Lack of Oxidative Phosphorylation and Low Mitochondrial Membrane Potential Decrease Susceptibility to Apoptosis and Do Not Modulate the Protective Effect of Bcl-xL in Osteosarcoma Cells*

We explored the role of low mitochondrial membrane potential (ΔΨm) and the lack of oxidative phosphorylation in apoptosis by assessing the susceptibility of osteosarcoma cell lines with and without mitochondrial DNA to staurosporine-induced death. Our cells without mitochondrial DNA had low ΔΨm and no functional oxidative phosphorylation. Contrary to our expectation, these cells were more resistant to staurosporine-induced death than were the parental cells. This reduced susceptibility was associated with decreased activation of caspase 3 but not with the mitochondrial permeability transition pore or cytochrome c release from the mitochondria. Apoptosis in both cell lines was associated with an increase in ΔΨm. Bcl-xL could protect both cell types against caspase 3 activation and apoptosis by a mechanism that does not appear to be mediated by mitochondrial function or modulation of ΔΨm. Nevertheless, we found that Bcl-xL expression can stimulate cell respiration in cells with mitochondrial DNA. Our results showed that the lack of functional oxidative phosphorylation and/or low mitochondrial membrane potential are associated with an antiapoptotic effect, possibly contributing to the development of some types of cancer. It also reinforces a model in which Bcl-xL can exert an antiapoptotic effect by stimulating oxidative phosphorylation and/or inhibiting caspase activation.

We explored the role of low mitochondrial membrane potential (⌬⌿m) and the lack of oxidative phosphorylation in apoptosis by assessing the susceptibility of osteosarcoma cell lines with and without mitochondrial DNA to staurosporine-induced death. Our cells without mitochondrial DNA had low ⌬⌿m and no functional oxidative phosphorylation. Contrary to our expectation, these cells were more resistant to staurosporine-induced death than were the parental cells. This reduced susceptibility was associated with decreased activation of caspase 3 but not with the mitochondrial permeability transition pore or cytochrome c release from the mitochondria. Apoptosis in both cell lines was associated with an increase in ⌬⌿m. Bcl-x L could protect both cell types against caspase 3 activation and apoptosis by a mechanism that does not appear to be mediated by mitochondrial function or modulation of ⌬⌿m. Nevertheless, we found that Bcl-x L expression can stimulate cell respiration in cells with mitochondrial DNA. Our results showed that the lack of functional oxidative phosphorylation and/or low mitochondrial membrane potential are associated with an antiapoptotic effect, possibly contributing to the development of some types of cancer. It also reinforces a model in which Bcl-x L can exert an antiapoptotic effect by stimulating oxidative phosphorylation and/or inhibiting caspase activation.
Programmed cell death, or apoptosis, plays an important role in normal development and is impaired in many types of cancer (1). The role of mitochondria in programmed cell death came into the spotlight after the discovery that antiapoptotic members of the Bcl-2 protein family localize mainly to the outer mitochondrial membrane (2)(3)(4)(5), and the proapototic BAX translocates to mitochondrial membranes after a death signal (6).
The participation of mitochondria in apoptosis has also been substantiated by a large number of reports describing proapoptotic mitochondrial alterations, such as the production of reactive oxygen species, the depletion of ATP, and the opening of the mitochondrial permeability transition pore (MPTP) 1 (7,8).
Opening of the MPTP causes a dissipation of the inner mitochondrial membrane potential (⌬⌿m), an increase in the matrix volume, and a consequential disruption of the outer mitochondrial membrane, leading to the release of intermembrane factors (9). This sequence of events is still controversial because many groups have observed either an increase in ⌬⌿m after an apoptotic stimulus (5) or cytochrome c release preceding a decrease in ⌬⌿m (10 -12). In any case, a large number of proapoptotic factors seem to be released from the mitochondrial intermembrane space early during the apoptotic cascade. Released cytochrome c, for example, interacts with Apaf-1 and activates cytosolic caspases (13,14). Caspases 2 and 9, the latter of which is the main activator of the cytoplasmic death executioner caspase 3, also seem to be present in mitochondria and, like cytochrome c, are released upon an apoptotic stimulus (9). The release of these factors from the mitochondria in some systems is blocked by cyclosporin A, an inhibitor of the MPTP. MPTP is a composite ion channel that spans the inner and outer mitochondrial membranes at sites of contact between them (15). The complex is formed by the adenine nucleotide translocator (inner membrane), the voltage-dependent anion channel (outer membrane), and cyclophilin D (matrix) and co-purifies with other apoptosis-regulating factors such as Bax but not Bcl-x L (16). MPTP, a regulator of mitochondrial Ca 2ϩ , pH value, ⌬⌿m, and volume, has been implicated in apoptosis of hepatocytes, neurons, and myocardiocytes induced by toxins, Ca 2ϩ , reactive oxygen species, and anoxia (17). It has been proposed that MPTP constitutes a point of integration of multiple apoptotic pathways because most of them facilitate its opening (18,19).
