochondrial membrane potential depolarizes with the transition from low to high oxygen The changes in mitochondrial oxygen consumption were likely associated with changes in mitochondrial membrane potential. Therefore we monitored mitochondrial membrane potential during the transition from low to high oxygen with the fluorescent cationic dye JC-1. JC-1 was shown through mitochondrial-specific inhibitors to be a sensitive indicator of mitochondrial membrane potential (Fig. 8a). Mitochondrial membrane potential underwent depolarization during the transition from low to high oxygen, a result consistent with an increase in the activity of processes that dissipate the proton gradient (cellular ATP use, proton leak) (Fig. 8b). To show that the depolarization was not an artifact of JC-1 fluorescence changing with oxygen concentration FCCP was added to Jurkat cells under low oxygen conditions to increase the rate of oxygen consumption to high oxygen rates, thereby mimicking the oxygen-concentration-sensitive changes in cellular oxygen consumption. FCCP at a

1600

S.E. Munns et al. / Free Radical Biology & Medicine 38 (2005) 1594–1603

Discussion

Fig. 8. (a) JC-1 is a sensitive indicator of mitochondrial membrane potential. The mitochondrial uncoupler FCCP (3.5 AM) depolarized membrane potential (5), whereas the mitochondrial ATP synthase inhibitor oligomycin (1 AM) increased membrane potential (n). Arrow indicates time of mitochondrial inhibitor addition. Inset shows expanded scale of JC-1 fluorescence from the time of oligomycin addition. (b) Mitochondrial membrane potential depolarized during the transition from low to high oxygen (5) relative to cells maintained at low oxygen (n). Reduced concentrations of FCCP (110 nM) added at low oxygen increased cellular oxygen consumption and depolarized mitochondrial membrane potential (o) to a similar extent compared to the transition from low to high oxygen. Arrow indicates beginning of the transition from low to high oxygen or time of FCCP addition (n = 3).

concentration of 110 nM mimicked the oxygen-concentration-sensitive changes in cellular oxygen consumption (data not shown) and caused mitochondrial membrane potential to decrease, similar to the depolarization measured during the low to high oxygen transition (Fig. 8b). Collectively these data indicate that mitochondrial membrane potential was most likely depolarized by the increase in oxygen consumption, related to elevated cellular ATP use and/or mitochondrial proton leak, at high oxygen.

The molecular mechanisms underlying the oxygenconcentration-sensitive changes in cellular oxygen consumption are not known. Our results point toward a mechanism involving changes in oxygen concentration affecting hydrogen peroxide production by mitochondria which, in turn, alters cellular ATP use, the majority of which was associated with intracellular calcium turnover and energy wastage through mitochondrial proton leak. Jurkat T cells were a useful model system to examine this metabolic response because they can be cultured and used in suspension. In suspension cell culture it is possible to directly monitor the oxygen concentration that cells are exposed to, whereas this is more difficult with attached cell cultures because of the considerable oxygen concentration gradients formed [39]. The dependence of cellular hydrogen peroxide production on oxygen concentration is consistent with other studies using mammalian cells that have measured hydrogen peroxide directly [25–30] or indirectly, through function-related processes [40–45]. Dismutation of superoxide anion from the mitochondrial electron transport chain has long been recognized as a significant source of cellular hydrogen peroxide [46] and, as indicated by our results, mitochondria are the source for the greater part of hydrogen peroxide responsible for the oxygen-concentration-sensitive changes in oxygen consumption in Jurkat cells. We did not examine the mitochondrial source of superoxide anion; however, complex I and/or complex III is a suggested site [47–51]. Hydrogen peroxide production has been observed to increase under hypoxic conditions in chick cardiomyocytes, findings contrary to our proposed model [52]. The relevance of these findings to our model is not clear, however, as intracellular ATP concentration and glycolytic rate were not monitored during hypoxia in the chick cardiomyocyte study. It is possible that the increased generation of hydrogen peroxide was the result of oxygen limitation to mitochondrial respiration, a condition known to predispose cells to hydrogen peroxide production [53]. In support of our model, a study on mammalian cardiac myocytes has shown hydrogen peroxide production to be reduced during hypoxia [30]. In addition, we have demonstrated that catalase (an enzyme specific for hydrogen peroxide degradation) mimicked low oxygen conditions and glucose oxidase (a specific generator of hydrogen peroxide) mimicked high oxygen conditions. Our observations indicate that hydrogen peroxide affected calcium turnover and there is considerable evidence for this concept in a variety of mammalian cells. For example, the L-type calcium channel in guinea pig cardiac myocytes responds to decreases in hydrogen peroxide concentration by attenuating basal calcium conductance across the plasma membrane [54]. Exposure of canine endothelial cells to exogenous hydrogen peroxide elevates intracellular calcium through influx of calcium across the

