Immunology and Cell Biology (2010) 88, 545–554 & 2010 Australasian Society for Immunology Inc. All rights reserved 0818-9641/10 $32.00 www.nature.com/icb
Granzyme B of cytotoxic T cells induces extramitochondrial reactive oxygen species production via caspase-dependent NADPH oxidase activation Juan I Aguilo´1, Alberto Anel1,5, Elena Catala´n1, Alvaro Sebastia´n1, Rebeca Acı´n-Pe´rez1, Javier Naval1, Reinhard Wallich2, Markus M Simon3 and Julia´n Pardo1,3,4,5 Induction of reactive oxygen species (ROS) is a hallmark of granzyme B (gzmB)-mediated pro-apoptotic processes and target cell death. However, it is unclear to what extent the generated ROS derive from mitochondrial and/or extra-mitochondrial sources. To clarify this point, we have produced a mutant EL4 cell line, termed EL4-q0, which lacks mitochondrial DNA, associated with a decreased mitochondrial membrane potential and a defective ROS production through the electron transport chain of oxidative phosphorylation. When incubated with either recombinant gzmB plus streptolysin or ex vivo gzmB+ cytotoxic T cells, EL4-q0 cells showed phosphatydylserine translocation, caspase 3 activation, Bak conformational change, cytochrome c release and apoptotic morphology comparable to EL4 cells. Moreover, EL4-q0 cells produced ROS at levels similar to EL4 under these conditions. GzmB-mediated ROS production was almost totally abolished in both cell lines by the pan-caspase inhibitor, Z-VAD-fmk. However, addition of apocynin, a specific inhibitor of nicotinamide adenine dinucleotide phosphate (NADPH) oxidases, led to a significant reduction of ROS production and cell death only in EL4-q0 but not EL4 cells. These data suggest that gzmB-induced cell death is accompanied by a caspase-dependent pathway of extra-mitochondrial ROS production, most probably through activation of NADPH oxidase. Immunology and Cell Biology (2010) 88, 545–554; doi:10.1038/icb.2010.5; published online 2 February 2010 Keywords: granzyme B; cell death; reactive oxygen species; NADPH oxidase
Cytotoxic T cells (Tc cells) are key components of the host immune system against intracellular parasites and tumors. Tc cells elicit their cytolytic potential either by ligation of death receptors or by granule exocytosis.1–4 During the latter process, perforin,5 and a family of serine-proteases, termed granzymes (gzm)6,7 are delivered to target cells. There gzmA and gzmB induce multiple distinct and overlapping pro-apoptotic pathways,8–11 leading to cell death.12–15 By using either purified gzmB or Tc cells from gzmA / mice (gzmB+Tc), it was shown that mitochondrial damage and caspase-3 activation are key steps in gzmB-induced apoptosis.11–13,16 Further studies revealed that gzmB induces two distinct pathways leading to activation of caspase-3: (a) a direct one, by processing of caspase-3 through cleavage of the proenzyme17 and (b) an indirect one, through a mitochondrial pathway, by processing Bid, a BH3 only member of the Bcl-2 family, to functionally active tBid,18 which activates Bak/Bax, resulting in cytochrome c (cyt c) release from mitochondria. Cyt c then forms a complex in the cytosol with Apaf-1, termed apoptosome, leading to sequential activation of caspase-9 and caspase-3.19 By using ex vivo-derived virus-specific gzmB+Tc, we have recently shown that
gzmB-induced cell death is accompanied by the production of reactive oxygen species (ROS). This process is critically dependent on activation of caspases-3 and -7, but independent of the mitochondrial pathway, including Bid, Bak, Bax and cyt c release.20 However, this and earlier related studies did not reveal the intracellular source of gzmBinduced ROS production.12,21–23 Moreover, it is not clear at what extent mitochondrial ROS are implicated in cell death induced by Tc cells and gzmB. A study reported that ROS scavengers were able to protect against lysis induced by Tc cells.23 However, later on it was shown that those scavengers may interfere with perforin function inhibiting gzm release into the cytosol of target cells.24 Owing to their redox activity, mitochondria are thought to be the major cytoplasmic organelle for ROS production, but other extra-mitochondrial sources such as the membrane-associated nicotinamide adenine dinucleotide phosphate (NADPH) oxidase of neutrophils, non-phagocytic analogs thereof or peroxisomes have been described.25,26 NADPH oxidase was discovered originally in neutrophils, in which it is a potent source of ROS during phagocytosis and has a critical role
de Bioquı´mica y Biologı´a Molecular y Celular, Universidad de Zaragoza, Zaragoza, Spain; 2Institut fu¨r Immunologie, Universita¨tsklinikum Heidelberg, Heidelberg, Germany; 3Metschnikoff Laboratory, Max-Planck Institute for Immunobiology, Freiburg, Germany and 4Fundacio´n Arago´n I+D/Gobierno de Arago´n, Zaragoza, Spain 5These authors share senior authorship. Correspondence: Dr J Pardo or Dr A Anel, Departamento de Bioquı´mica y Biologı´a Molecular y Celular, The University of Zaragoza, Pz San Francisco s/n, Zaragoza 50009, Spain. E-mails: [email protected]
or [email protected]
Received 22 September 2009; revised 21 December 2009; accepted 23 December 2009; published online 2 February 2010
gzmB induces extramitochondrial ROS JI Aguilo´ et al 546
in the innate host response against pathogens.27 However, recent studies have shown that non-phagocytic cells, including fibroblasts, endothelial cells, hepatocytes, lymphocytes and tumor cell lines also express ROS producing enzymes analogous to the phagocyte NADPH oxidase.26,28,29 The classical neutrophil NADPH oxidase comprises a membraneassociated complex composed of a catalytic moiety, gp91phox and one p22phox subunit and acquires its enzymatic activity on recruitment of three cytosolic subunits (p47phox p40phox and p67phox), together with Rac and Rap1 adapters.26 Recently, several homologs of gp91phox the key catalytic subunit of NADPH oxidase, and termed Nox for NADPH oxidase have been described. Using this new terminology, Nox2 represents the neutrophil gp91phox while Nox1, Nox3, Nox4 and Nox5 have been shown to be differentially expressed in non-phagocytic cells.26,30,31 In contrast to neutrophil oxidase, a significant proportion of the non-phagocytic NADPH oxidase(s) subunits in unstimulated cells are found intracellularly as preformed functional active complexes, associated with cytoskeleton.26 These non-phagocytic enzymes can be upregulated in response to various stimuli, such as growth factors and cytokines and the ROS produced may either serve as stimulus for gene transcription or lead to oxidative damage and cell death.26,30–32 To identify the cytosolic source of gzmB-induced ROS, we have established a mutant of the tumor cell line EL4 lacking mitochondrial DNA, termed EL4-r0 cells. Owing to deletion of mitochondrial genes EL4-r0 cells lack several components of the electron transport chain (ETC) of oxidative phosphorylation and are defective in generating mitochondrial ROS. We show here that purified gzmB or ex vivoderived virus-immune gzmB+Tc were able to induce high levels