Oxidant-induced inhibition of the plasma membrane Ca²⁺-ATPase in pancreatic acinar cells: role of the mitochondria

Abstract

Impairment of the normal spatiotemporal pattern of intracellular Ca²⁺ ([Ca²⁺]_i) signaling, and in particular, the transition to an irreversible “Ca²⁺ overload” response, has been implicated in various pathophysiological states. In some diseases, including pancreatitis, oxidative stress has been suggested to mediate this Ca²⁺ overload and the associated cell injury. We have previously demonstrated that oxidative stress with hydrogen peroxide (H₂O₂) evokes a Ca²⁺ overload response and inhibition of plasma membrane Ca²⁺-ATPase (PMCA) in rat pancreatic acinar cells (Bruce JI and Elliott AC. Am J Physiol Cell Physiol 293: C938–C950, 2007). The aim of the present study was to further examine this oxidant-impaired inhibition of the PMCA, focusing on the role of the mitochondria. Using a [Ca²⁺]_i clearance assay in which mitochondrial Ca²⁺ uptake was blocked with Ru-360, H₂O₂ (50 μM–1 mM) markedly inhibited the PMCA activity. This H₂O₂-induced inhibition of the PMCA correlated with mitochondrial depolarization (assessed using tetramethylrhodamine methylester fluorescence) but could occur without significant ATP depletion (assessed using Magnesium Green fluorescence). The H₂O₂-induced PMCA inhibition was sensitive to the mitochondrial permeability transition pore (mPTP) inhibitors, cyclosporin-A and bongkrekic acid. These data suggest that oxidant-induced opening of the mPTP and mitochondrial depolarization may lead to an inhibition of the PMCA that is independent of mitochondrial Ca²⁺ handling and ATP depletion, and we speculate that this may involve the release of a mitochondrial factor. Such a phenomenon may be responsible for the Ca²⁺ overload response, and for the transition between apoptotic and necrotic cell death thought to be important in many disease states.

acute pancreatitis is a serious and often fatal inflammatory disease primarily of pancreatic acinar cells and leads to autodigestion of the pancreas (6, 50). The major function of pancreatic acinar cells is the exocytotic secretion of powerful digestive enzymes, prepackaged as inactive zymogens. Impairment of exocytosis and/or its regulatory control mechanisms can lead to intracellular activation of zymogens (55), triggering a cascade of cellular events that culminate in autodigestion. Although several predisposing factors for pancreatitis have been identified, the major ones being alcoholism and gallstones, the underlying pathology, particularly the cellular triggering event, remains controversial. However, an impairment of the normal pattern of intracellular calcium ([Ca²⁺]_i) signaling, which controls exocytosis and/or the control of normal resting [Ca²⁺]_i, has emerged in recent years as a possible unifying underlying mechanism (18). In particular, an irreversible increase in [Ca²⁺]_i or “Ca²⁺ overload” has been shown to cause intracellular enzyme activation and is linked to the associated necrotic cell injury that is the hallmark of acute pancreatitis (50).

The transition between apoptotic and necrotic cell death has been suggested to be one of the most important features of acute pancreatitis, regardless of the causative agent or process (6). Apoptotic cell death is a carefully controlled event that leads to the “clean” dismantling of the cellular constituents and the ultimate removal of cell remnants by phagocytosis. In this respect, apoptosis in the pancreas can be protective, although in excess it can lead to the progressive loss of pancreatic tissue (and thus function) that is the hallmark of chronic pancreatitis (6). Necrosis, on the other hand, is a more uncontrolled event that leads to the “messy” and chaotic destruction of the cell, and it is characterized by vacuole formation, blebbing, and, ultimately, cell lysis. This leads to the release of potentially damaging cellular constituents, such as activated proteases, that can damage neighboring cells, resulting in a spiral of self-perpetuating tissue damage and inflammation. The Ca²⁺ overload response has been suggested to underlie this transition between apoptosis and necrosis (47) and thus represents an important mechanism in further understanding the pathology of acute pancreatitis.

Oxidative stress has been implicated in several models of acute and chronic pancreatitis and has been shown to induce both apoptosis and necrosis in pancreatic acinar cells (6). We have previously reported that increased oxidative stress (H₂O₂) profoundly altered the normal pattern of cholecystokinin (CCK)-evoked [Ca²⁺]_i signaling and resulted in an irreversible Ca²⁺ overload response in an increasing proportion of cells as H₂O₂ concentration was increased (9). This H₂O₂-induced Ca²⁺ overload also corresponded to a marked inhibition of the plasma membrane Ca²⁺-ATPase (PMCA), suggesting that this may, at least in part, be responsible for the oxidant-induced Ca²⁺ overload response (9). The aim of the present study was to further explore the mechanism of this oxidant-impaired inactivation of the PMCA, focusing on the role of the mitochondria. The results show that the H₂O₂-induced inactivation of the PMCA correlates with mitochondrial depolarization, but it appears to be largely independent of mitochondrial Ca²⁺ handling and ATP depletion. Furthermore, H₂O₂-induced PMCA inactivation was sensitive to inhibitors of the mitochondrial permeability transition pore (mPTP), suggesting that a factor or factors released from the mitochondria may inhibit the PMCA. Such a phenomenon may be responsible for the irreversible nature of the Ca²⁺ overload response and thus for the transition between apoptotic and necrotic cell death that is thought to underlie the pathology of acute pancreatitis.

MATERIALS AND METHODS

Cell isolation.

Small clusters of pancreatic acinar cells were isolated by collagenase-digestion of the pancreas of Sprague-Dawley rats by the method previously described (11), using 800 units of collagenase-P per gram of tissue (Roche Diagnostics), 0.12 mg/ml soybean trypsin inhibitor (Sigma), and 1% (wt/vol) bovine serum albumin (BSA, Fraction V, Sigma) in a HEPES-buffered physiological saline solution (HEPES-PSS) (composition in mM: 137 Na⁺, 4.7 K⁺, 0.56 Mg²⁺, 1.28 Ca²⁺, 143.54 Cl⁻, 1 HPO₄²⁻, and 10 HEPES, with a pH of 7.4). Following digestion, cells were resuspended in HEPES-PSS containing 0.1% BSA and kept on ice until use. Animals were humanely killed by an approved method as set out in Schedule 1 of the UK Animals Scientific Procedures Act 1986 (Certificate of Designation No 50/2506).

Digital imaging of fura-2 fluorescence.

Cells were loaded with 4 μM fura-2-AM (Invitrogen, Paisley, UK) in HEPES-PSS for 30 min at room temperature. Dye-loaded cells were allowed to adhere to a glass coverslip that formed the base of a gravity-fed perfusion chamber, continually perfused with HEPES-PSS with automatic valves for rapid switching of solutions (Harvard Apparatus, Kent, UK). All fluorescence imaging experiments were performed using an identical microscope and imaging system as previously described (1, 9). Briefly, this comprised a Nikon TE2000S microscope equipped with a CoolSNAP HQ CCD camera (Roper Scientific Photometrics, Tucson, AZ), and Cairn monochromator (Cairn Research, Kent, UK) controlled by MetaFluor imaging software (Molecular Devices, Downington, PA). Background-subtracted 340-nm and 380-nm fluorescence images were captured using 3 × 3 binning at a rate of 0.2 Hz to prevent excessive photobleaching and phototoxicity. The fura-2 fluorescence was calibrated into “estimated” [Ca²⁺]_i using the equation: [Ca²⁺]_i = K_d(R − R_min)/(R_max − R)(S_F380/S_B380), where K_d is the fura-2 dissociation constant, R is any given 340/380 ratio value, S_F380/S_B380 is the ratio of fluorescence measured for Ca²⁺-free and Ca²⁺-bound fura-2, and R_min and R_max are the minimum and maximum ratio values following in situ calibration experiments (9). All experiments were carried out at room temperature (20–22°C).

Measurement of mitochondrial membrane potential.