It also has been proposed that Bax, Bcl-2, and Bcl-x L exert at least part of their apoptosis-regulatory function by facilitating (Bax) or inhibiting (Bcl-2 and Bcl-x L ) pore opening and ⌬⌿m dissipation (20). However, mitochondria or oxidative phosphorylation (OXPHOS) may not be major players in some apoptosis systems. Caspase activation, for example, can occur without the release of mitochondrial intermembrane factors under certain conditions (21,22).
Despite the wealth of information generated in the last few years, cell death mechanisms and the involvement of mitochondria in different cell systems are still poorly understood. However, because of the undisputed importance of OXPHOS and the proposed effects of Bcl-x L at the level of ⌬⌿m and the MPTP, studies of cell death in cells harboring mitochondria with genetically defined defects in OXPHOS and ⌬⌿m can provide valuable clues to molecular mechanisms of apoptosis. Our study using osteosarcoma cell lines with and without mtDNA shows that the protective effect of Bcl-x L is not impaired by a low mitochondrial membrane potential and that the lack of functional OXPHOS does not predispose cells to apoptosis.

EXPERIMENTAL PROCEDURES
Cell Lines-The human osteosarcoma-derived cell line 143B(TK Ϫ ) containing mtDNA ( ϩ ) and its mtDNA-less derivative, 143B/206 0 , were cultured in Dulbecco's modified Eagle's medium-high glucose supplemented with 10% fetal bovine serum, 1 mM sodium pyruvate, and 50 g/ml uridine as described by King and Attardi (23). Jurkat cells (ATCC TIB-152, a lymphoblastoid T-cell line) were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum and penicillinstreptomycin. Chemicals were obtained from Sigma, with the exception of staurosporine (Roche Molecular Biochemicals).
Southern Blot-For determination of the ratio of mtDNA to nuclear DNA, 5 g of total DNAs were digested with PvuII. After electrophoresis and transfer, the Zeta-Probe membrane was hybridized with two probes simultaneously. One probe was a [␣-32 P]dCTP-labeled PCR fragment corresponding to mtDNA positions 8289 -10256. The second probe was a [␣-32 P]dCTP random prime-labeled 5.8-kb EcoRI insert from a plasmid containing the nucleus-encoded 18 S ribosomal RNA gene (24).
Bcl-x L Transfection-The full-length human Bcl-x L gene (25) cloned into plasmid pCDNA-3 (a gift from Dr. Lawrence Boise, University of Miami, Miami, FL) was subcloned into pIRES1neo vector (CLON-TECH). In this vector, the antibiotic neomycin exerts selective pressure on the whole expression cassette, and this allows the selection of colonies that express high levels of the gene of interest. Cells (143B and 206 0 ) were plated at ϳ70% confluence in 6-well tissue culture plates (Costar, Cambridge, MA) and transfected using the cationic lipid Lipo-fectAMINE Plus TM (Life Technologies, Inc.) as described by the manufacturer, using 1.5 g of plasmid DNA (pIRES1neo or pIRES1neo/Bclx L ), 4 l of LipofectAMINE reagent, and 6 l of PLUS reagent. After 48 h, the cells were selected in medium containing 1 mg/ml Geneticin TM (G418; Life Technologies, Inc.). Stable transfectants were isolated after 12 days under selection.
Cell Death Studies-The apoptotic death was assessed by two independent methods: 1) by using annexin V staining (cells were treated as indicated under "Results," labeled with annexin V-fluorescein isothiocyanate (PharMingen, San Diego, CA) and propidium iodide (Sigma), and analyzed in a FACScan flow cytometer (Becton Dickinson, Franklin Lakes, NJ)) and 2) by using a cell death detection enzyme-linked immunosorbent assay kit (Roche Biochemicals) based on the detection of nucleosome breakdown. The cell lines 143B and 206 0 (pIRES1neo or pIRES1neo/Bcl-x L ) were grown to ϳ90% confluence in 25-mm culture dishes. Controls or cells treated with 0.5 M staurosporine (STP) for 6 h in Dulbecco's modified Eagle's medium containing 1 mg/ml G418 at 37°C were washed, trypsinized, and collected in PBS. The protein concentration was adjusted to ϳ1.5 g/l for each cell type, and the nucleosomes released due to DNA fragmentation during apoptosis were measured using the photometric enzyme immunoassay of cytoplasmic histone-associated DNA fragments.
To study the role of the MPTP in STP-induced cell death, cells were pretreated for 30 min with 5 M of the mitochondrial permeability transition inhibitor cyclosporin A (CsA) and incubated for 6 h with 0.5 M STP. The assay was also performed using 5 M CsA and 50 M of the phospholipase A 2 inhibitor aristolochic acid (Sigma), which enhances and prolongs the effect of CsA (26,27).
