Cysteine Cathepsins Trigger Caspase-dependent Cell Death through Cleavage of Bid and Antiapoptotic Bcl-2 Homologues*

As a model for defining the role of lysosomal cathepsins in apoptosis, we characterized the action of the lysosomotropic agent LeuLeuOMe using distinct cellular models. LeuLeuOMe induces lysosomal membrane permeabilization, resulting in release of lysosomal cathepsins that cleave the proapoptotic Bcl-2 family member Bid and degrade the antiapoptotic member Bcl-2, Bcl-xL, or Mcl-1. The papain-like cysteine protease inhibitor E-64d largely prevented apoptosis, Bid cleavage, and Bcl-2/Bcl-xL/Mcl-1 degradation. The pancaspase inhibitor N-benzyloxycarbonyl-Val-Ala-Asp(OMe)fluoromethyl ketone failed to prevent Bid cleavage and degradation of anti-apoptotic Bcl-2 homologues but substantially decreased cell death, suggesting that cathepsin-mediated apoptosis in these cellular models mostly follows a caspase-dependent pathway. Moreover, in vitro experiments showed that one or more of the cysteine cathepsins B, L, S, K, and H could cleave Bcl-2, Bcl-xL, Mcl-1, Bak, and BimEL, whereas no Bax cleavage was observed. On the basis of inhibitor studies, we demonstrate that lysosomal disruption triggered by LeuLeuOMe occurs before mitochondrial damage. We propose that degradation of anti-apoptotic Bcl-2 family members by lysosomal cathepsins synergizes with cathepsin-mediated activation of Bid to trigger a mitochondrial pathway to apoptosis. Moreover, XIAP (X-chromosome-linked inhibitor of apoptosis) was also found to be a target of cysteine cathepsins, suggesting that cathepsins can mediate caspase-dependent apoptosis also downstream of mitochondria.

The major mechanism for removal of superfluous and potentially dangerous cells in eukaryotic organisms is apoptotic cell death. One of the central points of regulation of apoptotic cell death is mitochondria. They serve as sensors of cellular damage, and after mitochondrial membrane permeabilization (MMP) 4 they release a number of factors, such as cytochrome c and Smac/Diablo, which are critically involved in cell death signaling (1,2). As such, they are also centrally involved in the activation of caspases, a family of cysteine proteases that cleave a subset of cellular proteins, thereby signaling the execution phase of the apoptotic cell death program. Although caspases can also be activated in the absence of mitochondria after death receptor activation, in type 2 cells (cells in which Fas-mediated apoptosis can be blocked by overexpression of Bcl-2) the signal is amplified through caspase 8-mediated activation of Bid and subsequent MMP, thereby linking the extrinsic (death receptor pathway) and the intrinsic (mitochondrial) pathways (3). Regulation of apoptosis is achieved through the action of pro-and antiapoptotic Bcl-2 family members upstream of mitochondria (4,5) and by direct inhibition of caspases-3, -7, and -9 by XIAP (6,7).
In addition to mitochondria, there is considerable evidence that also other organelles, such as lysosomes and endoplasmic reticulum, are tightly linked with cell death (8,9). Lysosomes are acidic organelles containing a wide spectrum of hydrolytic enzymes that have a major role in intracellular protein recycling. Among them the most abundant and the best characterized are the cathepsins (10 -13). Most of the cathepsins are cysteine proteases, such as cathepsins B, L, S, V, C, F, K, X, and H, whereas cathepsin D is an aspartic protease, and cathepsin G a serine protease (12). Although cysteine cathepsins were long believed to be unstable at neutral pH, it is now clear that a number of them remain active at neutral pH for a certain amount of time, varying from minutes (cathepsin L (14)) to hours (cathepsins S (12)).
Initially lysosomes and lysosomal proteases were believed to be mainly involved in terminal steps of necrotic and autophagic cell death (15,16). However, more recently they were also found to be associated with apoptotic cell death (17)(18)(19)12). A critical step in the pathway is lysosomal rupture, often referred to as small-scale lysosomal leakage, which is followed by the release of lysosomal cathepsins into the cytosol (12,20). It has been * This work was supported by Ministry of Higher Education, Science, and shown that lysosomes are particularly sensitive toward oxidative stress (21)(22)(23)(24). In addition, LMP can be induced by anticancer agents, such as siramesine, a novel sigma-2 receptor ligand (25), cardenolide (UNBS1450) (26), biphosphinic palladacycle complex (27) and novel compounds that target p53independent apoptosis (28), by detergents including LeuLeuOMe, which was shown to have a protective effect in a model of graft versus host disease (29) and is currently in phase II clinical trials, and by a number of other stimuli (17, 30 -33).
