The metastasis suppressor gene C33/CD82 ... - The FASEB Journal

39 downloads 43 Views 2MB Size Report
Jan 13, 2004 - The metastasis suppressor gene C33/CD82/KAI1 induces apoptosis through reactive oxygen intermediates. Nicole Schoenfeld, Manuel K. A. ...
The FASEB Journal express article 10.1096/fj.03-0420fje. Published online Novermber 3, 2003.

The metastasis suppressor gene C33/CD82/KAI1 induces apoptosis through reactive oxygen intermediates Nicole Schoenfeld, Manuel K. A. Bauer, and Stefan Grimm Max-Planck-Institute for Biochemistry, 82152 Martinsried, Germany Corresponding author: Stefan Grimm, Max-Planck-Institute for Biochemistry, 82152 Martinsried, Germany. E-mail: [email protected] ABSTRACT Here we describe the isolation of C33/CD82/KAI1 in a screen for apoptosis-inducing genes. C33 is a gene that is downregulated in many metastatic tumor cells and the expression of which can attenuate the process of metastases formation in a variety of tumors. In accordance, we observed cell death induction by C33 in many different cell types. C33 seems to promote cell death by the generation of reactive oxygen intermediates (ROIs). These ROIs, however, are not derived from the mitochondrial respiratory chain as in most other scenarios leading to apoptosis. We observed that C33 renders cells sensitive to ROIs by causing the specific release of the intracellular antioxidant glutathione (GSH) from cells. Moreover, C33 activates the GTPase Cdc42, which mediates GSH release and apoptosis induction and allows to detect the formation of ROIs. Key words: cell death • metastasis suppressor gene • reactive oxygen intermediates • glutathione • Cdc42

A

poptosis induction and tumorigenesis are intimately linked. Since its initial description in tumor cells (1), the genetic program of apoptosis has been regarded as a determining factor in the progression of tumors. The downregulation of cell death observed in many different malignant cells is often caused by the activation of dominant oncogenes. Mutations in their counterparts, the tumor suppressor genes, confer a selective advantage to tumor cells as some of them mediate a signal for apoptosis and thereby protect the organism against cells undergoing malignant transformation (2). Inactivation of these genes disturb the homeostatic balance of cellular death and survival resulting in resistance to the many adverse conditions tumor cells have to withstand. Metastases formation is a decisive step in tumorigenesis that determines the clinical course of the disease. Many efforts have been directed toward a molecular understanding of this process, but only recent investigations have implicated some genes in this pathological condition (3). It is therefore especially important to understand how they achieve their effect. C33, a member of the tetraspanin protein-family (4), has been isolated because of its frequent downregulation in metastasis-forming prostate cancer cells (5). Its transfection into these cells leads to a marked reduction of metastases. Subsequently, it was shown that C33 is transcriptionally inactive in many other metastatic tumors and human cancer cell lines (5–9). Although its effect on metastases formation remained enigmatic, its localization in the plasma membrane suggests cellcell or cell-matrix-interactions to be involved (5).

Here we show that the metastasis suppressor gene C33 is a direct apoptosis inducer that causes the formation of ROIs for cell death. MATERIAL AND METHODS Materials The pan-caspase inhibitor zVAD was purchased from Enzyme Systems Products. Glutathionemonoethyl ester was from Bachem. All fine chemicals were from Sigma Inc. unless specified otherwise. A monoclonal antibody conjugated to phycoerythrin (Diaclone) was used to detect CD82 in stably and transiently transfected cells. The mouse monoclonal antibody γC11 was used to detect CD82 in Western blots (10). Cell culture Human embryonic kidney cells (293T), adherent HeLa cells, and MCF-7 cells were cultured in DMEM (Sigma) supplemented with 5% (293T) or 10% FCS, respectively. HeLa suspension cells were grown in Joklik-modified ME-Medium (Life Technologies). PC3 and LNCaP cells were kept in RPMI (Life Technologies) supplemented with 10% FCS. Generation of HeLa cells deficient in mitochondrial DNA (Hela ρ0 cells) was as described (11). If not stated otherwise, all cell culture experiments were performed under normoxic conditions. For transfecting 293T, MCF7, PC3, and LNCaP and suspension HeLa cells, the calcium phosphate coprecipitation method was used (12). As indicated, PC3 cells were also transfected with Superfect (Qiagen). Adherent HeLa cells were transfected with calcium phosphate or Effectene (Qiagen). For stable transfections, PC3 cells were treated with Effectene and selected with 500 µg/ml G418 (Calbiochem). The CD82-transfected cell pools were generated from two independent transfections. HeLa ρ0 cells were transfected with the Effectene lipofection reagent according to the instructions of the manufacturer. For the induction of apoptosis, HeLa, HeLa ρ0, and PC3 cells were incubated with different concentrations of TNF (Biomol) as described (13). DNA constructs The vector for caspase-2 was constructed by inserting the cDNA (a gift of S. Kumar, Adelaide, Australia) into the vector RcCMV. The p53 and luciferase expression vectors were obtained by subcloning luciferase or a human p53 cDNA (gift of M. Fritsche, Freiburg, Germany) into RcCMV. pmyc∆ was constructed by inserting oligos coding for the myc-epitope into pcDNA3 (Invitrogen). C33 deletions were generated by PCR using suitable primers with the Expand Long Template PCR System (Roche Diagnostics), which has proofreading activity. All PCR-products were cloned into pmycD and sequenced to verify the correct sequence. The human tetraspanin gene family members were isolated by PCR using suitable primers and cloned into a myc-tag vector. The correct sequence of all constructs was verified. A CD82-HA construct was used for the stable PC3 clone pools. Isolation of apoptosis-inducing genes mRNA was isolated from 10-wk-old CD1 mice and normalized as described previously (14). The cDNA was subcloned into a modified pcDNA3 vector in which the neomycin resistance gene was deleted. The screen for dominant apoptosis inducers was performed essentially as

