Sensitization to CD95 ligand-induced apoptosis in human ... - Nature

ã

Oncogene (2000) 19, 2338 ± 2345 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc

Sensitization to CD95 ligand-induced apoptosis in human glioma cells by hyperthermia involves enhanced cytochrome c release Mirjam Hermisson1, Bettina Wagenknecht1, Hartwig Wolburg2, Tamara Glaser1, Johannes Dichgans1 and Michael Weller*,1 1

Laboratory of Molecular Neuro-Oncology, Department of Neurology, University of TuÈbingen, School of Medicine, TuÈbingen, Germany; 2Institute of Pathology, University of TuÈbingen, School of Medicine, TuÈbingen, Germany

CD95L-induced apoptosis involves caspase activation and is facilitated when RNA and protein synthesis are inhibited. Here, we report that hyperthermia sensitizes malignant glioma cells to CD95L- and APO2L-induced apoptosis in the absence, but not in the presence, of inhibitors of RNA and protein synthesis. Hyperthermia does not alter CD95 expression at the cell surface and does not modulate the morphology of CD95-mediated cell death on electron microscopy. Bcl-2 gene transfer inhibits apoptosis and abrogates the sensitization mediated by hyperthermia. Hyperthermia does not overcome resistance to apoptosis conferred by the viral caspase inhibitor, crm-A, indicating the absolute requirement for the activation of crm-A-sensitive caspases, probably caspase 8, for apoptosis. CD95L-evoked DEVD-amc-cleaving caspase activity is enhanced by hyperthermia, suggesting that hyperthermia operates upstream of caspase processing to promote apoptosis. There is no uniformly enhanced processing of three caspase 3 substrates, poly-ADP ribose polymerase (PARP), protein kinase C (PKC) d and DNA fragmentation factor (DFF) 45. Yet, hyperthermia promotes CD95L-evoked DNA fragmentation. Interestingly, hyperthermia enhances the CD95L-evoked release of cytochrome c in the absence, but not in the presence, of CHX. In contrast, the reduction of the mitochondrial membrane potential is enhanced by hyperthermia both in the absence and presence of CHX, and enhanced cytochrome c release is not associated with signi®cantly enhanced caspase 9 processing. The potentiation of cytochrome c release at hyperthermic conditions in the absence of CHX is abrogated by Bcl-2. Thus, either hyperthermia or inhibition of protein synthesis by CHX potentiate cytotoxic cytokine-induced apoptosis. These pathways show no synergy, but rather redundance, indicating that CHX may function to promote apoptosis in response to cytotoxic cytokines by inhibiting the synthesis of speci®c proteins whose synthesis, function or degradation is temperature-sensitive. Oncogene (2000) 19, 2338 ± 2345. Keywords: CD95; hyperthermia; cytochrome c; glioma; mitochondria

*Correspondence: M Weller, Laboratory of Molecular NeuroOncology, Department of Neurology, University of TuÈbingen, School of Medicine, Hoppe-Seyler-Strasse 3, 72076 TuÈbingen, Germany Received 16 August 1999; revised 29 February 2000; accepted 29 February 2000

Introduction CD95 (Fas/APO-1) is a cell surface cytokine receptor molecule of the nerve growth factor/tumor necrosis factor (TNF) receptor superfamily. Activation of CD95 by agonistic antibodies or the natural ligand, CD95L, triggers apoptosis in susceptible target cells. Induction of apoptosis involves the activation of various caspases and mitochondrial cytochrome c release, acting in a positive feedback loop and resulting in the proteolytic degradation of multiple cellular substrates (Krammer, 1999; Schulze-Ostho et al., 1998). Conversely, inhibition of cytochrome c release from mitochondria is probably an important pathway of Bcl-2/Bcl-xLmediated prevention of apoptosis (Kluck et al., 1997; Kharbanda et al., 1997; Yang et al., 1997). While a reduction of the mitochondrial membrane potential often occurs during apoptosis (Kroemer, 1997), its signi®cance for the execution of the cell death program and speci®cally for cytochrome c release has remained controversial (Bossy-Wetzel et al., 1998; Green and Reed, 1998; Vander-Heiden et al., 1997). Like many other cancer cells, human malignant glioma cell lines express death receptors of the TNF receptor/CD95 family and are susceptible to CD95Land APO2L-mediated apoptosis. Yet, most glioma cell lines require the coexposure to CD95L or APO2L and inhibitors of RNA or protein synthesis such as actinomycin D (ActD) or cycloheximide (CHX) to die in response to death receptor ligation (Rieger et al., 1998; Weller et al., 1994). The putative cytoprotective proteins which inhibit CD95-mediated apoptosis in the absence of ActD or CHX have not been identi®ed. Bcl2 and inhibitor-of-apoptosis (IAP) proteins inhibit CD95L-induced apoptosis but were excluded as the endogenous cytoprotective proteins, synthesis of which is suppressed during ActD- or CHX-mediated potentiation of CD95L-induced apoptosis (Weller et al., 1995; Wagenknecht et al., 1999). Novel candidates include molecules which bind the death domains of the death receptors, such as silencer of death domain (SODD) (Jiang et al., 1999; Takayama et al., 1999). Since CD95 or related receptors are possible targets for immunotherapy of human malignant gliomas (Rieger et al., 1998; Weller et al., 1998), we have been particularly interested in potentiating CD95-mediated apoptosis of glioma cells, e.g., by coexposure to CD95L and cancer chemotherapy drugs (Roth et al., 1997) or by the transfer of potentially proapoptotic genes such as p53 (Pohl et al., 1999), bax (Naumann and Weller, 1998) or PTEN (Wick et al., 1999). Hyperthermia sensitizes human cancer cells to cancer chemotherapy in vitro. These eects of hyperthermia