The pathways downstream of lysosomal rupture are less well clarified. Several studies suggest that lysosomal membrane permeabilization-induced cathepsin leakage leads to MMP and subsequent release of proapoptotic factors followed by caspase activation (19, 31, 32, 34 -36), although some studies suggested that cell death could occur also without apparent activation of caspases (18,26,37,38). Initial studies in a cell-free system suggested Bid cleavage by lysosomal cathepsins as a possible link, as cathepsins failed to activate executioner caspases directly (34). This was later confirmed in several cellular models such as LeuLeuOMe-induced apoptosis in HeLa cells, with multiple cysteine cathepsins suggested to be responsible for the cleavage of Bid (32,39). However, apoptosis was found to be induced also in the absence of Bid cleavage in cellular models (35) and in an in vivo model of progressive myoclonus epilepsy (40), suggesting that cathepsins can induce apoptosis through cleavage of other cytosolic proteins, which however, remained unidentified.
To address these questions we have focused on the mechanism of cell death triggered by the lysosomotropic reagent LeuLeuOMe, currently in phase II clinical trials for allogeneic hematopoietic stem cell transplantation in a number of different cell lines. Lysosomal rupture was found to precede mitochondrial membrane permeabilization accompanied by major ultrastructural changes of mitochondria and subsequent cytochrome c release. In addition to cleavage of Bid in different cell models, we show here that the antiapoptotic proteins Bcl-2, Bcl-x L , and Mcl-1 are also substrates for cathepsins. Moreover, XIAP was also found to be a target of multiple cysteine cathepsins in several cell lines, suggesting that cathepsins can mediate caspase-dependent apoptosis both upstream and downstream of mitochondria.

EXPERIMENTAL PROCEDURES
Materials-The following antibodies were used: anti-Bcl-x L (H-62), anti-Bcl-2 (C-2), and anti-Bak (H-211) from Santa Cruz Biotechnology, anti-Mcl-1 (RC13) from Abcam, anti-cytochrome c from BD PharMingen, anti-actin from Sigma, anti-PARP antibodies recognizing caspase-cleaved PARP from Promega, anti-Bid from R&D Systems, anti-XIAP from BD Transduction Laboratories, and anti-cathepsin L, which were prepared in the laboratory (41). All horseradish peroxidaseconjugated secondary antibodies were from Sigma. M-PER mammalian protein extraction reagent was purchased from Pierce. The Bradford reagent was from Bio-Rad. The ECL detection kit was from Amersham Biosciences. E-64d was purchased from Peptide Inc., z-VAD-fmk, Ac-DEVD-trifluoromethylcoumarin and AC27P were from Bachem, whereas pepstatin A, CA-074 methyl ester (CA-074Me), and LeuLeuOMe were from Sigma. Cathepsins B, L, S, K, and H were prepared as described previously (42)(43)(44)(45). Annexin V-PE and 7-AAD for flow cytometry were purchased from BD Biosciences. Fluorescent organelle-specific probes Mitotracker CMXRos and Lysotracker Green  Induction of Apoptosis, Cell Lysis, and Western Blotting-Cells were cultured at 1 ϫ 10 6 cells/ml in 6-well plates overnight before the treatment with the lysosomotropic agent LeuLeuOMe at the final concentration of 1-2 mM depending on the cell line. In control experiments cells were incubated overnight in the presence of 20 M E-64d, a broad spectrum cathepsin and calpain inhibitor, 20 M z-VAD-fmk, a broad spectrum caspase inhibitor, 50 M pepstatin A, a general aspartic protease inhibitor, 15 M CA-074Me, a specific cathepsin B inhibitor, and 50 M AC27P, a specific calpastatin-derived calpain inhibitor before the addition of LeuLeuOMe. After 12-17 h cells were observed by light microscopy (Olympus IX71 magnification 40 and 60).
To prepare the extracts, cells were collected, pelleted by centrifugation at 1300 rpm for 5 min, and washed twice with phosphate-buffered saline. Whole-cell extracts were prepared in radioimmune precipitation assay buffer (50 mM Tris, pH 8.0, 100 mM NaCl, 0.1% (w/v) SDS, 1% (v/v) Nonidet P-40, 0.5% (w/v) deoxycholic acid, 1 mM EDTA) or M-PER mammalian protein extraction reagent. After 10 min of incubation on ice, insoluble material was removed by centrifugation at 14,000 rpm for 10 min. Cytosolic extracts were prepared as previously described (32). Total protein concentration was determined using the Bradford assay.