Isolation of Mitochondria-Mitochondrial and cytosolic fractions were obtained by disrupting the cells either by N 2 cavitation (28) or by Potter-Elvehjem homogenization followed by differential centrifugation as described previously (29). Postmitochondrial supernatants were centrifuged at 100,000 ϫ g (S-100) and used as a soluble cytosolic fraction.
Cell Respiration-Oxygen consumption in all of the cell lines was measured polarographically as described previously (30), with a few modifications. Cells were resuspended in Dulbecco's modified Eagle's medium containing no glucose and supplemented with 5% dialyzed serum. The respiration rates were determined as the rate of oxygen consumption per milligram of protein.
Immunoprecipitation-Cells were lysed in radioimmune precipitation buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, Complete TM Protease Inhibitor Mixture (Roche Molecular Biochemicals), and 1.0% Triton X-100) for 1 h at 4°C and centrifuged to remove insoluble material. Bcl-x L and associated proteins were immunoprecipitated with anti-Bcl-x L antibody (1:1000; a gift from Dr. Lawrence Boise) overnight at 4°C with constant agitation. This was followed by the addition of anti-mouse IgG-agarose beads (Sigma) and further incubation for 2 h. The immunoprecipitated proteins were incubated with 0.5 mg/ml activated biotin (EZ-Link TM Sulfo-NHS-LC-Biotin; Pierce) for 30 min at room temperature. The biotinylated proteins were solubilized in SDS-sample buffer, separated on a 15% polyacrylamide gel, and transferred onto polyvinylidene difluoride membrane. The membrane was either incubated with avidin-HRP and developed by a chemiluminescence system (Pierce) or used for immunoblotting experiments.
Immunoblotting-For analysis of the overexpression of Bcl-x L in the transfected clones, cells were grown in 25-cm 2 culture flasks, collected by trypsinization, and lysed in 1% SDS. About 10 g of protein were resolved by SDS-polyacrylamide gel electrophoresis (15% gels) and transferred onto polyvinylidene difluoride membrane (NEN Research Products, Boston, MA). Blots were blocked with 5% milk and probed with anti-Bcl-x L monoclonal antibody (1:2500 dilution) followed by a secondary anti-mouse antibody conjugated with HRP. The detection was performed using a Phototope-HRP Western Blot Detection Kit (New England Biolabs, Inc., Beverly, MA).
For the analysis of the distribution of cytochrome c, equivalent amounts of mitochondrial and cytosolic proteins (S-100) were used (20 g). The blots were probed with three different monoclonal antibodies: anti-cytochrome c (1.5 g/ml; PharMingen), anti-cytochrome c oxidase subunit IV (COX IV) antibody (1:500 dilution; a gift from Dr. Armand Miranda, Columbia University, New York, NY), and anti-superoxide dismutase (Cu/Zn; SOD-1; 1:100 dilution; PharMingen). After scanning the autoradiograms, band signals were quantified using the NIH Image 1.62 software package.
Immunocytochemical Staining-143B and 143B/206 0 cells (controls and Bcl-x L -transfected cells) were seeded on glass coverslips. Cells were fixed with 4% paraformaldehyde in PBS, permeabilized with methanol, and incubated for 1 h with monoclonal antibodies to either Bcl-x L or cytochrome c. A fluorescein isothiocyanate-conjugated AfiniPure donkey anti-mouse IgG (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) was used as a secondary antibody. For co-localization studies, MitoTracker™ (CMX-Ros; Molecular Probes, Eugene, OR) was applied to the cells (200 nM) 15 min before PBS wash and fixation. Coverslips were mounted onto glass slides with the ProLong TM Antifade kit (Molecular Probes). Fluorescence was inspected with a Wild Leitz photomicroscope (Fluovert, Heerbrugg, Switzerland).
Mitochondrial Membrane Potential (⌬ m )-⌬⌿m was estimated using 5,5Ј,6,6Ј-tetrachloro-1,1Ј,3,3Ј-tetraethylbenzimidazole carbocyanide iodide (JC-1; Molecular Probes). JC-1 is a fluorescent compound (excitation maximum, 490 nm) that exists as a monomer at low concentrations. At higher concentrations, JC-1 forms aggregates. Fluorescence of the monomer is green (emission, 527 nm), whereas that of the J aggregate is red (emission, 590 nm). Mitochondria with intact membrane potential (higher than 100 mV) concentrate JC-1 into aggregates that fluoresce red, whereas de-energized mitochondria cannot concentrate JC-1 and fluoresce green (31). Cells were grown in 25-cm 2 dishes until 90% confluence and either kept in the medium alone (control) or treated with 0.5 M STP for 6 h. Cells were washed with PBS and incubated at 37°C for 45 min with 6.5 M JC-1 in media in the dark. Cells were collected by trypsinization, washed in PBS, and resuspended in 500 l of PBS, and the samples were measured on a MPF-66 fluorescence spectrophotometer (Perkin-Elmer). The ratio of the reading at 590 nm to the reading at 527 nm (590:527 ratio) was considered as a relative ⌬⌿m value. Cells either untreated or treated with 0.5 M STP were also stained with 6.5 M (or 2 M) JC-1 for 40 min at 37°C. The aggregate (red) fluorescence was analyzed by FACS with the settings as described by Cossarizza et al. (32).
