Calcium Is a Key Signaling Molecule in β-Lapachone-mediated Cell Death

β-Lapachone (β-Lap) triggers apoptosis in a number of human breast and prostate cancer cell lines through a unique apoptotic pathway that is dependent upon NQO1, a two-electron reductase. Downstream signaling pathway(s) that initiate apoptosis following treatment with β-Lap have not been elucidated. Since calpain activation was suspected in β-Lap-mediated apoptosis, we examined alterations in Ca2+ homeostasis using NQO1-expressing MCF-7 cells. β-Lap-exposed MCF-7 cells exhibited an early increase in intracellular cytosolic Ca2+, from endoplasmic reticulum Ca2+ stores, comparable to thapsigargin exposures. 1,2-Bis-(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid-acetoxymethyl ester, an intracellular Ca2+ chelator, blocked early increases in Ca2+ levels and inhibited β-Lap-mediated mitochondrial membrane depolarization, intracellular ATP depletion, specific and unique substrate proteolysis, and apoptosis. The extracellular Ca2+ chelator, EGTA, inhibited later apoptotic end points (observed >8 h, e.g. substrate proteolysis and DNA fragmentation), suggesting that later execution events were triggered by Ca2+ influxes from the extracellular milieu. Collectively, these data suggest a critical, but not sole, role for Ca2+ in the NQO1-dependent cell death pathway initiated by β-Lap. Use of β-Lap to trigger an apparently novel, calpain-like-mediated apoptotic cell death could be useful for breast and prostate cancer therapy.

␤-Lapachone (␤-Lap) triggers apoptosis in a number of human breast and prostate cancer cell lines through a unique apoptotic pathway that is dependent upon NQO1, a two-electron reductase. Downstream signaling pathway(s) that initiate apoptosis following treatment with ␤-Lap have not been elucidated. Since calpain activation was suspected in ␤-Lap-mediated apoptosis, we examined alterations in Ca 2؉ homeostasis using NQO1-expressing MCF-7 cells. ␤-Lap-exposed MCF-7 cells exhibited an early increase in intracellular cytosolic Ca 2؉ , from endoplasmic reticulum Ca 2؉ stores, comparable to thapsigargin exposures. 1,2-Bis-(2-aminophenoxy)ethane-N,N,N,N-tetraacetic acid-acetoxymethyl ester, an intracellular Ca 2؉ chelator, blocked early increases in Ca 2؉ levels and inhibited ␤-Lap-mediated mitochondrial membrane depolarization, intracellular ATP depletion, specific and unique substrate proteolysis, and apoptosis. The extracellular Ca 2؉ chelator, EGTA, inhibited later apoptotic end points (observed >8 h, e.g. substrate proteolysis and DNA fragmentation), suggesting that later execution events were triggered by Ca 2؉ influxes from the extracellular milieu. Collectively, these data suggest a critical, but not sole, role for Ca 2؉ in the NQO1dependent cell death pathway initiated by ␤-Lap. Use of ␤-Lap to trigger an apparently novel, calpain-like-mediated apoptotic cell death could be useful for breast and prostate cancer therapy.
␤-Lap 1 is a naturally occurring compound present in the bark of the South American Lapacho tree. It has antitumor activity against a variety of human cancers, including colon, prostrate, promyelocytic leukemia, and breast (1)(2)(3). ␤-Lap was an effective agent (alone and in combination with taxol) against human ovarian and prostate xenografts in mice, with little host toxicity (4). We recently demonstrated that ␤-Lap kills human breast and prostate cancer cells by apoptosis, a cytotoxic response significantly enhanced by NAD(P)H:quinone oxidoreductase (NQO1, E.C. 1.6.99.2) enzymatic activity (5). 2 ␤-Lap cytotoxicity was prevented by co-treatment with dicumarol (an NQO1 inhibitor) in NQO1-expressing breast and prostate cancer cells (5). 2 NQO1 is a cytosolic enzyme elevated in breast cancers (6) that catalyzes a two-electron reduction of quinones (e.g. ␤-Lap, menadione), utilizing either NADH or NADPH as electron donors. Reduction of ␤-Lap by NQO1 presumably leads to a futile cycling of the compound, wherein the quinone and hydroquinone form a redox cycle with a net concomitant loss of reduced NAD(P)H (5).
Apoptosis is an evolutionarily conserved pathway of biochemical and molecular events that underlie cell death processes involving the stimulation of intracellular zymogens. The process is a genetically programmed form of cell death involved in development, normal turnover of cells, and in cytotoxic responses to cellular insults. Once apoptosis is initiated, biochemical and morphological changes occur in the cell. These changes include: DNA fragmentation, chromatin condensation, cytoplasmic membrane blebbing, cleavage of apoptotic substrates (e.g. PARP, lamin B), and loss of mitochondrial membrane potential with concomitant release of cytochrome c into the cytoplasm (7)(8)(9). Apoptosis is a highly regulated, active process that requires the participation of endogenous cellular enzymes that systematically dismantle the cell. The most well characterized proteases in apoptosis are caspases, aspartatespecific cysteine proteases, that work through a cascade that can be initiated by mitochondrial membrane depolarization leading to the release of cytochrome c and Apaf-1 into the cytoplasm (10), that then activates caspase 9 (11). Noncaspase-mediated pathways are less understood.