published (14), except that aliquots containing single bacteria clones were grown up. A novel 96well DNA isolation method allowed a considerably higher throughput (15). Apoptosis detection Apoptosis was quantified by PI staining and determining the amount of sub-G1-DNA with flow cytometry (FACS-Calibur, Becton-Dickinson; ref 16). When apoptosis was measured based on the phenotypic alterations, the ratio of transfected (GFP-positive) and morphologically apoptotic cells was determined in relation to all GFP-positive and therefore transfected cells. At least 250 cells were counted in each independent experiment. With 293T cells, phosphatidylserineexposure was detected using the annexinV-EGFP kit (Clontech Laboratories). HeLa and HeLa ρ0 cells were investigated with the annexinV-biotin staining protocol (PharMingen). “Percent specific apoptosis” was obtained by normalizing apoptosis to the percentage of transfected (GFP positive) cells. Differences in apoptosis induction were considered significant when the standard deviations did not overlap. For the analysis of cytochrome c release, the ApoAlert cell fractionating kit (Clontech) was used as recommended except that the cells were homogenized with 10 strokes of a douncer (Wheaton) and an antibody against Tim23 (BD Biosciences) was used (1:500 dilution). Detection of ∆Ψ and ROI The production of superoxide anions was determined essentially as described by Li et al. (17) using dihydroethidine (HE). The mitochondrial membrane potential was assayed by 3,3′dihexyloxacarbocyanine iodide [DiOC6(3)] (Molecular Probes Inc.) (18). Briefly, cells were harvested, washed with PBS, and incubated with 0.5 nM DiOC6(3) or 10 µM HE for 30 min (5% CO2) and analyzed by flow cytometry (FACS-Calibur, Becton-Dickinson). Lucigenin, a specific probe for superoxide anions, was used as described previously (19) after pooling three 10 cm plates transfected with the indicated plasmids. Its luminiscence was detected with a luminometer (Berthold). Glutathione concentration measurement Glutathione was measured by a commercial kit (ApoGSH glutathione detection kit, Biocat), and its fluorescence was detected with a Fluoroskan Ascent FL (Labsystems). The glutathione concentration in the medium was detected after reduction with 2mM DTT (20). Cdc42 and Rac1 activity assay A commercial kit [EZ-Detect Cdc42, Rac1 activation kit (Pierce)] was employed that uses a Pak1-GST fusion protein for the coimmunprecipition of GTP-bound and therefore activated Gproteins and their specific detection by Western blotting. Hypoxia experiments Cells were routinely cultured under a humidified atmosphere of 5% CO2–95% air (normoxia; incubator: Forma Scientific 3121); for exposure to hypoxia, cells were incubated under an atmosphere of 1% O2–5% CO2–94% N2 in a NuAire 2500E incubator in medium containing 1x minimal essential medium nonessential amino acids (Gibco BRL).

RESULTS Isolation of C33 Recently, we have established a screen for apoptosis-inducing genes (14, 20) and described several of the genes isolated so far (16, 22, 23, 24). With the use of this assay, we found the metastasis suppressor gene C33/KAI1/CD82 (5). The active clone of C33 was sequenced and recognized as containing the described wild-type sequence of C33. Expression of C33 led to rounding of transfected 293T cells and detachment from the substrate (Fig. 1A). FACS analysis detected the apoptosis-specific DNA degradation in C33-transfected cells (Fig. 1A). Phosphatidylserine exposure, a marker preceding the degradation of the genomic DNA and polyADP-ribose polymerase (PARP) cleavage, another hallmark of apoptotic cells, could both be observed (data not shown). Originally, C33 was found to repress metastases formation in prostate cancer cells (5). We therefore tested C33 for its apoptosis inducing capacity in the human prostate cell line PC3. Figure 1B displays that C33 was active for apoptosis induction in these cells as evident by the morphological changes that are typical for apoptotic cells. Quantitative analysis revealed a strong proapoptotic activity that was comparable with programmed cell death induced by RIP, a component of the TNF receptor complex (25). The same proapoptotic effect was also manifest in the prostate cell line LNCaP (data not shown) and in the human cervix carcinoma HeLa cell line (Fig. 1C). The extent of apoptosis could be reduced by addition of the pan-caspase inhibitor zVAD, which demonstrates the involvement of the caspase family of proapoptotic cysteine proteases in C33-induced cell death. To investigate whether caspase-3, a central mediator of proteolytic activity in apoptosis, is necessary for apoptosis, we made use of the mammary carcinoma cell line MCF-7 which is devoid of caspase3 (26). Even though C33 was as effective for apoptosis induction as caspase-2 in these cells (Fig. 1D), its proapoptotic activity remained caspase-dependent as zVAD could still repress apoptosis induction. We isolated the human homologue of C33 (designated CD82) and observed the same extent and kinetics of cell death induction as with the mouse C33 gene (Fig. 1E). Activation of mitochondria for apoptosis has been found to be coupled to the release of cytochrome c from this organelle, which can subsequently bind to Apaf-1 and activate the apoptosome, a proapoptotic protein complex (27). We tested HeLa cells that have been transfected with CD82 for the release of cytochrome c. Figure 1F reveals that CD82 and the positive control Bax led to an increase of free cytochrome c in the cytoplasm of transfected cells. To test whether the apoptosome is responsible for apoptosis, we tested Apaf-1-negative cells for cell death by CD82. Figure 1G demonstrates that the absence of this component of the apoptosome allows only marginal cell death induction by CD82. C33 is therefore able to induce a caspase- and Apaf-1-dependent apoptosis in a number of tumor cells, reflecting its broad activities as a metastasis suppressor (6– 8). CD82 mediates a proapoptotic signal As CD82 is downregulated in many tumor cell lines including PC3 cells (5), two clone pools of these cells with reconstituted CD82 expression were generated. According to FACS analysis, they constitutively expressed low levels of CD82 comparable to those in MCF7 cells, which have a normal expression of CD82 (Fig. 2A). Many different proapototic stimuli were tested on these cells including etoposide, actinomycin, anti-fas antibody, or TNF at various concentrations

and for different incubation times. However, none of these reagents led to an altered apoptosis induction under all examined conditions. TNF-mediated apoptosis is shown as an example in Fig. 2B. We then tested cycloheximide, a known apoptosis inducer in PC3 cells (28). With this reagent, we detected a more than threefold stronger cell death induction in both CD82expressing cell pools in comparison to the control-transfected cells (Fig. 2C). We also observed the cooperation between the transient expression of CD82 and cell death induction by cycloheximide (Fig. 2D). To assess the level of overexpression between transiently and stably transfected cells, we analyzed CD82 in a Western blot. Figure 2E reveals that the transient transfections led to an only moderate overexpression of CD82 in comparison with the stably transfected clones, which resulted in detectable apoptosis with 4 µg of CD82 plasmid. Specificity of the proapoptotic signal induced by CD82 CD82 is a member of the tetraspanin gene family that comprises ~20 different genes (4). To determine the specificity of the proapoptotic signal that is exerted by CD82, we isolated seven additional gene family members. All constructs were transfected into HeLa cells, and the equal expression of all genes was determined by FACS analysis (data not shown). Subsequently, we investigated their apoptosis induction. Figure 3A shows that while CO-029 and CD63 did not significantly induce apoptosis, the other family members led to cell death at different rates. Apoptosis induction varied from 14.4% (CD53) to 42.5% (CD151). While CD53 was relatively inefficient in apoptosis induction, the other members led to a strong development of cell death. In an additional control experiment, we also transfected expression plasmids for other integral plasma membrane proteins such as PDGF-R, EGF-R, integrin α3 and β1, and HER-2 but did not observe any signs of apoptosis induction (data not shown). To gain further insights as to how C33 induces apoptosis, we created deletions in C33 to map the apoptosis-inducing domains. C33 contains four transmembrane segments and three protruding loops, two of which reside on the extracellular side of the membrane (4). The longer extracellular loop at the C terminus is supposed to be heavily glycosylated and to be responsible for cell-substrate adhesion (5). Progressive deletions from the C-terminus revealed that the short intracellular domain did not contribute to apoptosis: its deletion (C33∆255) was as active as wild-type C33 (Fig. 3B). Interestingly, further deletion of the long, glycosylated extracellular loop, which was implicated in mediating the cellular attachment (construct C33∆119), did also not result in a different level of apoptosis induction. In contrast, deletion of the third transmembrane domain and the short intracellular loop (C33∆80) led to an almost complete abrogation of apoptosis activity. The N-terminal domain containing the signal sequence and the first transmembrane domain (C33∆49) was just as inefficient for apoptosis induction. FACS analysis of YFP-fusion proteins showed that all deletion mutants were equally expressed with only C33∆49 reaching one-third of the others’ expression level of the others (not shown). As C33 is a transmembrane protein, it was speculated that it is mediating its anti-metastatic effect by interacting with components of the basal membrane (5). However, our mapping data suggested that adherence is not obligatory for apoptosis induction. Consequently, we tested the requirement of substrate attachment for apoptosis induction by CD82. A derivative cell line of HeLa cells, which can grow in suspension, was transfected with an expression plasmid for C33 and p53 as a positive control. Figure 3C shows that C33 could still induce apoptosis in such nonadherent HeLa cells with an efficiency comparable to that of p53.

