A sustained increase in the cytosolic Ca2+ concentration ([Ca2+]i) can cause cell death. In this study, we found that, in cultured porcine aortic smooth muscle cells, endoplasmic reticulum (ER) stress, triggered by depletion of Ca2+ stores by thapsigargin (TG), induced an increase in the [Ca2+]i and cell death. However, the TG-induced death was not related to the [Ca2+]i increase but was mediated by targeting of activated Bax to mitochondria and the opening of mitochondrial permeability transition pores (PTPs). Once the mitochondrial PTPs had opened, several events, including collapse of the mitochondrial membrane potential, cytochrome c release, and caspase-3 activation, occurred and the cells died. TG-induced cell death was completely inhibited by the pan-caspase inhibitor Z-VAD-fmk and was enhanced by the Ca2+ chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA), suggesting the existence of a Ca2+-dependent anti-apoptotic mechanism. After TG treatment, Ca2+-sensitive mitogen-activated protein kinase (MAPK) activation was induced and acted as a downstream effector of phosphatidylinositol 3-kinase (PI 3-kinase). The protective effect of Z-VAD-fmk on TG-induced cell death was reversed by BAPTA, PD-098059 (an MAPK kinase inhibitor), or LY-294002 (a PI 3-kinase inhibitor). Taken together, our data indicate that ER stress simultaneously activate two pathways, the mitochondrial caspase-dependent death cascade and the Ca2+-dependent PI 3-kinase/MAPK anti-apoptotic machinery. The Bax activation and translocation, but not the [Ca2+]i increase, may activate mitochondrial PTPs, which, in turn, causes activation of caspases and cell death, whereas Ca2+-dependent MAPK activation counteracts death signaling; removal of Ca2+ activated a second caspase-independent death pathway.
- sarco(endo)plasmic reticulum calcium ion adenosine triphosphatase
- cytosolic calcium ion concentration
- mitogen-activated protein kinase
apoptosis, an active and highly regulated death process, is required for embryonic development and tissue homeostasis, and caspases play a central role in this process. Two main pathways have been shown to be responsible for apoptotic cell death, namely the extrinsic receptor-mediated and the intrinsic mitochondria-mediated signaling pathways (15, 37). The latter is associated with mitochondrial outer membrane permeabilization and the subsequent activation of caspase-9 and caspase-3, attributed to cytochrome c release from mitochondria (43), whereas the former is associated with caspase-8 activation (25). These two pathways might interact to cause cell death; in some cases, mitochondrial permeabilization is also involved in receptor-mediated cell death when caspase-8 activation following receptor activation results in targeting of activated Bax and/or Bak to mitochondria (19).
Many studies (reviewed in Ref. 13) have shown that, in addition to mitochondria, the endoplasmic reticulum (ER) is also involved in mediating apoptotic cell death. The ER is the major intracellular Ca2+ store and plays a dual role in Ca2+ regulation. It can release Ca2+ via the inositol trisphosphate (IP3) and ryanodine receptors to generate the Ca2+ signal during stimulation and can accumulate Ca2+ via sarco(endo)plasmic reticulum Ca2+ ATPase (SERCA) to buffer Ca2+ after stimulation, returning the cytosolic Ca2+ concentration ([Ca2+]i) to the resting state (2). In lymphocytes, Ca2+ released from the ER via the IP3 receptor is required for apoptosis (18). Recently, it was reported that cytochrome c released from mitochondria following mitochondrial permeabilization can bind to the IP3 receptor, preventing its autoinhibition by high [Ca2+]i and leading to enhanced ER Ca2+ release and a sustained [Ca2+]i increase; formation of the cytochrome c-IP3 receptor complex sensitizes the IP3 receptor and leads to a prolonged [Ca2+]i increase, which promotes cell death (3). Furthermore, any malfunction of the ER per se can also induce apoptotic cell death by so-called ER stress, for example, disruption of ER Ca2+ homeostasis or the accumulation of incorrectly folded proteins within the lumen (13).
The cell has evolved different pathways to respond to ER stress. The unfolded protein response alleviates the build-up of proteins in the ER lumen and includes increased expression of ER resident molecular chaperones, decreased general protein translation, and increased degradation of incorrectly folded proteins (30). Recently, it was shown that a survival response is also activated to counteract the adverse effects of ER stress. In human breast cancer MCF-7 cells, the protein kinase B (Akt)/inhibitor of apoptotic protein (IAP) and mitogen/extracellular signal-regulated kinase (MEK)/extracellular signal-regulated kinase (ERK) pathways are activated to protect cells from ER stress-induced death (17). When these stress adaptive responses are not sufficient to rescue cells, the apoptotic response is activated. Caspases-2, -3, -4, -7, -8, -9, and -12, Bcl-2 related proteins (Bax, Bak, PUMA, and Bim), and the ERK pathway have been shown to mediate ER stress-induced apoptosis (9, 14, 24). Thus the unfolded protein response, survival response, and apoptotic response are simultaneously activated following ER stress, and the balance between them decides whether the cell survives or dies.
