UV-induced apoptosis is mediated independent of caspase-9 in MCF-7 cells: a model for cytochrome c resistance.

The importance of the mitochondria in UV-induced apoptosis has become increasingly apparent. Following DNA damage cytochrome c and other pro-apoptotic factors are released from the mitochondria, allowing for formation of the apoptosome and subsequent cleavage and activation of caspase-9. Active caspase-9 then activates downstream caspases-3 and/or -7, which in turn cleave poly(ADP)-ribose polymerase (PARP) and other down-stream targets, resulting in apoptosis. In an effort to understand the mechanisms of Akt-mediated cell survival in breast cancer, we studied the effects of insulin-like growth factor (IGF)-I treatment on UV-treated MCF-7 human breast cancer cells. Apoptosis was induced in MCF-7 cells after UV treatment, as measured by caspase-7 and PARP cleavage, and IGF-I co-treatment protected against this response. Surprisingly caspase-9 cleavage was unchanged with UV and/or IGF-I treatment. Using MCF-7 cells overexpressing caspase-3 we have shown that resistance of caspase-9 to cleavage was not altered by the expression of caspase-3. Furthermore, overexpression of caspase-9 did not enhance PARP or caspase-7 cleavage after UV treatment. Because caspase-8 was activated with UV treatment alone, we believe that UV-induced apoptosis in MCF-7 cells occurs independently of cytochrome c and caspase-9, supporting the existence of a cytoplasmic inhibitor of cytochrome c in MCF-7 cells. We anticipate that such inhibitors may be overexpressed in cancer cells, allowing for treatment resistance.

Non-cytokine-mediated cellular stress, such as UV or chemical treatment, can initiate apoptosis through mitochondrial release of cytochrome c (8). In the caspase-9 initiated pathway, cytoplasmic cytochrome c triggers the formation of the apoptosome, a multi-protein complex containing cytochrome c, dATP, Apaf-1, and pro-caspase-9 (9,10). The presence of caspase-3 and/or -7 within the apoptosome may allow more optimal cleavage of caspase-9. Ultimately, activated caspase-9 serves as the initiator caspase, which may further amplify an apoptotic signal by activating caspase-8 and -2 upstream of the mitochondria (11). Downstream responses to caspase-9 include cleavage of caspase-3 and/or -7 and eventually poly(ADP)-ribose polymerase (PARP) (12).
Multiple mechanisms of resistance to apoptosis have been recently identified. By phosphorylating caspase-9, Akt inhibits its proteolytic activity (13). Further, resistance to mitochondrial initiated events has also been reported. For example, Bcl-2 and Bcl-XL can inhibit cytochrome c translocation (14 -16). Bcl-XL may also participate in binding to the apoptosome to inhibit its activity (14,17,18). The apoptosome and effector caspase activity can be further regulated by members of the IAP (inhibitor of apoptosis proteins) family of proteins, which can directly bind to caspases via BIR (baculoviral IAP repeat) protein domains. BIR domains 1 and 2 of IAPs bind and inhibit caspases-3 and -7, whereas the BIR3 domain binds to and specifically inhibits caspase-9 (19,20). These inhibitory effects can be offset by the stress-induced release of the mitochondrial protein, Smac/Diablo, which competitively binds to IAPs, relieving their inhibitory effects on caspases-3, -7, and -9 (21,22). Aberrations of Apaf-1 can also inhibit caspase-9 response. Methylation mediated transcriptional repression of Apaf-1 has been reported in metastatic melanomas and ovarian cancer cell lines (23,24), whereas reconstitution of Apaf-1 expression enhances apoptotic response to the chemotherapeutic drug doxorubicin (23).
Akt mediates survival by phosphorylating several substrates that are intimately involved in regulating programmed cell death. Akt phosphorylation of pro-caspase-9, a downstream target of p53, blocks cleavage of pro-caspase-9 and its subsequent activation (13,25). Thus, regulation of pro-caspase-9 is of particular importance in p53-mediated effects. Therefore, we initially set out to determine whether pro-caspase-9 may be an Akt target for survival of IGF-I sensitive breast cancer cells.
Since the discovery that MCF-7 human breast cancer cells do not express full-length caspase-3, because of a 47-base pair deletion within exon 3 of the caspase-3 gene, debate has existed regarding the ability of MCF-7 cells to undergo programmed cell death (26,27). More recently, the MCF-7 cell line has become a model for investigation of caspase-3-dependent and -independent effects. Although less studied, caspase-7 may convey many of the same effects as caspase-3, suggesting some redundancy between the two proteins. Here we show that MCF-7 cells undergo caspase-mediated apoptosis upon UV treatment, as indicated by caspase-7 and PARP cleavage. Most interesting is our observation that UV treatment can induce apoptosis via caspase-8 activation, independent of caspase-9. Additionally, in studying the potential mechanism for the lack of a caspase-9 response, we conclude that MCF-7 cells harbor a cytochrome c or apoptosome defect.