C33 generates ROIs independently of the respiratory chain C33 is a known p53 target gene with a typical p53 tandem binding motive in its promotor (29). As p53-induced apoptosis is mediated by ROIs (17), we wanted to know whether C33 induces the formation of oxygen radicals. For this, we tested C33-transfected cells with HE, which detects superoxide anions by specific fluorescence. After C33 transfection, we observed a strong induction of oxygen radicals peaking after 22 h compared with control cells (Fig. 4A). We have also observed an increase of DCFH fluorescence, an indicator of hydrogen peroxide (data not shown). In parallel, we performed apoptosis quantification by FACS analysis (Fig. 4B), which showed that C33-apoptosis occurred significantly later than ROI production. To test whether ROI generation is essential for apoptosis induction by C33, we used the superoxide scavenger Tiron. A 32 h-incubation with Tiron reduced C33-induced apoptosis by 33% (Fig. 4C), and a titration curve showed a dose-dependent reduction of CD82-mediated apoptosis (Fig. 4D). A comparison of Fig. 4C and D reveals that Tiron is more active as an apoptosis repressor if the extent of apoptosis is lowered to ~20%. We also observed an inhibition of apoptosis by 60% with the oxyradical scavenger EUK-8 (data not shown). Furthermore, cotransfection of the ROIscavenging enzymes catalase or Cu/Zn superoxide dismutase decreased C33-induced cell death by more than 50% (Fig. 4E). As C33 inactivation is crucial for metastases formation and as especially micro-metastases develop before angiogenesis and concomitantly oxygen supply has set in (30), we wanted to address the role of oxygen for apoptosis induction. Therefore, after transfection of C33, we kept HeLa cells in an environment with reduced (1%) oxygen (hypoxia). Figure 5A shows that C33 was still efficient as apoptosis inducer under such conditions. The respiratory chain is regarded as the main source of oxygen radicals for apoptosis induction (31). However, it is downregulated in metastasis-forming cells before angiogenesis sets in as it can react to a reduced amount of molecular oxygen (32). In the absence of a functional respiratory chain, cells are dependent on glycolysis, a metabolic pathway that is hyperactive in many tumor cells and required for metastases formation under hypoxic conditions (33). Thus, we were interested how cells that harbor an inactive respiratory chain would react on C33 transfection. To test this, we generated HeLa cells deficient in mitochondrial DNA. Such cells (HeLa ρ0) are incapable to synthesize the respiratory chain components encoded by the mitochondrial genome and produce ATP solely by glycolysis. Several reports have shown that ρ0 cells are less sensitive for apoptosis induction by various proapoptotic stimuli (34–36). A titration experiment revealed that HeLa ρ0 cells were more resistant to apoptosis by TNF than wild-type cells (Fig. 5B). In contrast, Fig. 5C shows that despite the absence of an intact electron transfer of mitochondria in HeLa ρ0 cells, C33 was just as active for apoptosis induction in these cells as in wild-type HeLa cells. Apoptosis induction by caspase-2 was similar in both cell types indicating equal capacities to undergo apoptosis. Furthermore, Fig. 5D shows that C33 expression should also produce ROIs in HeLa ρ0 cells. Even though the ROIs seem not to be derived from the mitochondrial respiratory chain, we investigated the activation of mitochondria for apoptosis induction. Mitochondria are regarded as important mediators of many proapoptotic signals (37). Common to pathways that stimulate mitochondria for apoptosis induction is the breakdown of the membrane potential (∆Y) over the inner mitochondrial membrane (38). In HeLa ρ0 cells, such a membrane potential is generated by the reverse transport of protons with concomitant ATP hydrolysis (39), which can be used for apoptosis (40). We therefore tested this membrane potential using DiOC6(3), which is taken up

by mitochondria with an intact membrane gradient. Figure 5E shows that upon CD82 expression the number of cells with a reduced mitochondrial membrane potential increased over time. CD82 causes the specific release of intracellular glutathione In an attempt to identify the enzymatic complex(es) responsible for the generation of ROIs by CD82, we performed a series of inhibition experiments. However, we did not observe a reduction of apoptosis with a number of inhibitors for known ROI-forming enzymes, including diphenyleneiodonium chloride (DPI, an inhibitor of NADPH oxidase), allopurinol (an inhibitor of xanthine oxidase), ibuprofen (a cyclooxygenase inhibitor), various inhibitors for phospholipase A and lipoxygenase, and 19 other described inhibitors of enzymatic activities involved in ROI formation (data not shown). Cotransfection experiments revealed that both the cytoplasmatic and the mitochondrial SOD could reduce apoptosis by C33 (not shown), suggesting that more than one source of ROIs is activated by C33. Further cotransfection experiments identified glutathione peroxidase-1 as a cell death repressor (Fig. 6A). This enzyme uses glutathione (GSH), a prominent intracellular ROI-scavanger, for the detoxification of ROIs. Because distinct apoptosis inducers stimulate cells to actively induce the efflux of this molecule (20, 41, 42), we investigated the concentration of glutathione in CD82-transfected cells. The left panel of Fig. 6B shows that the intracellular concentration of GSH significantly drops after transfection of CD82. This was not due to exhaustion of glutathione (that is present in mM concentrations in the cell) by CD82-produced ROIs since p53, that likewise generates ROIs (Fig. 4A) and induces apoptosis with similar kinetics (Fig. 4A and B), did not lead to an altered glutathione level. We have consistently observed that p53 and also Bax expression (not shown) did not reduce the cellular GSH concentration. When the supernatant of CD82-transfected cells was tested for the presence of GSH, we observed a corresponding quantitative increase of GSH (Fig. 4B). With these experiments having been performed in Hela cells, we wanted to investigate another cell system and detected a comparable loss of GSH from CD82-transfected PC3 cells (data not shown). To assess the contribution of this glutathione depletion on CD82-induced cell death, we incubated cells with increasing concentrations of buthionine sulfoximine (BSO), an inhibitor of GSH synthesis that leads to a pronounced reduction of GSH already after 10 h (not shown). BSO treatment is demonstrated in Fig. 6C to potentiate cell death by CD82. This chemical alone led only to a moderate cell death after 46 h at optimal conditions (Fig. 6D). To test whether the efflux of GSH is nevertheless necessary for cell death, we used glutathionemonoethyl ester, a membrane-permeable GSH form, which restores the CD82-mediated efflux of GSH from cells after 25 h (not shown). When CD82-transfected cells were incubated with 1.5 mM of this compound, we observed a pronounced reduction of ROIs (Fig. 6E) and a concomitant decrease in apoptosis induction (Fig. 6F). CD82 activates the GTPase Cdc42, which mediates glutathione efflux and apoptosis induction and allows ROI detection A recent report (43) suggested that C33 activates GTPases such as Cdc42. As some of these Gproteins have been implicated in proapoptotic signaling pathways (44, 45), we tested the activation of these enzymes. Figure 7A reveals that we could observe a potent stimulation of Cdc42 in an activity assay specific for this GTPase. This effect was independent of caspases as demonstrated by the use of the caspase inhibitor zVAD. In constrast, no activation was seen with the G-protein Rac-1 (not shown). We also detected a pronounced reduction of apoptosis by CD82 when a dominant-negative Cdc42 mutant was cotransfected (Fig. 7B) suggesting a