Sensitization to CD95L-induced cell death by hyperthermia M Hermisson et al

may relate to a facilitation of drug uptake or to a promotion of drug-induced metabolic eects in cancer cells. The eects of temperature on CD95L- or APO2L-induced apoptosis have not been studied in detail. The CD95L-related cytokine, TNF-a, is known to kill cancer cells more eciently at hyperthermic conditions (Klostergaard et al., 1992; Watanabe et al., 1988). Here, we examine the eect of hyperthermia on CD95L- and APO2L-mediated apoptosis and the death receptor-dependent signaling process in human malignant glioma cells.

2339

Results Hyperthermia sensitizes human malignant glioma cells to CD95L-induced apoptosis: dependence on protein synthesis and inhibition by Bcl-2 and crm-A Human glioma cells were exposed to CD95L at 37, 39 or 418C in the absence or presence of an inhibitor of protein synthesis, CHX. In the absence of CHX, there was a temperature-dependent augmentation of apoptosis in all cell lines (Figure 1a ± d, ®lled symbols). No apoptosis-promoting eect of hyperthermia became apparent when the glioma cells were co-treated with CD95L and CHX (open symbols). In the absence of CHX, the EC50 values for CD95L at 37, 39 and 418C were approximately 20, 10 and 5 U/ml for LN-18, 4150, 150 and 50 U/ml for LN-229, 33, 21 and 9 U/ ml for T98G, and 20, 13 and 8 U/ml for U87MG cells. In contrast, in the presence of CHX, there was no temperature-dependent change in the EC50 values for LN-18, T98G or U87MG cells, and heat (418C) even conferred minor protection from CD95L-induced apoptosis in the presence of CHX in LN-229 cells. Similar results to those shown in Figure 1a ± d were obtained when CHX was replaced by an inhibitor of RNA synthesis, ActD (data not shown). Time kinetic experiments showed that either hyperthermia or CHX promoted CD95L-induced apoptosis but that there was no further increase or acceleration of apoptosis when CHX was added at hyperthermic conditions (Figure 1e,f). Further, notably the results for LN-229 cells (Figure 1f) show that hyperthermia did not merely accelerate apoptosis but was actually required for apoptosis to occur. To examine whether hyperthermia altered the mode of CD95L-evoked cell death, electron microscopic analysis was performed at 8 and 16 h after CD95L exposure in the absence or presence of CHX and in parallel at 37 or 418C. These studies revealed no qualitative dierence in the morphology of cell death under all these conditions. Representative views of

Figure 1 Hyperthermia augments CD95L-induced apoptosis. (a ± d) LN-18 (a), LN-229 (b), T98G (c) or U87MG (d) cells were seeded in 96-well plates (104/well), allowed to attach for 24 h, and then treated with CD95L in the absence (®lled symbols) or presence (open symbols) of CHX (10 mg/ml) at 378C (rhomboids), 398C (squares) or 418C (triangles) for 16 h. (e ± f) LN-18 or LN-229 cells were treated with CD95L (LN-18: 15 U/ ml; LN-229: 20 U/ml) in the absence (straight lines, ®lled symbols) or presence (dashed lines, open symbols) of CHX, at 378C (rhomboids) or 418C (squares). (g,h) The cells were treated with APO2L in the absence (®lled symbols) or presence (open

symbols) of CHX (10 mg/ml) at 378C (rhomboids) or 418C (squares) for 16 h. (i,j) LN-18 or LN-229 neo (open symbols, dashed lines) or bcl-2-transfected (®lled symbols, straight lines) cells were treated with CD95L at 378C (rhomboids) or 418C (squares) for 16 h. (k,l) LN-18 or LN-229 puro (open symbols, dashed lines) or crm-A-transfected (®lled symbols, straight lines) cells were treated with CD95L at 378C (rhomboids) or 418C (squares) for 16 h. Survival was assessed by crystal violet staining. Data are expressed as mean percentages of survival and s.e.m. (n=3) relative to untreated cultures or CHX only-treated cultures maintained at the respective temperatures Oncogene