S.E. Munns et al. / Free Radical Biology & Medicine 38 (2005) 1594–1603

plasma membrane and by releasing intracellular endoplasmic reticulum calcium stores [32]. Similar responses toward exogenous hydrogen peroxide are seen in smooth muscle cells with the resultant increase in intracellular calcium sensitive to catalase inhibition [31]. Of particular relevance is the observation that hydrogen peroxide can enhance calcium release from the ryanodine receptor-associated channel through oxidation of redox-sensitive residues within the receptor protein complex [33–35,55]. In this study, ATP use related to intracellular calcium cycling was shown to increase with high oxygen. Exogenous hydrogen peroxide (50 AM) addition to Jurkat cells increased intracellular calcium concentration, confirming the proposed model of hydrogen peroxide-induced calcium leakage in Jurkat cells. However, no change in intracellular calcium concentration was detectable with indo-1 during the transition from low to high oxygen. Two possibilities may explain these results: 1. The indo-1 method was not sensitive enough to detect the change in intracellular calcium concentration induced by the transition from low to high oxygen. For calcium leakage to be the sole mechanism inducing the oxygen conformance response, a sustained increase in intracellular calcium concentration is required to produce the elevated steady-state oxygen consumption at high oxygen. However, increased ATP use by the ATPases regulating intracellular calcium concentration would limit the net rise in intracellular calcium. The indo-1 method could not detect an increase in intracellular calcium concentration induced by a 25 AM bolus addition of hydrogen peroxide. Using equations derived from a study on hydrogen peroxide metabolism in Jurkat cells, a 25 AM bolus dose of hydrogen peroxide would induce a 4 AM increase in intracellular hydrogen peroxide concentration [24]. Extracellular hydrogen peroxide concentrations for Jurkat cells are typically below 1 AM. Therefore, the increase in intracellular calcium concentration, induced by mitochondrial hydrogen peroxide production at high oxygen, may have been below the detection limits of the indo-1 method. 2. The calcium leakage induced by mitochondrial hydrogen peroxide production at high oxygen was matched via increased calcium clearance by hydrogen peroxideactivated ATPase(s). The Ca2+ ATPase from pulmonary smooth muscle has been shown to be activated via an oxidant-mediated (hydrogen peroxide) proteolytic process [56]. Hydrogen peroxide has additionally been shown to increase the V max of heart sarcolemmal Ca2+ ATPase [57]. Activation of Ca2+ ATPases by hydrogen peroxide provides a possible mechanism whereby ATP use related to calcium cycling can increase without a subsequent increase in intracellular calcium concentration. The mechanism through which hydrogen peroxide increases ATP use by ATPases associated with calcium leakage in Jurkat cells remains unidentified.