One of the aims of the current study was to use cultured aortic smooth muscle cells to elucidate the sequence of events in ER stress-induced cell death and determine the connection between the [Ca2+]i increase and possible cellular mediators. To determine whether the depletion of intraluminal Ca2+ or the sustained [Ca2+]i increase was critical for cell death, we used the SERCA inhibitor thapsigargin (TG) to induce ER stress. TG not only depletes Ca2+ stores but also increases the [Ca2+]i. We found that, with the exception of the increases in the [Ca2+]i, all TG-induced events, including Bax translocation, opening of the mitochondrial permeability transition pores (PTPs), collapse of the mitochondrial membrane potential, cytochrome c release, caspase-3 activation, and cell death, were insensitive to 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA) blockade of the [Ca2+]i increase. A Ca2+-sensitive anti-apoptotic response was simultaneously activated during ER stress in which mitogen-activated protein kinase (MAPK) played a critical role.
MATERIALS AND METHODS
Culture of porcine aortic smooth muscle cells.
Porcine aortic smooth muscle cells were prepared and cultured using the explant method as described previously (5). The cells that migrated from the explants were cultured in 100-mm dishes in DMEM supplemented with 10% FBS, 100 U/ml of penicillin, and 100 μg/ml of streptomycin and maintained at 37°C in an atmosphere of 95% air and 5% CO2. The cells were passaged by incubation with 0.05% trypsin/EDTA and were used for experiments during the first eight passages. According to the needs of the experiments, cells were plated in 60-mm dishes, 24-well plates, or 8-well chambers or on 24-mm cover slips. Unless otherwise stated, cells were bathed or suspended in a buffer consisting of (in mM) 150 NaCl, 5 KCl, 1 MgCl2, 2.2 CaCl2, 5 glucose, and 10 HEPES, pH 7.4 (“loading buffer”), for the experiments. All stages after dye loading were performed in loading buffer lacking Ca2+.
Culture of NG108–15 cells.
Neuroblastoma × glioma hybrid NG108–15 cells were cultured as described previously (6). In brief, cells (passage 30) were cultured in DMEM supplemented with 5% FBS, 100 μM hypoxanthine, 1 μM aminopterin, and 16 μM thymidine and were maintained at 37°C in an atmosphere of 95% air and 5% CO2.
Measurement of the [Ca2+]i.
The [Ca2+]i change in a single cell, grown on coverslips, was measured using the fluorescent Ca2+ indicator fura 2 in loading buffer as described previously (5, 16). After cells were incubated with 5 μM fura 2-AM in loading buffer at 37°C for 20 min, the coverslip was then mounted in a modified Cunningham chamber attached to the stage of a fluorescence microscope (model DMIRB; Leica). [Ca2+]i change was measured on-line under a computer-controlled dual-excitation fluorometric Ca2+ imaging system equipped with a high-speed scanning polychromatic light source (model C7773; Hamamatsu Photonics, Hamamatsu, Japan) and a CCD camera (HISCA; model C6790; Hamamatsu Photonics) controlled by Aquacosmos 2.5 software (Hamamatsu Photonics) to acquire quantitative fluorescence intensity data with excitation wavelength of 340 and 380 nm and emission wavelength of 505 nm. The sampling rate was 1 Hz, and seven cells from a coverslip were selected for experiments. The 340- to 380-nm fluorescence ratio (F340/F380) was used to monitor [Ca2+]i changes. All experiments were performed using at least 48 cells of different batches. The results of one representative experiment averaged from seven cells are illustrated graphically, and the mean ± SD values for the ratio increase, calculated for n cells, are given in the text.
Determination of the percentage of apoptotic cells.
Chromatin condensation during the process of apoptosis was examined using the chromatin-specific dye Hoechst 33258, as described previously (6, 21). Briefly, after the indicated time of treatment with TG in the presence or absence of other drugs, cells grown on coverslips were fixed for 10 min with 3% paraformaldehyde and stained for 20 min with 10 μM Hoechst 33258 in PBS at room temperature. Nuclear morphology was then examined on an Olympus IX-70 fluorescence microscope, and apoptotic and nonapoptotic nuclei were counted in five randomly chosen fields per coverslip. The cell number in each coverslip from five fields was between 40 and 70. The percentage of apoptotic cells was expressed as the mean ± SD for four to six experiments. Statistical differences between means were assessed using Student's t-test.
3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) reduction was determined as described previously (21). Briefly, cells in 24-well plates were treated with buffer or 10 μM TG in the absence or presence of 1 μM BAPTA for the indicated time, and then 20 μl of MTT (5 mg/ml) were added to each well. After incubation at 37°C for 1 h, the blue formazan reduction product produced due to the action of mitochondrial succinate dehydrogenase in living cells but not in dead cells or their lytic debris was dissolved in isopropanol and measured on an ELISA plate reader at 570 nm. Cell survival was expressed as the absorbance at the indicated time relative to the respective initial absorbance. The data presented are means ± SD for four independent experiments using different batches of cells.
After treatment with TG in the presence or absence of other drugs for the indicated time, cells plated in an eight-well chamber were washed two times in PBS, then fixed for 15 min at −20°C with ice-cold methanol. After being blocked for 30 min with PBS containing 3% BSA, the cells were incubated for 1 h with primary rabbit anti-Bax antibody (diluted 1:200; Santa Cruz Biotechnology, Santa Cruz, CA). After washing, Texas red-conjugated second antibody (diluted 1:200; Jackson ImmunoResearch Laboratories, West Grove, PA) was added for 1 h, with the mitochondrial marker MitoTracker green (1 μM; Molecular Probes, Eugene, OR) added for the last 20 min. The cells were then washed for 3 × 15 min with PBS; propidium iodide was added in the last wash as needed. Fluorescence microscopy was performed using a model LSM 510 Zeiss (Oberkochen, Germany) confocal laser scanning microscope. Experiments were repeated six times with similar results, and one representative result is shown in ⇓⇓Fig. 3.