We previously showed that apoptosis following ␤-Lap administration was unique, in that an ϳ60-kDa PARP cleavage fragment, as well as distinct intracellular proteolytic cleavage of p53, were observed in NQO1-expressing breast or prostate cancer cells (5). 2 These cleavage events were distinct from those observed when caspases were activated by topoisomerase I poisons, staurosporine, or administration of granzyme B (5,12,13). Furthermore, ␤-Lap-mediated cleavage events were blocked by administration of global cysteine protease inhibitors, as well as extracellular Ca 2ϩ chelators (12). Based on these data, we concluded that ␤-Lap exposure of NQO1-expressing breast and prostate cancer cells caused the activation of a Ca 2ϩ -dependent protease with properties similar to calpain; in particular, the p53 cleavage pattern of ␤-Lap-exposed cells was remarkably similar to the pattern observed after calpain activation (14,15). Ca 2ϩ is recognized as an important regulator of apoptosis (16 -21). The cytoplasmic Ca 2ϩ concentration is maintained at ϳ100 nM in resting cells by relatively impermeable cell membranes, active extrusion of Ca 2ϩ from the cell by plasma membrane Ca 2ϩ -ATPases, plasma membrane Na ϩ /Ca 2ϩ exchangers, and active uptake of cytosolic Ca 2ϩ into the endoplasmic reticulum (ER) by distinct Ca 2ϩ -ATPases. In contrast, the concentration of Ca 2ϩ in the extracellular milieu and in the ER is much higher (in the millimolar range). Evidence for involvement of Ca 2ϩ influx into the cytosol as a triggering event for apoptosis has come from studies with specific Ca 2ϩ channel blockers that abrogate apoptosis in regressing prostate following testosterone withdrawal (22). Other support for the involvement of Ca 2ϩ in apoptosis comes from the observation that agents that directly mobilize Ca 2ϩ (e.g. Ca 2ϩ ionophores or the sarcoplasmic reticulum Ca 2ϩ -ATPase pump inhibitor, thapsigargin, TG) can trigger apoptosis in diverse cell types (23)(24)(25)(26)(27). Inhibition of the sarcoplasmic reticulum Ca 2ϩ -ATPase pump by TG causes a transient increase in cytoplasmic Ca 2ϩ from ER Ca 2ϩ stores, and a later influx of Ca 2ϩ from the extracellular milieu, leading to the induction of apoptotic cell death (24,27,28). Consequently, emptying of intracellular Ca 2ϩ stores may trigger apoptosis by disrupting the intracellular architecture and allowing key elements of the effector machinery (e.g. Apaf-1) to gain access to their substrates (e.g. caspase 9). Ca 2ϩ has also been shown to be necessary for apoptotic endonuclease activation, eliciting DNA cleavage after many cellular insults (29 -31). Buffering intracellular Ca 2ϩ released from stored Ca 2ϩ pools (e.g. ER) with BAPTA-AM, or removal of extracellular Ca 2ϩ with EGTA, can protect cells against apoptosis (32,33). Therefore, increases in intracellular Ca 2ϩ levels appear to be important cell death signals in human cancer cells that might be exploited for anti-tumor therapy. Finally, Ca 2ϩ may act as a signal for apoptosis by directly activating key proapoptotic enzymes (e.g. calpain); however, these proteolytic responses are poorly understood. The role of Ca 2ϩ in cell death processes involving caspase activation has been examined in detail (28, 34 -36). However, the role of Ca 2ϩ in non-caspasedependent cell death responses is relatively unexplored.
Recent studies have suggested that alterations in mitochondrial homeostasis play an essential role in apoptotic signal transduction induced by cytotoxic agents (37,38). Various apoptotic stimuli have been shown to induce mitochondrial changes, resulting in release of apoptogenic factors, apoptosisinducing factor (39), and mitochondrial cytochrome c (9) into the cytoplasm. These changes are observed during the early phases of apoptosis in human epithelial cells, and were linked to the initial cascade of events, sending the cell to an irreversible suicide pathway. During high, sustained levels of cytosolic Ca 2ϩ , mitochondrial Ca 2ϩ uptake is driven by mitochondrial membrane potential to maintain Ca 2ϩ homeostasis in the cytosol. In de-energized mitochondria, Ca 2ϩ can be released by a reversal of this uptake pathway (40). These data, therefore, linked changes in Ca 2ϩ homeostasis and mitochondrial membrane potential to the initiation of apoptosis. Li et al. (41) reported that ␤-Lap caused a decrease in mitochondrial membrane potential with release of cytochrome c into the cytoplasm in a number of human carcinoma cell lines, shortly after drug addition. Other alterations in metabolism (e.g. ATP depletion) have not been examined in ␤-Lap-treated cells.
We previously characterized the activation of a novel cysteine protease in various breast cancer cell lines with properties similar to the Ca 2ϩ -dependent cysteine protease, calpain, after exposure to ␤-Lap (12). Using NQO1-expressing breast cancer cells, we show that ␤-Lap elicits a rise in intracellular Ca 2ϩ levels shortly after drug administration that eventually leads to apoptosis. This paper suggests a critical, but not sufficient, role for Ca 2ϩ in the cell death pathway initiated by NQO1-dependent bioactivation of ␤-Lap. Possible combinatorial effects (e.g. NAD(P)H depletion as well as intracellular calcium alterations) that initiate ␤-Lap-mediated apoptosis in NQO1-expressing breast cancer cells will be discussed.
Cell Culture-MCF-7:WS8 (MCF-7) human breast cancer cells were obtained from Dr. V. Craig Jordan, (Northwestern University, Chicago, IL). MDA-MB-468 cells were obtained from the American Type Culture Collection and transfected with NQO1 cDNA in the pcDNA3 constitutive expression vector as described previously (5). Tissue culture components were purchased from Life Technologies, Inc., unless otherwise stated. MCF-7 cells were grown in RPMI 1640 cell culture medium supplemented with 10% fetal bovine serum, in a 37°C humidified incubator with 5% CO 2 , 95% air atmosphere as previously described (2,5). For all experiments, log-phase breast cancer cells were exposed to 5 M ␤-Lap for 4 h (unless otherwise indicated), after which fresh medium was added and cells were harvested at various times post-treatment.
TUNEL Assay-Cells were seeded at 1 ϫ 10 6 cells/10-cm Petri dish and allowed to grow for 24 h. Log-phase cells were then pretreated for 30 min with 10 M BAPTA-AM, 3 mM EGTA, or 50 M dicumarol followed by a 4-h pulse of 5 M ␤-Lap, as described above, or 24 h treatment of 10 M ionomycin or 1 M staurosporine. Medium was collected from experimental as well as control conditions 24 h later, and attached along with floating cells were monitored for apoptosis using TUNEL 3Ј-biotinylated DNA end labeling via the APO-DIRECT kit (Pharmingen, San Diego, CA) as described (5). Apoptotic cells were analyzed and quantified using an EPICS XL-MCL flow cytometer that contained an air-cooled argon laser at 488 nm, 15 mW (Beckman Coulter Electronics; Miami, Fl), and XL-MCL acquisition software provided with the instrument.
Cell Growth Assays-MCF-7 cells were seeded at 5 ϫ 10 4 cells per well in a 12-well plate and allowed to attach overnight. The following day, log-phase cells were pretreated for 30 min with 5 M BAPTA-AM, followed by a 4-h pulse of ␤-Lap (0 -5 M). Drugs were removed and fresh medium added. Cells were allowed to grow for an additional 6 days. DNA content (a measure of cell growth) was determined by fluorescence using Hoechst dye 33258 as described (5) and changes in growth were monitored using a PerkinElmer HTS 7000 Plus Bio Assay Plate Reader (Norwalk, CT) with 360 and 465 nm excitation and emission filters, respectively. Data were expressed as relative growth, T/C (treated/control), using experiments performed at least twice.