1601

The residual oxygen consumption after oligomycin treatment was sensitive to changes in oxygen concentration. This indicates that energy wastage through mitochondrial proton leak is involved in the oxygen-concentrationsensitive changes in cellular oxygen consumption. The sensitivity of mitochondrial proton leak to oxygen concentration has previously been shown in isolated rat cardiomyocytes; however, the mechanism underlying this response remains undetermined [6]. The involvement of hydrogen peroxide and/or calcium-sensitive potassium channels in the oxygen-concentration-sensitive mitochondrial proton leak warrants investigation in light of the current findings [58]. Cytochrome oxidase has been proposed as an oxygen sensor capable of controlling the rate of cellular oxygen consumption [59,60]. It is suggested that electron flux through cytochrome oxidase is reduced at low oxygen concentrations leading to a reduction in cellular oxygen consumption. Under such conditions low oxygen would be expected to depolarize mitochondrial membrane potential due to the reduction in electron transport and its coupled translocation of protons from the mitochondrial matrix [59]. Mitochondrial membrane potential in Jurkat cells was elevated at low oxygen (compared to high oxygen), excluding the cytochrome oxidase model as an explanation of our data. A more likely explanation is that increase in cellular ATP use and/or mitochondrial proton leak at high oxygen decreases mitochondrial membrane potential, which, in turn, reduces the resistance for proton translocation from the mitochondrial matrix, thus increasing electron flow and subsequently the rate of cellular oxygen consumption. In summary, we have identified hydrogen peroxide derived from the mitochondria as a key mediator responsible for the oxygen-concentration-sensitive changes in cellular oxygen consumption in Jurkat cells. ATP use related to intracellular calcium turnover and energy wastage via mitochondrial proton leak were two cellular processes identified as contributing to the changes in oxygen consumption. The threshold and saturation behaviors of these cellular processes in response to changes in hydrogen peroxide remain to be defined. Nevertheless, as intracellular calcium turnover and mitochondrial proton leak are ubiquitous metabolic pathways for mammalian cells the regulation of these pathways by oxygen concentration, as described in Jurkat cells, may provide a model for understanding how metabolic rates of mammalian cells are affected by cellular oxygen supply.

Acknowledgments This work was supported by grants from The Neurotrauma Research Program and the National Heart Foundation.