EXPERIMENTAL PROCEDURES
Cell Culture and Treatments-MCF-7 cells were provided by Dr. C. Kent Osborne (University of Texas Health Science Center, San Antonio, TX), and other MCF-7 cells were provided by Dr. Heide Ford (University of Colorado Health Sciences Center, Denver, CO); HEK 293 cells were provided by Dr. Douglas Wolf (University of Colorado Health Sciences Center, Denver, CO). Cell lines were maintained in full serum medium (MCF-7 in Iscove's modified Eagle's medium (Mediatech, Herndon, VA) supplemented with 10% fetal bovine serum (Mediatech), antibiotics, and insulin; HEK 293 in Dulbecco's modified Eagle's medium (Mediatech) was supplemented with 10% fetal bovine serum (Mediatech) and antibiotics). In each experiment the cells were plated in full serum-containing media and cultured overnight at 5% CO 2 and 37°C. The cells were washed twice the next day in warm phosphate buffered saline (PBS; Biofluids, Rockville, MD) followed by overnight culture in serum-free medium. On the third day, the cells were treated with UV (UV-C, 10 J/m 2 ) in a SpectroLinker UV linker 100 (Spectronics, Westbury, NY) with lids removed and/or IGF-1 (obtained from the National Hormone and Pituitary Program, NIDDK, National Institutes of Health and Dr. A. F. Parlow) or with TNF-␣ (Alexis Biochemicals, San Diego, CA) and cycloheximide (Sigma-Aldrich), as indicated in figure legends. When applicable, pretreatments with LY294002 (Alexis Biochemicals) or Z-IETD-FMK (BD Pharmingen, San Diego, CA) were performed 40 or 30 min prior to stimulation with UV, respectively.
Stable and Transient Transfections-pcDNA3 caspase-3 and FLAGcaspase-9 vectors were graciously provided by Dr. C. Vincenz (University of Michigan). Generation of MCF-7 caspase-3 stable transfectants was performed by electroporation of 3.0 ϫ 10 6 MCF-7 cells in 200 l of medium containing 10 g of DNA. Forty-eight hours later, the cells were selected with 800 g/ml of G418 (Mediatech). Individual drugresistant colonies were isolated and expanded. Detection of clones overexpressing caspase-3 was performed by Western blot analysis using a caspase-3 antibody (Santa Cruz Biotechnology, Santa Cruz, CA).
For transient transfections with wild-type caspase-9 or Bcl-2, 1.0 ϫ 10 6 MCF-7 cells were plated in 6-cm 2 dishes and cultured overnight in full serum medium. The cells were washed twice the following day with warm PBS and then incubated with serum free Opti-MEM (Invitrogen) for 45 min at 37°C and 5% CO 2 . The cells were then transfected with 30 l of Plus reagent, 4 l of LipofectAMINE (Invitrogen), and 4 g of DNA (3 g of DNA of interest and 1 g of green fluorescent protein); 3 h later plates were supplemented with full serum medium. Twenty-four hours after transfection the cells were assessed for transfection efficiency by visualization of green fluorescent protein and then treated as mentioned above.
Preparation of Cell Lysates and Western Blot Analyses-At the indicated times, the cells were scraped and harvested by centrifugation. The plates were washed once with cold PBS followed by an additional centrifugation. The cells were then lysed in lysis buffer (20 mM Tris-HCl, 250 mM NaCl, 3 mM EDTA, 0.05% Nonidet P-40, 1 mM dithiothreitol, 0.368 mg/ml Na Orthovanadate, 5 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, and 17 g/ml aprotinin) followed by centrifugation at 13,000 to remove cellular debris. The protein concentrations were determined using a Bio-Rad D/C protein assay kit. Sixty micrograms of total cell lysate were resolved by 10% SDS-PAGE, unless stated otherwise, and transferred to nitrocellulose. Western blot analyses were performed using primary antibodies for caspase-3, caspase-7, caspase-8, caspase-9, cleaved caspase-9 D315 (Cell Signaling, Beverly, MA), PARP (BD Pharmingen), XIAP (BD Pharmingen), or tubulin (Sigma), and enhanced chemiluminescence (Applied Biosystems, Foster City, CA). Each Western blot shown is representative of at least three separate experiments.
Antibody Cross-linking and Immunoprecipitations-Immunoprecipitations were performed by first cross-linking antibody to protein Gagarose beads (Invitrogen). The antibody for caspase-9 (Santa Cruz Biotechnology) was incubated with protein G-agarose beads in PBS for 3 h at 4°C. The beads were washed six times in 0.18 M sodium borate (pH 8.0) before incubation with 0.18 M sodium borate (pH 8.0) containing 100 mM DMP (dimethylpimelimidate) for 2 h at room temperature. The beads were then washed three times with 0.2 M ethanolamine and incubated with 0.2 M ethanolamine for 2 h at room temperature. The cross-linked beads were washed in cold PBS and stored at 4°C.
For immunoprecipitations, the cells were treated and harvested as described above. For each treatment, 200 g of total cell lysate were incubated with 10 l of cross-linked beads in a total volume of 300 l of lysis buffer. Incubation was carried out for 3 h at 4°C; beads were then washed three times with lysis buffer before being resolved by 10% SDS-PAGE. Western blots were carried out as described above.