of ROS in EL4-r0 cells, independent of mitochondria, and suggest that NADPH oxidase is one source of ROS produced during gzmBinduced cell death in the mouse system. RESULTS Generation of mitochondrial DNA-deficient EL4 cells (EL4-q0) EL4 cells depleted of mitochondrial DNA, termed EL4-r0, were generated by propagating the parental EL4 cell line for at least 2 months at low doses of ethidium bromide, according to published protocols.33–35 As shown in Figure 1a, the level of mitochondrial DNA was reduced to near background levels in EL4-r0 compared with EL4 cells, corroborating previous findings with other cell lines.36 Moreover, unlike EL4 cells, EL4-r0 cells were unable to consume oxygen (Figure 1b), most probably because of the lack of several mitochondrial DNA-encoded electron transport carrier proteins, critical for oxidative phosphorylation. Treatment of the two cell lines with menadione, a ketone that displaces ubiquinone from the mitochondrial electron transport chain,37 led to a dose-dependent induction of oxygen superoxide (Figure 1c, left panel) and cell death (Figure 1c, middle panel) only in EL4, but not EL4-r0 cells. The selective specificity of menadione for mitochondria was ascertained by showing that menadione-induced ROS production in EL4 cells was inhibited, in a dose-dependent manner, by mito Q, a specific inhibitor of mitochondrial ROS,38 but not by its non-quinonic derivative, decyltriphenylphosphonium (Figure 1d). Moreover, the cytotoxic drug adaphostine, the sensitivity to which strictly depends on the generation of mitochondrial ROS,39 was only effective in EL4, but not in EL4-r0 cells (Figure 1c, right lower panel). These results clearly show that EL4-r0 cells are defective in mitochondrial ROS production. It is noteworthy that in EL4-r0 cells, staurosporine induces Bak conformational change, cyt c release and caspase 3 activation, despite the absence of mitochondrial ROS production (Supplementary Immunology and Cell Biology
Figure 1). The data indicate that mitochondrial apoptotic pathways are operative in these cells. Nevertheless, a prerequisite for further analysis of the intracellular source of gzmB-induced ROS production was to verify that deletion of mitochondrial DNA in EL4-r0 cells did not affect key proapoptotic processes implicated in gzmB-mediated cell death, including caspase activation, cyt c release and phosphatidylserine (PS) exposure.13,20,40,41 Both EL4 and EL4-r0 cells expressed similar amounts of Bid, caspase-3, cyt c, caspase-9, AIF, X-IAP or Cu/Zn SOD (Figure 1e). Thus, EL4-r0 cells are suitable targets to address the questions presented in this study. EL4 and EL4-q0 cells are similarly sensitive to gzmB (+SLO) or ex vivo virus-specific gzmB+Tc induced pro-apoptotic processes and cell death, including ROS production EL4 and EL4-r0 cells were incubated with recombinant (rec) gzmB in the presence of sublytic doses of streptolysin-O (SLO;42) and tested for proapoptotic markers, including loss of mitochondrial membrane potential (DCm), ROS production, PS exposure at the plasma membrane, caspase activation and nuclear fragmentation. As shown in Figure 2a, incubation of EL4 cells with recgzmB plus SLO leads to a significant loss of mitochondrial membrane potential (DCm), as monitored with DiOC6(3) (58%), compared with untreated EL4 cells (26%) or cells incubated with either SLO (28%) or recgzmB (21%) alone. In contrast, the lowered basal level of DCm observed in untreated EL4-r0 cells—because of the loss of their mitochondrial electron transport chain—was unaffected by recgzmB plus SLO (Figure 2a, lower panels). This lower basal DCm is reflected in a change in the region setting with respect to parental EL4 cells, because the analysis should reflect in each cell type the reduction in DCm induced specifically by gzmB over the basal level. Nevertheless, when incubated with recgzmB plus SLO, a similar percentage of both EL4 and EL4-r0 cells produced ROS. EL4 and EL4-r0 cells were also indistinguishable in their sensitivity to recgzmB/SLO-induced PS exposure on plasma membrane (Figure 2b), propidium iodide uptake (data not shown), nuclear apoptotic morphology (Figure 2c) and caspase 3 activation (Figure 2d). These effects were not observed in mock-treated EL4 and EL4-r0 cells or those incubated with either SLO or gzmB alone. Finally, both ROS production and PS exposure was similar in EL4-r0 and EL4 cells at all gzmB doses used (between 5 and 20 mg ml–1; Figure 2e). Incubation of EL4 and EL4-r0 cells with ex vivo-derived LCMVimmune Tc cells from gzmA / mice (gzmB+CTL20) led to significant DCm loss in peptide pulsed (75%) versus untreated (19%) EL4 cells (Figure 3a). Again, as observed before with rec gzmB plus SLO, no further reduction of the already lowered Cm in EL4-r0 cells was observed using gzmB+Tc under similar conditions. However, gzmB+Tc specifically induced ROS production in a similar percentage of EL4 ( gp33, 5%; +gp33, 48%) and EL4-r0 cells ( gp33, 11%; +gp33, 46%). GzmB+Tc also induced specific PS exposure on plasma membrane and caspase 3 activation in the presence but not in the absence of viral peptide, which was similar for both cell lines (Figures 3b and c). In addition they induced similar Bak conformational change and cyt c release in EL4 and EL4-r0 cells (Supplementary Figure 2). This indicated that the proteins involved in gzmBinduced cell death are not only expressed (Figure 1e), but functional in these cells. Moreover, the kinetics of ROS production and PS exposure induced by gzmB+Tc was similar in both cell lines (Figure 3d). Together, these experiments show that gzmB induces multiple proapototic processes, including plasma membrane disintegration, caspase 3 activation and ROS production in a cell line defective in mitochondrial ROS activity.
gzmB induces extramitochondrial ROS JI Aguilo´ et al 547
Figure 1 Characterization of EL4-r0 cells. (a) DNA was isolated from EL4, EL4-r0 and control r0 cells and quantitative PCR was performed as described in Methods. Data are shown as mean±s.e.m. of two independent experiments. (b) Oxygen consumption in EL4 and EL4-r0 cells was analyzed as described in Methods. Data are represented as milimol of oxygen consumed every minute by 1106 cells as mean±s.e.m. of two independent experiments. (c) EL4 and EL4-r0 cells were treated with different concentrations of menadione or of adaphostine, as indicated, for 18 h and ROS production or Trypan blue staining was analyzed as described in Methods. Data are represented as the mean±s.e.m. of two independent experiments. (d) EL4 cells were treated with 100 mM menadione for 1 h in the absence or presence of the ROS inhibitor MitoQ or the control inhibitor decyltriphenylphosphonium bromide (dTPP), as indicated, and ROS production was analyzed as described in Methods. Data are shown as mean± s.e.m. of two independent experiments. (e), lysates of EL4 and EL4-r0 cells were prepared and XIAP, Bid, cyt c, AIF, Cu/ZnSOD, caspase 3 and caspase 9 expression were analyzed by western blot as described in Methods. Tubulin was used as loading control.