Equal amounts of protein were loaded and resolved in 15% or 12.5% SDS-PAGE gels and electrotransferred to nitrocellulose membranes. Proteins were then visualized with ECL according to the manufacturer's instructions. To confirm equal protein loading, all immunoblots were also probed with actin. All experiments were repeated at least three times.
Apoptosis Quantification-Apoptosis was quantified by flow cytometry measurements of phosphatidylserine exposure and 7-AAD incorporation to measure membrane integrity and by measurements of the DEVDase activity of the caspases. Briefly, 100-l aliquots of cells were labeled with annexin V-PE and 7-AAD according to the manufacturer's instructions. The cells were then subjected to flow cytometry analysis using a FACScalibur flow cytometer (BD Biosciences) and analyzed with the CellQuest software. Alternatively, 50 g of protein of untreated and LeuLeuOMe-treated cells in the presence or absence of inhibitors were used to determine caspase activity by measuring the proteolytic cleavage of the fluorogenic substrate Ac-DEVD-trifluoromethylcoumarin (Bachem) as previously described (32).
Fluorescence Microscopy and Flow Cytometry Quantification of Mitochondria and Lysosomes-The integrity of mitochondria and lysosomes was monitored based on uptake of Mitotracker Red CMXRos and Lysotracker Green DND-26, respectively. Briefly, Mitotracker CMXRos was added to the cells at a final concentration of 30 nM. After 10 min of incubation at 37°C, cells were viewed with a fluorescence microscope (Olympus IX71; excitation 516 -555 nm, emission 574 -648 nm). In parallel, red fluorescence of 5000 cells per sample, corresponding to mitochondria, was determined by flow cytometry using the FL3 channel. Alternatively, Lysotracker Green DND-26 was added to the cells at a final concentration of 35 nM, and cells were observed after 5 min of incubation at 37°C under fluorescence microscope (excitation 461-500 nm, emission 511-560 nm). Green fluorescence of 5000 cells per sample, corresponding to lysosomes, was determined by flow cytometry using the FL1 channel.
Transmission Electron Microscopy-Growth medium was removed from the attached cells followed by a washing step with phosphate-buffered saline. Cells were then fixed with 4% (v/v) paraformaldehyde and 2% (v/v) glutaraldehyde, post-fixed in 1% OsO 4 , dehydrated in graded ethanol, and embedded in Epon812. Ultrathin sections were cut, counterstained with uranyl acetate and lead citrate, and then viewed in Philips CM 100 electron microscope. Essentially the same procedure was used also for detached cells, which were gently spun down before the washing step with phosphate-buffered saline.
DNA Constructs-The expression constructs for Bcl-xL (pcDNA3-FLAG-Bcl-xL) and Bak (pcDNA3-HA-Bak) were a gift from Dr. Lawrence Banks (The International Centre for Genetic Engineering and Biotechnology), whereas pCAGGSh-Bcl-2 was a gift from Dr. Peter Vandenabeele (University of Ghent) (accession number LMBP 4633). Bcl-2 cDNA was excised from pCAGGS using the two EcoRI sites and cloned into the pcDNA3 (Promega) vector at the same site. Bax, Bim EL, and Mcl-1 clones were all a gift from Dr. David Huang (The Walter and Eliza Hall Institute). Expression construct for Bax (pcDNA3-Bax) was made by excising Bax cDNA from pBSIISK using EcoRI and KpnI and cloning it into pcDNA3 at the same sites. Bim EL was cloned into pcDNA3 at EcoRI and NotI restriction sites after excising its cDNA from pBSIISK. Mcl-1 was cloned into pcDNA3 at cloning sites HindIII and EcoRI after excising Mcl-1 cDNA from pEF-FLAG.
In Vitro Transcription/Translation and in Vitro Cleavages of the Bcl-2 Homologues-In vitro transcription/translation of the Bcl-2 homologues Bcl-2, Bcl-xL, Bak, and Bax was performed using wheat germ lysate-TNT T7 Coupled Extract System and [ 35 S]methionine. The efficiency of expression was monitored by analyzing the translates by SDS-PAGE and autoradiography.
Cleavage assays of Bcl-2 homologues with different cathepsins were performed as described previously (32). The reactions were stopped by adding SDS sample buffer and boiling and analyzed by SDS-PAGE and autoradiography.