2340

CD95-mediated apoptosis in LN-229 cells are depicted in Figure 2. Further, ¯ow cytometry of LN-18 and LN229 cells maintained at 37, 39 or 418C for 1, 8 or 16 h showed that CD95 expression at the cell surface was unaected by hyperthermia (data not shown). When LN-18 and LN-229 cells were treated with APO2L, a cytotoxic cytokine related to CD95L, at 37 or 418C for 16 h in the absence or presence of CHX, hyperthermia enhanced APO2L-induced cell death in the absence, but not in the presence of CHX in both cell lines, too (Figure 1g,h). The hyperthermia-induced augmentation of apoptosis in LN-18 and LN-229 cells was completely abrogated by ectopic Bcl-2 expression (Figure 1i,j). Ectopic expression of the viral caspase inhibitor, crm-A, provided full protection from CD95L-induced apoptosis even under hyperthermic conditions (Figure 1k,l), con®rming the absolute requirement for caspase 8 activation during CD95Linduced apoptosis. Hyperthermia sensitizes human malignant glioma cells to CD95L-induced apoptosis up-stream of DEVD-amc-cleaving caspases We next asked whether hyperthermia facilitated apoptosis at the level or down-stream of caspase activation. Hyperthermia greatly enhanced Ac-DEVDamc cleavage as a measure of caspase activity when the glioma cells were exposed to CD95L in the absence of CHX (Figure 3a,c). In contrast, there was no signi®cant hyperthermia-mediated increase of caspase activity in CD95L-treated LN-18 in the presence of CHX (Figure 3b), and, consistent with the cytotoxicity data (Figure 1), there was a reduction rather than increase in CD95L-evoked caspase activity in LN-229 co-treated with CD95L and CHX at high temperature (Figure 3d). The levels of the active cleavage product of caspase 8, the proximate caspase during CD95L-dependent signaling, were reduced rather than increased at 418C in both cell lines (Figure 3e). This reduction in p18 levels occurred both in the absence and presence of CHX, suggesting that temperature-dependent changes in caspase 8 processing are not responsible for the enhanced sensitivity to apoptosis in the absence of CHX (Figure 1). Further, the levels of cleaved caspase 3 were unaltered by hyperthermia in both cell lines, both in the absence and presence of CHX. Immunoblot analysis for caspase 2 also showed unaltered or even reduced levels of the active cleavage product, p12/ caspase 2, at 418C. Three substrates of caspase 3 were also analysed by immunoblot analysis (Figure 3f). As previously reported (Wagenknecht et al., 1998), CD95L induced the cleavage of PARP to a 85 kD fragment in the glioma cells. PARP cleavage was not modulated by hyperthermia either in the absence or presence of CHX. Interestingly, a constitutively detected PARP fragment of 80 kD, that has been linked to physiological PARP turnover (Kameshita et Figure 2 Electron microscopic features of CD95L-induced apoptosis. LN-229 cells were treated with CD95L in the presence of CHX at 418C for 8 h (a) or 16 h (b,c). In a, the upper cell shows early apoptotic changes characterized by cytoplasmic blebbing, indented nucleus and irregularly shaped rough endoplasmic cisterns. The lower adjacent cell is still completely normal. In b, advanced apoptosis is characterized by vesiculation

Oncogene

and vacuolation of cytoplasm and the margination of condensed chromatin. In c, there is prominent nuclear blebbing and disintegration. Note multiple compartments containing chromatin and the central cytoplasm which contains masses of vehicles but lacks a nucleus


al., 1989), accumulated under hyperthermic conditions. The formation of the 80 kD PARP fragment was blocked by CHX, suggesting that new protein synthesis is required for physiological turnover of PARP. The cleavage of another substrate of caspase 3, PKCd, was unaltered in LN-18 cells, but consistently enhanced in LN-229 cells, by hyperthermia. This hyperthermia-induced enhancement of PKCd cleavage did not occur in LN-229 cells cotreated with CD95L and CHX. CD95L induced the cleavage of DFF45, too, but DFF45 cleavage was unaltered by hyperthermia in both cell lines (data not shown). To assess whether the apparently unaltered DFF45 processing at hyperthermic conditions would translate into unaltered levels of DNA fragmentation, despite enhanced cell death, we measured DNA fragmentation after exposure to CD95L in the absence or presence of CHX at 37 or 418C (Figure 3g). Hyperthermia signi®cantly enhanced CD95L-induced DNA fragmentation in the absence of CHX, consistent with the cytotoxicity data (Figure 1). As expected, coexposure to CD95L and CHX at 378C greatly enhanced DNA fragmentation compared with exposure to CD95L alone. However, CD95L-induced DNA fragmentation at 418C was reduced rather than enhanced by CHX, suggesting an enhanced turnover of proteins required for DNA fragmentation at 418C, synthesis of which is abolished by CHX.