1602

S.E. Munns et al. / Free Radical Biology & Medicine 38 (2005) 1594–1603

References [1] Semenza, G. L. Perspectives on oxygen sensing. Cell 98:281 – 284; 1999. [2] Wang, G. L.; Semenza, G. L. Oxygen sensing and response to hypoxia by mammalian cells. Redox Rep. 2:89 – 96; 1996. [3] Gnaiger, E.; Steinlechner-Maran, R.; Mendez, G.; Eberl, T.; Margreiter, R. Control of mitochondrial and cellular respiration by oxygen. J. Bioenerg. Biomembr. 27:583 – 596; 1995. [4] Noll, T.; Koop, A.; Piper, H. M. Mitochondrial ATP-synthase activity in cardiomyocytes after aerobic–anaerobic metabolic transition. Am. J. Physiol. 262:C1297 – C1303; 1992. [5] Schumacker, P. T.; Chandel, N.; Agusti, A. G. N. Oxygen conformance of cellular respiration in hepatocytes. Am. J. Physiol. 265:L395 – L402; 1993. [6] Casey, T. M.; Arthur, P. G. Hibernation in noncontracting mammalian cardiomyocytes. Circulation 102:3124 – 3129; 2000. [7] Arthur, P. G.; Giles, J. J.; Wakeford, C. M. Protein synthesis during oxygen conformance and severe hypoxia in the mouse muscle cell line C2C12. Biochim. Biophys. Acta 1475:83 – 89; 2000. [8] Hogan, M. C.; Kurdak, S. S.; Arthur, P. G. Effect of gradual reduction in O2 delivery on intracellular homeostasis in contracting skeletal muscle. J. Appl. Physiol. 80:1313 – 1321; 1996. [9] Chinet, A. E.; Mejsnar, J. Is resting muscle oxygen uptake controlled by oxygen availability to cells? J. Appl. Physiol. 66: 253 – 260; 1989. [10] Mimura, Y.; Furuya, K. Mechanisms of adaptation to hypoxia in energy metabolism in rats. J. Am. Coll. Surg. 181:437 – 443; 1995. [11] Gibbs, C. L.; Loiselle, D. S. Cardiac basal metabolism. Jpn. J. Physiol. 51:399 – 426; 2001. [12] Hochachka, P., Guppy, M., eds. Metabolic arrest and the control of biological time. Cambridge, MA: Harvard Univ. Press; 1987. [13] Petersen, T. D. P.; Hochachka, P. W.; Suarez, R. K. Hormonal control of gluconeogenesis in rainbow trout hepatocytes: regulatory role of pyruvate kinase. J. Exp. Zool. 243:173 – 180; 1987. [14] Grzelak, A.; Rychlik, B.; Bartosz, G. Light-dependent generation of reactive oxygen species in cell culture media. Free Radic. Biol. Med. 30:1418 – 1425; 2001. [15] Guppy, M.; Abas, L.; Arthur, P. G.; Whisson, M. E. The Pasteur effect in human platelets: implications for storage and metabolic control. Br. J. Haematol. 91:752 – 757; 1995. [16] Xie, Z.; Kometiani, P.; Liu, J.; Li, J.; Shapiro, J. I.; Askari, A. Intracellular reactive oxygen species mediate the linkage of Na+/K+ATPase to hypertrophy and its marker genes in cardiac myocytes. J. Biol. Chem. 274:19323 – 19328; 1999. [17] Mohanty, J. G.; Jaffe, J. S.; Schulman, E. S.; Raible, D. G. A highly sensitive fluorescent micro-assay of H2O2 release from activated human leukocytes using a dihydroxyphenoxazine derivative. J. Immunol. Methods 202:133 – 141; 1997. [18] Reers, M.; Smiley, S. T.; Mottola-Hartshorn, C.; Chen, A.; Lin, M.; Chen, L. B. Mitochondrial membrane potential monitored by JC-1 dye. Methods Enzymol. 260:406 – 417; 1995. [19] Grynkiewicz, G.; Poenie, M.; Tsien, R. Y. A new generation of Ca2+ indicators with greatly improved fluorescence properties. J. Biol. Chem. 260:3440 – 3450; 1985. [20] Arthur, P. G.; Ngo, C. T.; Wakeford, C. M. A perifusion system to control oxygen concentration in cell suspensions. Anal. Biochem. 263:208 – 213; 1998. [21] Casey, T. M.; Dufall, K. G.; Arthur, P. G. An improved capillary electrophoresis method for measuring tissue metabolites associated with cellular energy state. Eur. J. Biochem. 261:740 – 745; 1999. [22] Arthur, P. G.; Hochachka, P. W. Automated analysis of cellular metabolites at nanomolar to micromolar concentrations using bioluminescent methods. Anal. Biochem. 227:281 – 284; 1995. [23] Lakin-Thomas, P. L.; Brand, M. D. Stimulation of respiration by mitogens in rat thymocytes is independent of mitochondrial calcium. Biochem. J. 256:167 – 173; 1988.