Implication of NADPH oxidase-mediated superoxide anion generation in gzmB-induced cell death evidenced in EL4-q0 cells The data obtained with EL4-r0 cells suggest that gzmB-induced apoptosis is associated with ROS production derived from an extra-mitochondrial source(s). One possible candidate is the NADPH oxidase complex. To verify this assumption, we analyzed the expression of transcripts of the known Nox isoforms (Nox1–4) in EL4 and EL4-r0 cells using mouse neutrophils and embryonic fibroblasts as specificity controls (Figure 4a). As expected, mouse neutrophils expressed high amounts of the phagocytic Nox2 mRNA, but also Nox1 and Nox3 transcripts, although at much lower levels. Mouse embryonic fibroblasts expressed only Nox4 transcripts. EL4 and EL4-r0 cells expressed Nox1 and Nox3 transcripts at similar levels, indicating that ethidium bromide had not altered expression of these nuclear DNA-encoded genes. Moreover, p22phox expression was observed in every cell line tested. In addition, both cell lines expressed p22phox protein as tested by immunoblot (Figure 4a).
To assess the role of NADPH oxidase in ROS produced by EL4-r0 cells, two widely used inhibitors, diphenylene iodonium and apocynin, were evaluated.43 Diphenylene iodonium not only inhibited ROS produced by NADPH oxidase, but also mitochondrial ROS (data not shown). In contrast, apocynin inhibited NADPH-associated ROS production in phorbol 12-myristate 13-acetate-stimulated neutrophils (Figure 4b) but menadione-induced and mitochondria-associated ROS production in EL4 cells was unaffected (Figure 4c). Using gzmB+Tc, we first analyzed the effect of apocynin on the generation of ROS by EL4 and EL4-r0 cells. Apocynin did not alter the percentage of 2–HE+ EL4 cells (Figure 5a). In contrast, the inhibitor mediated a concentration-dependent reduction in ROS in EL4-r0 cells, with significant suppression at the highest concentration (2.4 mM). In the presence of 2.4 mM apocynin, the number of PS+ cells was also significantly reduced in EL4-r0, but not in EL4 cells exposed to gzmB+Tc (Figure 5b). Higher concentrations of apocynin (3–5 mM) resulted toxic against primary CTL as well as against the target cells (data not shown). Immunology and Cell Biology
gzmB induces extramitochondrial ROS JI Aguilo´ et al 548
NADPH oxidase activation by perforin/gzmB is dependent on caspase activation Neutophil Nox2 activation is mediated by the cytoplasmic adapter Rac1.26 Nox1 activation by tumor necrosis factor in murine L929
Immunology and Cell Biology
fibroblasts have been shown to be dependent on RIP kinase, but also on Rac1.30 However, Nox3 and Nox4 seem to be constitutively active, with no implication of the Rac1 adapter in their activity.44 We have tested the specific Rac1 inhibitor NSC23766,45 and observed that it did
gzmB induces extramitochondrial ROS JI Aguilo´ et al 549
Figure 3 Ex vivo gzmB+ Tc induce apoptosis and ROS generation in EL4-r0 cells. Gp33-pulsed EL4 and EL4-r0 cells were incubated with ex vivo virus-immune gzmB+Tc (MACS selected, 95%CD8+ cells) for different times (a–c, 2 h; d, 30 min, 1 h and 2 h) at 10:1 effector/target ratio. Subsequently, DCm loss (DiOC6(3)) and ROS generation (2-hydroxyethidine (2-HE), a, d) was analyzed by fluorescence-activated cell sorting (FACS). The settings are different in EL4-r0 cells to reflect their lower basal DCm. PS exposure on plasma membrane (annexin-V-FITC, b, d) and caspase 3 activation (c) were also analyzed by three-color flow cytometry in the cell population negative for CD8 expression as described in Methods. A representative experiment is shown in the left panels. Numbers correspond to the percentage of cells in each quadrant (a) or as marked by the horizontal bar (b). Specific quantitative data in the right panels are represented as the mean±s.e.m. of five (a) and four (b) independent experiments, respectively. Data in (d) are representative of two independent experiments.
Figure 2 Mouse recombinant gzmB induces apoptosis and ROS generation in EL4-r0 cells. EL4 and EL4-r0 cells were incubated with 5 mg ml–1 of gzmB in the presence or absence of a sub-lytic SLO dose (0.5 mg/ml–1) for 2 h. Subsequently, mitochondrial membrane potential (DiOC6(3)) and ROS production (2-hydroxyethidine (2-HE); a) were analyzed by fluorescence-activated cell sorting (FACS). The settings are different in EL4-r0 cells to reflect their lower basal DCm. Phosphatydylserine translocation (Annexin V-FITC; b) and caspase 3 activation (d) were also analyzed by FACS, and nuclear morphology (c) was analyzed by fluorescence microscopy after Hoechst staining as described in Methods by using a Nikon fluorescence microscope with software ACT-1. Original magnification 200. (a, b) A representative experiment is shown in the plots (a) and histograms (b). Numbers correspond to the percentage of cells in each quadrant (a) or as marked by the horizontal bar (b). Quantitative data in the graphs are represented as the mean±s.e.m. of three independent experiments. (c) Representative images are shown. (d) A representative experiment of two different independent experiments is shown. Numbers correspond to the percentage of cells as marked by the horizontal bar. (e) EL4 and EL4-r0 cells were incubated with different concentration of gzmB in the presence or absence of a sub-lytic SLO dose (0.5 mg ml–1) for 2 h. Subsequently, ROS production (2-hydroxyethidine (2-HE)) and phosphatydylserine translocation (Annexin V-FITC) was analyzed by FACS. Data are represented as the mean mean±s.e.m. of two independent experiments. Immunology and Cell Biology
gzmB induces extramitochondrial ROS JI Aguilo´ et al 550
Figure 4 EL4 and EL4-r0 cells express NADPH oxidase. (a) RNA was isolated from bone marrow-derived B6 mouse neutrophils, MEF, EL4, EL4-r0, the mouse CD8+ T cell line 1.3E6SN or EL4.F15 cells and Nox1-4 and p22 expression was analyzed by reverse transcriptase-PCR (left panels) as described in Methods. HPRT was used as housekeeping gene. In addition, p22phox protein expression was analyzed in EL4 and EL4-r0 cell lysates by immunoblot (right panels) as described in Methods. Actin was used as loading control. (b) Human primary neutrophils were isolated by Ficoll density centrifugation and oxidative burst induced by phorbol 12-myristate 13-acetate (PMA) in the presence or absence of different concentrations of diphenylene iodonium (DPI) or apocynin was measured by 2-hydroxyethidine (2-HE) staining and fluorescence-activated cell sorting (FACS). Data are presented as the ratio between the mean fluorescence intensities obtained in the presence and in the absence of PMA and are represented as the mean±s.e.m. of three independent experiments. (c) EL4 cells were treated with 10 mM Menadione for 8 h in the presence or absence of 4 mM apocynin and ROS production or Trypan blue staining was analyzed as described in Methods. Data are represented as the mean±s.e.m. of two independent experiments.