2341

Hyperthermia enhances the release of cytochrome c from mitochondria in CD95L-induced apoptosis Cytochrome c release from mitochondria may be a critical factor for eective activation of caspases during CD95L-induced apoptosis in nonglial cells (Li et al., 1998; Luo et al., 1998; Srinivasan et al., 1998). Therefore, we next asked whether CD95L evoked the release of cytochrome c in malignant glioma cells, too, and whether this release was enhanced by hyperthermia. At 2 ± 4 h after CD95L treatment, cytochrome c release was much more prominent at 418C than at 378C in both cell lines (Figure 4a). While the addition of CHX enhanced cytochrome c release at 4 h at 378C in both cell lines, no such eect was seen at 418C. Caspase 9, the proximate candidate caspase activated by cytosolic cytochrome c in association with apoptotic protease-activating factor (APAF-1) (Kuida et al., 1998; Hu et al., 1998), showed moderately enhanced processing at 418C in LN-18 cells in the absence, but not presence, of CHX. No such eect was seen in LN229 cells (Figure 4b). Since Bcl-2 abrogates the

Figure 3 Hyperthermia modulates CD95L-evoked caspase activity and DNA fragmentation. (a ± d) LN-18 or LN-229 cells were seeded in 96-well plates (104/well), allowed to attach for 24 h, and then treated for 4 h with CD95L in the absence (a,c) or presence (b,d) of CHX (10 mg/ml) at 378C (®lled rhomboids), 398C (open squares) or 418C (®lled triangles). DEVD-amc-cleaving caspase activity was assessed as described in Materials and methods. Data are expressed as mean optical densities (n=3). s.e.m. were below

10%. (e) LN-18 or LN-229 cells were treated for 4 h with 10 U/ ml (LN-18) or 100 U/ml (LN-229) CD95L in the absence of CHX, or with 5 U/ml (LN-18) or 20 U/ml (LN-229) in the presence of CHX (10 mg/ml) at the indicated temperatures. Control cells were untreated (left) or treated with CHX alone (right). Soluble protein lysates were prepared as described and 20 mg per lane subjected to SDS ± PAGE and immunoblot analysis for caspases 8, 3 and 2. (f) Lysates prepared as in e were examined for PARP and PKCd processing. PARP processing by caspases gives rise to a 85 kD fragment, PKCd to a 40 kD fragment. (g) LN-18 or LN-229 cells were untreated, exposed to CD95L, CHX or both at 378C (open bars) or 418C (black bars) (*P50.05, treated cells compared with untreated cells at the same temperature, ANOVA; +P50.05, 418C compared with 378C, t-test) Oncogene


2342

Figure 4 Hyperthermia enhances CD95L-induced cytochrome c release from mitochondria: prevention by Bcl-2. (a) LN-18 or LN229 cells were treated with 10 U/ml (LN-18) or 100 U/ml (LN-229) CD95L in the absence of CHX, or with 5 U/ml (LN-18) or 20 U/ml (LN-229) in the presence of CHX (10 mg/ml) for the indicated times at 378C or 418C. (b) The cells were treated as in a and examined for caspase 9 processing at 2 or 4 h after exposure to CD95L in the absence or presence of CHX at 378C or 418C. (c) LN18 or LN-229 neo or bcl-2 cells were treated as the wild-type cells in a for 2 or 4 h. Cytosolic cytochrome c was extracted as described and analysed by SDS ± PAGE and immunoblotting. (d,e) LN-18 or LN-229 cells were treated with 10 U/ml (LN-18) or 100 U/ml (LN-229) CD95L in the absence of CHX, or with 5 U/ml (LN-18) or 20 U/ml (LN-229) in the presence of CHX (10 mg/ ml) for 4 h at 378C or 418C. (d) The mitochondrial membrane potential was analysed by ¯ow cytometry, as described in Materials and methods. For illustration, CD95L-treated LN-229 cells are shown. The dotted lines on the left are untreated, unstained cells that are not labeled with CMXRos. Straight lines are untreated, but stained, control cells as control. Dashed lines are cells treated with mClCCP, as a positive control (middle). Straight fat lines are cells treated with CD95L. Note the shift of the fat line to the left at 418C compared with 378C, indicative of the loss of the mitochondrial membrane potential at this temperature. (e) The loss of the mitochondrial membrane potential at 378C (open bars) or 418C (black bars) after exposure to CD95L, CHX or both was quanti®ed by CellQuest software and is expressed as percentage of maximal (mClCCP-induced) loss. A representative experiment is shown. Similar results were obtained in three dierent experiments

sensitizing eect of hyperthermia (Figure 1i,j), we next asked whether Bcl-2 also abrogated the potentiation of cytochrome c release mediated by hyperthermia. As shown in Figure 4c, this was indeed the case. No signi®cant cytochrome c release occurred in LN-18-bcl2 cells, in contrast to the neo control cells. In LN-229bcl-2 cells, there was only weak cytochrome c release that was not enhanced under hyperthermic conditions. Hyperthermia enhances CD95L-induced reduction of the mitochondrial membrane potential in the absence or presence of CHX Cytochrome c release often correlates with a reduction of the mitochondrial membrane potential DCm (Kroemer, 1997). Therefore, we next asked whether DCm is reduced during CD95L-induced apoptosis in glioma cells and whether this reduction is modulated by hyperthermia. LN-18 and LN-229 cells were treated with CD95L in the absence or presence of CHX at 37 or 418C and the reduction of DCm measured after 4 h (Figure 4d). In both cell lines, hyperthermia greatly enhanced the CD95L-evoked reduction of DCm both in the absence and presence of CHX (Figure 4e). Since CHX-treated cells showed a distinct augmentation of the reduction of DCm at hyperthermia, whereas cytochrome c release was not enhanced under these conditions (Figure 4a), loss of the mitochondrial membrane potential and the release of cytochrome c from mitochondria seem not to be coupled events Oncogene