Caspase 3 Activity-The caspase activity was measured using ApoAlert TM CPP32/Caspase 3 Fluorescent Assay Kit (CLONTECH) based on the fluorometric detection of amino-4-trifluoromethyl coumarin after proteolytic cleavage of the synthetic substrate DEVD-amino-4-trifluoromethyl coumarin. Briefly, apoptosis was induced as described above. Cells were trypsinized and counted. About 0.5 ϫ 10 6 cells were used for proteolytic cleavage of the synthetic substrate. Units of caspase 3 (nmol amino-4-trifluoromethyl coumarin/h) were determined from a standard curve, and the values were normalized to protein content.
Statistical Analysis-The data were analyzed using SPSS software. Results are expressed as mean Ϯ S.D. Comparisons between multiple groups were performed by analysis of variance. Values with p Ͻ 0.050 were considered statistically significant.

RESULTS
Although cells devoid of mtDNA can grow normally in culture provided that a glucose-rich medium supplemented with pyruvate and uridine is used, their low mitochondrial membrane potential (⌬⌿m) could make them more susceptible to undergoing apoptosis. Moreover, the function of factors that can modulate ⌬⌿m during apoptosis, such as Bcl-x L , could be affected in these cells. We investigated the role of functional OXPHOS and ⌬⌿m in these processes by studying the susceptibility of osteosarcoma cells with and without mtDNA to STPinduced apoptosis.
Characterization of Cell Lines with and without mtDNA overexpressing Bcl-x L -The cell lines 143B and 143B/206 0 (hereafter referred to as 206 0 ) were transfected with a plasmid construct expressing a bicistronic RNA coding for Bcl-x L in the 5Ј segment and neo resistance in the 3Ј segment. The neo resistance can be expressed because of an internal ribosomal entry site between the two coding regions. This vector system (pIRES; CLONTECH) allows the expression of high levels of the cloned gene in cells selected by resistance to high concentrations of G418. Fig. 1 shows that individual clones selected in the presence of 1 mg/ml G418 expressed high levels of Bcl-x L (clones 143B/Bx-n and 206/Bx-n). Bcl-x L could not be detected in untransfected cells, suggesting that, if present, its expression is very low in these osteosarcoma cells. The expression pattern suggested that in both cell lines a significant fraction of the overexpressed Bcl-x L is localized to mitochondria because the Bcl-x L signal co-localizes with the mitochondrial marker MitoTracker™ (Fig. 1C). The same cell lines were transfected with a pIRES vector without Bcl-x L and selected under the same G418 conditions to be used as controls in the subsequent experiments.
Mitochondrial Membrane Potential (⌬⌿m), Cell Respiration, and Apoptotic Markers in Osteosarcoma Cells-Although cells with an impaired OXPHOS are expected to have a lower ⌬⌿m, it has been reported that some 0 cells are able to maintain a normal ⌬⌿m (33,34). We used the dye JC-1 to monitor ⌬⌿m, estimated as the 590:527 nm emission ratio (see "Experimental Procedures"). 206 0 and its Bcl-x L -overexpressing derivatives showed a relatively low ⌬⌿m, approximately 30% of the values obtained for the parental 143B cells. The overexpression of Bcl-x L increased the basal ⌬⌿m in 143B cells by approximately 35% (p ϭ 0.002 by analysis of variance) but not in 206 0 cells ( Fig. 2A). This increase in membrane potential was associated with a similar increase in cell respiration (p Ͻ 0.0001 by analysis of variance), showing that Bcl-x L can stimulate respiration even under nonapoptotic conditions (Fig. 2B). As expected, Bcl-x L did not increase the ⌬⌿m or respiration of 206 0 cells (Fig. 2, A and B). As described below, treating these cells for 6 h with the protein kinase inhibitor STP triggered apoptosis. Staurosporine treatment led to an increase in ⌬⌿m not only in 143B cells but also in 206 0 cells. This relative increase (approximately 50%) was less pronounced in 143B cell lines overexpressing Bcl-x L , mainly because of the higher basal ⌬⌿m observed in these cells. We have also analyzed changes in ⌬⌿m by FACS (using an emission/excitation filter for JC-1 aggregates; see "Experimental Procedures") and found that STP induced a similar increase in ⌬⌿m. Identical results were obtained using 2 or 6.5 M JC-1. The increase was in the mean of the cell population and not in a subset of cells (data not shown).