Mitochondrial Inner Membrane Potential-Mitochondrial inner membrane potential was assessed by a mitochondrial voltage-sensitive dye JC-1 (Intergen, Purchase, NY) per manufacturer instructions and flow cytometry (FACSCalibur, Becton Dickinson, University of Colorado Cancer Center Flow Cytometry Core facility, which is supported by NCI, National Institutes of Health Cancer Core Support Grant CA46934). The fluorescent mitochondrial probe, JC-1, was used to verify treatment mediated changes in mitochondrial membrane ⌬. In apoptotic cells, the JC-1 dye remains monomeric in the cytoplasm showing green fluorescence. Whereas in unstressed cells, the mitochondrial aggregate forms fluorescent red at 590 nm. The experiments were repeated three times, and representative results are shown.
Cytochrome c Immunostaining-The cells were plated at a density of 85,000 cells/25-mm diameter slides (Lab-Tek, Nalge Nunc, Naperville, IL) and fixed in PBS containing 2% paraformaldehyde. The cells were washed in PBS, permeabilized in 0.2% Triton X-100, and then blocked in 10% normal goat serum/PBS. The cells were incubated with cytochrome c antibody (clone 6H2.B4 (Pharmingen, San Diego, CA)) and then washed and incubated with mouse IgG Alexa Fluor 488 AB (Molecular Probes, Eugene, OR). The slides were imaged on a Nikon Diaphot TE200 microscope using a CoolSNAP-fx monochrome digital camera and Image-Pro Plus V4.1 software. The images are shown with color overlay. The experiments were repeated three times, and representative results are shown.
In Vitro Cleavage of Pro-caspase-9 -A plasmid containing cDNA encoding pro-caspase-9 (pcDNA3-FLAG-tagged caspase-9) was in vitro transcribed and translated in the presence of [ 35 S]methionine (Amersham Biosciences) using a coupled transcription/translation TNT kit (Promega, Madison, WI) according to the manufacturer's instructions. The protein was desalted and exchanged into buffer A (20 mM Hepes, pH 7.5, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EDTA, and 1 mM dithiothreitol) with Bio-Spin P-6 columns (Bio-Rad). The cell lysates were prepared as described above but were lysed with buffer A. The radiolabeled reactions consisted of 80 g of cell lysate, 5 l of 35 S-labeled pro-caspase-9, 1.5 mM dATP (Sigma-Aldrich), and 1.8 or 18.0 M horse heart cytochrome c (Sigma-Aldrich) in a 30-l total volume; the reactions for Western blot analysis were identical with the omission of 35 S-labeled pro-caspase-9. The reactions were allowed to proceed for 4 h at 30°C before being analyzed by SDS-PAGE and autoradiography (STORM 860, Molecular Dynamics, Amersham Biosciences) or Western blot (19). The experiments were repeated a minimum of two times, and representative results are shown.
Cytochrome c Sequencing-RNA was extracted from confluent 10cm 2 dishes of MCF-7 and HEK 293 cells using RNAwiz isolation reagent (Ambion, Austin, TX) per manufacturer instructions. 1 g of RNA was reverse transcribed with Moloney murine leukemia virus reverse transcriptase (Invitrogen) in 20 l of reaction mixture. Cytochrome c was amplified from the resulting cDNA (10 l) using platinum Pfx DNA polymerase (Invitrogen) and the primers 5Ј-gagtgttcgttgtgccagcg and 5Ј-gcccaacaaaatattctgtcagtc. cDNA was amplified in 35 cycles, consisting of denaturing for 15 s at 94°C, annealing for 30 s at 55°C, and primer extension for 60 s at 68°C. The PCR products were purified using Microcon PCR Centrifugal Filter devices (Ambion) per manufacturer instructions. The University of Colorado Cancer Center DNA Sequencing and Analysis Core Facility, which is supported by NCI, National Institutes of Health Cancer Core Support Grant CA46934, sequenced the DNA samples using an ABI Prism 3100 capillary automated sequencer (Applied Biosystems). Analyses of DNA sequences were done with Sequencher 3.1 (Gene Codes Corp., Ann Arbor, MI).

UV Induces Caspase-7 and PARP Cleavage in MCF-7 Cells-
Because UV irradiation frequently induces mitochondrial release of cytochrome c and cleavage of caspase-9, we exposed MCF-7 breast cancer cells to UV irradiation to induce apoptosis via PARP and caspase-7 cleavage and also to determine whether IGF-I co-treatment could inhibit UV-mediated cell death. We also pretreated cells with LY294002, which blocks the activity of phosphatidylinositol 3-kinase and downstream Akt. Measurement of PARP cleavage was used as a direct measure of apoptosis. Fig. 1 clearly shows that MCF-7 cells that are induced to undergo apoptosis via UV irradiation are protected by co-treatment with IGF-I, and this survival effect can be blocked by LY294002. Because MCF-7 cells are deficient in caspase-3, we also measured caspase-7 processing to determine whether it could function similarly to caspase-3 as an effector caspase that can cleave PARP (Fig. 1). Caspase-7 cleavage patterns were very similar to those of PARP, in that UV treatment resulted in caspase-7 cleavage, and IGF-I co-treatment inhibited its cleavage. These data support a previous study showing that caspase-7 catalytic activity can induce PARP cleavage (28). By assessing both caspase-7 and PARP cleavage, LY294002 not only reversed IGF-I-mediated survival, but it also enhanced cell death in the presence of UV irradiation, presumably by inhibiting survival signals mediated by phosphatidylinositol 3-kinase.