functional involvement of this GTPase in apoptosis. To investigate whether Cdc42 is necessary for the detection of ROIs induced by CD82, we cotransfected the dominant-negative Cdc42 and observed a potent reduction of ROIs (Fig. 7C). Finally, we analyzed whether Cdc42 mediates the release of cellular glutathione into the medium. The cotransfection of the dominant-negative Cdc42 prevented both the reduction of the cellular glutathione level and its detection in the supernatant of CD82-transfected cells (Fig. 7D). DISCUSSION Here we report the isolation of C33/CD82/KAI1, a known metastasis suppressor gene, in a screen for proapoptotic genes. Given the importance and the scarcity of information about molecular events leading to metastases formation, we further investigated this gene. Our data show that apoptosis induction by C33 was mediated by the generation of reactive oxygen intermediates (Fig. 4). These ROIs, however, did not originate from the respiratory chain as seen with HeLa ρ0 (Fig. 5D). This is unusual as the respiratory chain is regarded as the standard source of propaoptotic ROIs (34, 36). However, besides the mitochondrial respiratory chain, there is a number of different locations and enzymatic reactions that can produce oxygen radicals (46): The endoplasmatic reticulum has been shown to be a site of ROI production. In addition, the nuclear membrane as well as the plasma membrane contain superoxide-generating enzymatic activities of mostly unknown composition. The activity of both cytosolic and mitochondrial SOD to reduce apoptosis and our failure to diminish cell death by inhibitors of ROI-forming enzymes suggest that several enzymatic complexes contribute to the formation of this short-lived metabolite. However, we observed that CD82 induces the specific efflux of glutathione in transfected cells (Fig. 6B). GSH release alone is not sufficient for effective apoptosis, as BSO activates cell death after 46 h to only 14% (Fig. 6D). The efflux of GSH seems to be necessary for C33-mediated apoptosis, as we observed a reduced ROI content and diminished apoptosis induction (Fig. 6E and F), when we reconstituted the level of GSH. Although it remains to be seen whether CD82 activates specific transporters for the GSH release as suggested for the Fas receptor (19), the efflux of GSH could result in a loss of the cellular antioxidant defense, a prolonged half life of ROIs, and a stronger apoptosis induction. This could explain why several, however ineffective, ROI-generating enzymes can bring about cell death. We found that ROI detection and apoptosis induction was facilitated by the GTPase Cdc42 (Fig. 7). This signaling protein has been implicated in a number of apoptosis scenarios by cytotoxic T lymphocytes, by the Fas receptor (45), and by growth factor withdrawal (44). In addition, Cdc42 is a p53-target gene and mediates cell death by this tumor suppressor gene (47). Up to now, Cdc42 has been suggested to activate the NADPH oxidase (48), which, however, seems not to be stimulated by CD82 based on our inhibition studies. A model might explain why it is so advantageous for tumor cells to specifically inactivate C33 for metastases formation: after dissemination to distant tissues and before angiogenesis provides enough oxygen supply, metastases-forming cells encounter a hypoxic environment in micrometastases (30). Under this condition, the respiratory chain acts as a hypoxia sensor and reduces its activity while energy generation becomes dependent on glycolysis (49). Many proapototic

stimuli, e.g., TNF (50), ceramide (34), interferon-γ (35), and several cytostatic drugs (36, 51), do not properly function any more in such an environment because they rely on ROIs from the respiratory chain. Consequently, they cannot rescue the organism from metastases spread by inducing apoptosis. As the activity of C33 to generate ROIs is independent of the respiratory chain, it can still induce cell death (Fig. 5A). This could be caused by an apoptosis stimulus such as cycloheximide (see below). If C33 is then additionally inactivated, metastases formation can proceed. This model links a reduced potential of cells to undergo apoptosis to the process of metastases formation. Such a correlation was indeed observed numerous times (52, 53). In accordance, the importance of decreasing ROIs, for the process of metastases formation has been suggested by several reports: SOD expression is enhanced in tumor metastases of gastric cancers in vivo (54) and also confers increased metastasis formation in liver tumors in vitro (55). Although most cell death inducers were unaffected by the C33-status of the cell, we found that reconstituted C33 expression resensitizes PC3 cells to at least one specific apoptosis stimulus: cycloheximide (Fig. 2C and D). This specificity suggests that C33 is a component of a distinct endogenous apoptosis pathway. Many stress factors for metastasis-forming cells that could contribute to apoptosis induction are likely to be still unknown. However, it is conceivable that metastasis-forming cells in the blood or in foreign tissues lack appropriate growth or survival factors. This might downregulate translation in much the same way as cycloheximide does and lead to apoptosis induction via C33. Consequently, the inactivation of C33 might secure the survival of such tumor cells in distant tissues. Tetraspanin family members contain a long extracellular loop, which is heavily glycosylated. This led to the model that C33 regulates the adhesion of cells to the cellular substrate (5). A recent report showed that C33 expression alone is not sufficient for apoptosis induction but relied on glycosylation and was observable only after 10 days (56). In contrast, we found that C33 induces apoptosis in a wide variety of cells with a much faster kinetics (Fig. 1). In addition, and we also observed apoptosis induction in nonadherent cells (Fig. 3C). Finally, the deletion of the large extracellular loop in C33 that is a candidate substrate interaction- and glycosylationdomain had no effect on its efficiency to induce apoptosis (Fig. 3B). Although our data do not exclude an additional role of C33 for mediating the attachment of cells, they strongly suggest that C33 is able to stimulate an intracellular signaling pathway for apoptosis induction that is independent of substrate adherence. There are indeed indications for such an activity: Clustering of C33 by a monoclonal antibody has been shown to lead to an increase in intracellular Ca2+, protein tyrosine phosphorylation (57), and GTPase activation (43). Our results demonstrate the activity of C33/CD82/KAII as a direct apoptosis inducer via the generation of ROIs. It is hoped that this might lead to a better understanding of metastasis formation and possibly to a better therapeutic intervention. ACKNOWLEDGMENTS We thank Drs. Riesenberg, Klinikum Großhadern (Munich), for providing PC3 and LNCaP cells and T. Mak, Toronto, Canada, for the Apaf-1 negative cells. The expression constructs for PDGF-R, EGF-R, Cdc42 T17N, Rac1 T17N, and HER-2 were a gift of Dr. A. Ullrich, Martinsried. The vectors for catalase, glutathione peroxidase-1 and Cu/Zn superoxide dismutase were from Dr. S. Lenzen, Hannover. The monoclonal antibody γC11 against CD82 was a kind