Apoptosis was monitored in these cells by analyzing phosphatidylserine externalization (annexin V staining), propidium iodide uptake, and nucleosome breakdown. Preliminary results showed that cells treated with STP for 6 and 18 h underwent extensive nucleosome breakdown. Therefore, we used a 6-h STP treatment for all experiments described in this study. Annexin V staining and PI uptake were analyzed by FACS. Surprisingly, after 6 h of STP treatment, 143B cells showed higher annexin V staining and a higher PI uptake than the 206 0 derivative. Annexin V fluorescence in 206 0 cells was 57%, and PI uptake was 77% of the values observed in the parental 143B cells (Fig. 3A). The relative extent of nucleosome breakdown in these cells is illustrated in Fig. 3B. In agreement with the FACS results, 206 0 cells showed approximately 40% less nucleosome breakage than the parental 143B cells. Basal levels (i.e. without STP treatment) of nucleosome breakdown in the 206 0 and 143B cell lines were 2.6 Ϯ 0.1 and 4.0 Ϯ 0.2 OD/mg protein, respectively, and these values were increased as described in Fig. 3B after STP treatment to 13.2 Ϯ 0.5 and 41.2 Ϯ 6.5 OD/mg protein, respectively. Overexpression of Bcl-x L conferred strong protection against STP-induced apoptosis to both ϩ and 0 cells. On the other hand, CsA, an inhibitor of the MPTP, did not protect either 143B or 206 0 cells against STP-induced apoptosis. However, the protective effect of CsA could be observed in STP-treated Jurkat cells (Fig.  3B), suggesting that although mitochondrial permeability transition participates in the apoptotic cascade in some cell types, it does not seem to play such a role in osteosarcoma cells. The combination of aristolochic acid (an inhibitor of phospholipase A and a potentiator of CsA effect on mitochondrial permeability transition) and CsA was also ineffective in protecting 143B and 206 0 cells from apoptotic death (data not shown).
The lower levels of apoptosis observed in 206 0 cells were compatible with the levels of caspase 3 activation in these cells (Fig. 3D). Caspase 3 activation in 206 0 cells was less than half the activation observed in 143B cells after 6 h of treatment with STP. As expected, Bcl-x L strongly inhibited caspase 3 activation, thereby inhibiting apoptosis (Fig. 3D).
Cytochrome c Release Does Not Play a Major Role in the STP-induced Apoptotic Cascade of Osteosarcoma Cells with or without mtDNA-Cytochrome c release from the mitochondria has been shown to activate caspases and apoptosis in several cellular systems (11,12). We analyzed the cytochrome c release from osteosarcoma cells after 6 h of STP treatment. The redistribution of cellular factors was investigated by disrupting the cells either by homogenization or nitrogen cavitation (see "Experimental Procedures") and differential centrifugation. Mitochondrial and cytosolic (S-100) enriched fractions were analyzed by immunoblots. Disruption by Potter-Elvehjem homogenization resulted in a significant detection of cytochrome c in the S-100 fraction, even in the sample that was not treated with STP (data not shown). Therefore, we proceeded by isolating the cellular fractions by nitrogen cavitation, which has been reported to be less damaging to the outer mitochondrial membrane (28). Fig. 4 showed that treatment with STP did not trigger a significant cytochrome c release from 143B mitochondria. A slightly higher signal for cytochrome c was observed in the S-100 fraction of the 206 0 cells treated with STP. Notably, the mtDNA-less 206 0 cells showed significantly higher levels of cytochrome c in their mitochondria. For comparison, we included a nonspecific band recognized by the anti-cytochrome c antibody (the band labeled nsp) in Fig. 4A. Normalized to this unknown protein, mitochondria of our 206 0 cell line had approximately 3.5-fold higher levels of cytochrome c than did 143B cells. This increased amount of cytochrome c in our 206 0 cell line was also observed when total cell proteins were analyzed (data not shown). The enrichment of the cellular fractions was confirmed by using antibodies against a mitochondrial marker (COX IV) and a cytosolic marker (SOD-1). Steady-state levels of COX IV have been previously shown to be decreased in 0 cells (35), consistent with our findings.