Caspase-9 Cleavage Is Unchanged with UV and/or IGF-I Treatment-UV irradiation induces apoptosis among various cell types in a mitochondrial-dependent manner (8). Following irradiation, cytochrome c is released from the mitochondria into the cytosol, where it then complexes with Apaf-1, dATP, and pro-caspase-9 to form the apoptosome. Once bound in the apoptosome, pro-caspase-9 is processed to its active form, and it then activates effector caspases-3 and -7 to complete the apoptotic process. We sought to determine whether pro-caspase-9 is processed as expected in our MCF-7 cells and whether IGF-I could inhibit pro-caspase-9 cleavage. Again, UV treatment of MCF-7 cells induced PARP cleavage (Fig. 2). Despite the UVinduced PARP cleavage, we observed no detectable changes in the amount of caspase-9 cleavage following UV treatment. Also intriguing was the amount of caspase-9 cleavage present in the absence of stress (Fig. 2, Serum lane, negative control). Under normal conditions, pro-caspase-9 is not processed unless a stress signal induces cells to undergo apoptosis. In MCF-7 cells, pro-caspase-9 is processed to a small extent even in unstressed cells. We confirmed that the cleavage product observed in MCF-7 cell lysates corresponded to the manufacturer's positive lysate control and not the negative lysate control, ruling out the possibility that we were only detecting a nonspecific band (data not shown). Caspase-9 cleavage was also unaltered by IGF-I treatment, even though the cells were protected from apoptosis (compare PARP cleavage minus and plus IGF-I treatment).
Caspase-3 Expression Does Not Sensitize Caspase-9 to UV Treatment in MCF-7 Cells-The presence of some caspase-9 cleavage in untreated cells and the lack of a caspase-9 treatment-mediated increase in cleavage both suggest either a defect in endogenous pro-caspase-9 or the cellular machinery that processes it. We reasoned that if endogenous pro-caspase-9 of MCF-7 cells is defective, then transfected wild-type procaspase-9 could rescue the defect in caspase processing and should further enhance cleavage of the effector caspase-7. Because there is evidence to support a requirement for caspase-3 in the apoptosome for proper function (29), we also decided to address the role of caspase-3 in our model. To distinguish between these two possibilities, we transfected MCF-7 cells with wild-type pro-caspase-9 and reassessed caspase-9 cleavage in the presence and absence of UV irradiation (Fig. 3). Further, because a lack of endogenous caspase-3 could lead to a defect in apoptosome function and subsequent caspase-9 processing, we assessed caspase-9 cleavage in MCF-7 cells stably expressing wild-type caspase-3. As shown in Fig. 3A, transiently transfected pro-caspase-9 is partially processed in MCF-7 cells; however, it does not lead to enhanced processing of effector caspase-7 after UV treatment. Fig. 3B illustrates that caspase-9 cleavage products increase with transfection even in the absence of UV-induced stress. Together, these studies indicate that endogenous pro-caspase-9 is processed in a similar fashion as transfected wild-type protein, ruling out a defect in endogenous caspase-9 in MCF-7 cells and supporting the possibility that UV-mediated apoptosis in MCF-7 cells is caspase-9-insensitive. These data suggest an apoptosome defect in MCF-7 cells.
Other investigators have shown that overexpression of caspase-3 in MCF-7 cells enhances cleavage of pro-caspase-9 by using in vitro reactions with exogenous pro-caspase-9 and cytochrome c (27). Although the presence of caspase-3 is evident in the caspase-3 transfectants used in Fig. 3A (data not shown), processing of endogenous caspases-9 and -7 were unaffected by the presence of caspase-3 compared with parental MCF-7 cells. Again, these data indicate that the defect lies upstream of apoptosome formation, not within endogenous pro-caspase-9.
Stress Treatments Induce Cytochrome c Translocation in MCF-7 Cells-Because the initiator caspase-8 can also cleave Bid to tBid and tBid translocation can induce cytochrome c FIG. 1. UV induced PARP and caspase-7 cleavage can be inhibited by IGF-I effects on phosphatidylinositol 3-kinase. MCF-7 cells were plated and serum-starved as described under "Experimental Procedures." The following day cells were treated with UV irradiation (10 J/m 2 ) with or without IGF-I co-treatment (50 ng/ml). Where indicated, the cells were exposed to 50 M LY294002 for 40 min prior to UV irradiation and/or IGF-I treatment. The cells were lysed 6 h after UV and/or IGF-I treatment. The cell lysates were then subjected to Western blot analyses to detect parental or cleavage forms of PARP and caspase-7. The presence of the 85-kDa PARP and 20-kDa caspase-7 cleavage fragments indicate induction of apoptosis. SFM, serum-free medium.