not affect ROS generation induced by SLO/gzmB in EL4 or EL4-r1 cells (data not shown). Hence, the gzmB-induced ROS generation through NADPH oxidase activity could be rather related with the inhibition of the mechanisms that counterbalance the constitutive superoxide anion production mediated by Nox3. In light of data indicating that gzmB-induced ROS production in EL4 cells and wild-type mouse embryonic fibroblasts depends on active caspases-3 and -712,20,21 we investigated whether ROS production is similarly regulated in EL4-r0 cells. For this purpose, production of ROS and PS exposure induced by gzmB+Tc in EL4 and EL4-r0 cells was monitored in the presence of the general caspase inhibitor Z-VAD-fmk. Z-VAD-fmk substantially inhibited ROS generation and PS translocation in both cell types (Figure 6) suggesting that gzmB-induced extra-mitochondrial ROS also depends on the presence of the executioner caspases. DISCUSSION Mitochondria are not the sole source of ROS under normal and pathophysiological (such as cell death) conditions. Although ROS produced during gzmB-induced cell death12,21,22 has been attributed to mitochondrial disruption, the cellular source(s) of oxygen radicals has not been stringently evaluated. To this end, we have generated EL4-r0 defective in mitochondrial DNA and thus lacking a functional ETC to analyze their sensitivity to gzmB-induced ROS formation and susceptibility to cell death in the mouse system. We show here that EL4-r0 cells are quite capable of producing ROS during gzmBmediated cell death. Further, the evidence indicates the generated Immunology and Cell Biology
ROS is derived, at least in part, from NADPH oxidase and requires the presence of active caspases. The protocol that depletes mitochondrial DNA with ethidium bromide is well established.33–35 The depletion of mitochondrial DNA is associated with a loss of several ETC proteins.33 The lack of mitochondrial DNA in EL4-r0 cells was judged by PCR and by the inability of the cells to metabolize oxygen. The phenotype was substantiated by showing that EL4-r0 cells, in contrast to EL4, were insensitive to menadione and adaphostin, two cytotoxic drugs that activate mitochondrial ROS.37 These results together with the functionality of the multiple pro-apoptotic molecules involved in gzmBmediated apoptosis indicate that EL4-r0 cells are a good target to (i) study gzmB-induced ROS derived from extra-mitochondrial sources, and (ii) analyze the relevance of mitochondrial ROS during gzmB-induced cell death. Despite their mitochondrial defect, EL4-r0 cells were still sensitive to gzmB-induced pro-apoptotic processes and cell death. PS exposure in the plasma membrane, chromatin condensation, nuclear fragmentation and caspase-3 activation were similar to that observed in parental EL4 cells. However, in contrast to EL4 cells, the inherent lower mitochondrial membrane potential showed by EL4-r0 cells was not further affected by gzmB (plus SLO) or gzmB+Tc, which is in line with their deranged ETC. These results show that an intact ETC, including production of mitochondrial adenosine triphosphate and ROS, is dispensable for gzmB-induced apoptosis. At first sight, this contrasts with a recent study showing that HeLar0 cells were less sensitive to apoptosis elicited by drugs that exclusively engaged the
gzmB induces extramitochondrial ROS JI Aguilo´ et al 551
Figure 5 Ex vivo gzmB+ Tc-induced ROS and apoptosis are partially inhibited by a NADPH oxidase inhibitor in EL4-r0 cells, but not in EL4 cells. Gp33-pulsed EL4 and EL4-r0 cells were incubated with ex vivo virus-immune gzmB+Tc (MACS selected, 95%CD8+ cells) for 2 h at 10:1 effector/target ratio in the presence or absence of different concentrations of apocynin as indicated. Subsequently, ROS generation (2-hydroxyethidine (2-HE), a) and PS exposure on plasma membrane (annexin-V-FITC, b) were analyzed by three-color flow cytometry in the cell population negative for CD8 expression as described in Methods. Data are represented as the mean±s.e.m. of four independent experiments. *Po0.05; **Po0.01.