[24] Antunes, F.; Cadenas, E. Estimation of H2O2 gradients across biomembranes. FEBS Lett. 475:121 – 126; 2000. [25] Chance, B.; Sies, H.; Boveris, A. Hydroperoxide metabolism in mammalian organisms. Physiol. Rev. 59:527 – 605; 1979. [26] Turrens, J. F.; Freeman, B. A.; Crapo, J. D. Hyperoxia increases H2O2 release by lung mitochondria and microsomes. Arch. Biochem. Biophys. 217:411 – 421; 1982. [27] Archer, S. L.; Huang, J.; Henry, T.; Peterson, D.; Weir, E. K. A redox-based O2 sensor in rat pulmonary vasculature. Circ. Res. 73:1100 – 1112; 1993. [28] Turrens, J. F. Superoxide production by the mitochondrial respiratory chain. Biosci. Rep. 17:3 – 8; 1997. [29] Kroll, S. L.; Czyzyk-Krzeska, M. F. Role of H2O2 and hemecontaining O2 sensors in hypoxic regulation of tyrosine hydroxylase gene expression. Am. J. Physiol. 274:C167 – C174; 1998. [30] Hool, L. C.; Arthur, P. G. Decreasing cellular hydrogen peroxide with catalase mimics the effects of hypoxia on the sensitivity of the L-type Ca2+ channel to h-adrenergic receptor stimulation in cardiac myocytes. Circ. Res. 91:601 – 609; 2002. [31] Roveri, A.; Coassin, M.; Maiorino, M.; Zamburlini, A.; Van Amsterdam, F. T. M.; Ratti, E.; Ursini, F. Effect of hydrogen peroxide on calcium homeostasis in smooth muscle cells. Arch. Biochem. Biophys. 297:265 – 270; 1992. [32] Doan, T. N.; Gentry, D. L.; Taylor, A. A.; Elliott, S. J. Hydrogen peroxide activates agonist-sensitive Ca2+-flux pathways in canine venous endothelial cells. Biochem. J. 297:209 – 215; 1994. [33] Favero, T. G.; Zable, A. C.; Abramson, J. J. Hydrogen peroxide stimulates the Ca2+ release channel from skeletal muscle sarcoplasmic reticulum. J. Biol. Chem. 270:25557 – 25563; 1995. [34] Suzuki, Y. J.; Cleemann, L.; Abernethy, D. R.; Morad, M. Glutathione is a cofactor for H2O2-mediated stimulation of Ca2+-induced Ca2+ release in cardiac myocytes. Free Radic. Biol. Med. 24:318 – 325; 1998. [35] Oba, T.; Kurono, C.; Nakajima, R.; Takaishi, T.; Ishida, K.; Fuller, G. A.; Klomkleaw, W.; Yamaguchi, M. H2O2 activates ryanodine receptor but has little effect on recovery of releasable Ca2+ content after fatigue. J. Appl. Physiol. 93:1999 – 2008; 2002. [36] Makowska, A.; Zablocki, K.; Duszynski, J. The role of mitochondria in the regulation of calcium influx into Jurkat cells. Eur. J. Biochem. 267:877 – 884; 2000. [37] Lewis, R. S. Calcium signaling mechanisms in T lymphocytes. Annu. Rev. Immunol. 19:497 – 521; 2001. [38] Tiffert, T.; Lew, V. L. Kinetics of inhibition of the plasma membrane calcium pump by vanadate in intact human red cells. Cell Calcium 30:337 – 342; 2001. [39] Wolff, M.; Fandrey, J.; Jelkmann, W. Microelectrode measurements of pericellular PO2 in erythropoietin-producing human hepatoma cell cultures. Am. J. Physiol. 265:C1266 – C1270; 1993. [40] Fandrey, J.; Frede, S.; Jelkmann, W. Role of hydrogen peroxide in hypoxia-induced erythropoietin production. Biochem. J. 303: 507 – 510; 1994. [41] Weir, E. K.; Archer, S. L. The mechanism of acute hypoxic pulmonary vasoconstriction: the tale of two channels. FASEB J. 9:183 – 189; 1995. [42] Fandrey, J.; Frede, S.; Ehleben, W.; Porwol, T.; Acker, H.; Jelkmann, W. Cobalt chloride and desferrioxamine antagonize the inhibition of erythropoietin production by reactive oxygen species. Kidney Int. 51:492 – 496; 1997. [43] Canbolat, O.; Fandrey, J.; Jelkmann, W. Effects of modulators of the production and degradation of hydrogen peroxide on erythropoietin synthesis. Respir. Physiol. 114:175 – 183; 1998. [44] Hanson, E. S.; Foot, L. M.; Leibold, E. A. Hypoxia post-translationally activates iron-regulatory protein 2. J. Biol. Chem. 274: 5047 – 5052; 1999. [45] Neumcke, I.; Schneider, B.; Fandrey, J.; Pagel, H. Effects of pro- and antioxidative compounds on renal production of erythropoietin. Endocrinology 140:641 – 645; 1999.