Confocal Microscopy-MCF-7 cells were seeded at 2-3 ϫ 10 5 cells per 35-mm glass bottom Petri dishes (MatTek Corp., Ashland, MA) and allowed to attach overnight. Cells were rinsed twice in a Ca 2ϩ /Mg 2ϩ balanced salt solution (BSS, 130 mM NaCl, 5 mM KCl, 1.5 mM CaCl 2 , 1 mM MgCl 2 , 25 mM HEPES, pH 7.5, 5 mM glucose, 1 mg/ml bovine serum albumin) and loaded with the Ca 2ϩ -sensitive fluorescent indicator, fluo-4-AM (5 M), in BSS for ϳ20 -30 min at 37°C. Cells were rinsed twice in BSS and incubated for an additional 20 min at 37°C to allow for hydrolysis of the AM-ester. Cells were imaged with a Zeiss 410 confocal microscope (Thornwood, NY) equipped with a ϫ63 N.A. 1.4 oil immersion planapochromat objective at room temperature (the same results were observed at room temperature and 37°C). Confocal images of fluo-4 fluorescence were collected using a 488-nm excitation light from an argon/krypton laser, a 560-nm dichroic mirror, and a 500 -550 nm band-pass barrier filter. Three basal images were collected before drug addition (8 M ␤-Lap, Ϯ 50 M dicumarol or 200 nM TG). The mean pixel intensity was set to equal one for analyses of fold-increase in fluo-4 fluorescence intensity. Subsequently, images were collected after the indicated treatments at 90-s intervals. BAPTA-AM (20 M) was co-loaded with fluo-4-AM where indicated. Mean pixels were determined in regions of interest for individual cells at each time point.
Mitochondrial Membrane Potential Determinations-MCF-7 cells were seeded at 2.5-3 ϫ 10 5 cells per 6-well plate, and allowed to grow for 24 h. Log-phase cells were pretreated for 30 min with 10 M BAPTA-AM, 3 mM EGTA, or 50 M dicumarol followed by a 4-h pulse of 5 M ␤-Lap, unless otherwise indicated. Cells were trypsinized and resuspended in phenol red-minus RPMI medium for analyses. Cells were maintained at 37°C for the duration of the experiment, including during analyses. Prior to analyses, cells were loaded with 10 g/ml JC-1 for 9 -14 min and samples were analyzed using a Beckman Coulter EPICS Elite ESP (Miami, FL) flow cytometer. JC-1 monomer and aggregate emissions were excited at 488 nM and quantified using Elite acquisition software after signal collection through 525-and 590-nm band pass filters, respectfully. Shifts in emission spectra were plotted on bivariant dot plots, on a cell-by-cell basis, to determine relative mitochondrial membrane potential of treated and control cells.
ATP Measurements-Cells were seeded at 2.5 ϫ 10 5 cells per well in 6-well dishes and allowed to attach for 24 h. Fresh medium was added to the cells along with Ca 2ϩ chelators or dicumarol 30 min prior to ␤-Lap exposure (4 h unless otherwise indicated). Floating cells were collected, pelleted, and lysed in 1.67 M perchloric acid. Attached cells were lysed directly in 1.67 M perchloric acid. Following a 20-min incubation at room temperature, attached cells were scraped and transferred to corresponding microcentrifuge tube, cooled on ice for several minutes, and spun to pellet protein precipitates. Deproteinized samples were neutralized with 3.5 M KOH and HEPES/KOH (25 mM HEPES, 15 mM KOH, pH 8), and incubated on ice for 15 min. Precipitates were removed by centrifugation and samples stored at Ϫ20°C. Cell extracts were analyzed for ATP and ADP levels using a luciferase-based bioluminescent assay and rephosphorylation protocols, as described (42).
Western Blot Analyses-Whole cell extracts from control or ␤-Lapexposed MCF-7 cells were prepared and analyzed by SDS-polyacrylamide gel electrophoresis/Western blot analyses as previously described (2,5,12). Loading equivalence and transfer efficiency were monitored by Western blot analyses of proteins that are known to be unaltered by experimental treatments (2), and using Ponceau S staining of the membrane, respectively. Probed membranes were then exposed to x-ray film for an appropriate time and developed. Dilutions of 1:10,000 for the C-2-10 anti-PARP antibody (Enzyme Systems Products, Livermore, CA), and 1:2000 for anti-p53 DO-1 and anti-lamin B (Santa Cruz Biotechnology, Santa Cruz, CA) antibodies were used as described (2,12).

Ca 2ϩ Chelators Prevent ␤-Lap-induced
Apoptotic DNA Fragmentation and Protect against Cell Death-Log-phase MCF-7 cells were treated for 4 h with 5 M ␤-Lap, fresh medium was then applied, and cells were harvested 24 h later and analyzed for DNA fragmentation (i.e. apoptotic cells staining positive in a TUNEL assay). Treatment of MCF-7 cells with ␤-Lap resulted in Ͼ90% apoptotic cells (Fig. 1, A and B). However, MCF-7 cells exposed to a 30-min pretreatment with 10 M BAPTA-AM or 3 mM EGTA, followed by a 4-h pulse of 5 M ␤-Lap, exhibited only 20 or 39% apoptotic cells, respectively, in 24 h.
To examine whether BAPTA-AM could affect ␤-Lap lethality, we measured relative growth of MCF-7 cells with or without exposure to ␤-Lap, and in the presence or absence of BAPTA-AM. MCF-7 cells were treated for 30 min with 5 M BAPTA-AM, subsequently exposed to a 4-h pulse of ␤-Lap (1.5Ϫ5 M), and relative cell growth was measured 6 days later (Fig. 1C). The LD 50 dose of ␤-Lap in MCF-7 cells was ϳ2.5 M in colony forming assays, which correlated well with IC 50 relative growth inhibition, as measured by DNA content (2,5). At 1.5 M ␤-Lap, cells exhibited little or no toxicity. At ␤-Lap doses of 3 or 5 M, cells exhibited considerable toxicity, Ͼ90% growth inhibition, as previously reported (2,5). Toxicity was significantly prevented by 5 M BAPTA-AM pretreatment. BAPTA-AM pretreated cells exhibit only 44 and 73% growth inhibition after 3 or 5 M ␤-Lap treatments, respectively (Fig. 1C). BAPTA alone did not affect MCF-7 cell growth compared with untreated controls.