mitochondrial pathway(s) or acted through ER stress.46 However, the previous findings have shown that gzmB is able to induce multiple pro-apoptotic pathways independent of mitochondria explains the apparent discrepancy.12,20,41,47,48 In particular, as we showed before12,20,21 and again herein with EL4-r0 cells that caspases are critically involved in gzmB-induced apoptosis, independent of mitochondrial pathways. The latter findings are also supported by previous work showing that r0 cells are only resistant to the induction of apoptosis through oxidative damage, but still sensitive to caspasedependent apoptotic pathways, including those elicited by tumor necrosis factor and Fas.34,35 In light of their severe mitochondrial defect it was surprising that EL4-r0 cells produced ROS in response to gzmB (plus SLO) or gzmB+Tc at levels similar to those observed in EL4 parental cells. The present data thus indicate the induction of extra-mitochondrial ROS sources in EL4-r0 cells by gzmB and suggest that the ROS observed are generated, at least in part, by NADPH oxidases, most probably through activation of Nox1 and/or Nox3. This assumption is supported by the potential of apocynin, a specific inhibitor of NADPH oxidases, to partially inhibit ROS production in EL4-r0 cells. The possibility that apocynin interferes with other molecules involved in ROS metabolism is unlikely because of two reasons: apocynin (a) does not inhibit menadione-induced mitochondrial ROS production and (b) does not or only marginally affect ROS production induced by gzmB+Tc in the parental EL4 cell line. The latter observation also indicates that apocynin does not interfere with the potential of gzmB+Tc to induce cell death, thus excluding an unspecific interference with the cytotoxic potential of Tc cells. In contrast, apocynin was not able to completely block neither ROS production nor apoptosis in EL4-r0 cells. Higher apocynin concentrations could not be tested, because they resulted toxic by themselves. However, the results can also indicate that other ROS sources, different from mitochondria and NADPH oxidases, can be activated by gzmB. The mechanism underlying gzmB-mediated induction of ROS by NADPH oxidases in EL4-r0 cells is far from being resolved. However, the data obtained with the pan-caspase inhibitor, Z-VAD-fmk, clearly indicate that this process(es) is strictly dependent on the presence of functional caspases. As EL4-r0 cells express Nox1 and Nox3, but not the neutrophil associated Nox2, it is likely that caspases are regulators of ROS production mediated by Nox1 and/or Nox3 activity during gzmB-induced in EL4-r0 cells. This is in line with previous work
Figure 6 Ex vivo gzmB+ Tc -induced apoptosis and ROS generation in EL4-r0 cells is prevented by a caspase inhibitor. Gp33-pulsed EL4 and EL4-r0 cells were incubated with ex vivo virus-immune gzmB+Tc (MACS selected, 95%CD8+ cells) for 2 h at 10:1 effector/target ratio in the presence or absence of 100 mM ZVAD-fmk, as indicated. Subsequently, ROS generation (left panel) and PS exposure on plasma membrane (annexin-V-FITC, right panel) were analyzed by three-color flow cytometry in the cell population negative for CD8 expression as described in Methods. Data are represented as the mean±s.e.m. of four independent experiments. Immunology and Cell Biology
gzmB induces extramitochondrial ROS JI Aguilo´ et al 552
showing that ROS production mediated by NADPH oxidases occurring during cytotoxic drug-induced cell death was dependent on caspase 3.32 Furthermore, the finding that ROS production was blocked in EL4 and EL4-r0,cells by the pan-caspase inhibitor, ZVAD-fmk, suggests that active caspases are critical for gzmB-induced ROS generation from any cellular source, including NADPH oxidases, mitochondria and others. Although it is clearly established that gzmB-induced apoptosis is accompanied by the production of ROS in intact cells12,20,21 their putative source(s), including mitochondria, was not formally proven in previous studies. However, the present finding that ex vivo-derived gzmB+Tc are able to induce ROS in the absence of functionally active mitochondria, at least shows for the first time that mitochondrial ROS are not required for gzmB+Tc-mediated proapoptotic processes and cell death. How ROS, whether mitochondrial or cytosolic, contributes to gzmB-induced apoptosis, remains unclear.12,20,21,48,23 By using a panel of ROS scavengers, gzmB-induced cell death may be either totally23 or partially abolished.12,20,21,48 The discrepancies may be due to the different ROS scavengers used or to differences in the experimental protocols especially in those studies using human gzm in combination with mouse substrates and cells. In addition, as already reported,24 some of these scavengers could be affecting other processes, such as perforin assembly, and not only ROS generation. We emphasize here the use of a species-specific model to examine these problems49–51 in the absence of ROS scavengers. Although the mechanism that underlies how gzmB, through caspase activation, increases extra-mitochondrial ROS, requires additional study, the observation likely is important for Tc cell-mediated control of pathogens and tumors when defective mitochondrial metabolism exists. For example, as reported for several tumor cells52–54 the presence of low oxygen tension (poor blood supply) or during viral inhibition of mitochondrial respiration.55,56 METHODS Mouse strains Inbred C57BL/6 (B6), and mouse strains deficient for gzmA (gzmA / ), bred on the B6 background were maintained at the Max-Planck-Institut fu¨r Immunbiologie, Freiburg and analyzed for their genotypes as described.12,57 Male mice of 8 to 10 weeks of age were used in all experiments and were performed in accordance with the local ethical animal care commission.
Generation of ex vivo CD8+ cells Mice were infected with 105 pfu LCMV-WE intraperitoneal according to established protocols.12,57 On day 8 post-infection, CD8+ cells were positively selected from spleen using a-CD8-MicroBeads (Miltenyi Biotec, Bergisch Gladbach, Germany) with an autoMACS (Miltenyi Biotec) and resuspended in MEM/5% fetal calf serum before use in cytotoxic assays. Purity of selected CD8+ cells was assessed by fluorescence-activated cell sorting (FACS) staining and found to be between 95 and 98%.
Cell lines, cell culture and reagents EL4 cells were cultured in MEM supplemented with 10% fetal calf serum and 2-mercaptoethanol (10 5 M) at 37 1C, 7% CO2. EL4-r0 cells were generated as previously described.33,34 Selective elimination of mitochondrial DNA was achieved by long-term exposure to low concentrations (50 ng ml–1) of ethidium bromide. Selection and culture medium of EL4-r0 cells was also supplemented with glucose (4.5 mg ml–1), sodium pyruvate (0.1 mg ml–1) and uridine (50 pg ml–1). In some cases, the general caspase inhibitor Z-VAD-fmk (Bachem, Well am Rheim, Germany) at 100 mM, the NADPH oxidase inhibitor apocynin (Calbiochem, Madrid, Spain) at the indicated concentrations, the Rac1-specific inhibitor NSC23766 (Calbiochem) at 100 mM or the mitochondrial ROS inhibitor mitoQ or the control inhibitor decyltriphenylphosphonium bromide Immunology and Cell Biology
(Antipodean Pharmaceuticals Inc., Auckland, New Zealand) at the indicated concentrations were added to cell cultures 1 h before the specific stimulus.12
Mitochondrial DNA analysis Cells were harvested and the DNA was extracted by the phenol–chloroform method, as described elsewhere.34 Mitochondrial DNA amplification was performed by PCR, using the following set of primers: np 3148-3167 (forward) 5¢-CCTACTTCACAAAGCGCCTT-3¢ and np 3531-3550 (reverse) 5¢-CGAT GGTGAGAGCTAAGGTC-3¢.