LeuLeuOMe Triggers Cathepsin-dependent Apoptosis in
Eight Different Cell Lines-LeuLeuOMe accumulates in lysosomes, where it is converted to a membranolytic compound causing loss of lysosomal membrane integrity and subsequent apoptosis (46,47). LeuLeuOMe was initially shown to be extremely effective in killing various immune and myeloid tumor cells of bone-marrow origin and to efficiently protect against graft-versus-host disease in a mouse model of bone marrow transplantation (29,48). Later studies suggested that other cell types are susceptible to the compound but with lower sensitivity (32,49). Although the mechanism was not clarified, in HeLa cells apoptosis was found to proceed through cathepsin-dependent Bid cleavage and subsequent caspase activation (32). To investigate the generality of this mechanism, the effect of LeuLeuOMe was investigated on eight cancerous or noncancerous cell lines from different tissues to ensure wide coverage. Initially, all cell lines were incubated with 0.25-5.0 mM LeuLeuOMe for 10 -24 h followed by measurements of lactate dehydrogenase release using CytoTox-One TM assay (Promega) to investigate for its ability to induce cell death. All cells were found to be sensitive to LeuLeuOMe, which at higher concentrations and prolonged incubation induced necrotic cell death (not shown). LeuLeuOMe at 1 mM was found optimal to induce apoptotic cell death in all cell lines except in HeLa and CaCo-2 cells, where 2 and 3 mM concentration, respectively, was needed for the same effect. The optimal incubation time was found to be between 12-17 h, depending on cell line, and these conditions were used in all subsequent experiments. All cells exposed to LeuLeuOMe exhibited typical apoptotic morphology, such as cell shrinkage, rounding, and detachment from the surface, as observed by light microscopy (supplemental Fig. 1). In agreement with a previous study (32), E-64d and to a considerable extent also z-VAD-fmk, blocked apoptotic morphology, whereas the aspartic protease inhibitors pepstatin A and the specific calpastatin inhibitor AC27P had no effect on cellular morphology as well as on the increase of DEVDase activity (data not shown). Apoptotic cell death was confirmed by flow cytometry studies of phosphatidylserine exposure and membrane integrity ( Fig. 1A; supplemental Fig. 2), measurements of DEVDase activity representative of caspase activation (Fig. 1B), and by the generation of a specific PARP fragment, generated by caspases during apoptosis (Fig. 1C). Typically, 15-25% cells were apoptotic based on flow cytometry measurements ( Fig. 1A; supplemental Fig. 2). Increase in the percentage of apoptotic cells, caspase activity, and PARP cleavage could be completely prevented by both E-64d and z-VAD-fmk (Fig. 1, A-C), consistent with an essential role of cysteine cathepsins and the caspases in this model of cell death.
A further goal was to address the contribution of cathepsin B, the most abundant cysteine cathepsin in most cells, to LeuLeuOMe-induced apoptosis using a selective cell-permeable cathepsin B inhibitor CA-074Me. As can be seen in Fig. 1D, pretreatment of SH-SY5Y and HaCaT cells with CA-074Me diminished the percentage of apoptotic cells, in agreement with the suggestion that cathepsin B contributes to LeuLeuOMeinduced cell death.
LeuLeuOMe Induces Lysosomal Disruption Accompanied by Leakage of Cathepsins into the Cytosol-Release of lysosomal proteases to the cytosol is another characteristic of LeuLeuOMe-induced cell death (32). In agreement with these results, cathepsin L was detected in the cytosolic extracts of all cells after treatment with LeuLeuOMe, and this could not be prevented by any of the inhibitors including E-64d and z-VAD-fmk ( Fig. 2A). Moreover, using an intracellular fluorogenic general cathepsin substrate, MR-(FR) 2 , which enables detection of cathepsin activity in whole live cells by fluorescent microscopy, cathepsin activity could also be detected in the cytosol of whole live cells after LeuLeuOMe treatment, suggesting that lysosomal cathepsins retained activity in this location (not shown).
Because the Western blots did not provide any quantitative data, the next step was to quantify the number of cells with damaged lysosomes using LysoTracker Green DND-26, which enables detection by flow cytometry and also by fluorescent microscopy. As shown in Fig. 2B, control cells exhibited a punctate green fluorescent pattern of LysoTracker Green, suggesting that the dye accumulated in lysosomes (Fig. 2B, lower  panel). LeuLeuOMe treatment resulted in appearance of cells with substantially decreased green fluorescence due to the loss of lysosomal integrity in a number of green fluorescent cells ranging from 15% (SH-SY5Y) to 60% (NHDF), whereas in most of the others 25-30% pale cells were observed ( Fig. 2B; supplemental Fig. 3). Pretreatment of cells with either E-64d or z-VAD-fmk did not rescue lysosomal permeabilization, coinci- dent with the appearance of cathepsins in the cytosol. Very similar results were obtained also with acridine orange (not shown).