during CD95L-induced apoptosis in human malignant glioma cells. Discussion Human malignant gliomas are highly aggressive neoplasms which fail to respond to current treatment modalities. Cultured glioma cells are rather resistant to multiple proapoptotic stimuli, including cancer chemotherapy drugs and irradiation. In contrast, many glioma cells express death receptors of the TNF receptor/CD95 family and are susceptible to CD95Land APO2L-induced apoptosis (Rieger et al., 1998; Weller et al., 1994; Yount et al., 1998). Often, however, coexposure to these death ligands and to inhibitors of RNA or protein synthesis is required to sensitize glioma cells to apoptosis. In an eort to further explore a possible therapeutic role for cytotoxic cytokines in malignant glioma, we asked whether hyperthermia might overcome the partial resistance of these cells to CD95L-induced apoptosis. Here, we report that CD95L- and APO2L-induced apoptosis of glioma cells is greatly enhanced by hyperthermia (418C) (Figure 1). The enhanced sensitivity to apoptosis is not caused by a hyperthermia-mediated up-regulation of CD95 expression at the cell surface. Neither the levels of enzymatically active cleavage products of caspases 2, 3 and 8 nor the levels of cleavage products of various caspase 3 substrates,


PARP, PKCd and DFF45 (Figure 3) increased under hyperthermic conditions. Hyperthermia nevertheless enhanced the CD95L-induced cleavage of DEVD-amc (Figure 3a,c), indicative of a hyperthermia-dependent potentiation of the killing cascade upstream of caspase 3 activation. Thus, CD95L-evoked DNA fragmentation was also enhanced by hyperthermia even though immunoblot analysis failed to reveal increased caspase 3-dependent processing of DFF45, the putative down-stream mediator of DNA fragmentation during caspase-induced apoptosis (Liu et al., 1997). The activation of various caspases during apoptosis depends on the release of cytochrome c from mitochondria (Pastorino et al., 1998; Yang et al., 1997; Li et al., 1997). Cytochrome c acts as a cofactor of APAF-1, the human homolog of the cell death factor CED-4 in C. elegans, to activate caspase 9 (Kuida et al., 1998; Hu et al., 1998). Activation of caspase 9 may provide the critical link between mitochondrial events, e.g. cytochrome c release, and activation of the cytosolic caspase machinery (Slee et al., 1999). The present study indicates that hyperthermia facilitates mitochondrial cytochrome c release and that this eect may be responsible for the potentiation of cytotoxic cytokine-induced cell death by hyperthermia. However, we were not able to demonstrate a consistent increase in caspase 9 processing associated with enhanced cytochrome c release (Figure 4b). In that regard, it is interesting to note that caspase 9 has been suggested to exhibit activity in the absence of processing (Stennicke et al., 1999). Further, the limited role of caspase 9 in CD95L-induced apoptosis of glioma cells suggested by the inecient processing observed here (Figure 4b) is consistent with the unaltered sensitivity of thymocytes from caspase 97/7 mice to CD95-mediated apoptosis (Hakem et al., 1998). That crm-A-transfected glioma cells were completely resistant to CD95L-induced apoptosis even at hyperthermic conditions (Figure 1k,l), indicates: (i) the absolute requirement for caspase 8 for apoptosis; and (ii) that down-stream of mitochondria, a self-potentiating feedback loop connecting caspases 9, 8 and 3 (Slee et al., 1999) mediates apoptosis, with caspase 8, the putative target of crm-A, being an integral component of the killing cascade. Also, the abrogation of cell death by crm-A indicates that hyperthermia does not facilitate a second, caspaseindependent, nonapoptotic death pathway in glioma cells. Cytochrome c release from mitochondria has been shown to be associated with a reduction of the mitochondrial membrane potential, DCm (Kroemer, 1997). Several studies indicate that DCm reduction is not always associated with cytochrome c release, or vice versa (Bossy-Wetzel et al., 1998; Green and Reed, 1998; Vander-Heiden et al., 1997). We ®nd that the two events are dissociated during hyperthermia-mediated augmentation of CD95L-induced apoptosis. Cytochrome c release was only enhanced by hyperthermia in the absence of CHX (Figure 4a), which correlated with enhanced cell death. In contrast, hyperthermia enhanced the loss of the mitochrondrial membrane potential both in the absence and presence of CHX (Figure 4e) although death was not enhanced in the presence of CHX by hyperthermia. That the release of cytochrome c from mitochondria is the principal