gift of H. Conjeaud, Paris, France. Thanks also to A. Gewies and R. Gernert for helpful comments on the manuscript. This work was supported by the Bavarian Government, Roche Diagnostics, and Xantos Biomedicine AG. REFERENCES 1.

Kerr, J. F., Wyllie, A. H., and Currie, A. R. (1972) Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br. J. Cancer 26, 239–257

2.

McGill, G. (1997) Apoptosis in tumorigenesis and cancer therapy. Front. Biosci. 2, d353– d379

3.

Yoshida, B. A., Sokoloff, M. M., Welch, D. R., and Rinker-Schaeffer, C. W. (2000) Metastasis-suppressor genes: a review and perspective on an emerging field. J. Natl. Cancer Inst. 92, 1717–1730

4.

Maecker, H. T., Todd, S. C., and Levy, S. (1997) The tetraspanin superfamily: molecular facilitators. FASEB J. 11, 428–442

5.

Dong, J. T., Lamb, P. W., Rinker-Schaeffer, C. W., Vukanovic, J., Ichikawa, T., Isaacs, J. T., and Barrett, J. C. (1995) KAI1, a metastasis suppressor gene for prostate cancer on human chromosome 11p11.2. Science 268, 884–886

6.

Guo, X., Friess, H., Graber, H. U., Kashiwagi, M., Zimmermann, A., Korc, M., and Buchler, M. W. (1996) KAI1 expression is up-regulated in early pancreatic cancer and decreased in the presence of metastases. Cancer Res. 56, 4876–4880

7.

Takaoka, A., Hinoda, Y., Satoh, S., Adachi, Y., Itoh, F., Adachi, M., and Imai, K. (1998) Suppression of invasive properties of colon cancer cells by a metastasis suppressor KAI1 gene. Oncogene 16, 1443–1453

8.

White, A., Lamb, P. W., and Barrett, J. C. (1998) Frequent downregulation of the KAI1(CD82) metastasis suppressor protein in human cancer cell lines. Oncogene 16, 3143– 3149

9.

Yu, Y., Yang, J. L., Markovic, B., Jackson, P., Yardley, G., Barrett, J., and Russell, P. J. (1997) Loss of KAI1 messenger RNA expression in both high- grade and invasive human bladder cancers. Clin. Cancer Res. 3, 1045–1049

10. Lebel-Binay, S., Lagaudriere, C., Fradelizi, D., and Conjeaud, H. (1995) CD82, member of the tetra-span-transmembrane protein family, is a costimulatory protein for T cell activation. J. Immunol. 155, 101–110 11. King, M. P., and Attardi, G. (1989) Human cells lacking mtDNA: repopulation with exogenous mitochondria by complementation. Science 246, 500–503 12. Roussel, M. F., Rettenmier, C. W., Look, A. T., and Sherr, C. J. (1984) Cell surface expression of v-fms-coded glycoproteins is required for transformation. Mol. Cell. Biol. 4, 1999–2009

13. Deiss, L. P., Galinka, H., Berissi, H., Cohen, O., and Kimchi, A. (1996) Cathepsin D protease mediates programmed cell death induced by interferon-gamma, Fas/APO-1 and TNF-alpha. EMBO J. 15, 3861–3870 14. Grimm, S., and Leder, P. (1997) An apoptosis-inducing isoform of neu differentiation factor (NDF) identified using a novel screen for dominant, apoptosis-inducing genes. J. Exp. Med. 185, 1137–1142 15. Neudecker, F., and Grimm, S. (2000) High-throughput method for isolating plasmid DNA with reduced lipopolysaccharide content. Biotechniques 28, 107–109 16. Bauer, M. K. A., Schubert, A., Rocks, O., and Grimm, S. (1999) Adenine nucleotide translocase-1, a component of the permeability transition pore, can dominantly induce apoptosis. J. Cell Biol. 147, 1493–1502 17. Li, P. F., Dietz, R., and von Harsdorf, R. (1999) p53 regulates mitochondrial membrane potential through reactive oxygen species and induces cytochrome c- independent apoptosis blocked by Bcl-2. EMBO J. 18, 6027–6036 18. Rottenberg, H., and Wu, S. (1998) Quantitative assay by flow cytometry of the mitochondrial membrane potential in intact cells. Biochim. Biophys. Acta 1404, 393–404 19. Pervaiz, S., and Clement, M. V. (2002) Hydrogen peroxide-induced apoptosis: oxidative or reductive stress? Methods Enzymol. 352, 150–159 20. van den Dobbelsteen, D. J., Nobel, C. S., Schlegel, J., Cotgreave, I. A., Orrenius, S., and Slater, A. F. (1996) Rapid and specific efflux of reduced glutathione during apoptosis induced by anti-Fas/APO-1 antibody. J. Biol. Chem. 271, 15420–15427 21. Albayrak, T., and Grimm, S. (2003) A high-throughput screen for single gene activities: isolation of apoptosis inducers. Biochem. Biophys. Res. Commun. 304, 772–776 22. Mund, T., Gewies, A., Schoenfeld, N., Bauer, M. K., and Grimm, S. (2003) Spike, a novel BH3-only protein, regulates apoptosis at the endoplasmic reticulum. FASEB J. 17, 696–698 23. Gewies, A., and Grimm, S. (2003) UBP41 is a proapoptotic ubiquitin-specific protease. Cancer Res. 63, 682–688 24. Albayrak, T., Scherhammer, V., Schoenfeld, N., Braziulis, E., Mund, T., Bauer, M. K., Scheffler, I. E., and Grimm, S. (2003) The tumor suppressor cybL, a component of the respiratory chain, mediates apoptosis induction. Mol. Biol. Cell 14, 3082–3096 25. Stanger, B. Z., Leder, P., Lee, T. H., Kim, E., and Seed, B. (1995) RIP: a novel protein containing a death domain that interacts with Fas/APO-1 (CD95) in yeast and causes cell death. Cell 81, 513–523 26. Janicke, R. U., Walker, P. A., Lin, X. Y., and Porter, A. G. (1996) Specific cleavage of the retinoblastoma protein by an ICE-like protease in apoptosis. EMBO J. 15, 6969–6978