Determination of oxygen consumption For determination of the maximum respiration capacity, exponentially growing cells were collected by centrifugation, resuspended at 20 106 cells per ml in 1 ml of media and equilibrated at 37 1C. Each sample was transferred into a 1.5 ml water-jacketed chamber containing a small magnetic bar, and connected to a circulating water bath at 37 1C and a Clark-type oxygen electrode. Recording of oxygen consumption was carried out for 150 s. Then, 2, 4-dinitrophenol (Sigma, Madrid, Spain) at 2530 mM was added to uncouple the mitochondria in intact cells and oxygen consumption was monitored for 150 additional seconds to determine the maximum O2 consumption of each cell line.58
Human neutrophil generation and oxidative burst analysis Human neutrophils were separated from lymphocytes by Ficoll-Paque (GE Healthcare, Barcelona, Spain) centrifugation. Then, red blood cells were separated with 3% dextran with an average molecular weight of 400 000– 500 000 Daltons (Sigma). Finally, neutrophil purity was checked by flow cytometry using an anti-CD16b mAb conjugated with FITC (Genetex, Irvine, CA, USA), being that more than 90% in all cases. To test the inhibitory effect of diphenylene iodonium and apocynin, neutrophils were incubated with the inhibitors at the concentrations indicated in each case during one hour before adding 15 ng ml–1 phorbol 12-myristate 13-acetate. After 30 min of incubation at 37 1C, superoxide anion production was measured with 2-hydroxyethidine by fluorescence-activated cell sorting, and the mean fluorescent intensities were obtained with the CellQuest software.
Reverse transcriptase-PCR Total RNA was extracted from up to 5106 cells, using the QIAshredder spin columns, the RNeasy Mini Kit and the RNase-free DNase Kit (all from Qiagen, Hilden, Germany) according to manufacturer’s instructions, and specific transcripts were amplified. The sense/antisense primers for Hprt1 are described in.59 Sense/antisense primers for p22, Nox1, Nox2, Nox3 and Nox4 were respectively: (forward) 5¢-TTTCACACAGTGGTATTTCG-3¢ (reverse) 5¢-CGT AGTAATTCCTGGTGAGG-3¢; (forward) 5¢-ATATTTTGGAATTGCAGATGAA CA-3¢ (reverse) 5¢-ATATTGAGGAAGAGACGGTAG-3¢; (forward) 5¢-ACTCCT TGGGTCAGCACTGG-3¢ (reverse) 5¢- GTTCCTGTCCAGTTGTCTTCG-3¢; (forward) 5¢-GTGATAACAGGCTTAAAGCAGAAGGC-3¢ (reverse) 5¢-CCACT TTCCCCTACTTGACTT-3¢; and (forward) 5¢-CCTCATGGTTACAGCTTC TACCTACGC-3¢ (reverse) 5¢-TGACTGAGGTACAGCTGGATGTTCAC-3¢.
Immunoblot analysis Cell lysates from 106 cells were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Then, proteins were transferred to a nitrocellulose membrane and blocked with 5% fat-free milk. The following primary antibodies were incubated overnight at 4 1C: anti-XIAP, anti-Cyt c and antiCaspase 3 from BD Pharmingen (Madrid, Spain), anti-Bid from R&D systems (Minneapolis, MN, USA), anti-AIF and anti-p22phox (CYBA) from Sigma, anti-caspase 9 from Cell Signalling (Barcelona, Spain) and Cu/Zn superoxide dismutase from Fitzgerald (North Acton, MN, USA) at a 1/1000 dilution. Then, after a washing step and depending on the primary antibody used, blots were incubated during 1 h at room temperature with secondary anti-mouse or antirabbit IgG antibodies conjugated with alkaline phosphatase (Sigma) at a 1/5000 dilution, and washed again. Finally, they were revealed with a buffer containing Tris/HCl 0,2M pH 9,6, MgCl2 1 mM, 10% nitrotetrazoliumblue chloride and 1.5% 5-Bromo-4-chloro-3-indolyl phosphate p-toluidine.
gzmB induces extramitochondrial ROS JI Aguilo´ et al 553
Analysis of apoptotic processes induced by ex vivo Tc cells or by SLO+gzmB Cell death induced by ex vivo Tc cell was analyzed as described.12 EL4 cells were pre-treated with the LCMV-immunodominant peptide gp33 for 2 h before incubation with ex vivo-derived LCMV-immune Tc cell at 10:1 effector:target cell ratio for different times at 37 1C, 7%CO2. Subsequently, different apoptotic parameters were tested in the target population (CD8 ) by fluorescenceactivated cell sorting with a FACScan (BD) and CellQuest software as follows. PS exposure was analyzed by fluorescence-activated cell sorting as described using the annexin V-FITC kit from BD Pharmigen. The mitochondrial membrane potential was measured with the fluorescent probe 3,3¢-dihexyloxacarbocyanine iodide (DiOC6(3), Molecular Probes, Barcelona, Spain) and ROS generation with 2-hydroxyethidine (Molecular Probes) as described.12 Caspase-3 activation, Bak conformational change and cyt c release was monitored in the same CD8-negative population as described.20 Nuclear morphology was analyzed by Hoechst 33342 (25 ng ml–1; Molecular Probes) staining and fluorescence microscopy.60 As described,42 cells were treated with different concentrations of recombinant mouse gzmB in the presence of sublytic doses of SLO (1 mg /ml–1 for EL4 cells and 0.4 mg ml–1 for EL4-r0 cells), and the same parameters were analyzed.
Statistical analysis The statistical analysis of the difference between means of paired samples was performed using the paired t-test. The results are given as the confidence interval (p), and are considered significant when they are o0.05.
ACKNOWLEDGEMENTS This work was supported by grant SAF2004-03058, SAF2007-65144 and SAF2008-02139 from Ministerio de Educacio´n y Ciencia (MEC)/Fondo Social Europeo (Spain) and Grant PI076/08 by Gobierno de Arago´n. JIA was supported by a FPI fellowship from MEC (Spain). JP was supported by an Alexander von Humboldt fellowship (Germany) and Fundacio´n Agencia Aragonesa para Investigacio´n y Desarrollo (ARAID, Spain). We gratefully acknowledge Drs Jose´ Antonio Enrı´quez, Marı´a Angeles Alava and Marı´a Iturralde for scientific support; Dr Mike Murphy and Antipodean Pharmaceuticals Inc. for providing us with MitoQ and DTPP and for advice, S. Bahkdi for generous gift of SLO, Anton Grubisic, Aynur Ekiciler and Thomas Stehle for expert technical assistance and CJ Froelich for critical reading of the paper and advice.