S.E. Munns et al. / Free Radical Biology & Medicine 38 (2005) 1594–1603 [46] Boveris, A.; Cadenas, E. Mitochondrial production of superoxide anions and its relationship to the antimycin insensitive respiration. FEBS Lett. 54:311 – 314; 1975. [47] Turrens, J. F.; Alexandre, A.; Lehninger, A. L. Ubisemiquinone is the electron donor for superoxide formation by complex III of heart mitochondria. Arch. Biochem. Biophys. 237:408 – 414; 1985. [48] Liu, S. Cooperation of a reactive oxygen cycle with the Q cycle and the proton cycle in the respiratory chain—superoxide generating and cycling mechanisms in mitochondria. J. Bioenerg. Biomembr. 31: 367 – 376; 1999. [49] Pham, N. A.; Robinson, B. H.; Hedley, D. W. Simultaneous detection of mitochondrial respiratory chain activity and reactive oxygen in digitonin-permeabilized cells using flow cytometry. Cytometry 41: 245 – 251; 2000. [50] Raha, S.; McEachern, G. E.; Myint, A. T.; Robinson, B. H. Superoxides from mitochondrial complex III: the role of manganese superoxide dismutase. Free Radic. Biol. Med. 29:170 – 180; 2000. [51] Lui, Y.; Fiskum, G.; Schubert, D. Generation of reactive oxygen species by the mitochondrial electron transport chain. J. Neurochem. 80:780 – 787; 2002. [52] Duranteau, J.; Chandel, N. S.; Kulisz, A.; Shao, Z.; Schumacker, P. T. Intracellular signalling by reactive oxygen species during hypoxia in cardiomyocytes. J. Biol. Chem. 273:11619 – 11624; 1998. [53] Niknahad, H.; Khan, S.; O’Brien, P. J. Hepatocyte injury resulting

[54]

[55]

[56]

[57]

[58]

[59]

[60]

1603

from the inhibition of mitochondrial respiration at low oxygen concentrations involves reductive stress and oxygen activation. Chem.–Biol. Interact 98:27 – 44; 1995. Hool, L. C. Hypoxia increases the sensitivity of the L-type Ca2+ current to h-adrenergic receptor stimulation via a C2 regioncontaining protein kinase C isoform. Circ. Res. 87:1164 – 1171; 2000. Boraso, A.; Williams, A. J. Modification of the gating of the cardiac sarcoplasmic reticulum Ca2+-release channel by H2O2 and dithiothreitol. Am. J. Physiol. 267:H1010 – H1016; 1994. Ghosh, S. K.; Chakraborti, T.; Michael, J. R.; Chakraborti, S. Oxidantmediated proteolytic activation of Ca2+-ATPase in microsomes of pulmonary smooth muscle. FEBS Lett. 387:171 – 174; 1996. Kaneko, M.; Singal, P. K.; Dhalla, N. S. Alterations in heart sarcolemmal Ca2+-ATPase and Ca2+-binding activities due to oxygen free radicals. Basic Res. Cardiol. 85:45 – 54; 1990. Archer, S.; Michelakis, E. The mechanism of hypoxic pulmonary vasoconstriction: potassium channels, redox O2 sensors, and controversies. News Physiol. Sci. 17:131 – 137; 2002. Chandel, N. S.; Budinger, G. R.; Choe, S. H.; Schumacker, P. T. Cellular respiration during hypoxia. J. Biol. Chem. 272:18808 – 18816; 1997. Clementi, E.; Brown, G. C.; Foxwell, N.; Moncada, S. On the mechanism by which vascular endothelial cells regulate their oxygen consumption. Proc. Natl. Acad. Sci. USA 96:1559 – 1562; 1999.