Estimations of changes in mitochondrial membrane potential (ΔΨm) were made using the fluorescent dye tetramethylrhodamine methylester (TMRM), which accumulates in mitochondria driven by the highly negative inner mitochondrial membrane potential. Pancreatic acinar cells were preincubated with 100 nM TMRM for 15 min at 37°C, and 100 nM TMRM was added to all perfusion solutions during experiments. TMRM-loaded cells were excited with 545 ± 10 nm excitation light (50-ms exposure) at a rate of 0.2 Hz, and background-subtracted images were captured with no binning (to increase the spatial resolution) through a Fura/TRITC dual band dichroic (Chroma). Fluorescent signals from subcellular regions defined by high mitochondrial fluorescence were normalized to the average fluorescence from the initial 10 frames (expressed as F/F₀). Other experimental details were similar to those previously described (56).

Indirect measure of intracellular ATP concentration using Magnesium Green.

The Mg²⁺-sensitive fluorescent dye, Magnesium Green (MgGreen), was used to measure cytosolic ATP depletion, as described in previous studies (19, 34, 39). This is based on the principle that almost all cytosolic ATP exists as MgATP, such that ATP depletion causes an increase in free Mg²⁺ concentration ([Mg²⁺]_i). Since MgGreen has a K_d for Mg²⁺ of ∼0.9 mM, close to the resting [Mg²⁺]_i, this makes MgGreen useful as an indirect indicator of ATP concentration changes around the physiological range (39). In cardiac myocytes, similar MgGreen measurements of ATP have been shown to be comparable to those obtained using NMR methods (44). Pancreatic acinar cells were incubated with 4 μM MgGreen for 30 min at room temperature and excited with 496 ± 10 nm excitation light (50-s exposure) at a rate of 0.1 Hz. Background-subtracted images were captured using 5 × 5 binning (to increase signal detection) through a FITC dichroic (Chroma). Cells were selected on the basis of moderate fluorescence (neither very bright nor very low) to exclude cells with poor dye loading (very low fluorescence) or “unhealthy” cells with low initial cytosolic [ATP] and hence high MgGreen fluorescence. Fluorescent signals were normalized as relative fluorescence from the initial 10 frames (F/F₀).

Solutions.

In all imaging experiments, cells were perfused with a HEPES-PSS. Solutions containing H₂O₂ were made up fresh each day and were periodically assayed using a fluorimetric assay as described previously (9). For experiments with La³⁺ (1 mM) to block the PMCA, HPO₄²⁻ ions were replaced with Cl⁻ to prevent precipitation of La³⁺ salts. Stock solutions (1 mM) of Ru-360 (Calbiochem) were dissolved in deoxygenated water immediately before use.

Data analysis and experimental design.

In some experiments, an unpaired experimental design was used, whereby parameters such as normalized linear [Ca²⁺]_i clearance rate (see results) were compared between groups of experiments (e.g., control versus treatment). Where appropriate, an unpaired Student's t-test or Mann-Whitney test was used to determine statistical significance. In other experiments, a paired experimental design was used, whereby clearance rate in the presence of a treatment was compared with a control clearance time course in the same cells and expressed as a percentage of control. In these experiments, a one-sample t-test was then used to determine statistical significance within a group and a Mann-Whitney test was used for comparisons between groups (see results). For any given parameter analyzed, an “experimental average response” was determined from several cells in a particular experiment. These values were in turn averaged to give the true overall average and are expressed in the text as means ± SE.

RESULTS

Validation of PMCA activity assay.

To test the effects of oxidative stress on PMCA activity, we used a modified in situ [Ca²⁺]_i clearance assay. In this method, all other [Ca²⁺]_i clearance pathways were inhibited and PMCA activity was effectively isolated, similar to our previous studies in pancreatic and parotid acinar cells (1, 9). Briefly, this technique involved depleting the ER Ca²⁺ stores with the sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA) inhibitor, cyclopiazonic acid (CPA), in the absence of external Ca²⁺. This caused a small increase in [Ca²⁺]_i, presumably due to leak from the ER, which then returned to around or slightly below the baseline [Ca²⁺]_i, presumably due to Ca²⁺ extrusion by the PMCA. Readdition of external Ca²⁺ then caused a rapid increase in [Ca²⁺]_i (9), due to store-operated Ca²⁺ entry (SOCE) (51), which then reached a plateau, presumably due to a balance of Ca²⁺ influx and Ca²⁺ efflux (Ref. 9 and Fig. 1). Subsequent removal of external Ca²⁺ caused a rapid clearance of [Ca²⁺]_i, which was then used as a measure of PMCA activity (Ref. 9 and Fig. 1). In our previous study, readdition of “physiological” external Ca²⁺ (1.28 mM) led to a relatively small steady-state increase in [Ca²⁺]_i of around 50–100 nM (9), due to either limited SOCE or high [Ca²⁺]_i clearance activity. This therefore limited the measurement of clearance to a relatively low range of Ca²⁺ concentrations below 200 nM. To remedy this problem, in the current study, we modified the assay by adding back 20 mM external Ca²⁺, thereby increasing the driving force for Ca²⁺ entry and increasing [Ca²⁺]_i to much higher steady-state levels (459 ± 67 nM, Fig. 1A). This allowed [Ca²⁺]_i clearance, and thus PMCA activity, to be monitored over a large range of intracellular Ca²⁺ concentrations. [Ca²⁺]_i clearance was quantified by fitting the falling phase of the clearance to a single exponential decay. This gave an average time constant (τ) of 59 ± 10 s (n = 5, 116 cells, Fig. 1A) in untreated control cells. In our previous study under similar conditions (i.e., when SERCA was inhibited), the major [Ca²⁺]_i clearance pathway was found to be the PMCA. The basis of this conclusion was that the PMCA inhibitor, La³⁺ (1 mM), almost completely inhibited [Ca²⁺]_i clearance, while preincubation with the specific mitochondrial Ca²⁺ uptake inhibitor, Ru-360 (10 μM) (10, 36), had no significant effect (9). However, in the present study, since [Ca²⁺]_i clearance was monitored over a greater range of Ca²⁺ concentration, mitochondrial Ca²⁺ uptake might be expected to have a much greater contribution, due to the low affinity and high capacity of the uptake process (17). This was therefore investigated further in the present study by preincubating cells with 10 μM Ru-360 for 30 min immediately before beginning the clearance assay. Under these conditions, the average time constant (τ) was 94 ± 11 s (n = 8, 134 cells, Fig. 1B), a marginally significant increase (P = 0.054) compared with untreated control cells. It was also noticeable that the rate of increase in [Ca²⁺]_i was slowed, and the steady-state [Ca²⁺]_i, following addition of 20 mM Ca²⁺, was significantly lower in Ru-360-treated cells (steady-state [Ca²⁺]_i = 312 ± 27 nM) compared with untreated control cells ([Ca²⁺]_i = 459 ± 67 nM, P < 0.01).

Fig. 1.

Validation of plasma membrane Ca²⁺-ATPase (PMCA) activity assay. A and B: representative traces showing the intracellular Ca²⁺ ([Ca²⁺]_i) clearance assay (see results) in untreated fura-2-loaded pancreatic acinar cells (A) and cells treated with 10 μM Ru-360 for 30 min, to inhibit mitochondrial Ca²⁺ uptake (B). Cells were treated with 30 μM cyclopiazonic acid (CPA), as indicated by the arrow, in the absence of external Ca²⁺ (0 Ca²⁺) to deplete endoplasmic reticulum Ca²⁺. CPA was present in all solutions throughout the remainder of each experiment. [Ca²⁺]_i clearance was initiated by the addition of 20 mM external Ca²⁺ (20 Ca²⁺), followed by the subsequent removal of external Ca²⁺ (0 Ca²⁺). Arrows represent the steady-state [Ca²⁺]_i (see results) and the standardized [Ca²⁺]_i (300 nM) from which the linear rate of [Ca²⁺]_i clearance was measured (E). The inset in A represents the mean steady-state [Ca²⁺]_i (SS [Ca²⁺]_i) plotted against the mean linear clearance rate measured from the steady-state [Ca²⁺]_i value (SS rate, gray triangles) or standardized 300 nM value (Std rate, black circles). There was a clear correlation between SS rate and SS [Ca²⁺]_i (slope = 1.1 min⁻¹, which was significantly different from zero; P = 0.01), but there was no such correlation between Std rate and SS [Ca²⁺]_i (slope = 0.16 min⁻¹, which was not significantly different from zero; P = 0.16). C and D: inhibition of PMCA activity with 1 mM La³⁺ in control (C) and Ru-360-treated cells (D). EGTA (1 mM) was used to remove the La³⁺ and restore PMCA activity. E: mean standardized linear [Ca²⁺]_i clearance rate in untreated control cells (shown in A), compared with Ru-360-treated cells with or without La³⁺ (*statistical significance as assessed using an unpaired t-test).