Measurement of mitochondrial PTP opening.
The opening of mitochondrial PTPs was measured by determining the quench of the fluorescence of calcein, loaded into mitochondria, by Co2+ influx in mitochondria as described previously (33). Briefly, cells grown on cover slips were first loaded with calcein by incubation for 20 min at 37°C with 2 μM calcein-AM (Molecular Probes) in loading buffer. The cells were then treated with TG in the presence or absence of other drugs for the appropriate time in loading buffer supplemented with 1 mM CoCl2. After treatment, the cells were examined under a computer-controlled fluorescence microscope (model DMIRB; Leica) equipped with a HISCA CCD camera (model C6790; Hamamatsu Photonics) controlled by Aquacosmos 2.5 software (Hamamatsu Photonics) to acquire quantitative fluorescence intensity data with excitation and emission wavelengths of 485 and 535 nm, respectively. The fluorescence intensity (F) was obtained from the mean of 20–30 randomly chosen cells on each cover slip. The quench of the fluorescence was expressed as F/F0, where F0 is the initial calcein fluorescence intensity. The digital images of one representative experiment and the mean ± SD values for the fluorescence intensity calculated for five experiments using different batches of cells are shown in Fig. 4. Statistical differences between means were assessed using Student's t-test.
Measurement of the mitochondrial membrane potential.
To measure the mitochondrial membrane potential, the fluorescent dye JC-1 (Molecular Probes) was used. JC-1 exists as a monomer at low mitochondrial membrane potential and emits green fluorescence but forms aggregates and emits red fluorescence at a higher mitochondrial membrane potential. After treatment of cells with TG in the presence or absence of other drugs for the indicated time, cells grown on cover slips were incubated for 30 min at 37°C with 2 μM JC-1 and then, after washing with loading buffer, were examined under a confocal laser scanning microscope (model LSM 510; Zeiss) with respective excitation and emission wavelengths of 485 and 535 nm for monomer and 550 and 600 nm for aggregates, respectively. Experiments were repeated five times with similar results, and one representative result is shown in Fig. 5.
After treatment of cells with TG in the presence or absence of other drugs for the indicated time, cells grown in 60-mm dishes were detached, and the cell pellets were suspended in 200 μl of ice-cold lysis buffer (125 mM sucrose, 10 mM HEPES, 1 mM EGTA, 20 μg/ml of leupeptin, 5 μg/ml of aprotinin, 1 mM dithiothreitol, and 200 μM phenylmethylsulfonyl fluoride, pH 7.0) then homogenized by mechanical shearing by 20 passages through a Hamilton syringe. For cytochrome c blots, the homogenates were centrifuged at 800 g for 5 min at 4°C, and the supernatant was centrifuged again at 16,000 g for 30 min at 4°C; the final supernatant was designated the cytosolic fraction, and the final pellets, resuspended in lysis buffer, were designated the mitochondrial fraction. For Akt and MAPK blots, the homogenates were centrifuged at 25,000 g for 30 min, and the supernatant was used. For caspase-12 cleavage blots, the homogenate was used. Protein content was determined using the Bradford assay (Bio-Rad Laboratories, Hercules, CA). Aliquots of supernatants or resuspended pellets (10 μl, 15–20 μg) were mixed with an equal volume of twofold concentrated SDS-PAGE buffer and boiled for 10 min at 95°C. After electrophoresis, the proteins were electrophoretically transferred to nitrocellulose paper. Cytochrome c in the cytosol or mitochondria was analyzed by immunoblotting with a rabbit polyclonal anti-cytochrome c antiserum (diluted 1:1,000; Clontech Laboratories, Mountain View, CA). As controls, mouse monoclonal antibodies against α-tubulin (diluted 1:1,000; Oncogene Research Products, Boston, MA) or cytochrome c oxidase subunit IV (cox IV; 1:1,000; Clontech Laboratories), respective markers for the cytosol and mitochondria, were used. Bound antibody was detected using appropriate horseradish peroxidase-conjugated secondary antibodies (diluted 1:1,000) and enhanced chemiluminescence substrate (PerkinElmer Life Sciences, Boston, MA). The antibodies used for MAPK, Akt, and caspase-12 were rabbit polyclonal antibodies (all diluted 1:1,000) against p44/42 MAPK, phospho-p44/42 MAPK, Akt, phospho-Akt (Ser473) (all from Cell Signaling Technology, Beverly, MA), or caspase-12 (BioVision Research Products, Mountain View, CA). To confirm the detection of phospho-Akt, quiescent cells, obtained by culturing cells in 0.1% FBS for 24 h, were incubated for 10 min with 10 ng/ml of platelet-derived growth factor (PDGF). All experiments were repeated four times with similar results. Representative results are shown in Figs. 6 and 7.
Caspase-3 activity was measured using a caspase-3 assay kit (BD PharMingen, San Diego, CA) according to the manufacturer's instructions, as described previously (6). Briefly, after 3 h treatment of cells with TG in the presence or absence of other drugs as indicated, cells grown in 60-mm dishes were detached and lysed; next, 50-μl aliquots (0.2 mg of protein) of cell lysate were mixed with fluorogenic caspase-3 substrate and incubated at 37°C for 60 min; and the fluorescence of the liberated product was measured using a spectrofluorometer (CM System; Spex Industries, Edison, NJ). The emission fluorescence spectrum was scanned between 400 and 480 nm using an excitation wavelength of 380 nm; active caspase-3 results in considerable emission at 440 nm. Caspase-3 activity was expressed as fluorescence intensity (in arbitrary units) in control and TG-treated cells with or without the indicated drugs. Experiments were repeated five times using different batches of cells, and similar results were observed; one representative result is shown in Fig. 6.