FIG. 2. UV-induced apoptosis and IGF-I-mediated survival effects are caspase-9-independent. MCF-7 cells were plated and serum-starved as described under "Experimental Procedures." The following day cells were treated with UV irradiation (10 J/m 2 ) with or without IGF-I co-treatment (50 ng/ml). Six hours after treatment, the cells were harvested, and the lysates were subjected to Western blot analyses for PARP and pro-caspase-9 cleavage. SFM, serum-free medium.
release, we then tested whether there may be either an aberration in mitochondrial function, preventing proper release of cytochrome c, or a defect in the formation and/or function of the apoptosome. Using the JC-1 assay, we first sought to determine whether cellular stress results in mitochondrial membrane depolarization. Further, we wanted to determine whether cytochrome c is properly localized and released from mitochondria following UV irradiation. As a positive control for these assays, we used HEK 293 cells. Many investigators have confirmed that UV induces mitochondrial membrane changes and cytochrome c translocation in these cells. In the JC-1 assay, we used valinomycin treatment that generally induces mitochondrial membrane ⌬ changes in ϳ95-98% of either HEK 293 or MCF-7 cells (Fig. 4A). Although UV treatment also changed mitochondrial membrane ⌬, it was not as robust as with valinomycin treatment in either cell line. Next, we also examined cytochrome c cellular localization using cytochemistry. With this assay we observed that in untreated MCF-7 cells, cytochrome c was somewhat diffusely localized in the cytoplasm, whereas the pattern of staining was more characteristic of a mitochondrial, punctate staining in the HEK 293 cells (Fig.  4B). With UV or valinomycin exposure, cytochrome c becomes diffusely localized in the cytoplasm in both cell lines. These studies confirm that like HEK 293 cells, MCF-7 cells can respond to cellular stress by changing mitochondrial membrane ⌬ and inducing cytochrome c cytoplasmic translocation.
MCF-7 Cell Lysates Can Cleave Caspase-9 with Exogenous Cytochrome c-Given our evidence that cytochrome c release is stimulated by either valinomycin or UV in MCF-7 cells yet caspase-9 is unresponsive to treatment, we decided to determine whether the defect in pro-caspase-9 processing lies within the endogenous cytochrome c protein itself. Both MCF-7 and HEK 293 cell lysates were obtained from either untreated cells or cells exposed to valinomycin (to stimulate endogenous cytochrome c release) prior to harvesting. We then performed in vitro caspase-9 cleavage assays (Fig. 5A). Wild-type procaspase-9 was in vitro transcribed and translated in the presence of [ 35 S]methionine; this protein was then incubated with 80 g of MCF-7 or HEK 293 cell lysates described above. Exogenous cytochrome c was added to some reactions (as indicated) as a positive control to ensure the presence of functional, wild-type cytochrome c. The reactions were allowed to proceed for 4 h at 30°C before being subjected to SDS-PAGE and autoradiography. Untreated MCF-7 and HEK 293 cell lysates do not cleave 35 S-labeled pro-caspase-9 (Fig. 5A). When exogenous cytochrome c and dATP were added to the lysates, caspase-9 cleavage increased proportionate to the concentration of cytochrome c present in the reactions. Lysates from valinomycin-treated HEK 293 cells led to the presence of endogenous cytochrome c in the reaction and pro-caspase-9 cleavage, comparable with lysates treated with 18 M cytochrome c. In stark contrast, lysates from MCF-7 cells treated with valinomycin did not induce cleavage of caspase-9, even though endogenous cytochrome c levels were comparable with those seen in the HEK 293 valinomycin-treated cells. The observation that pro-caspase-9 can be cleaved only when exogenous cytochrome c is added to MCF-7 cell lysates suggests that endogenous Apaf-1 is functioning normally to form the apoptosome and that caspase-9 cleavage resistance results from an aberration in the endogenous cytochrome c of MCF-7 cells.
We then wanted to confirm our results indicating that endogenous cytochrome c cannot induce apoptosome formation and subsequent cleavage of pro-caspase-9. Further, we wanted to determine whether endogenous pro-caspase-9 cleavage could be enhanced when exogenous cytochrome c is added. To this end, the same reactions were performed as in Fig. 5A but in the absence of in vitro [ 35 S]methionine-labeled pro-caspase-9. In this instance, endogenous cleaved caspase-9 and cytochrome c in each reaction were analyzed by Western blot after performing the same reaction conditions as described for Fig. 5A. As illustrated in Fig. 5B, caspase-9 cleavage only occurred in MCF-7 lysates that contain exogenous cytochrome c. In contrast, HEK 293 cell lysates were capable of processing procaspase-9 when either endogenous or exogenous cytochrome c was present in the reactions, as confirmed by Western blotting with an antibody specific for cytochrome c (Fig. 5B). Because MCF-7 breast cancer cells may vary from source to source as a result of in vitro culturing conditions, we decided to obtain MCF-7 cells from an outside source to assess whether our observations can be generalized. Although a more punctate pattern of cytochrome c was observed by cytochemical analysis, we confirmed that the other MCF-7 cells showed an identical pattern of caspase-9 cleavage as our original MCF-7 cells (data not shown).