Similar to our previous studies (9, 10), treatment with H₂O₂ (0.1–1 mM) slowed the clearance to such an extent that the rate could no longer be fitted to an exponential decay. Therefore, we adopted an alternative method of quantifying clearance by fitting the first 60 s of the clearance time course by linear regression (see arrow and steady state, Fig. 1, A and B). However, the steady-state [Ca²⁺]_i varied markedly, even between control cells, but particularly following treatment with either H₂O₂ or mitochondrial inhibitors. Therefore, the data were standardized by measuring the linear clearance rate over a 60-s period from a starting value of 300 nM Ca²⁺ (see arrow and standardized, Fig. 1, A and B). This allowed direct comparisons to be made between control and treated cells and prevented the initial [Ca²⁺] from influencing the measured linear clearance rate. In untreated control cells, the mean steady-state clearance was 250 ± 44 nM/min_, whereas the standardized clearance in the same cells was 165 ± 20 nM/min (n = 5, 116 cells, Fig. 1, A and E). This suggests that [Ca²⁺] influences the linear clearance rate, and, indeed, closer analysis revealed that there was a clear correlation between the mean steady-state [Ca²⁺]_i and the mean linear clearance rate when measured at the steady-state [Ca²⁺]_i (see triangles within inset of Fig. 1A, slope = 1.1 min⁻¹, P < 0.01). This would be expected for clearance that follows a single exponential decay, viz. the linear rate is faster the higher the starting [Ca²⁺]_i from which the rate was measured. However, if the mean linear rate of clearance measured at the standardized value of 300 nM [Ca²⁺]_i was plotted against the same steady-state [Ca²⁺]_i in the same cells, there was no significant correlation (see circles within inset of Fig. 1A, slope = 0.16 min⁻¹, P = 0.16). This means that, on average, the linear clearance rate had slowed to approximately the same value by the time [Ca²⁺]_i had reached 300 nM regardless of the steady-state [Ca²⁺]_i the cell had attained. This suggests that prior exposure to high [Ca²⁺] does not differentially regulate [Ca²⁺]_i clearance as would be expected if the PMCA exhibited memory, at least over the time course of the clearance assay.

Using this method of analysis it was found that, on average the standardized clearance rate in Ru-360-treated cells (115 ± 14 nM/min, n = 8, 96 cells, Fig. 1, B and E) was significantly less than untreated control cells (165 ± 20 nM/min, n = 5, 116 cells, P = 0.02, Fig. 1, A and E). This therefore suggests that when [Ca²⁺]_i clearance is measured over a much greater concentration range, mitochondrial Ca²⁺ uptake contributes to clearance and is independent of the cells attaining a lower steady-state [Ca²⁺]_i. However, under these conditions, the residual [Ca²⁺]_i clearance in Ru-360-treated cells would suggest that the relative contribution of mitochondrial Ca²⁺ uptake is much less than the PMCA. This was tested further by adding the PMCA inhibitor La³⁺ (1 mM) at the steady-state [Ca²⁺]_i before the removal of external Ca²⁺ to initiate clearance. La³⁺ inhibits Ca²⁺ influx with μM affinity and inhibits Ca²⁺ efflux with mM affinity (5, 48). Therefore, in control cells, addition of 1 mM La³⁺ caused an initial rapid decrease, presumably due to an immediate inhibition of Ca²⁺ entry (48), followed by an abrupt slowing of the rate of decrease of [Ca²⁺]_i as La³⁺ accumulated to sufficiently high enough concentrations to inhibit the PMCA (Fig. 1C). The [Ca²⁺]_i continued to slowly decline during the La³⁺ exposure at a constant low rate (25 ± 3 nM/min, n = 4, 38 cells; Fig. 1C and mean data in Fig. 1E). This slow La³⁺-insensitive clearance was unaffected by the subsequent removal of external Ca²⁺ (Fig. 1C) until EGTA (1 mM) was applied to chelate the La³⁺ from the cells, a treatment that restored PMCA activity. The La³⁺-insensitive [Ca²⁺]_i clearance was presumably due to mitochondrial Ca²⁺ uptake, because it was greatly reduced in Ru-360-treated cells (8 ± 2 nM/min, n = 4, 52 cells; Fig. 1D and mean data in Fig. 1E, P = 0.004).

Collectively, these data provide convincing evidence that when [Ca²⁺]_i is relatively high (200–500 nM), mitochondrial Ca²⁺ uptake as well as PMCA activity contributes to the measured [Ca²⁺]_i clearance. Furthermore, Ru-360 effectively blocked this mitochondrial Ca²⁺ uptake, consistent with previous studies (9, 10, 36). Therefore, to prevent mitochondrial Ca²⁺ uptake (and thus any possible subsequent mitochondrial Ca²⁺ release) from contaminating the measured [Ca²⁺]_i clearance and thus the assessment of PMCA activity, all subsequent experiments were performed in Ru-360-treated cells.

H₂O₂ inhibits the PMCA activity in the presence of mitochondrial Ca²⁺ uptake inhibitors.

We have previously demonstrated that either pretreatment or acute treatment of pancreatic acinar cells with H₂O₂ (>100 μM) inhibited the PMCA activity (9). However, it was necessary to further examine the effects of H₂O₂ using the modified [Ca²⁺]_i clearance assay employed in the present study. When acutely applied 2–5 min before the addition of 20 mM external Ca²⁺, H₂O₂ was even more effective than in our previous study (9), causing a concentration-dependent inhibition and, ultimately, inactivation of the PMCA (Fig. 2). At 50 μM, H₂O₂ reduced the clearance rate from 115 ± 14 nM/min to 47 ± 10 nM/min (P = 0.001, n = 8, 96 cells), whereas 100 μM reduced the clearance to 22 ± 7 nM/min (P < 0.001, n = 5, 56 cells, Fig. 2). At 0.5 and 1 mM, H₂O₂ reduced the clearance to 26 ± 4 nM/min (P < 0.001, n = 6, 72 cells) and 23 ± 7 nM/min (P < 0.001, n = 5, 47 cells), respectively. It must be noted, however, that in many of the cells treated with 0.5 and 1 mM H₂O₂, clearance could not be measured from the standardized 300 nM, because the clearance slowed to such an extent that the [Ca²⁺]_i rarely fell to this value. Rather than exclude these cells from analysis, linear clearance was measured from the steady-state [Ca²⁺]_i, which was much above 300 nM. According to the data from untreated control cells, measurement of clearance from a higher steady-state [Ca²⁺]_i value would result in a higher rate due to the nature of the exponential decay. However, in most cases, clearance slowed to such an extent that it resembled a linear rather than exponential relationship. Therefore, extrapolation of the rate to 300 nM would presumably be the same as if it were measured at the steady-state [Ca²⁺]_i, thereby removing this confounding problem. In addition, in cells where [Ca²⁺]_i exceeds 500 nM and the fura-2 ratio approaches the R_max, the calibrated [Ca²⁺]_i values become less reliable, which may contribute to the high variability of the computed clearance rates. Finally, many cells treated with high concentrations of H₂O₂ underwent rapid cell lysis, causing dye leakage (as indicated by the arrows in Fig. 2, D and E) and making quantification difficult. Nevertheless, these technical and analytical difficulties, if anything, are likely to underestimate the inhibitory effectiveness of high concentrations of H₂O₂ on PMCA activity. In conclusion, these data suggest that in the presence of mitochondrial Ca²⁺ uptake inhibitors (Ru-360), H₂O₂ substantially inhibited PMCA activity, even at the lowest concentration tested (50 μM), and completely inactivated the pump at higher concentrations (>100 μM).