LeuLeuOMe Induces Cytochrome c Release from Mitochondria and Loss of Mitochondrial Membrane Potential-Because lysosomal disruption was found to precede mitochondrial cytochrome c release after LeuLeuOMe treatment in HeLa cells (32), we next investigated cytochrome c release and loss of mitochondrial membrane potential in other cells. In agreement with the earlier study, cytochrome c was found in the cytosol in all cells (Fig. 3A). The release was completely prevented by E-64d, but not by z-VAD-fmk, consistent with cysteine cathepsins acting upstream of caspases in this model.
In the next step the loss of ⌬⌿m was monitored using the mitochondria-specific dye Mitotracker Red CMXRos, which decreases its red fluorescence upon breakdown of ⌬⌿m and enables quantification of cells with damaged mitochondria. The number of cells with decreased red fluorescence increased to 19 -32% after LeuLeuOMe treatment ( Fig. 3B; supplemental Fig. 4), correlating with the number of cells with significantly damaged lysosomes (see above). Preincubation with E-64d, but not z-VAD-fmk, prevented loss of ⌬⌿m, consistent with other results (see above) indicating that cysteine cathepsins act upstream of caspases in this model. This decrease of red fluorescence in LeuLeuOMe-treated cells is also clearly seen by fluorescence microscopy (Fig. 3B, lower panels; supplemental Fig. 4, insets). Equivalent results were obtained with the potentiometric dye tetramethylrhodamine ester perchlorate (data not shown).
Having shown a major loss of ⌬⌿m in this model, we next studied changes of mitochondrial ultrastructure during LeuLeuOMe-triggered apoptosis by transmission electron microscopy. In all cells exposed to LeuLeuOMe a large number of dysfunctional mitochondria with swollen morphology and loss of cristae and inner structure, indicative of inner membrane permeabilization, were found (Fig. 3C). In the presence of E-64d, mitochondrial ultrastructure and integrity were preserved, but not in the presence of z-VAD-fmk, suggesting that mitochondrial breakdown is a critical step in the pathway.

The Bcl-2 Family Members Bcl-2, Bcl-xL, Mcl-1, and Bid Are Substrates of Cysteine Cathepsins in LeuLeuOMe-triggered Cell
Death-To identify possible links between lysosomes and mitochondria, we initially focused on Bid, which was found previously to be cleaved in HeLa cells after LeuLeuOMe treatment (32). In agreement with these results Bid was found to be cleaved into a ϳ15-kDa truncated variant (tBid) in MCF-7, HaCaT, SH-SY5Y, CaCo-2, and HepG2 cells, and this was prevented by E-64d but not by z-VAD-fmk ( Fig. 4A; supplemental Fig. 5), suggesting that Bid is a general cathepsin target in apoptosis. However, in HEK293 cells tBid could not be detected, whereas in NHDF cells Bid appears to be degraded (supplemental Fig. 5), suggesting that in addition to Bid there are other cellular targets of cathepsins important for cell death progression.
Other Bcl-2 homologues in addition to Bid are known to be critical regulators of cell death upstream of mitochondria (50 -52), so we examined cleavage of the Bcl-2 homologues Bcl-2, Bcl-x L , and Mcl-1 in the eight cell lines. As seen in Fig. 4B, at least one of those proteins was found to be cleaved in most of the cell lines investigated, although it should be noted that Bcl-2 and Bcl-x L were not detected in some cell lines. A band corresponding to full-length Bcl-2 disappeared in HeLa, MCF-7, and

SH-SY5Y cells, Mcl-1 disappeared in MCF-7 and SH-SY5Y cells, and
Bcl-x L disappeared in HaCaT, NHDF, and CaCo-2 cells. We show only the regions of the Western blots containing full-length proteins, because no smaller fragments could be detected for any of them. It is very likely that they were degraded in the cells examined under the conditions used or that the cleaved fragments are not recognized by the antisera. None of the three was found to be cleaved in HepG2 and HEK293 cells, although they were all detected in both cell lines (Table 1). Similar to Bid cleavage, degradation of Bcl-2, Bcl-x L , and Mcl-1 could be prevented by E-64d but not by z-VAD-fmk (Fig. 4B), suggesting that cysteine cathepsins were responsible for the cleavage. In an additional experiment aimed at preventing apoptosis, Bcl-2 was transiently overexpressed in MCF-7 and SH-SY5Y cells. However, no significant decrease in LeuLeuOMe-induced apoptosis was observed in transfected cells, although the levels of Bcl-2 were clearly increased in the transfected cells (Fig. 4C).