1 Ka¨gi D, Vignaux F, Ledermann B, Bu¨rki K, Depraetere V, Nagata S et al. Fas and perforin pathways as major mechanisms of T cell-mediated cytotoxicity. Science 1994; 265: 528–530. 2 Russell JH, Ley TJ. Lymphocyte-mediated cytotoxicity. Annu Rev Immunol 2002; 20: 323–370. 3 Rouvier E, Luciani MF, Golstein P. Fas involvement in Ca2+-independent T cellmediated cytotoxicity. J Exp Med 1993; 177: 195–200. 4 Lieberman J. The ABCs of granule-mediated cytotoxicity: new ewapons in the arsenal. Nature Rev Immunol 2003; 3: 361–370. 5 Bolitho P, Voskoboinik I, Trapani JA, Smyth MJ. Apoptosis induced by the lymphocyte effector molecule perforin. Curr Opin Immunol 2007; 19: 339–347. 6 Simon MM, Krammer MD. Granzyme A. Methods Enzymol 1994; 244: 68–79. 7 Tschopp J. Granzyme B. Methods Enzymol 1994; 244: 80–87. 8 Bredemeyer AJ, Lewis RM, Malone JP, Davis AE, Gross J, Townsend RR et al. A proteomic approach for the discovery of protease substrates. Proc Natl Acad Sci USA 2004; 101: 11785–11790. 9 Adrain C, Duriez PJ, Brumatti G, Delivani P, Martin SJ. The cytotoxic lymphocyte protease, granzyme B, targets the cytoskeleton and perturbs microtubule polymerization dynamics. J Biol Chem 2006; 281: 8118–8125. 10 Goping IS, Sawchuk T, Underhill DA, Bleackley RC. Identification of a-tubulin as a granzyme B substrate during CTL-mediated apoptosis. J Cell Sci 2006; 119: 858–865. 11 Lieberman J, Fan Z. Nuclear war: the granzyme A-bomb. Curr Opin Immunol 2003; 15: 553–559. 12 Pardo J, Bosque A, Brehm R, Wallich R, Naval J, Mu¨llbacher A et al. Apoptotic pathways are selectively activated by granzyme A and/or granzyme B in CTL-mediated target cell lysis. J Cell Biol 2004; 167: 457–468. 13 Trapani JA, Sutton VR. Granzyme B: pro-apoptotic, antiviral and antitumor functions. Curr Opin Immunol 2003; 15: 533–543.
14 Bleackley RC. A molecular view of cytotoxic T lymphocyte induced killing. Biochem Cell Biol 2005; 83: 747–751. 15 Froelich CJ, Metkar SS, Raja SM. Granzyme B-mediated apoptosis—the elephant and the blind men? Cell Death Differ 2004; 11: 1–3. 16 Thomas DA, Scorrano L, Putcha GV, Korsmeyer SJ, Ley TJ. Granzyme B can cause mitochondrial depolarization and cell death in the absence of BID, BAX and BAK. Proc Natl Acad Sci USA 2001; 98: 14985–14990. 17 Darmon AJ, Ley TJ, Nicholson DW, Bleackley CR. Cleavage of CPP32 by granzyme B represents a critical role for granzyme B in the induction of target cell DNA fragmentation. J Biol Chem 1996; 271: 21709–21712. 18 Barry M, Heibein JA, Pinkoski MJ, Lee SF, Moyer RW, Green DR et al. Granzyme B short-circuits the need for caspase-8 activity during granule-mediated CTL killing by directly cleaving Bid. Mol Cell Biol 2000; 20: 3781–3794. 19 Li P, Nijhawan D, Budihardjo I, Srinivasula SM, Ahmad M, Alnemri ES et al. Cytochrome c and dATP-dependent formation of Apaf-1/caspase-9 complex initiates apoptotic protease cascade. Cell 1997; 91: 479–489. 20 Pardo J, Wallich R, Martin P, Urban C, Rongvaux A, Flavell RA et al. Granzyme Binduced cell death exerted by ex vivo CTL: discriminating requirements for cell death and some of its signs. Cell Death Differ 2008; 15: 567–579. 21 Heibein JA, Barry M, Motyka B, Bleackley RC. Granzyme B-induced loss of mitochondrial inner membrane potential and cytochrome c release are caspase independent. J Immunol 1999; 163: 4683–4693. 22 MacDonald G, Shi L, VandeVelde C, Lieberman J, Greenberg AH. Mitochondriadependent and -independent regulation of granzyme B-induced apoptosis. J Exp Med 1999; 189: 131–144. 23 Martinvalet D, Zhu P, Lieberman J. Granzyme A induces caspase-independent mitochondrial damage, a required first step for apoptosis. Immunity 2005; 22: 355–370. 24 Fellows E, Gil-Parrado S, Jenne DE, Kurschus FC. Natural killer cell–derived human granzyme H induces an alternative, caspase-independent cell-death program. Blood 2007; 110: 544–552. 25 Jezek P, Hlavata L. Mitochondria in homeostasis of reactive oxygen species in cell, tissues, and organism. Int J Biochem Cell Biol 2005; 37: 2478–2503. 26 Li JM, Shah AM. ROS Generation by nonphagocytic NADPH oxidase: potential relevance in diabetic nephropathy. J Am Soc Nephrol 2003; 14: S221–SS26. 27 Lambeth JD. NOX enzymes and the biology of reactive oxygen. Nature Rev Immunol 2004; 4: 181–189. 28 Bokoch GM, Knaus UG. NADPH oxidases: not just for leukocytes anymore. Trends Biochem Sci 2003; 28: 502–508. 29 Jackson SH, Devadas S, Kwon J, Pinto LA, Williams MS. T cells express a phagocytetype NADPH oxidase that is activated after T cell receptor stimulation. Nature Immunol 2004; 5: 818–827. 30 Kim YS, Morgan MJ, Choksi S, Liu ZG. TNF-induced activation of the Nox1 NADPH oxidase and its role in the induction of necrotic cell death. Mol Cell 2007; 26: 675–687. 31 Carmona-Cuenca I, Herrera B, Ventura JJ, Roncero C, Ferna´ndez M, Fabregat I. EGF blocks NADPH oxidase activation by TGF-b in fetal rat hepatocytes, impairing oxidative stress, and cell death. J Cell Physiol 2006; 207: 322–330. 32 Simizu S, Takada M, Umezawa K, Imoto M. Requirement of caspase-3(-like) proteasemediated hydrogen peroxide production for apoptosis induced by various anticancer drugs. J Biol Chem 1998; 273: 26900–26907. 33 King MP, Attardi G. Human cells lacking mtDNA: repopulation with exogenous mitochondria by complementation. Science 1989; 246: 500–503. 