Fig. 2.

H₂O₂ (50–1,000 μM) inhibited PMCA activity under conditions in which mitochondrial Ca²⁺ uptake was inhibited. Pancreatic acinar cells were pretreated with 10 μM Ru-360 for 30 min, to inhibit mitochondrial Ca²⁺ uptake, immediately before the start of the [Ca²⁺]_i clearance assay in all experiments. H₂O₂ was added 2–5 min before the addition of 20 mM external Ca²⁺. Representative traces show control cells (A) and the effects of 50 μM (B), 100 μM (C), 500 μM (D), and 1 mM (E) H₂O₂. Gray arrows indicate the point at which CPA was added, which was then present throughout the experiment. Black arrows in D and E show rapid loss of dye due to cell lysis. F: mean data showing the standardized linear [Ca²⁺]_i clearance rate following treatment with varying concentrations of H₂O₂ (*P < 0.001, as assessed using an unpaired t-test).

It was also observed that the initial increase in [Ca²⁺]_i following the addition of 20 mM external Ca²⁺ (Ca²⁺ repletion) was slower in the presence of H₂O₂ (Fig. 2, B–E) compared with control (Fig. 2A). This is classically believed to be a much better representation of store-operated Ca²⁺ entry (SOCE), compared with the steady-state [Ca²⁺]_i. This is because the opposing contribution of Ca²⁺ clearance pathways (e.g., PMCA) will be minimal during this initial increase in [Ca²⁺]_i but then begin to counter the Ca²⁺ entry as [Ca²⁺]_i increases. This therefore slows the apparent rate of increase in [Ca²⁺]_i until it reaches a new steady state where there is a balance between Ca²⁺ entry and Ca²⁺ efflux.

Effects of mitochondrial inhibitors on PMCA activity.

The mitochondria are central to the control of multiple cellular processes including many Ca²⁺ transport pathways (22) and are the major fuel source for ATP-consuming processes, including the PMCA. It is therefore possible that the PMCA has an intimate dependence on functional mitochondria. In addition, oxidants have been shown extensively to impair mitochondrial function (30, 54). We therefore examined the effects of three mechanistically distinct mitochondrial inhibitors to test if these have similar effects to H₂O₂. CCCP is a protonophore that dissipates proton gradients across organelle membranes, most notably the inner mitochondrial membrane. This causes rapid depolarization of the mitochondrial membrane potential (Ψm) and uncouples ATP synthesis from oxidative metabolism (22). Antimycin-A is an inhibitor of complex-III of the electron transport chain of mitochondria and therefore inhibits ATP synthesis (23). However, antimycin-A has also been shown to cause mitochondrial release of reactive oxygen species and to slowly depolarize Ψm (27, 43). Oligomycin is an inhibitor of the F₁/F₀-ATP synthase, which thereby inhibits ATP synthesis and is not known to affect Ψm (34). All these agents were added (similarly to H₂O₂ in previous experiments) 2–5 min before the addition of 20 mM Ca²⁺ (Fig. 3, B–D). CCCP (4 μM) and antimycin-A (0.5 μM) had very similar inhibitory effects on PMCA activity, significantly reducing [Ca²⁺]_i clearance from 115 ± 14 nM/min to 58 ± 7 nM/min (P = 0.003, n = 7, 118 cells, Fig. 3, Bi, Bii, and E) and 51 ± 11 nM/min (P = 0.01, n = 5, 53 cells, Fig. 3, C and E), respectively. However, CCCP had variable effects when applied to cells following ER depletion with CPA in the absence of external Ca²⁺ before the addition of 20 mM Ca²⁺. In 60% of cells (n = 4, 71 cells), CCCP increased [Ca²⁺]_i (Fig. 3Bi), whereas in the remaining 40% of cells (n = 3, 47 cells), CCCP had no effect (Fig. 3Bii). This suggests that CCCP may be causing mitochondrial Ca²⁺ release in 60% of cells; however, antimycin-A failed to have such an effect in any cell (Fig. 3C). This therefore suggests that CCCP may have nonspecific effects, such as dissipating proton gradients across other organelle membranes, that may facilitate Ca²⁺ leak. Oligomycin, on the other hand, had no significant effect on [Ca²⁺]_i clearance (163 ± 14 nM/min, n = 7, 66 cells Fig. 3, D and E) compared with control (115 ± 14 nM/min, P = 0.09).

Fig. 3.

Antimycin-A and CCCP, but not oligomycin, inhibited PMCA activity. As in Figs. 1 and 2, cells were pretreated with Ru-360 (10 μM for 30 min) to inhibit mitochondrial Ca²⁺ uptake in all experiments. Representative traces show control cells (A) and the effects of 4 μM CCCP (Bi and Bii), 0.5 μM antimycin-A (C), and 10 μM oligomycin (D). Arrows indicate the point at which CPA was added, which was then present throughout the experiment. In 60% of cells, CCCP evoked an increase in [Ca²⁺]_i when added in the absence of external Ca²⁺ (Bi) but had no effect in the remaining 40% of cells (Bii). E: mean data showing the standardized linear [Ca²⁺]_i clearance rate following treatment with each mitochondrial inhibitor (*P < 0.05, as assessed using an unpaired t-test). Con, control; Anti, antimycin; Oligo, oligomycin.

Effects of H₂O₂ and mitochondrial inhibitors on ATP depletion.

Because of the intimate functional relationship between the mitochondria and the PMCA, it is tempting to speculate that H₂O₂-induced inhibition of the PMCA is due to ATP depletion. However, all the mitochondrial inhibitors used in the above experiments, including oligomycin, are likely to inhibit mitochondrial ATP synthesis regardless of mechanism. Therefore, to test this further, we measured ATP depletion indirectly using MgGreen, similar to previous studies (19, 34, 39). As a positive control, and to determine the maximum dynamic range of the dye, cells were treated with an “ATP-depletion” cocktail of 100 μM carbachol (CCh), 2 mM iodoacetate, and 10 μM oligomycin. CCh was used at a maximal concentration (100 μM) to rapidly increase [Ca²⁺]_i, thereby activating Ca²⁺-ATPases (and presumably other Ca²⁺-dependent processes) that will likely consume ATP. Iodoacetate was used to inhibit glycolytic ATP production, while oligomycin inhibits mitochondrial ATP production (4). Addition of CCh, iodoacetate, or oligomycin alone is likely to deplete ATP only slightly, because either glycolysis or oxidative metabolism should be able to compensate and maintain cytosolic ATP. However, the combined use of all three reagents together is designed to rapidly and maximally deplete ATP, similar to previous studies (4). The effects of all other reagents (H₂O₂, antimycin-A, CCCP, and oligomycin) used singly were then compared with this maximum ATP depletion in an attempt to normalize the MgGreen responses.