mechanism underlying hyperthermia-induced augmentation of apoptosis, is compatible with the strong protection from apoptosis mediated by Bcl-2 (Figure 1). This is because prevention of mitochondrial events, probably chie¯y cytochrome c release, is now thought to be a major mechanism underlying Bcl-2-dependent inhibition of apoptosis (Green and Reed, 1998; Mignotte and Vayssiere, 1998). The potentiation of apoptosis by hyperthermia in the absence, but not in the presence, of CHX suggests that the pathways of enhancing apoptosis by hyperthermia, or inhibition of protein synthesis by CHX, are antagonistic or redundant. We propose that the underlying mechanisms of sensitization are redundant and that hyperthermia may therefore obviate the need for concurrent inhibition of protein synthesis to induce apoptosis. This hypothesis might have important implications for cancer therapy. Hyperthermia is already used in clinical practice as a sensitizer to chemotherapy and radiotherapy (Sminia et al., 1994). Further, hyperthermia has been reported to enhance the cytotoxicity of TNFa in lymphoid cells even though the molecular basis of this eect has remained obscure (Klostergaard et al., 1992; Watanabe et al., 1988). Hyperthermia may become a useful adjunct to fully exploit the therapeutic potential of cytotoxic cytokines such as CD95L or APO2L, and enhanced cytochrome c release from mitochondria may be the biological event mediating this eect.

2343

Materials and methods Reagents and cell lines ActD and CHX were purchased from Sigma (St. Louis, MO, USA). Acetyl-Asp-Glu-Val-Asp-chloromethylcoumarin (AcDEVD-amc) was obtained from Biomol (Plymouth Meeting, PA, USA). CMXRos and m-chlorophenolhydrazon (mClCCP) were from Molecular Probes (Eugene, OR, USA). CD95L was obtained from the supernatant of CD95L-transfected N2A murine neuroblastoma cells (Roth et al., 1997). APO2L was kindly provided by Dr A Ashkenazi (Genentech South, San Francisco, CA, USA). LN-18, LN-229, U87MG and T98G human malignant glioma cells were kindly provided by Dr N de Tribolet (Lausanne, Switzerland). Glioma cell lines expressing murine bcl-2 were generated as described (Weller et al., 1995). These cells were compared with neo control cells which harbor the same plasmid without the bcl-2 insert. Crm-A cell lines were obtained using the Flag-crm-A-puro construct and were compared with puro control cells (Wagenknecht et al., 1998). The cells were maintained in DMEM containing 10% fetal calf serum, 2 mM glutamine and 1% penicillin/streptomycine. Viability assay Glioma cell viability was measured by crystal violet staining. The cell culture medium was removed and surviving cells stained with 0.5% crystal violet in 20% methanol for 10 min at room temperature. The plates were washed extensively under running tap water, air-dried and optical density values read in an ELISA reader at 550 nm wave length. Electron microscopy Glioma cells treated as indicated were ®xed in 2.5% glutaraldehyde (Paesel-Lorei, Frankfurt, Germany) in Hank's modi®ed salt solution (HMSS), post®xed in 1% OsO4 in 0.1 M cacodylate buer, scraped o from the plastic and Oncogene


2344

dehydrated in ethanol. The 70% ethanol step was saturated with uranyl acetate for contrast enhancement. Dehydration was completed in propylene oxide. The specimens were embedded in Araldite (Serva, Heidelberg, Germany). Ultrathin sections were produced on a FCR Reichert Ultracut ultramicrotome (Leica, Bensheim, Germany), mounted on pioloform-coated copper grids and contrasted with lead citrate. Specimens were analysed and documented with an EM 10A electron microscope (Zeiss, Oberkochen, Germany). DEVD-amc-cleaving caspase activity The cells were seeded in 96-well plates (10 000 cells per well) and allowed to attach for 24 h. The cells were treated with CD95L as indicated and lysed in lysis buer containing 25 mM TRIS-HCl (pH 8.0), 60 mM NaCl, 2.5 mM EDTA and 0.25% NP40 for 10 min. Ac-DEVD-amc (12.5 mM), diluted in PBS, was added and incubated at 378C for 10 min. Caspase activity was measured for 1 h using a CytoFluor 2350 Millipore ¯uorimeter at 360 nm excitation and 480 nm emission wave lengths. Immunoblot analysis Soluble protein lysates were obtained from subcon¯uent glioma cell cultures. SDS ± PAGE with electroblotting was performed as described (Weller et al., 1994). The rabbit polyclonal PARP antibody was purchased from Boehringer (Mannheim, Germany), rabbit polyclonal PKCd and goat polyclonal caspase-2 antibodies from Santa Cruz Biotechnology (Santacruz, CA, USA), mouse monoclonal caspase 3 antibody from Transduction Laboratories (Lexington, KY, USA), mouse monoclonal cytochrome c antibody from PharMingen (San Diego, CA, USA). The caspase 8 antibody C15 (mouse monoclonal) was kindly provided by Dr PH Krammer (Heidelberg, Germany). The caspase 9 antibody was kindly provided by Dr Y Lazebnik (Cold Spring Harbour, NY, USA). The DFF antibody was kindly provided by Dr X Wang (Howard-Hughes Medical Institute, Dallas, TX, USA). The secondary antibodies, protein A and anti-mouse IgG, were purchased from Amersham (Braunschweig, Germany), anti-goat antibody was from Santa Cruz. Enhanced chemoluminescence (ECL, Amersham) was used for detection. Measurement of DNA fragmentation The cells were treated as indicated and pelleted by centrifugation (7 min, 1200 r.p.m., 48C). The pellet was dissolved in lysis buer (10 mM TRIS-HCl pH 7.5, 10 mM EDTA, 0.2% Triton X-100) and lysed for 10 min at 48C. The lysates were centrifuged at 13 000 r.p.m. and 48C. Supernatant was taken for the measurement of fragmented DNA. The pellet with the intact DNA was again dissolved in lysis buer and sonicated for 20 s. RNase (100 mg/ml) was added to all samples. RNA was digested for 2 h at 378C. All samples were diluted in buer containing 5 mM TRIS-HCl pH 7.5, 0.5 mM EDTA and 0.5 mM ethidium bromide and measured for the content of DNA using a CytoFluor 2350 Millipore ¯uorimeter at 530 nm excitation and 620 nm emission wave lengths.