Ca 2ϩ Chelators Do Not Block Apoptosis Induced by Other
Agents-It was possible based on the data in Fig. 1 that calcium chelators may block ␤-Lap-mediated apoptosis by sequestering calcium required for the activation of apoptotic endonucleases. We, therefore, examined both intra-and extracellular Ca 2ϩ chelators for their ability to prevent apoptosis in NQO1-transfected MDA-468 (MDA-468-NQ3) cells induced by ␤-Lap, ionomycin (which induces Ca 2ϩ -mediated cell death (36)), and staurosporine (STS, which inhibits protein kinase C and works via a caspase-mediated cell death pathway (43,44)). We used MDA-468-NQ3 cells to assay for caspase-mediated endonuclease activation and DNA fragmentation since they express the endonuclease-activating caspase 3, unlike MCF-7 cells (45). We previously demonstrated that MDA-468-NQ3 cells responded similarly to ␤-Lap as MCF-7 cells (Fig. 2 and Ref. 5). EGTA significantly protected MDA-468-NQ3 cells against ionomycininduced apoptosis, but not against STS-induced apoptosis (Fig.  2). MDA-468-NQ3 cells treated for 24 h with 10 M ionomycin exhibited 49% apoptotic cells, whereas, MDA-468-NQ3 cells pretreated for 30 min with 3 mM EGTA followed by a 24-h exposure to ionomycin exhibited only 4% apoptotic cells. Cells treated for 24 h with 1 M STS in the absence or presence of 3 mM EGTA exhibited 56 and 46% apoptosis, respectively. BAPTA-AM (10 M) did not significantly block apoptosis induced by ionomycin. BAPTA-AM pretreatment of STS-exposed MDA-468-NQ3 cells did not significantly decrease apoptosis (p Ͻ 0.4) compared with cells exposed to STS alone; the modest effect of BAPTA-AM on STS-induced apoptosis may reflect the Ca 2ϩ dependence of the apoptotic endonucleases involved in this response. Neither BAPTA-AM nor EGTA alone elicited apoptotic responses at the doses used in the aforementioned experiments (Figs. 1B and 2). Furthermore, preliminary data suggest that DFF45 (ICAD) was cleaved in NQO1-expressing MCF-7 or MDA-468-NQ3 cells at 8 h after ␤-Lap treatment, in a temporal manner corresponding to the induction of apoptosis (data not shown). Cleavage of DFF45, an endogenous inhibitor of the magnesium-dependent and Ca 2ϩ -independent apoptotic endonuclease, DFF40 (CAD), suggests that DFF40 is activated following treatment with ␤-Lap. Taken together with results in Fig. 1, these data strongly suggest that a rise in intracellular Ca 2ϩ levels is part of a critical signaling pathway for the induction of apoptosis in NQO1-expressing human breast cancer cells following ␤-Lap exposure.
Exposure of NQO1-expressing MCF-7 Cells to ␤-Lap Results in Increased Intracellular Ca 2ϩ -We next directly examined whether intracellular Ca 2ϩ levels were increased in log-phase MCF-7 cells after ␤-Lap treatment using the cell-permeant intracellular Ca 2ϩ indicator dye, fluo-4. Cells were loaded with 5 M fluo-4-AM, and where indicated, 20 M BAPTA-AM, incubated for ϳ25 min to allow for the dye to permeate cells, rinsed, and then incubated for an additional ϳ20 min for hydrolysis of the AM-ester. Following drug addition, images were collected every 90 s for ϳ60 min using confocal microscopy. Three basal images were recorded before drug addition and average pixels per cell were determined (indicative of fluo-4 fluorescence and, therefore, basal intracellular Ca 2ϩ levels) and used for analyses over time. The fluorescence of basal images were averaged and set to equal one; fold increases were determined from changes in fluo-4 fluorescence over control.
After exposure to 8 M ␤-Lap, MCF-7 cells exhibited an ϳ2-fold increase in fluo-4 fluorescence from 4 to 9 min, after which time Ca 2ϩ levels returned to basal levels in a majority of cells examined (43 of 50, 86%) (Fig. 3A). The rise in intracellular Ca 2ϩ levels in MCF-7 cells following ␤-Lap exposure was prevented by preloading cells with BAPTA-AM (20 M) (Fig.  3B). Interestingly, not all ␤-Lap-exposed MCF-7 cells were affected by pretreatment with BAPTA-AM; 3 of 26 cells (12%) exhibited a rise in intracellular Ca 2ϩ levels after exposure to ␤-Lap despite the presence of this Ca 2ϩ chelator. However, BAPTA-AM pretreated MCF-7 cells that did exhibit a rise in intracellular Ca 2ϩ levels following ␤-Lap treatment exhibited a similar, but delayed Ca 2ϩ increase (10 -20 min), as compared with ␤-Lap-exposed MCF-7 cells in the absence of BAPTA-AM (4 -9 min). This may be due to a saturation of the chelator or heterogeneity of the tumor cell population. These results are consistent with previous reports that the buffering capacity of BAPTA-AM may be overwhelmed with time (34,46). Higher doses of BAPTA-AM were not used due to toxicity caused by the drug alone (data not shown).