We previously showed that, in porcine aortic smooth muscle cells, the [Ca2+]i increase in response to stimulation by G protein-coupled receptor agonists, including ATP, bradykinin, lysophosphatidic acid, and sphingosylphosphorylcholine, is mainly due to Ca2+ release from intracellular Ca2+ stores (5). Thus, in these cells, the intracellular Ca2+ store plays an important role in Ca2+ homeostasis. In the present study, we examined whether cell death was induced after depletion of the intracellular Ca2+ stores by the SERCA inhibitor TG and whether the increase in the [Ca2+]i, resulting from depletion of the intracellular Ca2+ stores, was required to cause cell death. To simplify the interpretation of the results, the entry of Ca2+ via store-operated Ca2+ channels that are activated following depletion of the intracellular Ca2+ store was minimized by removal of extracellular Ca2+. Thus all experiments were performed in loading buffer lacking Ca2+. As shown in Fig. 1A, trace a, the [Ca2+]i, expressed as F340/F380, increased from the basal level of 0.98 ± 0.12 (n = 52) to a peak level of 1.38 ± 0.15 (n = 52) within 30 s after addition of 10 μM TG and then fell to a new basal level of 1.11 ± 0.13 (n = 52) over the next 180 s. Pretreatment of cells with the pan-caspase inhibitor Z-VAD-fmk had no effect on the basal [Ca2+]i or the TG-induced [Ca2+]i increase, with the basal, peak, and second basal [Ca2+]i being 0.96 ± 0.14 (n = 48), 1.40 ± 0.11 (n = 48), and 1.06 ± 0.10 (n = 48), respectively (Fig. 1A, trace b). Incorporation of the Ca2+ chelator BAPTA in cells by incubation with the permeable BAPTA-AM for 5 min efficiently prevented the [Ca2+]i increase induced by TG in both control and Z-VAD-fmk-treated cells (Fig. 1, traces c and d). These data show that the TG-induced [Ca2+]i increase is caspase independent.
We next examined the effect of TG on cell death and whether cell death was dependent on the TG-induced [Ca2+]i increase. Figure 1B shows the time course of cell death determined by chromatin condensation and clumping after cells were exposed to TG in the presence or absence of BAPTA and/or Z-VAD-fmk. Apoptotic cell death increased with exposure time to TG, being 49 ± 3% (n = 6) after 5 h of TG treatment compared with 6 ± 2% (n = 6) in control cells (Fig. 1Ba). In cells treated simultaneously with TG and Z-VAD-fmk, the percentage of apoptotic cells at 5 h decreased significantly to 23 ± 3% (n = 6; Fig. 1Bb). In contrast, cell death at 5 h increased slightly, but significantly, when BAPTA was incorporated into cells to 69 ± 4% (n = 6; Fig. 1Bc). BAPTA or Z-VAD-fmk per se did not cause cell death (Fig. 1B, b and c). Unexpectedly, the inhibitory effect of Z-VAD-fmk on TG-induced cell death was prevented when BAPTA was added simultaneously, with cell death being 53 ± 2% (Fig. 1Bd compared with Fig. 1Bb). These results show that TG-induced cell death is dependent on caspase activation and is insensitive to the [Ca2+]i increase and that there may be a second caspase-independent death pathway activated by Ca2+ removal.
To ensure that TG action specifically occurs at the level of inhibiting SERCA rather than influencing other cellular components, the effect of other agents, either inhibiting SERCA or causing ER stress, on cell death was examined. Cyclopiazonic acid is a SERCA inhibitor, structurally different from TG (39). Brefeldin A and tunicamycin both induce ER stress with distinct mechanism from that of TG. Brefeldin A specifically blocks translocation of proteins from ER to Golgi apparatus (40), whereas tunicamycin inhibits the N-linked glycosylation of protein (38). Figure 2A demonstrates the effect of cyclopiazonic acid, brefeldin A, and tunicamycin on apoptotic cell death in the absence or presence of BAPTA. Cyclopiazonic acid (100 μM)-induced apoptotic cell death was indistinguishable in the absence or presence of BAPTA, being 50 ± 3% (n = 4) and 50 ± 4% (n = 4) (Fig. 2Aa compared with Fig. 2Ab), respectively, after 5 h of cyclopiazonic acid treatment compared with 9 ± 2% (n = 4) and 12 ± 3% (n = 4) in control cells. Brefeldin A and tunicamycin display similar results, with the cell death being 60 ± 3% (n = 4) and 48 ± 2% (n = 4) at 5 h exposure to brefeldin A (100 μM) and 61 ± 5% (n = 4) and 63 ± 4% (n = 4) for tunicamycin (10 μM) in the absence and presence of BAPTA, respectively. Thus ER stress-induced cell death is independent of the change of [Ca2+]i; regardless of how ER stress is caused by interference of Ca2+ homeostasis with cyclopiazonic acid, by interference of N-glycosylation with tunicamycin, or by interference of membrane trafficking with brefeldin A. We further confirmed this phenomenon by determining cell viability with MTT assay. As shown in Fig. 2B, MTT reduction by succinate dehydrogenase in living cells was reduced following exposure of smooth muscle cells to TG. Again, incorporation of BAPTA had no effect on TG's action. The decrease of cell viability induced by TG in control and BAPTA-treated cells was nonsignificantly different. We next determined whether this Ca2+ independence of TG-induced cell death is still seen in other cell types. Previously, we have shown that TG induced [Ca2+]i increase and cell death in neuroblastoma × glioma hybrid NG108–15 cells (6). In the current study, we found that, in NG108–15 cells, TG-induced apoptotic cell death was also insensitive to the increase of [Ca2+]i, since the incorporation of BAPTA had no effect on TG-induced cell death, as shown in Fig. 2C.