The ability of MCF-7 cells to process pro-caspase-9 only in the presence of exogenous cytochrome c indicates either a defect in the endogenous cytochrome c protein or the presence of an inhibitor specific to cytochrome c that can be overcome by the addition of excess protein. Thus, cytochrome c was sequenced using a reverse transcription-PCR-generated product from MCF-7 cells and compared with the human, wild-type cytochrome c. No mutations or truncations were observed in the MCF-7-derived cytochrome c (data not shown), confirming its wild-type sequence and further supporting the function of a cytoplasmic, cytochrome c inhibitor expressed in MCF-7 breast cancer cells.

FIG. 3. Overexpression of caspase-3 or -9 does not enhance UV-mediated apoptosis in MCF-7 cells.
Parental MCF-7 cells and caspase-3 stable transfectants were transiently transfected with increasing amounts of pro-caspase-9 as described under "Experimental Procedures." 24 h after transfection the cells were irradiated (10 J/m 2 ) and then harvested 6 h later. A, cell lysates were subjected to SDS-PAGE and Western blot analyses to determine the extent of procaspase-7 (p35) processing to a p20 fragment. The treatment-induced appearance of the p20 cleaved caspase-7 fragment indicates induction of apoptosis. B, pro-caspase-9 (p47) expression and cleavage to a p35 fragment was assessed with and without UV treatment. Western analysis of tubulin indicates protein loading in each lane.
Overexpression of Bcl-2 Does Not Inhibit UV-mediated Apoptosis-Our results thus far indicated that MCF-7 cells do not utilize the intrinsic pathway for UV-mediated apoptosis. We wanted to use other approaches to confirm these observations and also to assess what point in the pathway contributes to a lack of caspase-9 activity. Bcl-2 function is a critical inhibitor of the intrinsic pathway in two potential ways. First, Bcl-2 (and Bcl-XL) can homodimerize to inhibit apoptosis by maintaining mitochondrial membrane potential in the presence of an apoptotic stimuli (15,16). Second, Bcl-2 (and Bcl-XL) can bind within the apoptosome to inhibit its function (17,18,30). To begin these studies untransfected MCF-7 cells were treated with UV and compared with MCF-7 cells that were transfected with either empty vector or vector containing Bcl-2 and then later treated with TNF-␣ or UV. Again, caspase-7 cleavage was used as a measure of apoptosis. Fig. 6A shows that mock transfected cells were slightly more sensitive to apoptosis compared with untransfected control cells. Cells transfected with Bcl-2 experienced a modest degree of protection from both TNF-␣-and UV-mediated apoptosis when overexpressing Bcl-2 compared with the mock transfected cells. Bcl-2 and ␣-tubulin Western blot analyses confirmed Bcl-2 overexpression of transfectants and even loading of each sample, respectively. These data again indicate that the intrinsic pathway does not significantly contribute to UV-mediated apoptosis.
XIAP Binding to Caspase-9 Is Unaltered by UV Treatment-Recent attention has focused on XIAP binding and inhibition of caspase-9 activity as an important mechanism for caspase-9 resistance to apoptotic stimuli. Therefore, we decided to evaluate whether XIAP may be contributing to caspase-9 resistance. We were unsuccessful at using XIAP as an siRNA target to determine whether inhibition of XIAP would lead to caspase-9 sensitivity, thus we tested whether binding XIAP to  caspase-9 may be altered with either IGF-I and/or UV treatment. Fig. 6B illustrates that XIAP was co-immunoprecipitated with a caspase-9 antibody, and its binding did not significantly change with either IGF-I or UV treatment at various time points. Western blot analyses of the same samples using whole cell lysates showed that cleaved caspase-9 and XIAP expression were also unchanged. In conclusion, we believe that the intrinsic pathway is inactive in MCF-7 cells, and this defect may be a result of an inhibitor of cytochrome c or the apoptosome.
UV Treatment Induces PARP and Caspase-8 Cleavage Independent of Caspase-9 -Despite our results showing that the intrinsic pathway does not transmit an apoptotic signal in MCF-7 cells, our studies also confirm that MCF-7 cells can undergo UV-induced apoptosis. Typically, UV irradiation induces cells to undergo apoptosis via mitochondrial release of cytochrome c from the intermembrane space into the cytosol. Cytochrome c in the cytosol, in the presence of dATP, complexes with Apaf-1 and pro-caspase-9, resulting in caspase-9 autocatalytic activity. Although poorly characterized, other studies have suggested that caspase-8 could be indirectly involved in UV-mediated cell death by subsequent cytokine release and receptor activation or via a positive feedback loop subsequent to caspase-9 activation (31)(32)(33). In an effort to identify the caspase-9 independent pathway induced by UV irradiation, we asked whether UV could induce cleavage of the initiator caspase-8, which could then cleave caspase-7. As a positive control for caspase-8 activation, MCF-7 cells were treated for 4, 6, or 8 h with TNF-␣ and cycloheximide. As shown in Fig. 7, TNF-␣ induced activation of caspase-8 and PARP cleavage, indicating that the cells were undergoing apoptosis as expected. Interestingly, UV irradiation also activated caspase-8, although to a lesser extent than with TNF-␣. PARP cleavage was also less robust with UV treatment at the dose used (10 J/m 2 ). Again, caspase-9 cleavage was insensitive to either treatment, indicating that apoptosis was occurring independently of caspase-9 and that pro-caspase-8 may be the initiator caspase with UV treatment. Because catalytically active caspase-8 is known to cleave the effector caspase-7, we conclude that UV-mediated apoptosis can occur independently of mitochondrial release of cytochrome c and subsequent activation of caspase-9 (5, 6).