Since the ER is a major store of Ca 2ϩ in the cell, we tested if the initial rise in intracellular Ca 2ϩ levels after exposure of MCF-7 cells to ␤-Lap was due to release of Ca 2ϩ from this organelle. If ␤-Lap exposure led to release of Ca 2ϩ stored in the ER, then TG (a sarcoplasmic reticulum Ca 2ϩ -ATPase pump inhibitor) administration should not cause additional Ca 2ϩ release. Similarly, if the sequence of drug administration were reversed, additional Ca 2ϩ release would also not be observed. When ␤-Lap was added after TG-induced depletion of ER Ca 2ϩ stores, no measurable rise in intracellular Ca 2ϩ levels occurred in 25 of 27 (93%) cells analyzed (Fig. 3C). Similarly, when TG was added to cells after ␤-Lap, only 1 of 18 (6%) cells that initially responded to ␤-Lap exhibited a rise in intracellular Ca 2ϩ levels following subsequent TG administration (Fig. 3D). At the end of the experiment, all cells analyzed remained responsive to ionomycin. Thus, cells exposed to ␤-Lap and/or TG were still capable of altering Ca 2ϩ levels, and the Ca 2ϩ indicator dye was not saturated. We noted that the increase in fluo-4 fluorescence (2-3-fold over basal levels, Fig. 3A) in MCF-7 cells observed after exposure to ␤-Lap was comparable to that elicited by TG (1.5-2.5-fold over basal levels, Fig. 3C), further suggesting that the two agents mobilized the same ER pool of Ca 2ϩ . All cells analyzed started with comparable basal levels of Ca 2ϩ and appeared to load equal amounts of the indicator dye, as determined by basal fluorescence (measured by pixels per cell) at the beginning of each analysis; relative basal fluo-4 fluorescence for each experiment in Fig. 3   Loss of Mitochondrial Membrane Potential After ␤-Lap Is Attenuated by Intracellular, but Not Extracellular, Ca 2ϩ Chelation-Mitochondrial membrane potential was previously shown to drop from a hyperpolarized state to a depolarized state after treatment of various human cancer cells with ␤-Lap (41). A drop in mitochondrial membrane potential in ␤-Laptreated cells was accompanied by a concomitant release of cytochrome c into the cytosol (41). To explore whether early changes in intracellular Ca 2ϩ levels were upstream of mitochondrial changes in NQO1-expressing breast cancer cells, log phase MCF-7 cells were pretreated for 30 min with either 10 M BAPTA-AM or 3 mM EGTA and then exposed to 5 M ␤-Lap for 4 h. Prior to analyses, cells were loaded with JC-1, a cationic dye commonly used to monitor alterations in mitochondrial membrane potential (47,48). Mitochondrial depolarization measurements using JC-1 were indicated by a decrease in the red/green fluorescence intensity ratio (a movement of events from upper left to lower right, Fig. 4), as seen following a 10-min treatment with the potassium ionophore, valinomycin (100 nM), which causes a collapse of mitochondrial membrane potential by uncoupling mitochondrial respiration (Fig. 4e) (49); cells in the upper left-hand quadrant exhibited high mitochondrial membrane potential, whereas, cells in the lower right-hand quadrant have low mitochondrial membrane potential and are depolarized. Cells in the upper right-hand quadrant exhibited intermediate membrane potential. Mitochondrial membrane potential decreased in MCF-7 cells in a time-and dose-dependent manner following exposure to ␤-Lap (Figs. 4, a-d, and data not shown). By 4 h, the majority of ␤-Lap-treated MCF-7 cells exhibited low mitochondrial membrane potential (53%), while the majority of control cells maintained high mitochondrial membrane potential (51%) (Fig. 4, b, a and g, f, respectively). This drop in mitochondrial membrane potential observed 4 h after treatment with ␤-Lap (low, 53%) was abrogated by pretreatment with BAPTA-AM (low, 23%), but not by EGTA (low, 48%) (Fig. 4, g-i, respectively). Pretreatment with 10 M BAPTA-AM prevented the decrease in mitochondrial membrane potential (low, 23%); however, BAPTA-AM did not maintain ␤-Lap-exposed cells in a high-potential state (high, 28%) as observed in control untreated cells (high, 51%). Approximately half of the BAPTA-AM-exposed cells were in an intermediate membrane potential state (45%) (Fig. 4h). We noted, however, that BAPTA-AM or EGTA exposures alone caused depolarization of the mitochondria, with a majority of the cells residing in the same intermediate energized state as observed following BAPTA-AM and ␤-Lap (Fig. 4, j-k). Therefore, BAPTA-AM prevented mitochondrial depolarization induced by ␤-Lap to the same extent as in cells treated with BAPTA-AM alone. Pretreatment with 3 mM EGTA did not affect the loss of mitochondrial membrane potential caused by ␤-Lap (low 48%), implying that an early rise in intracellular Ca 2ϩ levels from intracellular stores was sufficient to cause a drop in mitochondrial membrane potential, and that extracellular calcium was not needed for these effects in ␤-Lap-treated cells (Fig. 4, h-i).
Loss of ATP After ␤-Lap Is Attenuated by Intracellular Ca 2ϩ Chelation-The bioactivation of ␤-Lap by NQO1 is thought to lead to a futile cycling between quinone and hydroquinone forms of the compound, presumably due to the instability of the hydroquinone form of ␤-Lap (5). This futile cycling led to depletion of NADH and NADPH, electron donors for NQO1 in in vitro assays (5). Exhaustion of reduced enzyme co-factors may be a critical event for the activation of the apoptotic pathway in NQO1-expressing cells following ␤-Lap exposure. We, therefore, measured intracellular ATP and ADP in log-phase MCF-7 cells after various doses and times of ␤-Lap (using a luciferasebased bioluminescent assay (42)). Intracellular ATP levels were reduced in MCF-7 cells after treatment with ␤-Lap in a dose-and time-dependent manner (Fig. 5A). At all doses of ␤-Lap above the LD 50 of the drug (ϳ2.5 M) in MCF-7 cells (2), intracellular ATP levels were reduced by Ͼ85% at 4 h, the time at which drug was removed (Fig. 5A, left); the loss of ATP correlated well with ␤-Lap-induced cell death in MCF-7 cells (Fig. 1C). ADP levels remained relatively unchanged after various doses of ␤-Lap, however, the [ATP]/[ADP][P i ] ratio decreased dramatically. Intracellular ATP levels began to drop to 70% of control levels 2 h after 5 M ␤-Lap exposure, the time at which ␤-Lap began to elicit mitochondrial membrane depolarization (Figs. 5, A, right, and 4, c). ATP levels continued to drop to 8% of control levels by 4 h after drug exposure (Fig. 5A,  right). In contrast, ADP levels remained relatively unchanged during the course of the experiment, with an increase at 30 min (172% control levels) that returned to control levels by 1 h post-treatment. Cellular ATP levels in ␤-Lap-treated cells did not appear to recover to normal levels within the 6 -24-h interval after drug removal (data not shown).
Loss of ATP following ␤-Lap was prevented by a 30-min pretreatment with an intracellular Ca 2ϩ chelator, but not an extracellular Ca 2ϩ chelator (Fig. 5B). At 4 h, pretreatment with 10 or 30 M BAPTA-AM elicited only 58 and 43% ATP loss, respectively, compared with ␤-Lap alone (92% loss). The extracellular Ca 2ϩ chelator, EGTA, did not significantly affect the loss of ATP, nor [ATP]/[ADP][P i ] ratio observed in MCF-7 cells after ␤-Lap treatment (Fig. 5B). Exposure of MCF-7 cells to TG (200 nM) did not elicit decreases in ATP or ADP levels 4 h after drug exposure, compared with untreated control cells.