Bax is a member of the proapoptotic Bcl-2 family and is activated and translocated to the mitochondria in response to death stimuli, causing cytochrome c release from mitochondria (1). We therefore examined whether Bax activation was induced by TG and characterized the underlying mechanism. Cells were treated with TG in the presence and absence of BAPTA or cyclosporin A, a mitochondrial PTP inhibitor. To localize Bax mitochondrial staining during apoptosis, we double stained cells with anti-Bax antibody and the mitochondrial marker MitoTracker green. The antibody we used preferentially recognizes the active conformation of Bax, the epitope mapping to the NH2 terminus (11). As shown in Fig. 3, immunostaining of cells treated for 1 h with TG demonstrated increased activated Bax immunoreactivity compared with control cells. In addition, the pattern of activated Bax staining in TG-treated cells matched that of MitoTracker green staining, whereas negligible amounts of activated Bax colocalized with MitoTracker green in control cells. No red fluorescence resulting from activated Bax was detected in control cells, so, to confirm red fluorescence was indeed detectable in these cells, propidium iodide was incorporated, and the positive result is shown in Fig. 3, inset. Bax activation and the appearance of Bax mitochondrial staining were independent of the concomitant increase in the [Ca2+]i induced by TG, since incorporation of BAPTA in cells failed to prevent the increase in Bax immunoreactivity in TG-treated cells. Similarly, colocalization of activated Bax with MitoTracker green was still seen after cotreatment of cells with TG and cyclosporin A. These data indicate that exposure to TG induces Bax activation and that TG-induced Bax activation is not reduced by the decrease of [Ca2+]i and inhibition of PTPs.
Activation of mitochondrial PTPs is involved in the intrinsic mitochondrial death pathway. We next determined whether the opening of mitochondrial PTPs was activated by TG treatment and whether this was involved in TG-induced cell death. Because the mitochondrion is the only organelle impermeable to Co2+, the opening of mitochondrial PTPs can be quantified by the quench of mitochondrial calcein fluorescence following Co2+ influx in mitochondria after mitochondrial permeabilization (31). In Fig. 4A, the calcein fluorescence remaining in cells in the presence of 1 mM Co2+ was mainly due to dye in the mitochondria seen in cells treated with buffer, BAPTA, Z-VAD-fmk, or cyclosporin A (Fig. 4A, a–d). In these groups of cells, the calcein fluorescence was slightly and steadily quenched by Co2+ as the incubation time increased, with the fluorescence decrease being 8–24% at 12 min (Fig. 4A, a′–d′). When the cells were treated with TG, the quench of the calcein fluorescence was significantly increased to about 62% at 12 min (Fig. 4A, e vs. e′), indicating permeabilization of the mitochondria. The TG-induced fluorescence quench was inhibited by cyclosporin A (Fig. 4A, h and h′) but not by BAPTA or Z-VAD-fmk (Fig. 4A, f, f′, g, and g′), with about 93, 36, or 40% of the fluorescence remaining at 12 min, respectively. Figure 4B shows the statistical data for TG-induced mitochondrial PTP opening in the presence of the indicated drugs. These results show that mitochondrial PTP opening is insensitive to the [Ca2+]i increase and caspase activation.
The opening of mitochondrial PTPs dissipates the mitochondrial membrane potential, which is maintained by the integrity of the inner membrane and respiration. We therefore examined the effect of TG on the mitochondrial membrane potential in the absence or presence of BAPTA or cyclosporin A. As shown in Fig. 5A, TG treatment induced a shift from the aggregate (red) to the monomeric (green) form of the dye, JC-1, indicating loss of membrane potential (Fig. 5A, right) compared with untreated control cells (Fig. 5A, left), which had an abundance of JC-1 aggregate-containing cells. The TG-induced increase in green fluorescence and the concomitant disappearance of red fluorescence was not seen at 3 min (Fig. 5A, d and d′), but appeared after 10 min of TG treatment (Fig. 5A, e and e′). Simultaneous treatment of cells with TG and cyclosporin A resulted in less of a change in mitochondrial membrane potential, as indicated by less green fluorescence and more red fluorescence (compare the panels in center with those in top in Fig. 5B), whereas BAPTA incorporation in cells before TG treatment had a negligible effect on the TG-induced mitochondrial membrane potential change (compare the panels on bottom with those on top in Fig. 5B). These data show that TG causes a collapse of the mitochondrial membrane potential that is insensitive to the [Ca2+]i change but is mediated by the opening of mitochondrial PTPs.