Addition of the ATP-depletion cocktail caused a slow increase in MgGreen fluorescence (48 ± 4% F/F₀, n = 7, 34 cells), with an average latency of ∼8 min (2–25 min range; Fig. 4A, black trace). In addition, the time to reach a maximum was highly variable between cells, ranging from 11 to 39 min (∼30 min on average). In some cells, there was a small but rapid spikelike increase in fluorescence as the ATP-depletion cocktail was applied that rapidly declined back to baseline levels before the slower secondary increase in fluorescence. This may have reflected an initial spikelike increase in [Ca²⁺]_i caused by CCh, since MgGreen is sensitive to Ca²⁺, with a reported K_d of 4 μM (39), and CCh-evoked changes in [Ca²⁺]_i are very likely to reach micromolar concentrations at the peak of the response. However, the slower Ca²⁺ overload-type response evoked by H₂O₂ (9) or by metabolic inhibition is only likely to reach submicromolar [Ca²⁺]_i (500–700 nM) (9). Nevertheless, this was tested further by incubating cells with 10 μM BAPTA-AM for 30 min at room temperature. BAPTA-AM loads into the cells, similarly to fura-2, and rapidly buffers and thus markedly attenuates changes in [Ca²⁺]_i (12). Consistent with our previous assertion, the slow increase in MgGreen fluorescence evoked by the ATP-depletion cocktail was unaffected by preincubation with BAPTA-AM (58 ± 4% F/F₀; n = 9, 29 cells, Fig. 4A, gray trace) compared with control cells (52 ± 3% F/F₀; n = 10, 42 cells, Fig. 4A, black trace). This was despite markedly attenuating the CCh-evoked [Ca²⁺]_i response (240 ± 84 nM in BAPTA-loaded cells, n = 5, 22 cells) compared with control cells (1,088 ± 197 nM, n = 8, 38 cells). This therefore provides evidence that any contamination of the H₂O₂-evoked or ATP-depletion cocktail-evoked increase in MgGreen fluorescence by [Ca²⁺]_i changes is likely to be minimal and thus represents a reliable readout of ATP depletion. Moreover, it was also noticed that a number of cells exhibited a rapid increase in MgGreen fluorescence followed by a rapid decline in fluorescence that fell well below the resting fluorescence (as indicated by gray arrow in Fig. 4A, inset). This is consistent with rapid dye loss and thus cell lysis. Therefore these cells were excluded from analysis. The increase in MgGreen fluorescence in these cells reached a plateau level that was well above the maximum reached in “healthy” cells, suggesting that in these cells the MgGreen fluorescence had reached saturation, possibly due to increased Mg²⁺ permeability that preceded total cell lysis. These results were broadly in line with experiments in cardiac myocytes (39).

Fig. 4.

Effects of H₂O₂, CCCP, antimycin, and oligomycin on ATP depletion (ATP-dep). Pancreatic acinar cells were loaded with 4 μM Magnesium Green (MgGreen) for 30 min at room temperature to indirectly measure cytosolic ATP concentration. Representative traces show the relative MgGreen fluorescence (F/F₀; arbitrary units) in response to the “ATP depletion” cocktail with (gray trace) or without (black trace) preincubation with 10 μM BAPTA (A), 50 μM H₂O₂ (B), 500 μM H₂O₂ (C), 4 μM CCCP (D), 0.5 μM antimycin (E), and 10 μM oligomycin (F). The ATP depletion cocktail consisted of 100 μM carbachol (CCh), 10 μM oligomycin, and 2 mM iodoacetate and was used as a positive control to induce maximum ATP depletion. Cells that exhibited a rapid increase followed by a rapid decrease below the baseline fluorescence (see inset in A) were excluded from analysis because these likely represented cells undergoing cell lysis. Vertical bar represents 0.3 F/F₀ and horizontal bar represents 5 min. G: mean data were quantified and normalized by expressing the change in F/F₀ as a percentage of the ATP depletion cocktail response (*P < 0.05, as assessed using a one-sample t-test).

The effect of H₂O₂ or the mitochondrial inhibitors was quantified for each intervention by expressing the fluorescence change as a percentage of the maximum increase in MgGreen fluorescence evoked by the ATP-depletion cocktail (see Fig. 4). At the high concentration of 500 μM, which caused almost complete inactivation of the PMCA activity, H₂O₂ produced a significant increase in MgGreen fluorescence (16 ± 3% F/F₀, n = 4, 12 cells, P < 0.05 one-sample t-test). However, this was still significantly less than that produced by the ATP-depletion cocktail (54 ± 5% of maximum cocktail response, P = 0.01, Fig. 4C). CCCP (4 μM) and antimycin-A (0.5 μM) also significantly increased the MgGreen fluorescence by 22 ± 6% (n = 4, 12 cells; P = 0.001, Fig. 4D) and 18 ± 6% (n = 4, 12 cells; P < 0.001, Fig. 4E) of the maximum cocktail response, respectively. Oligomycin (10 μM), which had no effect on PMCA activity (see Fig. 3D), also increased the MgGreen fluorescence by 18 ± 4% of the maximum ATP-depletion cocktail response (n = 6, 17 cells; P < 0.001, Fig. 4F). Furthermore, the lower concentration of H₂O₂ tested (50 μM), which also inhibited PMCA activity, had no significant effect on MgGreen fluorescence (2 ± 1% of maximum cocktail response, n = 4, 9 cells, Fig. 4B). These data therefore suggest that H₂O₂-induced inhibition of PMCA activity can occur independently of ATP depletion. Moreover, the ATP depletion caused by oligomycin, which was equivalent to CCCP and antimycin-A, did not inhibit PMCA activity.

Effects of H₂O₂ and mitochondrial inhibitors on mitochondrial membrane potential.

In an attempt to elucidate the mechanism of the effects of H₂O₂ and mitochondrial inhibitors on PMCA activity, we next tested the effects of all these agents on mitochondrial membrane potential (Ψm), using the membrane potential-sensitive fluorescent dye, TMRM, similar to previous studies (19, 27, 56). TMRM accumulates in the mitochondria based on the highly negative membrane potential and redistributes following depolarization. Following loading, TMRM fluorescence exhibited a punctate distribution consistent with mitochondrial accumulation (see image A associated with Fig. 5, A–F ; see also Refs. 19, 27, and 56). As expected, CCCP (4 μM) caused a rapid decrease in the relative TMRM fluorescence and the punctate fluorescence became rapidly diffuse (see image B, Fig. 5A). On average, CCCP rapidly decreased fluorescence by 47 ± 6% of the initial fluorescence (F/F₀, n = 21, 123 cells). CCCP was therefore used as a positive control in TMRM experiments, and the subsequent effect of other reagents (H₂O₂, antimycin-A, and oligomycin) was quantified by normalizing them to the CCCP response. Antimycin-A (0.5 μM) caused a slow but substantial decrease in TMRM fluorescence to 68 ± 4% of the CCCP response (n = 6, 34 cells), which was significantly different when compared with corresponding time-matched control (8 ± 2%, n = 6, 84 cells, P < 0.001 as assessed by Mann-Whitney test). Over a minimum of 10 min exposure of antimycin-A, a corresponding loss of punctate TMRM fluorescence was also seen (see image B, Fig. 5B). However, oligomycin (10 μM) had no significant effect on TMRM fluorescence (7 ± 2%, n = 6, 28 cells), when compared to the time-matched control, nor did the drug affect the punctate distribution of TMRM fluorescence (see image B, Fig. 5C). H₂O₂ also caused a slow concentration-dependent decrease in TMRM fluorescence (50 μM = 35 ± 5%; n = 4, 56 cells; Fig. 5D and 500 μM = 53 ± 3%; n = 5, 36 cells, P < 0.001), which also corresponded to a loss of punctate TMRM fluorescence (image B, Fig. 5, D and E). These data suggest that CCCP rapidly depolarizes Ψm, while H₂O₂ and antimycin-A slowly depolarize the mitochondria and oligomycin had no effect on mitochondrial membrane potential in acinar cells.

Fig. 5.