Ca 2ϩ Chelators Prevent ␤-Lap-induced Proteolysis-We previously showed that apoptosis in various breast cancer cell lines induced by ␤-Lap was unique, causing a pattern of PARP and p53 intracellular cleavage events distinct from those in-duced by caspase activating agents (12). After ␤-Lap treatment, we observed an ϳ60-kDa PARP cleavage fragment and specific cleavage of p53 in NQO1-expressing breast cancer cells. Furthermore, we showed that this proteolysis in ␤-Lap-treated cells was the result of activation of a Ca 2ϩ -dependent protease with properties similar to -calpain (12). PARP and p53 proteolysis in ␤-Lap-exposed, NQO1-expressing cells was prevented by pretreatment with the extracellular Ca 2ϩ chelators, EGTA and EDTA, in a dose-dependent manner (at 8 and 24 h) (Ref. 12, and data not shown). Additionally, PARP, p53, and lamin B proteolysis induced at 24 h in MCF-7 cells following ␤-Lap treatment were abrogated by pretreatment with 10 or 30 M BAPTA-AM (Fig. 6). These data strongly suggest that a Ca 2ϩdependent pathway and potentially a Ca 2ϩ -dependent protease are operative in ␤-Lap-mediated apoptosis.
A simple explanation for the aforementioned results could be that BAPTA blocks bioactivation of ␤-Lap by NQO1 in a manner similar to that of dicumarol (5). However, BAPTA (free acid) did not affect the enzymatic activities of NQO1 using standard enzymatic assays (data not shown) (5). The free acid (active) form of BAPTA, instead of its ϪAM ester form, was used in these assays since intracellular accumulation of this Ca 2ϩ chelator was not necessary and was physiologically relevant in the in vitro enzyme assay. Using ␤-Lap as a substrate, NQO1 enzymatic activity in the presence of 10 mM BAPTA (a dose of the free acid form of BAPTA that was Ͼ1000-fold higher than that used in the experiments of Figs. 1-6) was reduced by Ͻ20%. Thus, BAPTA-AM did not affect the activity of NQO1, a two-electron reductase required for ␤-Lap cytotoxicity (5). We conclude that BAPTA-AM prevents ␤-Lap-induced apoptosis by blocking Ca 2ϩ -mediated signaling events via chelating intracellular Ca 2ϩ .
␤-Lap Bioactivation by NQO1 Is Critical for Ca 2ϩ -mediated Signaling-We previously reported that cells expressing NQO1 are more sensitive to the cytotoxic effects of ␤-Lap (5). 2 NQO1 is inhibited by dicumarol, which competes with NADH or NADPH for binding to the oxidized form of the enzyme. Dicumarol thereby prevents reduction of quinones (50,51). We demonstrated that dicumarol attenuates ␤-Lap-mediated proteolysis of apoptotic substrates (e.g. PARP and p53), apoptosis, and survival in NQO1-expressing cells (5). 2 As expected, increases in intracellular Ca 2ϩ levels in NQO1-expressing human cancer cells elicited by ␤-Lap were abrogated by co-treatment with 50 M dicumarol in 26 of 27 cells (96%) examined (Fig. 7A, lower  panel). The ability of dicumarol to inhibit increases in intracellular Ca 2ϩ levels was greater than that observed with BAPTA-AM, where intracellular Ca 2ϩ level increases were prevented in only 89% of cells examined (Fig. 3B). Thus, NQO1 was critical for the rise in intracellular Ca 2ϩ levels observed in MCF-7 cells after ␤-Lap exposure.
Mitochondrial membrane depolarization induced by ␤-Lap was also abrogated by pretreatment with dicumarol (Fig. 7B). By 4 h, the majority of ␤-Lap-treated cells exhibited low mitochondrial membrane potential (58%), while very few control cells were depolarized (9%) (Fig. 7B). Pretreatment with dicumarol attenuated this response to ␤-Lap, with only 34% being depolarized. The inability of dicumarol to prevent mitochondrial depolarization in 34% of ␤-Lap-treated cells was probably due to the high background of control cells (20%) that were depolarized after exposure to dicumarol alone. In comparison with intracellular Ca 2ϩ buffering, BAPTA-AM elicited only a minor depolarization of the mitochondria on its own (low, 14%) and thus was able to elicit a greater protective effect (Fig. 4B); only 23% of cells exposed to BAPTA-AM and ␤-Lap exhibited low mitochondrial membrane potential as compared with ␤-Lap exposed cells in the presence of dicumarol (34%).
The dramatic loss of intracellular ATP in MCF-7 cells following ␤-Lap exposure was inhibited by a 30-min pretreatment with 50 M dicumarol (Fig. 7C). ␤-Lap-treated MCF-7 cells pretreated with dicumarol exhibited only 34% loss of intracellular ATP, compared with 92% loss after ␤-Lap treatment alone (Fig. 7C). ADP levels were not altered by any of the treatments used, however, the [ATP]/[ADP][P i ] ratio decreased dramatically in ␤-Lap-treated cells, and was only partially decreased with dicumarol pretreatment alone, as compared with control untreated cells.
Dicumarol also abrogated DNA fragmentation induced by ␤-Lap in MCF-7 cells. MCF-7 cells exhibited 94% apoptosis following ␤-Lap exposure that was prevented by a 30-min pretreatment with 50 M dicumarol; only 6% of the cells staining positive in a TUNEL assay at 24 h post-treatment (Fig. 7D). These data are consistent with prior results (5), and correlate well with the survival protection afforded by dicumarol to ␤-Lap-treated cells. Dicumarol did not induce DNA fragmentation on its own. These data are consistent with the protection from apoptosis observed with either intra-and extracellular Ca 2ϩ chelators. BAPTA-AM or EGTA protected ␤-Lap exposed MCF-7 cells from apoptosis (Fig. 1, A and B). Collectively, these data implicate the bioactivation of ␤-Lap by NQO1 as a critical step in the rise of intracellular Ca 2ϩ levels following ␤-Lap exposure, and thus ␤-Lap-mediated downstream apoptotic events.