To determine the role of caspase-8 in UV-induced apoptosis FIG. 7. TNF-␣ and UV treatment both result in caspase-8 and PARP cleavage, whereas caspase-9 cleavage is unchanged. To determine whether UV treatment can result in activation of the initiator caspase-8, MCF-7 cells were serum-starved for 18 h before treatment with either UV irradiation (10 J/m 2 ) or TNF-␣ (100 ng/ml) and cycloheximide (1 g/ml). At the times indicated, the cells were harvested. The cell lysates were subjected to SDS-PAGE and Western blot analyses to determine the extent of PARP, caspase-9, and caspase-8. Western analysis of tubulin indicates similar protein loading in each lane.

FIG. 6. Bcl-2 overexpression does not affect UV-mediated apoptotic responses, and XIAP binding to caspase-9 is unaltered by UV treatment.
A, MCF-7 cells were transfected with either empty vector (mock) or Bcl-2 as described under "Experimental Procedures" or not transfected (control). Fortyeight hours post-transfection apoptosis was induced by TNF-␣ (100 ng/ml) or UV irradiation (10 J/m 2 ). Six hours later the cells were harvested, and the cell lysates were subjected to SDS-PAGE and Western blot analyses to determine the extent of caspase-7 cleavage and Bcl-2 expression. Western analysis of tubulin indicates similar protein loading in each lane. B, MCF-7 cells were plated and serumstarved. The following day cells were treated with UV irradiation (10 J/m 2 ). At the times indicated, the cells were harvested and lysed. The lysates were immunoprecipitated with cross-linked anticaspase-9 beads as described under "Experimental Procedures." Immunoprecipitation reactions (top panel) as well as 60 g of whole cell lysates from the same samples (bottom panel) were then subjected to SDS-PAGE and Western blot analysis. Pro-caspase-9 Western analysis was used as a loading control for immunoprecipitations. SFM, serum-free medium. more directly, we took advantage of a commercially available cell-permeable caspase-8 inhibitor, Z-IETD-FMK. As shown in Fig. 8, treatment with Z-IETD-FMK dramatically reduced apoptosis following treatment with either TNF-␣ or UV irradiation in MCF-7 cells, as measured by PARP and pro-caspase-7 cleavage. Cleavage of pro-caspase-8 was unaffected with Z-IETD-FMK treatment, as expected, because the mechanism of action of the inhibitor is to block the active site of caspase-8 rather than inhibiting pro-caspase-8 processing (data not shown). These results demonstrate a critical role for caspase-8 in both TNF-␣-and UV-induced apoptosis pathways in MCF-7 cells.

DISCUSSION
Under normal circumstances, apoptotic pathways are tightly regulated in organisms to assure appropriate growth and development. Clearly, aberrations of apoptosis have been documented in many diseases such as autoimmunity and cancer. In an effort to identify the underlying mechanisms of these diseases, executioners of apoptosis such as caspases, Smac/Diablo, and pro-apoptotic Bcl-2-related proteins have been described, along with proteins that inhibit apoptotic responses to cellular stress including IAPs and anti-apoptotic Bcl-2-related proteins. Much attention has focused on the function of caspase-9 and its requirement of the apoptosome for activity. Caspase-9 is of particular interest because it initiates cell death in response to non-cytokine-mediated cellular stress. The tightly regulated function of caspase-9 is crucial for development of the central nervous system; embryos of caspase-9 and Apaf-1 knockout mice show gross abnormalities in the brain and central nervous system (reviewed by Ranger et al. (34)).
In contrast, aberrations in mitochondrial apoptotic signals or caspase-9 function are better tolerated and more frequently observed in cancer cells resulting in treatment resistance (35). Because activated Akt phosphorylates pro-caspase-9 to inhibit its activity (13) and because caspase-9 is required downstream of p53 for p53-mediated apoptosis (25,36), caspase-9 has become an important cancer target. Loss of p53 function or p53mediated responses are both commonly observed in various forms of cancer; thus understanding the potential mechanisms of caspase-9 resistance to stress stimuli is of utmost importance in offsetting treatment resistance. To this end, significant progress has been made to identify the mechanisms for Bcl-2 or Bcl-XL-mediated cell survival and treatment resistance. One mechanism for these effects is the ability of these proteins to inhibit mitochondrial membrane potential and cytochrome c release (15,37), as well as their ability to bind and inhibit the activity of the apoptosome (14,17,18). Thus, overexpression of these proteins alters cellular response to stress at the level of the mitochondria and/or the apoptosome. Bcl-2 overexpression in MCF-7 cells had little effect after UV treatment, indicating either that the intrinsic pathway is not significantly contributing to the induction of apoptosis in these cells or that Bcl-2 is not inhibiting apoptosome function. IAPs can also inhibit apoptosis via binding to caspases-3, -7, and -9; the BIR3 domain of XIAP inhibits caspase-9 by direct binding to caspase-9 within the apoptosome (19,20,38). In our model, endogenous XIAP binding to caspase-9 did not significantly change with UV treatment. Because XIAP can also bind and inactivate caspase-7, we do not believe that role for XIAP is supported in this model because caspase-7 appears to be cleaving PARP in the absence of caspase-3. We show that in MCF-7 cells, other mechanisms of resistance must also exist that affect cytochrome c function. To our knowledge, this is the first report identifying an aberration in cytochrome c function after mitochondrial release, resulting in caspase-9 unresponsiveness.