Our data indicated that the TG-induced Bax activation and translocation to the mitochondria, opening of mitochondrial PTPs, collapse of the mitochondrial membrane potential, and cell death were all insensitive to the [Ca2+]i increase. We next determined whether TG induced cytochrome c release and caspase-3 activation and the effect of BAPTA and cyclosporin A. As shown in Fig. 6, cytochrome c release was seen after 1 h treatment with TG and increased further at 2 h (Fig. 6A). No cytochrome c release was seen in control cells under the same experimental conditions. The cytochrome c seen in the cytosolic fraction was not the result of mitochondrial contamination, since the mitochondrial marker cox IV was not detected, whereas the cytosolic marker α-tubulin was present at a similar density in all cytosolic fractions. Although a considerable amount of cytochrome c remained in the mitochondrial fractions after 2 h treatment with TG, it was less than that in control cells. TG-induced cytochrome c release was not significantly changed by cotreatment of the cells with Z-VAD-fmk or BAPTA (Fig. 6B), the induced cytochrome c release being 0.92- or 1.27-fold, respectively, of that in cells treated with TG alone. The extent of cytochrome c release was similar when cells were treated with TG and BAPTA plus Z-VAD-fmk, being 1.33-fold that seen with TG alone. A graph depicting the degree of change of densitometric data relative to that in cells treated with TG alone from four experiments is shown in Fig. 6B, right. In contrast, TG-induced cytochrome c release was markedly inhibited by cyclosporin A, with virtually no cytochrome c being seen in the cytosolic fraction in all groups (Fig. 6B). As shown in Fig. 6C, caspase-3 activity, indicated by the fluorescence emitted at 440 nm, increased after 3 h treatment with TG compared with in control cells treated with buffer alone (TG vs. buffer), and this increase was only reduced slightly after incorporation of BAPTA (TG vs. TG + BAPTA). BAPTA alone had no effect on caspase-3 activation (BAPTA vs. buffer; Fig. 6Ca). These data suggest that the [Ca2+]i increase is not a prerequisite for caspase-3 activation after treatment of cells with TG. Our data further showed that Z-VAD-fmk completely blocked the TG-induced activation of caspase-3 regardless of the presence or absence of BAPTA (Fig. 6Cb), as did cyclosporin A (Fig. 6Cc). Figure 6D shows the results for cell death measured by chromatin condensation. TG-induced cell death was significantly inhibited by cyclosporin A, with the death rate being 4 ± 0.96% (n = 4) in the presence of cyclosporin A and 71 ± 15.7% (n = 4) in its absence. Cyclosporin A per se had no effect on cell death under the same experimental conditions. Because prolonged ER stress causes cleavage of caspase-12 (26), we also examined the effect of TG on caspase-12 cleavage in smooth muscle cells. As shown in Fig. 6E, lanes 5–8, in the absence of TG, basal cleavage of caspase-12 occurred, which was not affected by BAPTA or cyclosporin A but was inhibited by Z-VAD-fmk. In the presence of TG (Fig. 6E, lanes 1–4), additional cleavage of caspase-12 was seen (Fig. 6E, lane 4 vs. lane 8), which was insensitive to BAPTA but was inhibited by cyclosporin A or Z-VAD-fmk. The corresponding mean value of statistical quantitative densitometric data of each band relative to that in control cells without any treatment from three experiments is indicated in Fig. 6E, bottom.
In addition to the unfolded protein response, ER stress also activates certain survival elements to prevent ER stress-induced apoptotic signaling (6, 17). The data in Fig. 1B support this notion by suggesting the existence of a Ca2+-dependent prosurvival mechanism that resists ER stress-induced cell death in smooth muscle cells. After removal of Ca2+ by BAPTA, a second caspase-independent death pathway was activated. We therefore examined possible components of this second signaling pathway during ER stress. First, we examined whether TG activated MAPKs (p44 and p42) and Akt and the effect of BAPTA. As shown in Fig. 7A, phosphorylation of p44/42 was induced after 15 min exposure to TG (Fig. 7A, lane 2 vs. lane 1), and this effect was Ca2+-dependent, since it was decreased in the presence of BAPTA (Fig. 7A, lane 2 vs. lane 5). MAPKs are phosphorylated by MAPK kinase; to confirm the involvement of MAPK, the MAPK kinase inhibitor PD-098059 was used and markedly inhibited TG-induced p44/p42 phosphorylation (Fig. 7A, lanes 7–9). To explore the relationship between phosphatidylinositol 3-kinase (PI 3-kinase) and MAPK, cells were incubated with TG in the presence of the PI 3-kinase inhibitor LY-294002, which resulted in a significant decrease in TG-induced p44/p42 phosphorylation (Fig. 7A, lane 2 vs. lane 11), showing that the TG-induced MAPK activation was PI 3-kinase dependent. The pan-caspase inhibitor Z-VAD-fmk had little or no effect on the TG-induced phosphorylation of p44/p42 (Fig. 7A, lane 2 vs. lane 3). In contrast, levels of phospho-Akt (Ser473) were not affected by 15 min treatment with TG with or without BAPTA, PD-098059, or LY-294002. To confirm the ability of our system to detect phospho-Akt, a positive control was used in which phospho-Akt was detected after quiescent cells were treated for 10 min with 10 ng/ml of PDGF. In addition, because sphingosine 1-phosphate has been shown to inhibit ceramide-mediated programmed cell death via MAPK activation (8), we also examined the effect of two sphingosine kinase inhibitors, N,N-dimethylsphingosine and dl-threodihydrosphingosine, on the TG-induced phosphorylation of p44/p42 and found that neither had an effect (data not shown). The statistical quantitative data from four experiments for p44 MAPK phosphorylation analyzed by densitometry are shown in Fig. 7B. To characterize the underlying mechanism of the second signaling pathway during ER stress described in Fig. 1B, we next examined whether inhibition of TG-induced MAPK activation masked the protective effect of Z-VAD-fmk on TG-induced cell death. As shown in Fig. 7C, cell death induced by 5 h treatment with TG was unaffected in the presence of PD-098059 or LY-294002, being 50 ± 0.5% (n = 4) and 54 ± 9% (n = 4), respectively, compared with 49 ± 9% (n = 4) in cells treated with TG alone (Fig. 7C, b and c vs. a); in contrast Z-VAD-fmk was again protective, with cell death being 15 ± 5% (n = 4; Fig. 7Cd). However, the protective effect of Z-VAD-fmk on TG-induced cell death was completely inhibited by addition of PD-098059 or LY-294002, with cell death being fully restored to 56 ± 11 or 46 ± 5%, respectively (Fig. 7C, e and f vs. d), results consistent with the effect of PD-098059 and LY-294002 in blocking MAPK phosphorylation.