Effects of mitochondrial inhibitors and H₂O₂ on the mitochondrial membrane potential (ΔΨm). Representative traces show the relative tetramethylrhodamine methylester (TMRM) fluorescence (F/F₀; arbitrary units) and corresponding TMRM fluorescent images of time-matched control (A) and cells treated with 0.5 μM antimycin-A (B), 10 μM oligomycin (C), 50 μM H₂O₂ (D), 500 μM H₂O₂ (E), and 500 μM H₂O₂ in the presence of 5 μM cyclosporin (F). A decrease in relative TMRM fluorescence represents mitochondrial depolarization. Inset images in A–C were taken at the corresponding points on the trace as indicated by the arrows. Image A from each experiment indicates a punctate TMRM fluorescence distribution consistent with mitochondrial staining, from which fluorescence changes were measured. In each experiment, 4 μM CCCP was used as a positive control to evoke maximum mitochondrial depolarization. G: mean data were quantified and normalized by expressing the change in F/F₀ as a percentage of the CCCP response (*P < 0.05, as assessed using a Mann-Whitney test). CS-A, cyclosporin-A.

One possible explanation for the H₂O₂-induced mitochondrial depolarization is due to the opening of the mPTP. This was therefore tested using cyclosporin-A, which binds cyclophilin-D, a component of the mPTP, and has been shown to potently inhibit the mPTP (8). After preincubation of cells for 30 min and in the continued presence of 5 μM cyclosporin-A, the mitochondrial depolarization evoked by 500 μM H₂O₂ was completely abolished (1.5 ± 0.3% of the CCCP response), compared with the effect of 500 μM H₂O₂ alone (53 ± 3% of the CCCP response; P < 0.001). In fact, this was found to be statistically lower than that produced during time-matched controls (8 ± 2%, P = 0.03), suggesting that there may have been a slight hyperpolarization of the mitochondrial membrane potential. The data suggest that inhibition of the mPTP protects the mitochondria from oxidant-induced depolarization and support the idea that H₂O₂ mediates the opening of the mPTP.

Inhibition of the mPTP partially prevents the H₂O₂-induced inhibition of the PMCA.

So far, our data suggest that H₂O₂-evoked inhibition of PMCA activity correlates with mitochondrial depolarization but is unlikely to be due to mitochondrial Ca²⁺ handling or to ATP depletion. The opening of the mPTP has been shown to mediate the release of a variety of pro-apoptotic factors, such as cytochrome c, which then activate a cascade of apoptotic pathways (27, 40). It is therefore tempting to speculate that a similar mechanism may be responsible for the inhibition of the PMCA during H₂O₂ treatment. We tested this hypothesis using the mPTP inhibitor, cyclosporin-A (8), to determine whether this could protect the PMCA from inhibition by H₂O₂. Cylosporin-A was applied in combination with CPA at the beginning of the experiment and was then present throughout. In an attempt to improve the reproducibility, and reduce the variability of the clearance assay, we adopted a paired experimental design whereby an initial control [Ca²⁺]_i clearance phase (R₁, Fig. 6A) was compared with a second [Ca²⁺]_i clearance phase (R₂, Fig. 6A) during which the cells were exposed to H₂O₂ (see Fig. 6, B–D). The linear rate of [Ca²⁺]_i clearance, measured from a starting Ca²⁺ of 300 nM (see dotted line, Fig. 6, B–D) as in previous experiments, was then normalized by expressing the second clearance rate as a percentage of the first (R₂/R₁ × 100%, Fig. 6E mean data), thus allowing each cell to act as its own control. Data using cyclosporin-A (Fig. 6C) were then compared with a time-matched control (Fig. 6A) and a positive control, whereby cells were treated with H₂O₂ during the second clearance phase in the absence of cyclosporin-A (Fig. 6B). The first observation from these data was that cyclosporin-A, when applied alone during the first clearance phase (see R₁, Fig. 6C), did not significantly affect the clearance (τ = 86 ± 8 s; linear rate = 145 ± 33 nM/min, n = 5, 58 cells) as compared with the equivalent clearance phase (see R₁, Fig. 6A) during the time-matched control (τ = 91 ± 9 s; linear rate = 137 nM/min ± 27, n = 5, 74 cells). This suggests that inhibition of the mPTP, which is thought to mediate mitochondrial Ca²⁺ release under certain conditions, has no effect on [Ca²⁺]_i clearance under the conditions of these experiments (i.e., when mitochondrial Ca²⁺ uptake had been prevented with Ru-360). Second, during time-matched control experiments, the second clearance phase (R₂, Fig. 6A) was not significantly different from the first clearance phase (R₁, Fig. 6A; 101 ± 3% n = 5, 74 cells, as assessed by one-sample t-test), making this an ideal method for normalizing the data. Furthermore, analysis of these normalized data revealed that, as expected, H₂O₂ dramatically inhibited [Ca²⁺]_i clearance to 15 ± 4% of the initial control clearance phase (Fig. 6B; n = 4, 59 cells; P < 0.001 as assessed by one-sample t-test) compared with the equivalent time-matched control (Fig. 6A; P < 0.001, as assessed by Mann-Whitney test). More importantly, however, in the presence of cyclosporin-A, the extent of the H₂O₂-evoked inhibition of [Ca²⁺]_i clearance was significantly reduced to 39 ± 6% of the initial control clearance phase (Fig. 6C; n = 5, 58 cells, P = 0.02 as assessed by Mann-Whitney test). However, clearance under these conditions was still significantly lower than the time-matched control, where cells were neither treated with H₂O₂ nor cyclosporin-A (P < 0.001 as assessed by one-sample t-test).

Fig. 6.

Inhibition of the mitochondrial permeability transition pore partially protects the H₂O₂ induced inhibition of the PMCA. Representative traces show a paired experimental design whereby two repetitive [Ca²⁺]_i clearance phases (R₁ and R₂) were invoked by successive application of 20 mM (20 Ca²⁺) and zero external Ca²⁺ (0 Ca²⁺). A: a time-matched control. B: effect of 500 μM H₂O₂ applied during the second clearance phase (R₂). C: combined effect of 500 μM H₂O₂ and 5 μM cyclosporin-A (applied at the same time as CPA and throughout the experiment). D: effect of 500 μM H₂O₂ following preincubation with 50 μM bongkrekic acid for 30 min. E: mean data were quantified and normalized by expressing the linear rate (standardized at 300 nM Ca²⁺) during the second clearance phase as a percentage of the linear rate during the first clearance phase (R₂/R₁ × 100%). *P < 0.05, statistical significance as assessed using a paired one-sample t-test; †P < 0.05, statistical significance as assessed using a Mann-Whitney test.

Since cyclosporin-A has been reported to have multiple molecular targets (41), further experiments were carried out using the highly specific and functionally distinct mPTP inhibitor, bongkrekic acid (21). This inhibits the adenine nucleotide translocase, responsible for ATP/ADP transfer between the mitochondrial matrix and cytosol, and is also another molecular component of the mPTP (21). Pancreatic acinar cells were preincubated with 50 μM bongkrekic acid for 30 min before beginning the [Ca²⁺]_i clearance assay. Using a similar paired experimental design, cells were treated with H₂O₂ during the second clearance phase and compared with the first clearance phase (Fig. 6D and mean data in Fig. 6E). Consistent with experiments using cyclosporin-A, preincubation with bongkrekic acid markedly attenuated the H₂O₂-induced inhibition of [Ca²⁺]_i clearance (65 ± 7% of the initial control clearance phase; Fig. 6D; n = 7, 64 cells; P < 0.001) compared with the effect of H₂O₂ alone (15 ± 4% of the initial control clearance phase) but significantly lower than time-matched control (P = 0.002, as assessed by one-sample t-test). These data therefore suggest that inhibition of the mPTP partially protects the PMCA from inhibition by H₂O₂. Moreover, since inhibition of the mPTP was found to fully protect the mitochondrial from H₂O₂-induced depolarization, any residual inhibition of the PMCA by H₂O₂ is likely to be independent of mitochondria.