27. Srinivasula, S. M., Ahmad, M., Fernandes-Alnemri, T., and Alnemri, E. S. (1998) Autoactivation of procaspase-9 by Apaf-1-mediated oligomerization. Mol. Cell 1, 949–957 28. Rokhlin, O. W., Bishop, G. A., Hostager, B. S., Waldschmidt, T. J., Sidorenko, S. P., Pavloff, N., Kiefer, M. C., Umansky, S. R., Glover, R. A., and Cohen, M. B. (1997) Fasmediated apoptosis in human prostatic carcinoma cell lines. Cancer Res. 57, 1758–1768 29. Mashimo, T., Watabe, M., Hirota, S., Hosobe, S., Miura, K., Tegtmeyer, P. J., RinkerShaeffer, C. W., and Watabe, K. (1998) The expression of the KAI1 gene, a tumor metastasis suppressor, is directly activated by p53. Proc. Natl. Acad. Sci. USA 95, 11307– 11311 30. Ellis, L. M., and Fidler, I. J. (1996) Angiogenesis and metastasis. Eur. J. Cancer 32A, 2451– 2460 31. Lenaz, G. (1998) Role of mitochondria in oxidative stress and ageing. Biochim. Biophys. Acta 1366, 53–67 32. Chandel, N. S., Budinger, G. R., Choe, S. H., and Schumacker, P. T. (1997) Cellular respiration during hypoxia. Role of cytochrome oxidase as the oxygen sensor in hepatocytes. J. Biol. Chem. 272, 18808–18816 33. Dang, C. V., and Semenza, G. L. (1999) Oncogenic alterations of metabolism. Trends Biochem. Sci. 24, 68–72 34. Quillet-Mary, A., Jaffrezou, J. P., Mansat, V., Bordier, C., Naval, J., and Laurent, G. (1997) Implication of mitochondrial hydrogen peroxide generation in ceramide- induced apoptosis. J. Biol. Chem. 272, 21388–21395 35. Sun, I. L., Sun, L. E., and Crane, F. L. (1996) Cytokine inhibition of transplasma membrane election transport. Biochem. Mol. Biol. Int. 38, 175–180 36. Jia, L., Allen, P. D., Macey, M. G., Grahn, M. F., Newland, A. C., and Kelsey, S. M. (1997) Mitochondrial electron transport chain activity, but not ATP synthesis, is required for druginduced apoptosis in human leukaemic cells: a possible novel mechanism of regulating drug resistance. Br. J. Haematol. 98, 686–698 37. Green, D. R., and Reed, J. C. (1998) Mitochondria and apoptosis. Science 281, 1309–1312 38. Zamzami, N., Brenner, C., Marzo, I., Susin, S. A., and Kroemer, G. (1998) Subcellular and submitochondrial mode of action of Bcl-2-like oncoproteins. Oncogene 16, 2265–2282 39. Buchet, K., and Godinot, C. (1998) Functional F1-ATPase essential in maintaining growth and membrane potential of human mitochondrial DNA- depleted rho degrees cells. J. Biol. Chem. 273, 22983–22989 40. Jiang, S., Cai, J., Wallace, D. C., and Jones, D. P. (1999) Cytochrome c-mediated apoptosis in cells lacking mitochondrial DNA. Signaling pathway involving release and caspase 3 activation is conserved. J. Biol. Chem. 274, 29905–29911

41. Jardine, H., MacNee, W., Donaldson, K., and Rahman, I. (2002) Molecular mechanism of transforming growth factor (TGF)-beta1-induced glutathione depletion in alveolar epithelial cells. Involvement of AP-1/ARE and Fra-1. J. Biol. Chem. 277, 21158–21166 42. Ghibelli, L., Coppola, S., Rotilio, G., Lafavia, E., Maresca, V., and Ciriolo, M. R. (1995) Non-oxidative loss of glutathione in apoptosis via GSH extrusion. Biochem. Biophys. Res. Commun. 216, 313–320 43. Delaguillaumie, A., Lagaudriere-Gesbert, C., Popoff, M. R., and Conjeaud, H. (2002) Rho GTPases link cytoskeletal rearrangements and activation processes induced via the tetraspanin CD82 in T lymphocytes. J. Cell Sci. 115, 433–443 44. Bazenet, C. E., Mota, M. A., and Rubin, L. L. (1998) The small GTP-binding protein Cdc42 is required for nerve growth factor withdrawal-induced neuronal death. Proc. Natl. Acad. Sci. USA 95, 3984–3989 45. Subauste, M. C., Von Herrath, M., Benard, V., Chamberlain, C. E., Chuang, T. H., Chu, K., Bokoch, G. M., and Hahn, K. M. (2000) Rho family proteins modulate rapid apoptosis induced by cytotoxic T lymphocytes and Fas. J. Biol. Chem. 275, 9725–9733 46. Cross, A. R., and Jones, O. T. (1991) Enzymic mechanisms of superoxide production. Biochim. Biophys. Acta 1057, 281–298 47. Thomas, A., Giesler, T., and White, E. (2000) p53 mediates bcl-2 phosphorylation and apoptosis via activation of the Cdc42/JNK1 pathway. Oncogene 19, 5259–5269 48. Bonizzi, G., Piette, J., Schoonbroodt, S., Greimers, R., Havard, L., Merville, M. P., and Bours, V. (1999) Reactive oxygen intermediate-dependent NF-kappaB activation by interleukin-1beta requires 5-lipoxygenase or NADPH oxidase activity. Mol. Cell. Biol. 19, 1950–1960 49. Lahiri, S. (2000) Historical perspectives of cellular oxygen sensing and responses to hypoxia. J. Appl. Physiol. 88, 1467–1473 50. Sidoti-de Fraisse, C., Rincheval, V., Risler, Y., Mignotte, B., and Vayssiere, J. L. (1998) TNF-alpha activates at least two apoptotic signaling cascades. Oncogene 17, 1639–1651 51. Suzuki, S., Higuchi, M., Proske, R. J., Oridate, N., Hong, W. K., and Lotan, R. (1999) Implication of mitochondria-derived reactive oxygen species, cytochrome C and caspase-3 in N-(4-hydroxyphenyl)retinamide-induced apoptosis in cervical carcinoma cells. Oncogene 18, 6380–6387 52. Glinsky, G. V., Glinsky, V. V., Ivanova, A. B., and Hueser, C. J. (1997) Apoptosis and metastasis: increased apoptosis resistance of metastatic cancer cells is associated with the profound deficiency of apoptosis execution mechanisms. Cancer Lett. 115, 185–193 53. McConkey, D. J., Greene, G., and Pettaway, C. A. (1996) Apoptosis resistance increases with metastatic potential in cells of the human LNCaP prostate carcinoma line. Cancer Res. 56, 5594–5599