34 Gamen S, Anel A, Montoya J, Marzo I, Pin˜eiro A, Naval J. mtDNA depleted U937 cells are sensitive to TNF- and Fas-induced cytotoxicity. FEBS Lett 1995; 376: 15–18. 35 Anel A, Gamen S, Alava MA, Schmitt-Verhulst AM, Pin˜eiro A, Naval J. Role of oxidative damage and ICE-like proteases in Fas-based cytotoxicity exerted by effector T cells. Int Immunol 1996; 8: 1173–1183. 36 Acı´n-Pe´rez R, Bayona-Bafaluy MP, Ferna´ndez-Silva P, Moreno-Loshuertos R, Pe´rezMartos A, Bruno C et al. Respiratory complex III is required to maintain complex I in mammalian mitochondria. Mol Cell 2004; 13: 805–815. 37 Laux I, Nel A. Evidence that oxidative stress-induced apoptosis by menadione involves Fas-dependent and Fas-independent pathways. Clin Immunol 2001; 101: 335–344. 38 Murphy MP, Smith RAJ. Targeting antioxidants to mitochondria by conjugation to lipophilic cations. Annu Rev Pharmacol Toxicol 2007; 47: 629–656. 39 Dasmahapatra G, Rahmani M, Dent P, Grant S. The tyrphostin adaphostin interacts synergistically with proteasome inhibitors to induce apoptosis in human leukemia cells through a reactive oxygen species (ROS)–dependent mechanism. Blood 2006; 107: 232–240. 40 Cullen SP, Martin SJ. Mechanisms of granule-dependent killing. Cell Death Differ 2008; 15: 251–262. 41 Sedelies KA, Ciccone A, Clarke CJ, Oliaro J, Sutton VR, Scott FL et al. Blocking granule-mediated death by primary human NK cells requires both protection of mitochondria and inhibition of caspase activity. Cell Death Differ 2008; 15: 708–717. 42 Pardo J, Wallich R, Ebnet K, Iden S, Zentgraf H, Martin P et al. Granzyme B is expressed in mouse mast cells in vivo and in vitro and causes delayed cell death independent of perforin. Cell Death Differ 2007; 14: 1768–1779. 43 Brandes RP. Role of NADPH oxidases in the control of vascular gene expression. Antioxidants Redox Signal 2003; 5: 803–811. 44 Ueno N, Takeya R, Miyano K, Kikuchi H, Sumimoto H. The NADPH oxidase Nox3 constitutively produces superoxide in a p22phox-dependent manner. J Biol Chem 2005; 280: 23328–23339.
Immunology and Cell Biology
gzmB induces extramitochondrial ROS JI Aguilo´ et al 554 45 Desire L, Bourdin J, Loiseau N, Peillon H, Picard V, De Oliveira C et al. RAC1 inhibition targets amyloid precursor protein processing by gamma-secretase and decreases Abeta production in vitro and in vivo. J Biol Chem 2005; 280: 37516–37525. 46 Kwong JQ, Henning MS, Starkov AA, Manfredi G. The mitochondrial respiratory chain is a modulator of apoptosis. J Cell Biol 2007; 179: 1163–1177. 47 Metkar SS, Wang B, Ebbs ML, Kim JH, Lee YJ, Raja SM et al. Granzyme B activates procaspase-3 which signals a mitochondrial amplification loop for maximal apoptosis. J Cell Biol 2003; 160: 875–885. 48 Goping IS, Sawchuk T, Rieger A, Shostak I, Bleackley RC. Cytotoxic T lymphocytes overcome Bcl-2 inhibition: target cells contribute to their own demise. Blood 2008; 111: 2142–2151. 49 Casciola-Rosen LA, Garcia-Calvo M, Bull HG, Becker JW, Hines T, Thornberry NA et al. Mouse and human granzyme B have distinct tetrapeptide specificities and abilities to recruit the bid pathway. J Biol Chem 2007; 282: 4545–4552. 50 Cullen SP, Adrain C, Lu¨thi AU, Duriez PJ, Martin SJ. Human and murine granzyme B exhibit divergent substrate preferences. J Cell Biol 2007; 176: 435–444. 51 Kaiserman D, Bird CH, Sun J, Matthews A, Ung K, Whissock JC et al. The major human and mouse granzymes are structurally and functionally divergent. J Cell Biol 2006; 175: 619–630. 52 Isidoro A, Martinez M, Fernandez PL, Ortega AD, Santamaria G, Chamorro M et al. Alteration of the bioenergetic phenotype of mitochondria is a hallmark of breast, gastric, lung and oesophageal cancer. Biochem J 2004; 378: 17–20.
53 Mayr JA, Meierhofer D, Zimmermann F, Feichtinger R, Kogler C, Ratschek M et al. Loss of complex I due to mitochondrial DNA mutations in renal oncocytoma. Clin Cancer Res 2008; 14: 2270–2275. 54 Meierhofer D, Mayr JA, Foetschl U, Berger A, Fink K, Schmeller N et al. Decrease of mitochondrial DNA content and energy metabolism in renal cell carcinoma. Carcinogenesis 2004; 25: 1005–1010. 55 Derakhshan M, Willcocks MM, Salako MA, Kass GE, Carter MJ. Human herpesvirus 1 protein US3 induces an inhibition of mitochondrial electron transport. J Gen Virol 2006; 87: 2155–2159. 56 Saffran HA, Pare JM, Corcoran JA, Weller SK, Smiley JR. Herpes simplex virus eliminates host mitochondrial DNA. EMBO Rep 2007; 8: 188–193. 57 Pardo J, Balkow S, Anel A, Simon MM. The differential contribution of granzyme A and granzyme B in CTL-mediated apoptosis is determined by the quality of target cells. Eur J Immunol 2002; 32: 1980–1985. 58 Hofhaus G, Shakeley RM, Attardi G. Use of polarography to detect respiration defects in cell cultures. Methods Enzymol 1996; 264: 476–483. 59 Martin P, Wallich R, Pardo J, Mu¨llbacher A, Munder M, Modollel M et al. Quiescent and activated mouse granulocytes do not express granzyme A and B or perforin: similarities or differences with human polymorphonuclear leukocytes? Blood 2005; 106: 2871–2878. 60 Go´mez-Benito M, Balsas P, Carvajal-Vergara X, Pandiella A, Anel A, Marzo I et al. Mechanism of apoptosis induced by IFN-a in human myeloma cells: role of Jak1 and Bim and potentiation by rapamycin. Cell Signalling 2007; 19: 844–854.
Supplementary Information accompanies the paper on Immunology and Cell Biology website (http://www.nature.com/icb)
Immunology and Cell Biology