The localization of cytochrome c after 6 h of STP treatment was also analyzed by light microscopy. Cells were double stained with a monoclonal anti-cytochrome c antibody and the mitochondrial marker MitoTracker™ (Fig. 4, B-K). Fig. 4, B and C, shows the mitochondrial pattern of the cytochrome c FIG. 3. Apoptosis is impaired in 0 cells. A, cell lines stably transfected with pIRES or pIRES/Bcl-x L were treated with staurosporine (0.5 g/ml, 6 h), stained with annexin V and PI, and analyzed by FACS. B, cells were treated as described above, and nucleosome breakdown was measured as described under "Experimental Procedures." The graph illustrates the increase in nucleosome breakdown relative to untreated cells. A and B show that 206 0 cells have impaired apoptosis. As expected, both ϩ and 0 cells were protected by the overexpression of Bcl-x L . C shows that osteosarcoma cells undergoing STP-induced apoptosis are not protected by CsA, suggesting that the MPTP does not play a major role in this system. Nuclear apoptosis of Jurkat cells, on the other hand, was strongly inhibited by CsA. D, caspase 3 activity was determined as described under "Experimental Procedures." Activation of caspase 3 (U ϭ nmol amino-4-trifluoromethyl coumarin/ h/mg protein) was also impaired in 206 0 cells. Error bars, the S.D. of the mean of three independent determinations. staining in 143B and 206 0 cells. STP treatment in both cell lines led to profound changes in morphology, but the cytochrome c signal (green staining) still co-localized with the mitochondria (red staining) in both cell lines. Fig. 4, D and E, exemplifies unusual cells in which the co-localization was not striking. Even in these cells, most cytochrome c co-localized with the mitochondria, but a slightly more diffuse cytochrome c staining was observed in the cell found in the center of Fig. 4, D and E. However, these same panels illustrate how the mitochondrial network (MitoTracker™ staining) may also show a diffuse pattern of staining after apoptosis (cells found in the lower part of the Fig. 4, D and E).
Bcl-x L Interacts with Caspase 9 in Osteosarcoma Cells-We also examined the pattern of Bcl-x L interactions with other cellular factors by co-immunoprecipitation of cellular extracts with an anti-Bcl-x L antibody. Immunoprecipitated proteins were labeled with biotin, separated by SDS-polyacrylamide gel electrophoresis, and analyzed with avidin-HRP by chemiluminescence as described under "Experimental Procedures." The pattern of proteins co-immunoprecipitated with Bcl-x L was very similar in extracts from 143B/Bx1 and 206/Bx6 (Fig. 5). One of the co-immunoprecipitated proteins had a molecular mass of 47,000 Da and could represent caspase 9. The identity of this band was confirmed by immunoblotting the same membrane with an anti-caspase 9 antibody (Fig. 5). DISCUSSION The Role of ⌬⌿m in Apoptosis-The sequence of events after an apoptotic stimulus has been the subject of much controversy, but the emerging picture suggests that it varies depending on the stimuli and cell system. Using mostly lymphoblastoid cell lines, a large number of studies have shown that the collapse of the mitochondrial membrane potential and the redistribution of cytochrome c are early steps in the apoptotic Bcl-x L was immunoprecipitated. The immunoprecipitated proteins were biotinylated, resolved by 15% SDS-polyacrylamide gel electrophoresis, and developed with avidin-HRP using a chemiluminescence system. The same membrane was stripped and subsequently incubated with anti-caspase 9 and anti-Bcl-x L antibodies. The left panel shows that among several proteins co-immunoprecipitated with Bcl-x L , there was a M r 47,000 band. This band was also detected by the anti-caspase 9 antibody (right panel). The presence of Bcl-x L was also demonstrated by immunoblotting. ment of the subcellular fractions was evident by the use of a mitochondrial marker (COX IV) and a cytosolic marker (SOD-1). B-K, co-localization studies in 143B (B and D--G) and 206 0 cell lines (C and H-K) cells before (B and C) and after STP treatment (D--K). Cells were grown on coverslips and stained with MitoTracker™ and with an anti-Bcl-x L monoclonal antibody as described under "Experimental Procedures." cascade and are, in many cases, necessary for the activation of cytosolic caspases (11,12,36). Mitochondrial permeability transition has been implicated as the cause of ⌬⌿m collapse and ultimately cytochrome c release (37)(38)(39), either by inducing mitochondrial swelling and rupture of the outer membrane (40) or by creating pores large enough for the extrusion of folded polypeptides (41). Other studies reported an increase in ⌬⌿m after a death stimulus, with ⌬⌿m decreasing later in the death process (5,42). We analyzed ⌬⌿m in osteosarcoma cells with and without mtDNA that overexpressed or did not overexpress Bcl-x L before and after a STP death stimulus. JC-1 was chosen because it has been described as a reliable probe for analyzing ⌬⌿m changes in intact cells, whereas other probes capable of accumulating in the mitochondria show a lower sensitivity (e.g., rhodamine 123) or noncoherent behavior due to a high sensitivity to changes in the plasma membrane potential (e.g., 3-3Ј-dihexilocarbocyanide iodide; Ref. 43). Changes in the plasma membrane potential do not seem to affect the JC-1 status (43). Using osteosarcoma cell lines with (143B ϩ ) and without mtDNA (206 0 ), with the latter having a markedly reduced ⌬⌿m, we observed a clear increase (double) in the mitochondrial membrane potential after 6 h of STP treatment ( Fig. 2A). Interestingly, the increase in ⌬⌿m after STP treatment was proportionally the same for 143B and 206 0 cells, arguing against the involvement of OXPHOS function in the and still found a similar increase in ⌬⌿m. We do not have an explanation for these differences, but it is possible that because the 0 cell lines were obtained independently (by ethidium bromide treatment), they may have a critical difference in a pivotal nuclear gene involved in apoptosis. Because Jiang et al. (44) did not study the parental cell line, we do not know whether their 0 cell behaved abnormally.