54. Malafa, M., Margenthaler, J., Webb, B., Neitzel, L., and Christophersen, M. (2000) MnSOD expression is increased in metastatic gastric cancer. J. Surg. Res. 88, 130–134 55. Nonaka, Y., Iwagaki, H., Kimura, T., Fuchimoto, S., and Orita, K. (1993) Effect of reactive oxygen intermediates on the in vitro invasive capacity of tumor cells and liver metastasis in mice. Int. J. Cancer 54, 983–986 56. Ono, M., Handa, K., Withers, D. A., and Hakomori, S. (1999) Motility inhibition and apoptosis are induced by metastasis-suppressing gene product CD82 and its analogue CD9, with concurrent glycosylation. Cancer Res. 59, 2335–2339 57. Lebel-Binay, S., Lagaudriere, C., Fradelizi, D., and Conjeaud, H. (1995) CD82, tetra-spantransmembrane protein, is a regulated transducing molecule on U937 monocytic cell line. J. Leukoc. Biol. 57, 956–963 Received May 1, 2003; accepted October 10, 2003.

Fig. 1

Figure 1. Apoptosis induction in various cell lines by C33. A) C33 expression leads to phenotype changes of apoptosis in 293T cells. A C33 expression plasmid or a control plasmid together with a GFP vector was transfected into human 293T cells. After 22 h, fluorescence microscopy pictures were taken (left panel). Right panel: quantification of C33 apoptosis induction. An expression plasmid encoding C33 or luciferase (Control) was transfected into 293 cells. After 42 h, apoptosis was quantified by flow cytometry of cells with sub-G1 DNA content. Results were normalized to transfection efficiencies, which were obtained by cotransfecting green fluorescent protein (GFP). Means ± SD of 3 independent experiments each are shown. PC3 prostate cells (B), as well as HeLa cells (C), display the apoptosis-specific phenotype alterations after transfection of a C33 expression vector. An expression vector for luciferase was used in control experiments. A quantitative apoptosis evaluation (means ± SD of 3 independent experiments) based on phenotypic alterations is shown at right with each cell line 30 (PC3) or 22 h (HeLa) posttransfection, respectively. Luciferase (Luc) was used as negative controls and RIP or caspase-2 as positive controls. HeLa cells were additionally treated with the caspase inhibitor zVAD (50 µM added 13 h posttransfection). D) C33 apoptosis is caspase dependent but caspase-3 independent. MCF-7 breast carcinoma cells that are deficient in caspase-3 were used to evaluate the importance for caspase-3 activity for C33 apoptosis induction. Expression vectors for luciferase (Control), caspase-2, or C33 were transfected into MCF-7 cells. The pan-caspase inhibitor zVAD was added at a concentration of 45 µM 11 h posttransfection. After 26 h, apoptosis was quantified as in B. E) Comparison of mouse C33 and human CD82 for apoptosis induction. Equal amounts of expression constructs for C33 or CD82 were transfected into HeLa cells together with a GFP expression plasmid; 47 h later apoptosis was quantified by FACS analysis. Means ± SD of apoptotic cells from 4 independent experiments are shown. F) Cytochrome c is released from mitochondria in CD82-expressing HeLa cells. Cytosolic and mitochondrial extracts (2.5 µg each) of luciferase-, Bax-, and CD82-transfected cells were probed after 36 h in Western blots for presence of Tim23, cytochrome c, and β-actin. G) Apaf-1 is required for efficient apoptosis by CD82. Indicated genes (luciferase or CD82, the human homologue of C33) were transfected into wild-type (WT) mouse embryonic fibroblasts or Apaf-1-deficient cells. Apoptosis was quantified based on phenotypic alterations after 36 h. Means ± SD of apoptotic cells from 3 independent experiments are shown.

Fig. 2

Figure 2. Effect of CD82 expression on apoptosis induction by different stimuli. A) Expression of CD82 in a clone pool of stably transfected PC3 cells in comparison to contol-transfected PC3 cells and MCF7 cells. Cells were decorated with a phycoerythrin conjugated antibody against CD82 and investigated by FACS analysis in comparison to untreated cells. In FACS histograms, gray area is untreated cells and dark line is distribution of stained cells. B) CD82 does not influence TNF-induced cell death. PC3 cells that lack CD82 expression and two clone pools of PC3 cells with reconstituted CD82 expression were tested for apoptosis induction at different TNF concentrations after 45 h. Means ± SD of 3 independent FACS experiments are shown. C) Cycloheximide-induced apoptosis is potentiated by CD82 expression. CD82-expressing and -nonexpressing PC3 cells from A were treated with cycloheximide (0.25 µg/ml). After 30 h, cells were investigated for apoptosis by FACS analysis. Means ± SD of 3 independent measurements are shown. D) CD82caused apoptosis cooperates with cycloheximide in induction of cell death in a transient expression assay. PC3 cells were transiently transfected with a GFP expression plasmid together with a CD82 expression plasmid using lipofection; 4 h after transfection cycloheximide (CHX) was added and incubated for 30 h after which FACS analysis was used to determine apoptosis. Means ± SD of 3 experiments that were normalized to GFP expression are shown. E) Expression levels of CD82 in transient and stably transfected PC3 cells. Equal amounts of protein (20 µg) of 2 stably transfected cell pools as well as transiently transfected cells were investigated in a Western blot with a monoclonal antibody against CD82. The amount of transfected CD82 expression plasmid in 10 cm plates is indicated for each transient transfection (~40% transfection efficiency). Loading control with an antiserum against β-actin is also shown.

Fig. 3

Figure 3. Characterization of C33 cell death signal. A) Different apoptosis induction by human tetraspanin gene family members. Human genes of the tetraspanin gene family, including the C33 homologue CD82 were isolated by PCR. Forty-seven hours after transfection into HeLa cells, the extent of apoptosis induction was determined by FACS and compared with apoptosis induced by RIP, a component of TNF receptor complex. Means ± SD of 4 independent experiments for each construct are shown. Apoptosis induction was normalized to transfection efficiencies as measured by a cotransfected GFP-construct. B) Mapping the domain responsible for apoptosis induction in C33. C-terminal deletions of C33 at left were generated, and 1 µg of each construct together with 0.25 µg of GFP plasmid was transfected into HeLa cells. After 44 h, apoptosis was quantified using FACS analysis. The extent of apoptosis induction was normalized to the transfection efficiencies as measured by GFP-fluorescence. Data are means ± SD from 3 independent experiments. C) C33 can induce apoptosis in nonadherent cells. HeLa suspension cells were transfected with 3 µg expression plasmid for C33, p53, or β-gal (Control). Each sample was cotransfected with an expression plasmid for GFP, which was used to normalize for transfection efficiencies. Apoptosis was measured by FACS analysis after 52 h. Means ± SD of 3 independent experiments are shown.