DISCUSSION
When homeostatic mechanisms for regulating cellular Ca 2ϩ are compromised, cells may die, either by necrosis or apoptosis (20,21,36). We demonstrated that bioactivation of ␤-Lap by NQO1 induced cell death in a manner that was dependent upon Ca 2ϩ signaling (Figs. 1-6). ␤-Lap can be reduced by NQO1 and FIG. 5. ATP depletion after ␤-Lap treatment is Ca 2؉ dependent. Intracellular ATP and ADP levels were measured using a luciferase-based bioluminescent assay. A, cells were treated with the indicated dose of ␤-Lap for 4 h or were treated with 5 M ␤-Lap for the time indicated, and harvested for ATP analyses. ATP levels were expressed as nanomoles of ATP per 10 6 cells. Purified ATP was used as a standard to determine intracellular ATP concentrations. B, cells were either pretreated or untreated with the indicated Ca 2ϩ chelators for 30 min prior to drug addition, and ␤-Lap (5 M) was then added for 4 h. Cells were harvested for analyses following ␤-Lap exposure. Results represent the average of at least three independent experiments, Ϯ S.E. Student's t test for paired samples, experimental group compared with drug alone are indicated (*, p Ͻ 0.05; **, p Ͻ 0.01).
FIG. 6. Intracellular Ca 2؉ chelators prevent apoptotic proteolysis after ␤-Lap treatment. Apoptotic proteolysis was measured in MCF-7 cells exposed to a 4-h pulse of 5 M ␤-Lap, with or without a 30-min pretreatment of the indicated dose of BAPTA-AM. Whole cell extracts were prepared 24 h after drug addition, and analyzed using standard Western blotting techniques with antibodies to PARP, p53, and lamin B. Shown is a representative Western blot of whole cell extracts from experiments performed at least three times. may undergo futile cycling between quinone and hydroquinone forms (␤-Lap-Q and ␤-Lap-HQ, Fig. 8), presumably depleting NADH and/or NADPH in the cell (5). We theorize that depletion of NAD(P)H, along with a rise in intracellular Ca 2ϩ levels in response to ␤-Lap, activate a novel caspase-independent apoptotic pathway, as described in this paper and previously (2,5,12). The rise in intracellular Ca 2ϩ appears to be dependent upon the bioactivation of ␤-Lap by NQO1, suggesting a critical and necessary signaling role for Ca 2ϩ in the downstream apoptotic pathway induced by this drug. Dicumarol completely abrogated intracellular Ca 2ϩ changes (Fig. 7), as well as apoptosis and survival, following ␤-Lap exposure of NQO1-expressing cells (5). 2 When increases in intracellular Ca 2ϩ levels were directly prevented by pretreatment with BAPTA-AM, downstream apoptotic responses, as well as lethality, caused by ␤-Lap were prevented; when corrected for BAPTA-AM affects alone, ␤-Lap-induced apoptosis, proteolysis, and lethality were essentially blocked by preventing early Ca 2ϩ release from ER stores. Thus, correcting for the BAPTA-AM affects alone, the role of Ca 2ϩ in ␤-Lap-mediated apoptosis may be more significant that that revealed by the data shown. These data strongly mean Ϯ S.E. of at least three independent experiments. Student's t test for paired samples, experimental groups compared with ␤-Lap exposure alone are indicated (*, p Ͻ 0.005). DC, 50 M dicumarol. In cells that express NQO1, ␤-Lap is reduced from the quinone (␤-lap-Q) to the hydroquinone (␤-lap-HQ) form in a futile cycle that results in dramatic losses of NAD(P)H (5). During the metabolism of ␤-Lap by NQO1, Ca 2ϩ is subsequently released from the ER causing a rise in cytosolic Ca 2ϩ levels by an as yet unknown mechanism. To maintain low cytoplasmic Ca 2ϩ levels, we theorize that mitochondria sequester Ca 2ϩ and numerous cellular ATPases probably function to pump Ca 2ϩ out of the cytosol. This leads to mitochondrial membrane depolarization and ATP hydrolysis, respectively (Figs. 4 and 5). Sustained depolarization of the mitochondrial membrane leads to further loss of ATP and prevents ATP synthesis by inhibiting respiration. The loss of ATP disrupts ionic homeostasis within the cell and thereby allows extracellular Ca 2ϩ to enter the cell down its concentration gradient (see "Discussion"). The secondary rise in cytosolic Ca 2ϩ levels leads to protease (presumably activation of calpain or a calpainlike protease) and, thus, endonuclease (DFF40) activation, ultimately resulting in apoptosis.
suggest that DNA fragmentation, mitochondrial membrane depolarization, ATP loss, and apoptotic proteolysis were a consequence of the increase in intracellular Ca 2ϩ levels (Figs. 1-6 and 8). Interestingly, the cell death pathway induced by ␤-Lap was quite distinct from that observed after exposure to TG, an agent known to specifically cause release of Ca 2ϩ from ER stores and mediate caspase-dependent apoptosis (24,28,33,52). Thus, Ca 2ϩ release was necessary for ␤-Lap-induced cytotoxicity, but apparently not sufficient for the unique apoptotic responses induced by ␤-Lap.
␤-Lap and TG-induced Similar Ca 2ϩ Responses, but Different Patterns of Apoptosis-␤-Lap elicited an early rise in intracellular Ca 2ϩ levels from the same ER store as released by TG, however, subsequent cell death processes were remarkably different between the two compounds. TG is known to cause transient increases in intracellular Ca 2ϩ levels, however, these were insufficient to induce apoptosis. Much like ␤-Lap, Ca 2ϩ was needed from the extracellular milieu, along with a sustained increase in intracellular Ca 2ϩ levels, for TG-induced apoptosis (23) in MCF-7 cells (27). Depolarization of the mitochondrial membrane potential and loss of intracellular ATP in cells exposed to ␤-Lap, may have prevented plasma membrane Ca 2ϩ pumps and ER Ca 2ϩ pumps from functioning and maintaining Ca 2ϩ homeostasis. This, in turn, may have facilitated Ca 2ϩ leakage down its concentration gradient into the cytosol, providing a secondary and sustained elevation of Ca 2ϩ that initiated a protease cascade(s) and ultimately caused apoptosis after exposure to ␤-Lap. This is consistent with what we observed in NQO1-expressing cells after ␤-Lap treatment and co-administration of Ca 2ϩ chelators. Buffering intracellular Ca 2ϩ with BAPTA-AM partially abrogated all of the downstream events induced in MCF-7 cells by ␤-Lap (and thus prevented secondary Ca 2ϩ entry by buffering the initial rise in cytosolic Ca 2ϩ ). In contrast, extracellular chelation by EGTA only prevented those events initiated by secondary Ca 2ϩ entry (e.g. protease activation and DNA fragmentation). Thus, a secondary rise in intracellular Ca 2ϩ levels after exposure to ␤-Lap seems probable, and necessary, for protease activation and DNA fragmentation as was observed for TG-induced caspasemediated apoptosis (23,27). However, a secondary influx of Ca 2ϩ does not appear to be necessary for reduction in mitochondrial membrane potential or loss of intracellular ATP after ␤-Lap exposure, since EGTA did not prevent these responses.