In further attempts to identify additional cathepsin substrates, Bak and Bim were also tested, but they were not found to be processed or degraded in any of the cellular models tested (supplemental Fig. 6) Although Bax would also be a suitable candidate, it was not tested, as it was not found to be cleaved by any of the cathepsins in vitro (see below).
XIAP Is Substrate of Cysteine Cathepsins in LeuLeuOMe-triggered Cell Death-Because quite a few antiapoptotic proteins were found to be cathepsin targets, we next investigated XIAP, the major inhibitor of caspases, which regulates apoptosis downstream of mitochondria, as another potential cathepsin target. Indeed, the intensity of the 54-kDa XIAP band was found to be reduced in both cell lines tested, SH-SY5Y and CaCo-2, after LeuLeuOMe treatment in an E-64d-dependent manner (Fig. 5), suggesting that XIAP was partially degraded by cysteine cathepsins.
Bcl-2 Family Proteins and XIAP Are Cathepsin Substrates in Vitro-To validate the cellular data and to identify the cathepsin(s) responsible for the cleavage of Bcl-2 homologues and  (Fig. 6), thereby confirming cellular data. Moreover, with the exception of Bcl-x L , which was only degraded by cathepsin L, all other proteins were found to be cleaved by multiple cathepsins, suggesting that cathepsins have redundant roles in the system.

DISCUSSION
The lysosomal/endosomal system, which has a major role in intracellular protein degradation and recycling, has been increasingly linked with cell death processes. Although this may to some extent be due to the presence of highly potent were immunoblotted for Bcl-2, Bcl-xL, and Mcl-1 detection. Degradation of Bcl-2 after lysosomal disruption was detected in MCF-7 and SH-SY5Y, and E-64d but not z-VAD-fmk prevented the degradation. Bcl-xL degradation occurred in HaCaT cells after treatment with LeuLeuOMe and could be prevented by E-64d but not by z-VAD-fmk. Degradation of Mcl-1 was observed in MCF-7 cells. Actin was used as the loading control. C, after transfection with Bcl-2, MCF-7 cells were collected and stained with annexin V-PE and 7-AAD for flow cytometry analysis. Both transfected and nontransfected cells were treated with LeuLeuOMe (2, 4), whereas in control experiments LeuLeuOMe treatment was omitted (1, 3). Each plot represents a typical experiment of 5000 cells. Immunoblotting shows overexpression of Bcl-2 after transfection (right, lanes 3 and 4).
proteases in lysosomes leading to extensive cell damage, more controlled or limited damage of lysosomes has been suggested to lead to apoptotic cell death (20,36,53). Here we show that limited damage of lysosomes using the lysosomotropic compound LeuLeuOMe results in translocation of active cathepsins to the cytosol followed by MMP, cytochrome c release, and subsequent caspase activation (Figs. 1-3). Essentially the same mechanism was observed in eight unrelated cell lines, including those of fibroblast, keratinocyte, hepatocyte, neuroblastoma, and breast cancer origin, suggesting that the mechanism used by LeuLeuOMe is likely highly conserved. The only distinction seems to be the different sensitivity of cells toward LeuLeuOMe, where NK cells and other bone marrow-derived cells were considerably more sensitive from the others (47)(48)(49) with no difference found between cancerous and noncancerous cell lines. Similar findings, i.e. LMP leading to MMP and subsequent caspase activation, were obtained by different lysosomotropic stimuli in different cellular models, such as by photodamage of murine hepatoma 1c1c7 cells (54), hydroxychloroquine, ciprofloxacin, or norfloxacin in HeLa and Jurkat cells and in various mouse embryonic fibroblasts (31,35) and in reactive oxygen species-induced cell death in neutrophils in a model of bacteria infection (39). These findings lead us to conclude that the mechanism of lysosomal cell death is substantially conserved downstream of LMP regardless of the model used.