In this study, we showed that the ER stress induced by Ca2+ depletion by TG caused cell death and an associated [Ca2+]i increase. In general, organelles accumulate a substantial amount of Ca2+. It is believed that cytosolic Ca2+, especially when the concentration is high, is sequestered into not only the ER but also mitochondria via the mitochondrial Ca2+ uniporter. Intracellular organelles are therefore involved in shaping the [Ca2+]i signal after stimulation to avoid the cytotoxicity caused by high [Ca2+]i. [Ca2+]i overload plays a central role in cell death. However, in the present study, we obtained different results. The opening of mitochondrial PTPs induced by TG was not affected by pretreatment of the cells with BAPTA (Fig. 4), which prevented the [Ca2+]i increases (Fig. 1). Furthermore, except for Bax activation and translocation (Fig. 3), all the TG-induced cellular events examined, including collapse of the mitochondrial membrane potential (Fig. 5), cytochrome c release, caspase-3 activation, and cell death (Fig. 6), were sensitive to cyclosporin A, indicating that mitochondrial PTP opening is a central event in ER stress-induced cell death. The fact that Bax activation and translocation was the only TG-induced cellular event that was insensitive to cyclosporin A treatment (Fig. 3) rules out the possibility that Bax activation/translocation occur downstream of mitochondrial PTPs opening. TG-induced Bax activation/translocation not only was independent of the opening of mitochondrial PTPs but also of the [Ca2+]i increase, since it was not affected by BAPTA (Fig. 3). Taken together, our data suggest that Bax activation/translocation and mitochondrial PTP opening, rather than the [Ca2+]i increases, are the initial triggers in the ER stress-induced death pathway.
Our data show that cytochrome c release from mitochondria occurred after 1 h of TG treatment of smooth muscle cells. Cytochrome c release from mitochondria proceeds by a two-step process: its dissociation from cardiolipin in the inner mitochondrial membrane and its release following permeabilization of the outer mitochondrial membrane (29, 32). Permeabilization of the outer mitochondrial membrane occurs after Bax activation and translocation into mitochondria and is attributable to Bax oligomer pores formed in the mitochondria (12). The opening of mitochondrial PTPs is another recognized mechanism for outer mitochondrial membrane permeabilization and is associated with osmotic swelling of mitochondrial matrix and the subsequent rupture of the outer mitochondrial membrane (7, 15). These two mechanisms might interact in causing outer mitochondrial membrane permeabilization. After activation/translocation, Bax binds to the voltage-dependent anion channel (VDAC), a component of the mitochondrial PTPs, and forms a hybrid pore with VDAC to allow cytochrome c release (35). Bax also interacts with adenine nucleotide translocase (ANT), another component of the mitochondrial PTPs, and enhances its activity in controlling cell death (23). In our study, it is possible that both the Bax activation/translocation to mitochondria and the opening of mitochondrial PTPs were responsible for the outer mitochondrial membrane permeabilization following TG treatment. However, our data also showed that, in the presence of cyclosporin A, little TG-induced cytochrome c release was seen, suggesting that release is predominantly the result of mitochondrial PTP opening and that release through Bax oligomer pores in the outer mitochondrial membrane is negligible. It is also likely that, in porcine aortic smooth muscle cells, TG initially induces Bax activation/translocation and that the migration of Bax to mitochondria in turn activates the opening of mitochondrial PTPs by its interaction with VDAC or ANT.