DISCUSSION

We have previously demonstrated that increased oxidative stress (H₂O₂) altered the normal pattern of CCK-evoked [Ca²⁺]_i signaling and gradually resulted in an irreversible Ca²⁺ overload response in an increasing proportion of cells as H₂O₂ concentration was increased (9). This H₂O₂-induced Ca²⁺ overload also corresponded to a marked inhibition of the PMCA, which occurred at H₂O₂ concentrations above 100 μM (9). Using a modified assay, whereby [Ca²⁺]_i clearance was monitored over a much greater range of [Ca²⁺]_i concentrations, the present study confirms that H₂O₂ inhibited PMCA activity (9). The present results show that mitochondrial depolarization, independent of mitochondrial Ca²⁺ handling and ATP depletion, correlates with and may be responsible for the H₂O₂-induced inactivation of the PMCA. This phenomenon was sensitive to the mPTP inhibitors, cyclosporin-A and bongkrekic acid, suggesting that opening of the mPTP may be partly responsible for the inhibition of the PMCA, perhaps via the release of a mitochondrial factor or factors. Such a phenomenon is of interest because it may be one of the earliest, or most critical, events responsible for determining the irreversible nature of the Ca²⁺ overload response and thus the transition between apoptotic and necrotic cell death. In the pancreas this transition may be key in the progression to acute pancreatitis as previously argued (6, 50).

In the current study, whereby [Ca²⁺]_i clearance was measured over a much greater range of [Ca²⁺]_i concentrations, H₂O₂ was much more effective compared with our previous study (9), suggesting that at least a component of the H₂O₂-induced inhibition of the PMCA is Ca²⁺ dependent. This interpretation depends on all major [Ca²⁺]_i clearance pathways other than PMCA being inactive under the conditions of our [Ca²⁺]_i clearance measurements. In our experiments, SERCA was inhibited by CPA, and the Na⁺/Ca²⁺-exchanger is not thought to be expressed in pancreatic acinar cells to levels sufficient to influence Ca²⁺ extrusion (42). Mitochondrial Ca⁺ uptake, on the other hand, due to its low affinity and high capacity (17), is likely to make a significant contribution to [Ca²⁺]_i clearance when [Ca²⁺]_i is elevated to such high levels (300–600 nM). Therefore, either inhibition of mitochondrial Ca²⁺ uptake or activation of mitochondrial Ca²⁺ release (or both) by H₂O₂ could conceivably contribute to the observed [Ca²⁺]_i clearance when [Ca²⁺]_i is high. Indeed, the specific mitochondrial Ca²⁺ uptake inhibitor, Ru-360, significantly inhibited the measured [Ca²⁺]_i clearance and reduced the La³⁺-insensitive [Ca²⁺]_i clearance to negligible values, suggesting that mitochondrial Ca²⁺ uptake does contribute.

It was notable that Ru-360 slowed the increase in [Ca²⁺]_i and significantly reduced the steady-state [Ca²⁺]_i following addition of 20 mM external Ca²⁺. The rate of [Ca²⁺]_i increase under these conditions is classically believed to be governed by SOCE, and the steady-state [Ca²⁺]_i is thought to be due to a balance of SOCE and [Ca²⁺]_i clearance (51). This observation may therefore seem counterintuitive, as one might expect inhibition of mitochondrial Ca²⁺ uptake to have the opposite effect, i.e., a larger and faster increase in [Ca²⁺]_i because there is less clearance to oppose the influx. However, studies have shown that SOCE, specifically that via Ca²⁺ release-activated Ca²⁺ current (I_CRAC), is inhibited by mitochondrial Ca²⁺ uptake inhibitors (28) that increase the local [Ca²⁺] close to the CRAC channels, thus facilitating Ca²⁺-dependent inactivation of I_CRAC (60). The effect of Ru-360 on Ca²⁺ repletion is consistent with Ru-360 inhibiting mitochondrial Ca²⁺ uptake in the present study and provides further validation that all subsequent [Ca²⁺]_i clearance experiments on Ru-360-treated cells most likely represent PMCA activity. Nevertheless, although Ru-360 has proved effective in inhibiting mitochondrial Ca²⁺ uptake in the present study (see Fig. 1, B and D) and in previous studies (10, 36), we cannot completely rule out the possibility that Ru-360 is not 100% effective in every cell and that there is some residual mitochondrial Ca²⁺ uptake. However, mitochondrial Ca²⁺ uptake (La³⁺-insensitive clearance) only represented ∼15% of the total clearance in untreated control cells, which was then reduced to <5% in Ru-360-treated cells. Therefore, any residual mitochondrial Ca²⁺ uptake from Ru-360-treated cells is unlikely to contribute to such a dramatic inhibition of [Ca²⁺]_i clearance observed with H₂O₂ and suggests that an alternative mechanism is responsible for the H₂O₂-induced inhibition of PMCA activity that is independent of mitochondrial Ca²⁺ handling.

The current study also demonstrated that H₂O₂ inhibited the rate of increase in [Ca²⁺]_i following Ca²⁺ repletion, which is classically believed to be governed by SOCE. This is either due to a direct effect of H₂O₂ on the SOCE channels, most notably the candidate SOCE channel, Orai, or an indirect effect on the coupling of STIM (the ER depletion sensor) with Orai (or transient receptor potential channels) (32). An alternative explanation is that inhibition of the PMCA indirectly inhibits SOCE due to the accumulation of [Ca²⁺] close to the channels, causing Ca²⁺-dependent inactivation, similar to that described for mitochondria.

The most obvious explanation for the inhibitory effect of both H₂O₂ and mitochondrial inhibitors on PMCA activity is that inhibition of mitochondria causes ATP depletion. PMCA activity is critically dependent on ATP, and the PMCA has a high-affinity catalytic site (K_m ∼ 3 μM) and a low-affinity regulatory site (K_m ∼ 145 μM) (52). This raises two important questions. First, by how much would ATP levels have to drop in such a short amount of time (5–10 min) to account for the dramatic inhibition of PMCA activity observed with H₂O₂ and the mitochondrial inhibitors? Second, would inhibition of mitochondrial metabolism alone be enough to deplete ATP levels sufficiently to inhibit the PMCA activity, especially if glycolytic ATP production remains active, for example, in experiments with CCCP and antimycin-A?

The absolute ATP dependency of the PMCA is controversial and has recently been suggested to be more complex than proposed in earlier studies (24). This is partly due to the fact that most studies to determine ATP dependency of the PMCA have used cell-free in vitro assays. This makes extrapolation to experiments using intact cells difficult, especially when one considers the numerous cytosolic factors and plasma membrane interactions that might also influence PMCA activity. Nevertheless, several lines of evidence suggest that severe ATP depletion can inhibit PMCA activity (4, 15, 19). However, it remains unclear whether ATP depletion from inhibition of mitochondrial metabolism alone would be sufficient to inhibit PMCA activity. Indeed, it has been suggested that the PMCA has its own localized glycolytic ATP supply that may render it largely insensitive to inhibition of mitochondrial metabolism (49).

In the present study, we used MgGreen fluorescence to measure ATP levels and found that maximum ATP depletion could only be achieved with an ATP-depletion cocktail designed to maximize ATP consumption and simultaneously inhibit all ATP production. In addition, even using this ATP depletion cocktail, the rise in MgGreen fluorescence (fall in ATP concentration) took 11–39 min to reach a maximum with a latency of ∼8 min. This suggests that ATP depletion is unlikely to be responsible for the dramatic inhibition of the PMCA that occurred within 5–10 min. All three mitochondrial inhibitors (CCCP, antimycin-A, and oligomycin) depleted ATP levels by similar amounts (∼20%), consistent with inhibition of mitochondrial metabolism. ATP depletion by the high concentration of H₂O₂ (500 μM) was greater than the mitochondrial inhibitors (∼55%), suggesting that at this concentration, H₂O₂ may also partially inhibit glycolysis, consistent with previous studies (33). However, the ATP depletion observed with these agents was always much less than the maximum ATP depletion achieved with the ATP-depletion cocktail. If one extrapolates these data and assumes an initial [ATP]_i of 1 mM (3) and an approximately linear relationship between MgGreen fluorescence and [ATP]_i, this suggests that high concentrations of H₂O₂ would, at the most, only decrease ATP to ∼500 μM. This raises the question as to whether this would be sufficient to inhibit the PMCA. In addition, at the lower concentration (50 μM), H₂O₂ had no effect on ATP depletion yet significantly inhibited PMCA activity by ∼50%. Furthermore, oligomycin caused similar ATP depletion to CCCP and antimycin-A, yet had no effect on PMCA activity, suggesting that this level of ATP depletion is insufficient to inhibit the PMCA. Collectively these data strongly imply that ATP depletion alone is unlikely to be responsible for the inhibition of the PMCA activity by H₂O₂. The only caveat to these conclusions is that acidic phospholipids, such as phosphatidylserine (PS), increase the ATP sensitivity of the PMCA (45) and loss of PS from the membrane decreases the affinity for ATP (K_m ∼ 5–10 mM) (53). Therefore, H₂O₂-induced depletion of PS from the membrane (13) or membrane externalization of PS (31) may be sufficient to render the PMCA highly sensitive to only mild ATP depletion. However, these exciting possibilities are highly speculative and remain to be determined.