Most of the pathways involving LMP converge at the level of mitochondria, raising a critical question about the possible links between LMP and MMP. Our data suggest that cysteine cathepsins have a major role in LeuLeuOMe-induced apoptosis through proteolytic activation of Bid and degradation of antiapoptotic Bcl-2 family members Bcl-2, Bcl-x L , and Mcl-1 as well as the caspase inhibitor XIAP. However, to understand how cleavage of Bcl-2 proteins by cathepsins can trigger cell death, one has to consider to the function of Bcl-2 proteins as pro-and antiapoptotic regulators (Fig. 7). Antiapoptotic Bcl-2 family proteins represent a critical control point on the apoptotic pathway upstream of mitochondria, and their function is antagonized by "BH3-only" pro-apoptotic members of the family (Bid, Bad, Bim, Puma, Noxa, etc.). Bax and Bak integrate the progression of apoptosis, although based on gene ablation studies, they may have at least partially redundant roles despite their different localization in viable cells (5). The current theory suggests that the BH-3 only proteins induce Bak and Bax to permeabilize the outer mitochondrial membrane, resulting in release of various proapoptotic factors, including cytochrome c, Smac/ DIABLO, and the serine protease Omi/HTRA2 (2). The apoptotic switch is often achieved by transcriptional activation of BH3-only proteins, such as Puma, Noxa, and Bim (4) or, apropos to our study, by direct proteolytic activation of Bid, which FIGURE 5. LeuLeuOMe-triggered lysosomal disruption resulted in XIAP degradation. 50 and 75 g of extract from SH-SY5Y and CaCo-2 cells, respectively, untreated (controls) and treated were immunoblotted for XIAP are shown. Degradation of XIAP after lysosomal disruption triggered by LeuLeuOMe could be prevented by E-64d but not by z-VAD-fmk. Actin was used as a loading control.

TABLE 1 Cleavage of the Bcl-2 homologues in distinct cell lines following LeuLeuOMe-induced LMP
Experimental conditions were as described under "Experimental Procedures." (ߛ), degraded; -, intact; ND, not detected when 75-100 g of protein was loaded.  directly activates Bax and Bak, thereby bypassing the prosurvival Bcl-2 protein checkpoint (55).

HeLa MCF-7 SH-SY5Y HaCaT CaCo-2 NHDF HEK293 HepG2
Bid clearly seems to be an important target of cysteine cathepsins (Fig. 4A), which is in agreement with other in vitro (32,34) or cellular studies (32,39,54). Although Reiners et al. (54) were not able to prevent tBid generation with M concentrations of the cathepsin inhibitor z-FA-fmk, it is likely that this concentration of inhibitor was insufficient to block the cathepsins as Blomgran et al. (39) were able to block Bid cleavage in neutrophils with a substantially higher concentration of the same inhibitor. The latter report is in agreement with our results demonstrating that 15 M or higher concentrations of E-64d, which is a more efficient inhibitor of cysteine cathepsins than Z-FA-fmk (56), completely abolished Bid cleavage.
Our data lead to the suggestion that degradation of prosurvival Bcl-2 proteins by itself can induce apoptosis, in a sense resembling the mechanism used by BH3-only proteins, i.e. sequestration of prosurvival Bcl-2 proteins. However, whereas a few BH3-only proteins such as Bad and Noxa have limited ability to sequester different prosurvival Bcl-2 homologues (57,58), cathepsins show much less selectivity, as a number of them can degrade all of the Bcl-2 homologues in vitro (Fig. 6), resembling more Bim and Puma. This suggests that cathepsins are redundant in the system. Cathepsin L appears to be the most potent among them (Fig. 6), although under in vivo conditions this could be partially compromised with the low stability of cathepsin L at neutral pH (12,14), suggesting that cathepsin B also plays an important role due to its abundance and reasonably good stability. This is also supported by the finding that the selective cathepsin B inhibitor CA-074 diminished apoptosis in several cell lines (Fig. 1D) and by previous studies that demonstrate the importance of cathepsin B for apoptosis in other models (17,18). Degradation of prosurvival Bcl-2 proteins and apoptosis progression could be also amplified through their sequestration by BH3-only proteins, such as Bim, which is known to be often transcriptionally up-regulated during apoptosis (5). However, no difference in Bim level was observed, suggesting that cathepsins are the major players upstream of mitochondria in our model.
Cysteine cathepsins can also be involved in apoptosis regulation downstream of mitochondria by degradation of XIAP, the major endogenous inhibitor of caspases (7). Intracellular inhibitors of cathepsins (stefins and serpins; Refs. 59 and 60) probably represent one of the major regulatory sites in lysosomal disruption paradigms by providing a threshold to prevents adventitious proteolytic activity. At least in LeuLeuOMe-induced cell death this barrier is relatively high, since a considerable number of lysosomes in a cell need to be damaged to enter apoptosis, and minor damage seems to be repaired probably by autophagy. 5 Although XIAP degradation by cathepsins may be less efficient than XIAP neutralization by Smac/Diablo and Omi, it is likely that cathepsins amplify the effect of the latter, thereby lowering the apoptotic threshold.