There have been some conflicting reports on the influence of a functional OXPHOS in apoptosis. Marchetti et al. (33) studied human leukemia cells U937 ϩ and 0 treated with tumor necrosis factor ␣ plus cyclohexamide. However, their 0 cells had a normal ⌬⌿m, despite the lack of OXPHOS. The authors found that the treatment led to a decrease in ⌬⌿m for both the ϩ and 0 cell lines. They also found that the 0 variant underwent apoptosis at rates that were similar to those of the parental ϩ cells. These observations contrast with the cells studied here. However, our 0 cells had ⌬⌿m that was only 30% of the parental ϩ , and this may explain why they were more resistant to apoptosis. Resistance to tumor necrosis factor ␣-induced apoptosis was observed in other 0 cell lines, and it was linked to an inhibition in nuclear factor B transcription (44,45). Other studies reported that OXPHOS inhibitors trigger apoptosis (46). However, it is important to maintain a distinction between experiments performed with OXPHOS inhibitors and those performed with genetically determined 0 cells. Not only can OXPHOS poisons lead to variable increases in reactive oxygen species production, but they also have potent deathinducing effects that are unrelated to OXPHOS inhibition (47).
Vander Heiden et al. (5) reported an increase in ⌬⌿m after a 6-h treatment of Jurkat cells with STP. They attributed this increase to swollen mitochondria maintaining an intact ⌬⌿m. Similar observations have been reported recently by Samali et al. (48) and in p53-induced apoptosis by Pei-Feng et al. (49). The STP-induced increase in ⌬⌿m in our system cannot be caused by increases in proton pumping by OXPHOS complexes in the 0 cells and is more likely due to changes in ionized molecules inside the mitochondria (50,51). Although we do not have an explanation for the increase in ⌬⌿m after STP treatment, we could speculate that STP, by inhibiting kinase reactions, is increasing the ATP concentration in the cytosol. These increased levels could reverse the ADP/ATP exchange by the adenine nucleotide translocator (i.e. moving ATP inside the mitochondria), leading to an increase in the net negative charge in the mitochondrial matrix. A similar explanation has been proposed as the mechanism responsible for maintaining a basal ⌬⌿m in 0 mitochondria (34).
We were not able to observe a major redistribution of cytochrome c after STP treatment. Some cytochrome c was found to redistribute to the cytosolic fraction of STP-treated 206 0 cells, but it is possible that this is due to the fact that 206 0 mitochondria are more fragile than 143B mitochondria and that cytochrome c is dissociated from the respiratory complexes (Fig. 4A). This dissociation would make cytochrome c more prone to "leak" into the cytosolic fraction during purification. Adachi et al. (28,52) reported that cytochrome c was found to be dissociated from respiratory complexes after an apoptotic stimulus, with possible implications on cell death. Cytochrome c from 206 0 cells is not only completely dissociated from respiratory complexes, but it is also present at higher than normal levels. We do not have an explanation for this increase, but it can be due either to a feedback mechanism designed to compensate for the lack of OXPHOS or to the aneuploid nature of these cells (53) that may lead to increases in gene dosage. Although we cannot discard the possibility that low levels of cytochrome c redistribution may amplify apoptosis in our system, we do not believe this redistribution is significant because STP-induced apoptosis was less prominent in 0 cells, which showed (possibly) higher levels of redistribution. The insensitivity to CsA protection observed in our cells is compatible with the hypothesis that mitochondrial swelling and cytochrome c release are not determining events in STP-induced death of osteosarcoma cells with or without functional OXPHOS. It is possible that the lower susceptibility to apoptosis is related to a limited availability of ATP in 0 cells. Although these cells seem to produce adequate amounts of ATP for growth in glucose-rich media, the cytosolic ATP pool may be reduced because of the internalization of ATP to maintain a minimum ⌬⌿m required for life (34). Many steps involved in the apoptotic cascade require cytosolic and nuclear ATP (54) and may be impaired in 0 cells undergoing a death stimulus. It has been shown that if ATP levels drop below a certain threshold, nuclear apoptosis is blocked (55,56). This would be compatible with our results, which showed reduced nucleosome breakdown in STP-treated 206 0 cells. Moreover, we also observed a decrease in other apoptotic markers in STP-treated 206 0 cells (i.e. annexin V staining and PI uptake).