Flow cytometry for CD95 cell surface expression The cells were washed in PBS and then incubated in ¯ow cytometry buer (1% bovine serum albumin (BSA), 0.01% sodium azide in PBS) containing 10% sheep serum for 20 min at 48C. After centrifugation, the cells were resuspended in ¯ow cytometry buer with 1 mg/ml anti-CD95 antibody (mouse IgG1, Immunotechnology, Hamburg, Germany) or 1 mg/ml mouse IgG1 as a control. After 1 h incubation the cells were washed and then incubated with FITC-labeled sheep anti-mouse IgG (Sigma), diluted 1 : 256, for 20 min at 48C. The cells were washed, ®xed in 1% formaldehyde and analysed measured for cell surface expression of CD95 by ¯ow cytometry. The level of expression was calculated as the speci®c ¯uorescence index (SFI) derived from the ratio of ¯uorescent signal obtained with the speci®c CD95 antibody and an isotype control antibody. Measurement of cytochrome c release The cells were washed with PBS and lysed with MSH (mannitol, sucrose, HEPES) buer plus digitonin (210 mM Dmannitol, 70 mM sucrose, 10 mM HEPES, 200 mM EGTA, 5 mM succinate, 0.15% BSA, 40 mg/ml digitonin) at 48C. After lysis, the supernatant was removed and centrifuged immediately for 10 min at 13 000 g. An equal volume of 10% trichloroacetic acid was added to the supernatant. Samples were kept at 7208C for at least 30 min. After another centrifugation (15 min at 13 000 g), the pellets were dissolved in sample buer (50 mM TRIS-HCl, pH 6.8, 2% SDS, 0.1% bromphenolblue, 10% glycerol) and then analysed for cytochrome c content by SDS ± PAGE and immunoblot analysis. Measurement of DCm by flow cytometry Five6105 cells were incubated with 200 nM CMXRos for 20 min in culture medium at 378C and 5% CO2. As a positive control for DCm loss, the cells were incubated with mClCCP (1 mM), a protonophore disrupting DCm. Cells were centrifuged at 48C (1200 r.p.m., 5 min) and ®xed in 1% paraformaldehyde in PBS, pH 7.4, prior to analysis by ¯ow cytometry. Data acquisition and analysis were performed using the CellQuest software. Statisical analysis The quantitative data are from triplicate experiments repeated three times with similar results. EC50 values for CD95L-induced apoptosis in various transfected cell lines were determined by linear regression and compared by t-test.

Acknowledgments Supported by grants from the FortuÈne program of the University of TuÈbingen Medical School (M Hermisson), Deutsche Forschungsgemeinschaft (We 1502/3-2) and the IZKF TuÈbingen to M Weller. We would like to thank EvaMaria Knittel for skilful technical assistance in electron microscopy.

References Bossy-Wetzel E, Newmeyer D and Green D. (1998) EMBO J., 17, 37 ± 49. Green D and Reed J. (1998). Science, 281, 1309 ± 1312. Hakem R, Hakem A, Duncan GS, Henderson JT, Woo M, Soengas MS, Elia A, de la Pompa JL, Kagi D, Khoo W, Potter J, Yoshida R, Kaufman SA, Lowe SW, Penninger JM and Mak TW. (1998). Cell, 94, 339 ± 352. Oncogene

Hu Y, Benedict M, Wu D, Inohara N and NuÂnez G. (1998). Proc. Natl. Acad. Sci. USA, 95, 4386 ± 4391. Jiang Y, Woronicz JD, Liu W and Goeddel DV. (1999). Science, 283, 543 ± 546. Kameshita I, Mitsuuchi Y, Matsuda M and Shizuta Y. (1989). Jacobson MK and Jacobson E. (eds). New York: Springer-Verlag, pp. 71 ± 75.