A number of recent studies have linked caspase-12 activation to ER stress-induced apoptosis (26). The involvement of Bax in ER stress-induced caspase-12 activation has also been suggested (42). In these studies, ER stress-induced caspase-12 activation was independent of mitochondrial damage. In contrast, it has also been reported that caspase-12 is not required for the induction of ER stress-induced apoptosis (27, 28). Our data show that TG-induced cytochrome c release was unaffected by the pan caspase inhibitor Z-VAD-fmk, ruling out the involvement of caspases upstream of cytochrome c release in the death pathway (Fig. 6B). Cleavage of caspase-12 occurred after 3 h treatment of cells with TG and was insensitive to cotreatment with BAPTA but highly sensitive to cyclosporin A (Fig. 6E), suggesting that it was an end product of the mitochondrial caspase-9 and -3 cascade rather than a primarily cleaved product in the ER. Consistent with this is the finding that TG-induced caspase-3 activation was completely abolished by cyclosporin A, indicating dependence of caspase-3 activation on cytochrome c release (Fig. 6C). Moreover, TG-induced cell death was barely detectable after inhibition of the mitochondrial PTPs by cyclosporin A and of caspases by Z-VAD-fmk (Figs. 1B and 6D). Thus, in porcine aortic smooth muscle cells, TG treatment predominantly activates the intrinsic mitochondrial pathway to cell death, which is cyclosporin A sensitive and caspase dependent. Activation of caspase-12 was mediated by either Bax activation/translocation or mitochondrial PTP opening. Recently, it was reported that, in neuronal cells, ER stress induces activation of caspase-12 via caspase-3, consistent with our finding (20). The question remains how Bax activation/translocation occurs during ER stress in porcine aortic smooth cells.
Many studies (4, 34) have shown the involvement of Bax in ER stress-induced apoptosis, although the trigger for Bax activation has not been identified. Glycogen synthase kinase-3β (GSK3β) has been shown to be the molecular link between ER stress and Bax activation. Many studies (22, 36) support the notion that ER stress results in increased dephosphorylation of Akt and GSK3β, and the latter is therefore activated and, in turn, activates Bax via phosphorylation or caspase-3, leading to cell death; in contrast, growth factor increases levels of phospho-Akt and phospho-GSK3β via PI 3-kinase activation and reduces susceptibility to cell death. In myocardial H9c2 cells, a significant reduction in phospho-Akt levels is seen after 3 h of TG treatment (41). In the present study, we found that MAPK phosphorylation occurred after 15 min of TG treatment (Fig. 7A). To identify the upstream kinase involved in MAPK phosphorylation, we measured Akt phosphorylation at the same time point and found that phospho-Akt levels were not altered (Fig. 7A). This lack of an increase in phospho-Akt may explain the activation/translocation of Bax after TG treatment in porcine aortic smooth muscle cells. In addition to GSK3β, the opening of mitochondrial PTPs is reported to cause Bax translocation and multimerization; in Cos 7 cells, mitochondrial PTP opening causes relocation of cytosolic Bax to the outer mitochondrial membrane, which is responsible for cytochrome c release not associated with mitochondrial swelling (10). In our study, TG-induced Bax activation/translocation was insensitive to cyclosporin A (Fig. 3), ruling out the possibility that it was downstream of mitochondrial PTP opening.
In addition to the death signaling triggered by TG, our data suggest that an anti-apoptotic pathway was activated in parallel and that this was Ca2+ sensitive. First, not only was TG-induced cell death enhanced by BAPTA but also the protective effect of Z-VAD-fmk on TG-induced cell death was completely abolished by BAPTA, suggesting that a Ca2+-sensitive pathway counteracted cell death and was required for cell survival. Blockade of the antiapoptotic pathway by Ca2+ chelation causes the activation of another caspase-independent death pathway. So, if the Ca2+ is chelated, cells die via caspase-independent death pathway even when the caspase-dependent death pathway is inhibited by caspase inhibition (Fig. 1B). Second, LY-294002 and PD-098059 mimicked the action of BAPTA in reversing the protective effect of Z-VAD-fmk on TG-induced cell death, further suggesting that PI 3-kinase and MAPK are involved in this anti-apoptotic pathway (Fig. 7C). Third, Ca2+-sensitive MAPK phosphorylation was induced by TG and was inhibited by LY-294002 (Fig. 7A), suggesting that PI 3-kinase is involved upstream of MAPK phosphorylation. Thus we showed that the PI 3-kinase/MAPK cascade acts as the Ca2+-dependent anti-apoptotic pathway that is simultaneously activated by TG to overcome the death signal. Recently, it was shown that, in the MCF-7 human breast cancer cell line, PI 3-kinase/Akt and MEK/MAPK are acutely activated in ER stress and subsequently prevent the apoptosis response, and the authors speculated that Ca2+ is a potential linker between ER stress and PI 3-kinase activation (17). In the present study, we provided evidence for Ca2+ dependency of the survival response. In addition, sphingosine 1-phosphate has been shown to antagonize ceramide-mediated cell death via MAPK activation (8). We have previously shown that, in NG108–15 cells, caspase-3 is activated by 3 h treatment with either TG or the protonophore FCCP, but TG-induced cell death is not seen until 7 h of treatment, whereas FCCP-induced cell death is seen at 3 h, because TG also activates sphingosine kinase to generate sphingosine 1-phosphate, which delays cell death (6). In the present study, TG-induced MAPK phosphorylation was not inhibited by two sphingosine kinase inhibitors (data not shown), ruling out involvement of sphingosine 1-phosphate in MAPK activation.
In conclusion, we have provided evidence that the ER stress induced by depletion of intracellular Ca2+ stores activates two signaling pathways, leading to cell death or cell survival. The death signaling is mediated by the mitochondrial caspase death cascade and is insensitive to the [Ca2+]i, whereas the survival signaling is mediated by the PI 3-kinase/MAPK cascade and is Ca2+-dependent; removal of Ca2+ activates a second caspase-independent death pathway.
This work was supported by National Science Council Grants NSC93–2311-B033–002 (T.-Y. Chin) and NSC93–2320-B016–043 (S.-H. Chueh), Taiwan, Republic of China.
We thank Dr. Thomas Barkas for helpful discussion.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society