Mechanisms of Bcl-x L Protection in Osteosarcoma Cells-We also found that Bcl-x L increases the basal ⌬⌿m in 143B cells. However, this effect was not observed in 206 0 cells ( Fig. 2A). This specific effect of Bcl-x L on 143B cells suggests that in contrast to STP, this increase may be associated with changes in OXPHOS. This hypothesis was corroborated by the observation that cell respiration was increased proportionally in Bclx L -transfected 143B cells (Fig. 2B). Bcl-x L has not been previously shown to enhance respiration under normal conditions. The protective effect of Bcl-x L was independent of its increase in the basal ⌬⌿m because 206 0 cells overexpressing Bcl-x L were also protected against STP-induced death. Although Vander Heiden et al. (5) also described an increase in ⌬⌿m associated with STP treatment, their results differed from ours in that they found that overexpression of Bcl-x L led to a decrease in ⌬⌿m after a 6-h STP treatment of Jurkat cells. We found that the increase in ⌬⌿m was still observed after STP treatment of Bcl-x L overexpressors, although this relative increase was proportionally less marked than in the parental 143B cells mainly because of the higher basal ⌬⌿m observed in the latter cell line. However, in 206 0 cells, the ⌬⌿m increase observed after STP treatment did not differ between the parental cell and the Bcl-x L overexpressors. Recently, Vander Heiden et al. (42) showed that the pro-B cell FL5.12 undergoing interleukin 3 withdrawal displayed increased ⌬⌿m due to the failure of the voltage-dependent anion channel/adenine nucleotide translocator complex to maintain ATP/ADP exchange. The authors suggested that Bcl-x L would protect cells by facilitating ATP/ADP exchange, thereby maintaining a lower ⌬⌿m. This would be protective because it would counteract the apoptosis-associated increase in ⌬⌿m, which, in their model, is caused by an inhibition of oxidative phosphorylation. As mentioned above, we could not observe a decrease in ⌬⌿m associated with Bcl-x L overexpression in our cells, but we did detect an enhanced respiration in cells overexpressing Bcl-x L that would be compatible with stimulation in ATP/ADP exchange. Our results do not support the concept that hyperpolarization induced by a death-promoting agent is caused by a block in ATP/ADP exchange concomitant with continuing proton pumping (42) because the STP-induced ⌬⌿m increases observed in the 0 cell would not fit this model. On the other hand, our results support their model in which Bcl-x L regulates mitochondrial respiration, possibly by stimulating ATP/ADP exchange. The main protective effect of Bcl-x L in our cells, however, does not seem to be related to the maintenance of low ⌬⌿m because 206 0 cells were protected against apoptosis even though their ⌬⌿m was still increased by STP treatment.
How is Bcl-x L protecting osteosarcoma cells? As described above, our results suggest that the protective effect of Bcl-x L in these cells is only partially associated with modulation of OX-PHOS function, but not with absolute ⌬⌿m values or cytochrome c release as reported in other systems (57). These differences reinforce the view that the participation of these factors in apoptosis is complex and variable (58). Although protection against ⌬⌿m collapse has been shown to be a major component of Bcl-x L function in some systems (59,60), it is likely that the protective effect observed in our cells is related mostly to direct inhibition of caspase activation as reported previously (61,62). This inhibition may be associated with direct interaction of caspases with Bcl-x L (63). Bcl-2 has been shown to act downstream of cytochrome c release by interaction with either caspases or Apaf-1 (64). Brustugun et al. (65) showed that apoptosis induced by cytochrome c injection was inhibited by Bcl-2, indicating that a significant level of protection by antiapoptotic members of the Bcl-2 family does not occur by inhibiting cytochrome c release. Bcl-2 and Bcl-x L are known to prevent apoptosis by heterodimerizing with proapoptotic proteins such as Bax and Bad (2, 66) that may participate in the opening of mitochondrial pores (27,67). However, Bcl-x L has also been shown to protect cells by a BAX-independent mechanism (68). Bcl-x L has been recently shown to interact with caspase 9 (69), which can be found in the mitochondria (9,48). We also found that Bcl-x L interacts with caspase 9 in our system. Although our experiments did not address this mechanism in detail, they suggest that interactions between Bcl-x L and caspase 9 may play a role in the protection observed in our cells.
Perhaps the most import implications of our findings relate to the role of mtDNA mutations recently identified in colorectal cancers (70). Some of these mutations were predicted to be null mutations, completely impairing OXPHOS. It is conceivable that these cells may behave as 0 cells, being more resistant to apoptosis and therefore more aggressive and untreatable forms of cancer. Further work using cells with genetically defined OXPHOS defects should help verify this hypothesis.