Fig. 4

Figure 4. C33 leads to production of reactive oxygen radicals during apoptosis. A) C33 induces oxidation of

hydroethidine to ethidium (ET) during apoptosis. After HeLa cells were transfected with 2 µg expression plasmid for C33, p53, or control vector (Luc), ROI generation was measured by staining with hydroethidine, which detects superoxide radicals in cells. Results are percentage of fluorescence intensity compared with control-transfected cells. Means ± SD of 4 independent experiments are shown. B) C33 leads to a progressive accumulation of apoptotic cells. Aliquots of transfections described in A were removed at indicated time points, and extent of apoptosis was assessed by cells with sub-G1 DNA content using a FACS. Means ± SD of 4 independent experiments are shown. C) The antioxidant Tiron reduces C33-mediated apoptosis; 13 h after transfection of C33 or β-gal, Tiron was added at a concentration of 300 µM. After 32 h, apoptosis was quantified by FACS analysis. D) Titration curve of Tiron. Increasing concentrations of Tiron as indicated were added to CD82-transfected HeLa cells that were harvested 36 h later. Means ± SD of 3 independent FACS experiments are shown. E) Cotransfection of catalase or Cu/Zn superoxide dismutase diminishes cell death by C33 expression. β-gal (100 ng) or C33 (50 ng) and expression vectors for β-gal or catalase (Cat) and superoxide dismutase (SOD) (50 ng each) were cotransfected. Means ± SD of 4 independent experiments are shown, which were obtained by FACS after PI staining 44 h after transfection.

Fig. 5

Figure 5. C33 can induce apoptosis under hypoxia conditions and in cells lacking an intact respiratory chain but still activates mitochondria for apoptosis induction. A) Apoptosis induction by C33 is efficient under oxygen deprivation. HeLa cells were transfected with 100 ng C33 or β-gal expression vector and kept in 1% oxygen. Apoptosis was quantified by PI staining and FACS analysis after 39 h. Means ± SD of 4 independent experiments are shown. B) HeLa ρ0 cells display a diminished apoptosis induction after TNF treatment. Different concentrations of TNF were used to induce apoptosis in WT HeLa and ρ0 HeLa cells that are deficient in the mitochondrial DNA content. After 45 h, cell death was quantified by FACS analysis. C) C33 efficiently induces apoptosis in HeLa ρ0 cells. WT HeLa and HeLa ρ0 cells were transfected with an expression plasmid (100 ng) for C33, caspase-2, or β-gal together with a GFP vector (25 ng). Apoptosis quantification was performed 46 h posttransfection by annexin V-FACS. Apoptosis induction was normalized for transfection efficiency. Means ± SD of 3 independent experiments are shown. D) C33 leads to ROI production in HeLa ρ0 cells. C33 or a control vector encoding luciferase was transfected into HeLa ρ0 cells, and ROI production was measured by hydroethidine oxidation after 20 h. Results are percentage of fluorescence intensity compared with control-transfected cells. Means ± SD of 3 independent experiments are shown. E) C33 expression causes a reduction in mitochondrial membrane potential (∆Ψ). HeLa cells were transfected with a plasmid for p53, CD82, or a control vector (Luc). After the indicated incubation times, aliquots were tested for DiOC6(3) fluorescence in a FACS analysis. Means ± SD of cells in 3 independent experiments with each condition are shown.

Fig. 6

Figure 6. Involvement of the cellular antioxidant glutathione in CD82 apoptosis. A) Glutathione peroxidase can

reduce C33-induced cell death. Expression vectors for CD82 and β-gal or glutathione peroxidase-1 (GPx) were cotransfected at a ratio of 1:4. Apoptosis was quantified by FACS after 46 h. Means ± SD of 3 independent experiments are shown. B) CD82 leads to the release of intracellular glutathione. HeLa cells were transfected with equal amounts of indicated plasmids for luciferase, p53, or CD82. After indicated times, cells were harvested and glutathione level in cells (left panel) and in the medium (right panel) was determined. Given are percent values relative to intracellular GSH concentration of control (β-gal)-transfected cells. For GSH measurements in the medium, its concentration in control medium was subtracted. C) Inhibition of GSH synthesis leads to augmented apoptosis by CD82. Increasing concentrations of BSO, an inhibitor of the enzymatic generation of glutathione, were added to CD82-transfected cells. After 36 h, apoptosis was quantified by FACS. D) Inhibition of glutathione synthesis leads to weak apoptosis induction. HeLa cells were incubated with 200 µM BSO; 46 h later apoptosis was determined by FACS analysis. E) GSH addition reduces ROI detection. After transfection of indicated plasmids, 1.5 mM of glutathione-monoethyl ester, a membrane-permeable GSH form, was added to cells and ROIs were detected with lucigenin 14 h later. F) GSH reduces apoptosis by CD82. CD82transfected cells were treated with 1.5 mM glutathione-monoethyl ester. Apoptosis was quantified by FACS analysis after 36 h.

Fig. 7

Figure 7. GTPase Cdc42 is activated by CD82, mediates apoptosis induction and efflux of glutathione, and allows ROI detection. A) Cdc42 is activated by CD82. HeLa cells were transfected with indicated plasmids for luciferase or CD82. In 1 transfection zVAD, a pan-caspase inhibitor, was added. Sixteen hours later extracts were prepared, and activated (GTP-bound) Cdc42 was coimmunprecipitated from equal quantities of cell lysates with a GST-Pak1 protein that associates only with activated Cdc42, which was subsequently detected in an immunoblot using a specific antibody. Untransfected extracts were treated with GTP or GDP to activate or inactivate Cdc42. B) A dominant-negative Cdc42 reduces apoptosis induction by C33. An expression vector for CD82 and luciferase or dominant-negative CDC42 T17N mutant was cotrasfected at a ratio of 1:4. After 46 h, apoptosis was quantified by FACS analysis. C) A dominant-negative Cdc42 reduces ROI-formation. HeLa cells were transfected with expression vector for luciferase, CD82, and luciferase or CD82 and a dominant-negative Cdc42 mutant (at a ratio: 1:1). After 14 h, superoxide anions were determined by lucigenin. Means ± SD of 3 independent experiments are shown. D) A dominant-negative Cdc42 mutant inhibits efflux of GSH from CD82-transfected cells. HeLa cells were transfected with the indicated plasmids for luciferase, CD82 and luciferase, or CD82 and dominant-negative Cdc42 at a ratio of 1:1. Twenty-five hours after transfection, cells were harvested, and glutathione level in cellular extracts and in the medium was determined. Given are percent values relative to intracellular GSH concentration of control (β-gal)-transfected cells.