Kharbanda S, Pandey P, Scho®eld L, Israels S, Roncinske R, Yoshida K, Bharti A, Yuan Z-M, Saxena S, Weichselbaum R, Nalin C and Kufe D. (1997). Proc. Natl. Acad. Sci. USA, 94, 6939 ± 6942. Klostergaard J, Leroux E, Siddik Z, Khodadadian M and Tomasovic S. (1992). Cancer Res., 52, 5271 ± 5277. Kluck R, Bossy-Wetzel E, Green D and Newmeyer D. (1997). Science, 275, 1132 ± 1136. Krammer P. (1999). Adv. Immunol., 71, 163 ± 210. Kroemer G. (1997). Cell Death Dier., 4, 443 ± 456. Kuida K, Haydar T, Kuan C-Y, Gu Y, Taya C, Karasuyama H, Su M-S, Rakic P and Flavell R. (1998). Cell, 94, 325 ± 337. Li F, Srinivasan A, Wang Y, Armstrong R, Tomaselli K and Fritz L. (1997). J. Biol. Chem., 272, 30299 ± 30305. Li H, Zhu H, Xu C and Yuan J. (1998). Cell, 21, 491 ± 501. Liu X, Zou H, Slaughter C and Wang X. (1997). Cell, 89, 175 ± 184. Luo X, Budihardjo I, Zou H, Slaughter C and Wang X. (1998). Cell, 94, 481 ± 490. Mignotte B and Vayssiere J-L. (1998). Eur. J. Biochem., 252, 1 ± 15. Naumann U and Weller M. (1998). Int. J. Cancer, 77, 645 ± 648. Pastorino J, Chen S-T, Tafani M, Snyder J and Farber J. (1998). J. Biol. Chem., 273, 7770 ± 7775. Pohl U, Wagenknecht B, Naumann U and Weller M. (1999). Cell. Physiol. Biochem., 9, 29 ± 37. Rieger J, Naumann U, Glaser T, Ashkenazi A and Weller M. (1998). FEBS Lett., 427, 124 ± 128. Roth W, Fontana A, Trepel M, Reed J, Dichgans J and Weller M. (1997). Cancer Immunol. Immunother., 44, 55 ± 63. Schulze-Ostho K, Ferrari D, Los M, Wesselborg S and Peter M. (1998). Eur. J. Biochem., 254, 439 ± 459.

Slee E, Harte M, Kluck R, Wolf B, Casiano C, Newmeyer D, Wang H, Reed J, Nicholson D, Alnemri E, Green D and Martin S. (1999). J. Cell Biol., 144, 281 ± 292. Sminia P, van der Zee J, Wondergem J and Haveman J. (1994). Int. J. Hyperthermia, 10, 1 ± 30. Srinivasan A, Li F, Wong A, Kodandapani L, Smidt R, Krebs J, Fritz L, Wu J and Tomaselli K. (1998). J. Biol. Chem., 273, 4523 ± 4529. Stennicke HR, Deveraux QL, Humke EW, Reed JC, Dixit VM and Salvesen GS. (1999). J. Biol. Chem., 274, 8359 ± 8362. Takayama S, Xie Z and Reed JC. (1999). J. Biol. Chem., 274, 781 ± 786. Vander-Heiden M, Chandel N, Williamson E, Schumacker P and Thompson C. (1997). Cell, 91, 627 ± 637. Wagenknecht B, Schulz J, Gulbins E and Weller M. (1998). Cell Death Dier., 5, 894 ± 900. Wagenknecht B, Glaser T, Naumann U, KuÈgler S, Isenmann S, BaÈhr M, Korneluk R, Liston P and Weller M. (1999). Cell Death Dier., 6, 370 ± 376. Watanabe N, Niitsu Y, Umeno H, Sone H, Neda H, Yamauchi N, Maeda M and Urushizaki I. (1988). Cancer Res., 48, 650 ± 653. Weller M, Frei K, Groscurth P, Krammer P, Yonekawa Y and Fontana A. (1994). J. Clin. Invest., 94, 954 ± 964. Weller M, Malipiero U, Aguzzi A, Reed J and Fontana A. (1995). J. Clin. Invest., 95, 2633 ± 2643. Weller M, Kleihues P, Dichgans J and Ohgaki H. (1998). Brain Pathol., 8, 285 ± 293. Wick W, Furnari F, Naumann U, Cavenee W and Weller M. (1999). Oncogene, 18, 3936 ± 3943. Yang J, Liu X, Bhalla K, Kim C, Ibrado A, Cai J, Peng T, Jones D and Wang X. (1997). Science, 275, 1129 ± 1132. Yount G, Haas-Kogan D, Levine K, Aldape K and Israel M. (1998). Cancer Res., 58, 3819 ± 3825.

2345

Oncogene