Given the importance of cytochrome c in cell respiration and metabolism, one would expect that cells would not survive gross changes in protein structure or mitochondrial function. Indeed, we did not observe any alterations in the DNA sequence of cytochrome c derived from MCF-7 cells. Given that MCF-7 cells were able to respond to cellular stress by releasing mitochondrial cytochrome c but that only exogenous cytochrome c induced notable caspase-9 cleavage changes, we believe that our data are consistent with the presence of a cytoplasmic inhibitor of cytochrome c. In reviewing the literature to compare our findings, we note that many investigators have assessed pro-caspase-9 cleavage in MCF-7 cells using a cell-free system (4,33) or microinjection of exogenous cytochrome c, concluding that MCF-7 cells are cytochrome c-insensitive (39). Typically, exogenous forms of pro-caspase-9, dATP, and cytochrome c are used in these reactions to detect caspase cleavage with high specificity and sensitivity, but they are not designed to test the function of endogenous caspase-9 or cytochrome c. Although microinjection of cytochrome c tests the function of endogenous proteins downstream of the mitochondria, it also does not ascertain endogenous cytochrome c function. Further, by adding exogenous cytochrome c, both assays may increase cytochrome c concentrations beyond that which any endogenous inhibitor can bind.
Other investigators report that caspase-9 transfection of MCF-7 cells induces apoptosis, independent of treatment (9,40). We also have observed treatment-independent apoptosis with MCF-7 cells overexpressing pro-caspase-9 but only after reaching very high expression levels. In the studies described herein, we specifically expressed levels of caspase-9 that did not lead to apoptosis in the untreated transfectants to best mimic the in vivo functions of stress-induced caspase-9 activity. Finally, some groups have described treatment-induced cleavage of endogenous caspase-9 using MCF-7 cells (41). Such disparate results may occur from in vitro selection conditions because it is well accepted that the MCF-7 cell line may vary depending on in vitro culture conditions. We have confirmed our findings using MCF-7 cells from another source but are unable to study all sources. Despite these potential variations, we propose that MCF-7 cells offer a unique opportunity to study apoptotic mechanisms that occur independent of cytochrome c.
These considerations do not detract from the novelty of our findings, in that alterations in cytochrome c function have been greatly overlooked to date, whereas other mechanisms of resistance to apoptosis have been rapidly advanced. Because cytochrome c is a necessary component of the apoptosome, its loss of function would have notable consequences to chemically induced apoptosis in most cells. Surprisingly, our findings with MCF-7 cells suggest that this may not be the case in our particular model. Our data are consistent with a cytochrome c-independent, caspase-dependent pathway conveying UV-mediated apoptosis. We propose that caspase-8 can function as the initiator caspase in these circumstances. MCF-7 cells neither express full-length caspase-3 to induce a positive feedback loop (33) nor appear to induce apoptosome formation in vivo. Further, caspase-2 cleavage was unchanged with UV treatment (data not shown). Thus, our data support that caspase-8 activation by UV treatment results in caspase-7 and PARP cleavage, without the requirement of tBid to induce cytochrome c translocation and pro-caspase-9 cleavage (see Ref. 42 for review).
We were also intrigued by the observation that MCF-7 cells harbor cleaved caspase-9 in an unstressed state. We propose two possible explanations for this observation. First, in the presence of a cytochrome c inhibitor, the level of cleaved procaspase-9 may be below the threshold necessary for induction of apoptosis, although with pro-caspase-9 transfection we observed significantly higher levels of cleavage product without any apparent changes in downstream caspase-7 or PARP. Another explanation may lie in the observations made by Alnemri and co-workers (4, 10) using an Apaf-1 mutant (Apaf-530) that lacks its WD-40 repeats and thus cannot bind cytochrome c. Experiments using this Apaf-1 mutant show that it oligomerizes and processes pro-caspase-9 in the absence of cytochrome c and dATP. However, processed caspase-9 is not released from the apoptosome. Although not directly comparable, these findings may explain why we observe some cleaved caspase-9 in unstressed MCF-7 cells that may have no effect on apoptosis if not released from the apoptosome. On the other hand, deletion of the WD-40 domain may simply result in the loss of an inhibitory function on Apaf-1 itself (4), suggesting that this mutant better describes the importance of the WD-40 domain on Apaf-1 structure than on the importance of cytochrome c binding. In fact, little is known structurally about how cytochrome c binds to Apaf-1 to form a functional apoptosome or the consequences of cytochrome c inhibition in this model. However, studies to identify the cytoplasmic, cytochrome c inhibitor and to characterize its biological function(s) are currently underway in our laboratory.