Although cathepsins are relatively powerful and nonselective proteases, they are not able to kill cells in the absence of caspases, as demonstrated by the use of z-VAD-fmk at low concentrations. The most plausible explanation is that cathepsins are slowly inactivated in the cytosol by pH-induced unfolding and/or oxidation and are, therefore, active for a limited time, which enables them to cleave only a small subset of proteins. This is consistent with the finding that none of the proapoptotic proteins with the exception of Bid was found to be cleaved by cathepsins during apoptosis. Bax was not found to be cleaved at all, whereas Bak and Bim, which were found to be degraded in vitro by different cathepsins (Fig. 6), were not found to be cleaved in any of the cell lines. At least for Bak this could be explained by its membrane localization masking the cleavage site(s). In contrast to cathepsins, caspases are designed to work in the cytosolic conditions for a prolonged time, which enables them to efficiently process proteins and execute the cell death program. Moreover, caspases differ markedly from the cathepsins in their ability to process protein substrates. Caspases as extremely specific proteases cleave their substrates only at a small number of sites, thereby clearly modulating the function of substrates by limited proteolysis (61). In contrast, cathepsins are primarily degrading enzymes with a broad specificity (62) and destroy their substrates, although exceptions such as limited proteolysis of Bid leading to its activation are also known. Cathepsins and caspases, thus, nicely demonstrate two different ways of protease signaling, i.e. caspases often activating other proteins by limited proteolysis and cathepsins primarily inactivating them by stepwise degradation (63).
Based on these results we propose the following mechanism of lysosomal pathway to apoptosis (Fig. 7). After lysosome destabilization, active cysteine cathepsins are released in cytosol, where they process Bid and/or degrade prosurvival Bcl-2 homologues, thereby liberating the pro-apoptotic activity 5 U. Repnik, S. Ivanova, and B. Turk, unpublished data. of Bak and Bax. This is then followed by MMP and subsequent cytochrome c release, which leads to caspase activation and execution of cell death program. At the same time, cathepsins amplify the apoptotic signal by degrading XIAP. This is still a simplified model, and cathepsins probably process several other protein substrates during lysosome-induced apoptosis. However, this is probably not a very big number or they are not critical, as cells largely survive when caspases are blocked. Moreover, the number of substrates cleaved varies from cell to cell, and not all are required to be cleaved for successful apoptosis progression. It is clear that lysosomal cathepsins primarily have a critical role in the initiation of the death signaling, whereas the executioner role lies with the caspases and, therefore, such cell death cannot be called caspase-independent cell death. This mechanism could nicely explain the data of Blomgran et al. (39), who found that cathepsin-mediated Bid cleavage was associated with down-regulation of Mcl-1 in reactive oxygen species-mediated neutrophil apoptosis. Such a mechanism could probably also explain cell death induced by cathepsins in the apparent absence of Bid cleavage such as in cathepsin B-dependent cerebellar apoptosis in stefin B-deficient animals (40), although an experimental confirmation is still lacking.
However, this mechanism does not explain all the models of lysosomal cell death, including those where cathepsin D has been suggested to play a major role and where cysteine cathepsins are implicated in the tumor necrosis factor-␣ pathway (17, 64 -66). It is, therefore, likely that more cellular substrates of cathepsins exist which await identification and which would help us to understand the lysosomal pathways to cell death. However, it is clear that cathepsins play a major role in several processes, one of them likely being reactive oxygen speciesmediated cell death in neurodegeneration and aging (12,19).
In conclusion, we describe how cysteine cathepsins are major initiator proteases in lysosome-mediated cell death. After their release from the lysosomes, they were found to activate caspases indirectly through degradation of antiapoptotic Bcl-2 family members and proteolytic activation of Bid, engaging the mitochondrial pathway of apoptosis. Moreover, cathepsins were found also to cleave XIAP, suggesting that they are involved in the control of apoptosis both upstream and downstream of mitochondria. Furthermore, because increased levels of antiapoptotic Bcl-2 members and inhibitors of apoptosis were found in a number of cancers (67,68), these results could explain why disrupting lysosomes is a very potent way of cancer cell elimination, especially since cysteine cathepsins are also often up-regulated in cancer (13).