Data from the present study do not dispute the importance of ATP depletion in acute pancreatitis (6) nor do they rule out the possibility that “complete” ATP depletion can inhibit PMCA activity (4). However, our data do suggest that, under conditions of metabolic inhibition or high oxidant stress when ATP depletion is only partial, non-ATP-dependent mechanisms can also inhibit PMCA activity, thus revealing a potentially novel mechanism.

So far, our data suggest that H₂O₂-evoked inhibition of the PMCA is independent of mitochondrial Ca²⁺ handling and unlikely to be due to ATP depletion, at least over the time course of the [Ca²⁺]_i clearance assay (5–15 min). Interestingly, all the reagents tested that inhibited the PMCA also caused mitochondrial Ψm depolarization (H₂O₂, CCCP, and antimycin-A, but not oligomycin). This suggests that collapse of the mitochondrial Ψm may be responsible, at least in part, for the inhibition of the PMCA activity or that both phenomena share a common cause. Previous studies have also shown that oxidants, CCCP, and antimycin-A all cause mitochondrial Ψm depolarization in pancreatic acinar cells (19, 27), whereas oligomycin has no appreciable effect in adrenal chromaffin cells (34), consistent with the present study. This H₂O₂-induced mitochondrial Ψm depolarization is likely to be due to opening of the mPTP (27, 40) since this was completely prevented by the mPTP inhibitor, cyclosporin-A. Moreover, cyclosporin-A (8) and the functionally distinct mPTP inhibitor, bongkrekic acid, partially prevented the H₂O₂-induced inhibition of the PMCA, suggesting that this is partly mediated by the opening of the mPTP.

Opening of the mPTP is likely to cause dissipation of mitochondrial inner membrane ion gradients, most notably Ca²⁺ and/or H⁺. Under normal conditions, dissipation of Ca²⁺ gradients would cause substantial mitochondrial Ca²⁺ release and presumably mask the clearance by the PMCA. However, under the conditions of our [Ca²⁺]_i clearance assay, whereby mitochondrial Ca²⁺ uptake is inhibited by Ru-360, opening of the mPTP is unlikely to cause substantial mitochondrial Ca²⁺ release and therefore has minimal effects on the apparent [Ca²⁺]_i clearance. Dissipation of the mitochondrial H⁺ gradient could also conceivably inhibit the PMCA as Ca²⁺ efflux is coupled to H⁺ influx (14). However, previous studies in pancreatic acinar cells have shown that CCCP has no appreciable effect on bulk cytosolic pH (57), although local pH changes, for example if mitochondria are in close proximity to the plasma membrane (35), cannot be ruled out.

Another possibility is that opening of the mPTP causes the release of some mitochondrial factor that directly binds to or activates a secondary pathway and thus inhibits the PMCA. The release of mitochondrial factors have been well characterized and are strongly implicated as one of the early events during apoptosis (6). Although cytochrome c is perhaps the best characterized factor to be released from the mitochondria during apoptosis, several other released proteins have also been identified, including Smac/DIABLO, HtrA2, Endo G, and AIF (apoptosis-inducing factor) (6). Furthermore, the mitochondrial-specific lipid, cardiolipin (26), which under normal situations couples to cytochrome c, is also released from the mitochondria during apoptosis and is another of the acidic phospholipids reported to modulate the PMCA activity (45). Binding and/or regulation of Ca²⁺ transport proteins by mitochondrially released proteins is not an unprecedented phenomenon. For example, cytochrome c (7), Bax/Bak (46), and Bcl-2/Bcl-xL (16, 58) have all been shown to bind to, and regulate, inositol 1,4,5-trisphosphate receptors, and Bcl-2 can also regulate SERCA (38). This therefore provides a large scope for potential candidate mitochondrially derived PMCA-inhibitory factors.

Most experimental assays of apoptosis are over several hours following a pro-apoptotic stimulus. This suggests that the release of mitochondrial proteins, although one of the earlier events during apoptosis, may occur over a similar timeframe. However, some studies suggest much faster and/or abrupt changes, for example, in cytochrome c release (29). In pancreatic acinar cells, cytochrome c has also been reported to be released into the cytosol within 2 min of treatment with the oxidant menadione (27), a result that fits well with the time course of our [Ca²⁺]_i clearance assay. In addition, inhibition of the PMCA and the irreversible Ca²⁺ overload response observed in our previous (9) and current studies are more likely to facilitate necrotic rather than apoptotic cell death. However, more recently, the distinction between these two cell death pathways has become less clear, and there is now evidence that necrosis and apoptosis share common mechanisms (37). In particular, and of relevance to the current study, the mPTP has been suggested to have a role in both apoptotic and necrotic cell death (37). The mPTP is thought to be made up of a complex of proteins including the voltage-dependent anion channel, adenine nucleotide translocase, the molecular target of bongkrekic acid (21), and cyclophilin-D (20), the molecular target of cyclosporin-A (8). Transgenic studies in which cyclophilin-D has either been knocked out or overexpressed have suggested that this protein is critical in controlling necrotic cell death rather than apoptotic cell death (2). In particular, cyclophilin-D knockout mice were found to be protected against ischemia-reperfusion injury, and isolated myocytes were resistant to oxidant-induced mitochondrial swelling, permeability transition, Ca²⁺-overload, and necrotic cell death (2). Conversely, cyclophilin-D overexpressing cells were unusually more susceptible to mitochondrial swelling and spontaneous cell death (2). These studies therefore provide compelling evidence that opening of the mPTP is important for oxidant-induced Ca²⁺ overload and necrotic cell death and thus are broadly in line with our conclusion that this may also mediate inhibition of the PMCA.

It is important to note that in current experiments, cyclosporin-A only partially protected the PMCA from inhibition despite maximally protecting the mitochondria from depolarization. In addition, the inhibitory effect of high concentrations of H₂O₂ was much greater than the effect of CCCP, which also caused “full” depolarization of the mitochondrial Ψm. This suggests that opening of the mPTP by H₂O₂ is not the only mechanism responsible for the inhibition of the PMCA. These additional mechanisms could include direct oxidation of critical thiol groups within the PMCA or oxidation of CaM (59), which reduces the Ca²⁺-dependent regulation of the PMCA (25).

To summarize, the present study demonstrates that oxidative stress causes an inhibition of PMCA activity that is independent of mitochondrial Ca²⁺ handling and is unlikely to be due to ATP depletion. The study also reveals a potentially novel mechanism whereby part of the oxidant-induced inhibition of the PMCA involves mitochondrial depolarization and opening of the mPTP, suggesting that a mitochondrially released factor or factors may be responsible. Such a phenomenon could represent one of the critical events underlying the Ca²⁺ overload response that leads to necrotic cell death. Apart from being potentially important during the pathogenesis of acute pancreatitis, this may also represent a general feature of cell death in other cell types.

GRANTS

This work was supported by a Biotechnology and Biological Sciences Research Council grant awarded to J. I. E. Bruce.