Although MCF-7 cells treated with ␤-Lap had similar calcium responses, as do TG-exposed cells, ␤-Lap-exposed cells exhibited a very different pattern of apoptosis than TG-treated cells. ␤-Lap-exposed cells exhibit loss of intracellular ATP and a decrease in the [ATP]/[ADP][P i ] ratio. In contrast, TG-exposed cells did not exhibit loss of ATP (Fig. 5, and as reported by Ref. 53). Our data suggest that in contrast to TG where ATP-dependent caspase activation results in cell death (28,33,34,54), an ATP-independent protease is activated after exposure to ␤-Lap. Ca 2ϩ may regulate apoptosis by activating Ca 2ϩdependent protein kinases and/or phosphatases leading to alterations in gene transcription. However, with the rapid loss of intracellular ATP after exposure to ␤-Lap (2-4 h, Fig. 5), ␤-Lap-mediated cell death unlikely involves stimulated kinases or phosphatases or new protein synthesis. Instead, indirect kinase inhibition, due to ATP depletion, along with continued phosphatase activity is likely. Consistent with this notion, we found dramatic de-phosphorylation of pRb in cells exposed to ␤-Lap at 3 h (2), a time consistent with loss of ATP following exposure to this drug. Furthermore, loss of ATP at 2 h may also be responsible for inhibition of NF-B activation induced by tumor necrosis factor-␣ in ␤-Lap pre-exposed cells (55), since significant loss of ATP would prevent proteosome-mediated IB degradation. Thus, Ca 2ϩ -dependent loss of ATP in NQO1-expressing cells following ␤-Lap treatment may explain the reported pleiotropic effects of this agent.
␤-Lap-exposed cells also exhibited a very different pattern of substrate proteolysis compared with that observed after TG (2,12,28). We previously showed that ␤-Lap elicited a unique cleavage of PARP (ϳ60-kDa fragment), compared with the classical caspase-3-mediated fragmentation of the protein (ϳ89 kDa) observed after TG exposure (data not shown and Ref. 28). In a variety of NQO1-expressing cells exposed to ␤-Lap, atypical PARP cleavage was inhibited by the global cysteine protease inhibitors, iodoacetamide and N-ethylmaleimide, as well as the extracellular Ca 2ϩ chelators, EGTA and EDTA (12). In addition, ␤-Lap-mediated apoptotic responses were insensitive to inhibitors of caspases, granzyme B, cathepsins B and L, trypsin, and chymotrypsin-like proteases (12). In contrast, classic caspase inhibitors blocked TG-induced caspase activation and apoptosis (28). Caspase activation, as measured by procaspase cleavage via Western blot analyses, does not occur following ␤-Lap exposures. 3 Thus, protease activation after ␤-Lap treatment appears to be Ca 2ϩ -dependent, or alternatively, is activated by another protease or event that is Ca 2ϩdependent (Figs. 1-6 and Ref. 12).
Loss of Reducing Equivalents Is Also Necessary for ␤-Lapmediated Apoptosis, Similar to Menadione-mediated Apoptosis-Menadione is a quinone that can be detoxified by NQO1 two-electron reduction. However, menadione can also be reduced through two, one-electron reductions via other cellular reductases (56), thus eliciting menadione's toxic effects. Menadione toxicity, elicited via two, one-electron reductions, exhibited many similarities to ␤-Lap-mediated, NQO1-dependent, toxicity (5). These included: (a) elevations in cytosolic Ca 2ϩ (57,58); (b) NAD(P)H depletion (5,59,60); (c) ATP depletion (Ͻ0.1% control) 3 (61-63); and (d) mitochondrial membrane potential depolarization 3 (64). We previously demonstrated that menadione caused similar substrate proteolysis (p53 and atypical PARP cleavage) in NQO1-deficient cells, or at high doses in cells that express NQO1 where detoxification processes were over-ridden (5). 3 The semiquinone form of menadione can undergo spontaneous oxidation to the parent quinone (59,63,65,66); a pattern similar to the futile cycling observed after ␤-Lap bioactivation by NQO1 (5). Loss of reducing equivalents, such as NADH, due to the futile cycling of menadione may cause inactivation of the electron transport chain with the concomitant loss of mitochondrial membrane potential, and thus, loss of ATP (67,68). These responses were also observed in MCF-7 cells exposed to ␤-Lap (Figs. 4 and 5). Extensive mitochondrial Ca 2ϩ accumulation can also mediate mitochondrial depolarization (69,70). Thus, Ca 2ϩ sequestration may elicit mitochondrial membrane depolarization and consequent ATP depletion in cells exposed to ␤Ϫlap. These data further suggest that Ca 2ϩ is necessary for ␤-Lap-mediated cell death, but other factors are apparently needed for the initiation of the novel execution apoptotic pathway observed in cells treated with this compound.
The rise in intracellular Ca 2ϩ appears to be dependent on the bioactivation of ␤-Lap by NQO1, suggesting a critical and necessary signaling role for Ca 2ϩ in the downstream apoptotic pathway induced by this drug. These data suggest that DNA fragmentation, mitochondrial membrane depolarization, ATP loss, and apoptotic proteolysis were a consequence of the increase in intracellular Ca 2ϩ levels. Work in our laboratory is focused on elucidating the signaling response(s) that elicits ER Ca 2ϩ release following ␤-Lap bioactivation by NQO1. The cell death pathway induced by ␤-Lap is quite distinct from that observed after exposure to TG, and ␤-Lap-mediated apoptosis exhibited many similarities to menadione-mediated apoptosis. These observations further suggest that early release of Ca 2ϩ from ER stores, as well as influx of Ca 2ϩ from the extracellular milieu are necessary, but not sufficient for the novel apoptotic execution pathway induced by ␤-Lap. Thus, changes in Ca 2ϩ homeostasis in conjunction with the presumed loss of reducing equivalents are both necessary and sufficient for ␤-Lap-mediated apoptosis. We propose that development of ␤-Lap for treatment of human cancers that have elevated NQO1 levels (e.g. breast and lung) is warranted (6). Since most clinical agents used to date kill cells by caspase-dependent and p53-dependent pathways, and many cancers evade death by altering these pathways, development of agents that kill by specific targets (NQO1-mediated) and in p53-and caspase-independent manners are needed.