Endoplasmic reticulum stress induces calcium-dependent ... - Nature

Oncogene (2008) 27, 285–299

& 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00 www.nature.com/onc

ORIGINAL ARTICLE

Endoplasmic reticulum stress induces calcium-dependent permeability transition, mitochondrial outer membrane permeabilization and apoptosis A Deniaud1,3, O Sharaf el dein1,3, E Maillier1, D Poncet2, G Kroemer2, C Lemaire1 and C Brenner1 1

Universite´ de Versailles/SQY, CNRS UMR 8159, 45, Versailles, France and 2IGR, CNRS UMR 8125, 39, Villejuif, France

The accumulation of Ca2 þ in the mitochondrial matrix can stimulate oxidative phosphorylation, but can also, at high Ca2 þ concentrations, transmit and amplify an apoptotic signal. Here, we characterized the capacity of physiological stimuli (for example, histamine and inositol1,4,5-triphosphate) and inducers of endoplasmic reticulum (ER) stress (for example, A23187, thapsigargin and tunicamycin) to release Ca2 þ from ER stores, induce mitochondrial Ca2 þ accumulation, and trigger cell death in human cervix and colon carcinoma cell lines. Sustained Ca2 þ accumulation in the mitochondrial matrix induced by ER stress triggered signs of proapoptotic mitochondrial alteration, namely permeability transition, dissipation of the electrochemical potential, matrix swelling, relocalization of Bax to mitochondria and the release of cytochrome c and apoptosis-inducing factor from mitochondria. In contrast, rapid and transient accumulation of Ca2 þ induced by physiological stimuli failed to promote mitochondrial permeability transition and to affect cell viability. The specificity of this apoptosis pathway was validated in cells using a panel of pharmacological agents that chelate Ca2 þ (BAPTA-AM) or inhibit inositol-1,4,5trisphosphate receptor (IP3R; 2-aminoethoxydiphenyl borate), voltage-dependent anion channel (VDAC) (4,40 diisothiocyanatostilbene-2,20 -disulfonate, NADH), the permeability transition pore (cyclosporin A and bongkrekic acid), caspases (z-VAD-fmk) and protein synthesis (cycloheximide). Finally, we designed an original cell-free system in which we confronted purified mitochondria and ER vesicles, and identified IP3R, VDAC and the permeability transition pore as key proteins in the ER-triggered proapoptotic mitochondrial membrane permeabilization process. Oncogene (2008) 27, 285–299; doi:10.1038/sj.onc.1210638; published online 13 August 2007 Keywords: permeability transition; mitochondrion; cell death

Correspondence: Dr C Brenner, Universite´ de Versailles/SQY, CNRS UMR 8159, LGBC, 45, avenue des Etats-Unis, 78035 Versailles, France. E-mail: [email protected] 3 These authors contributed equally to this work. Received 23 November 2006; revised 10 April 2007; accepted 16 May 2007; published online 13 August 2007

Introduction The endoplasmic reticulum (ER) is dedicated to synthesis, folding, maturation and transport of nascent proteins. Functional and structural perturbations affecting the ER entail a specific stress response, which in severe conditions, can activate the apoptotic program and/or arrest the cell cycle (Schroder and Kaufman, 2005). ER stress causes cell death through a variety of distinct signals including the activation of transcription factors (Reddy et al., 2003), ER-resident caspases (Breckenridge et al., 2003), proteins from the Bax/Bcl2 family (Scorrano et al., 2003; White et al., 2005), ER channels such as the inositol-1,4,5-trisphosphate (IP3) receptor (IP3R) (Luo et al., 2005; Oakes et al., 2005) and the ryanodine receptor (Hajnoczky et al., 2000a, b). Frequently, ER stress leads to the release of Ca2 þ from the ER lumen. This induces the electrophoretic accumulation of Ca2 þ influx from the cytosol (Ca2 þ c) or through the so-called ‘microdomains’ (which are contacts between ER and mitochondria), in the mitochondrial matrix. A protracted elevation of matrix calcium (Ca2 þ m), beyond a critical threshold, is one of the strongest inducers of the proapoptotic mitochondrial membrane permeabilization (MMP) (for review, see Rizzuto and Pozzan, 2006). MMP is the critical event for the intrinsic pathway of apoptosis, which marks the frontier between death and life, the ‘point-ofno-return’ (for review, see Green and Kroemer, 2004). MMP involves permeabilization of the outer mitochondrial membrane as well as a profound cristae remodeling, resulting in the release of proapoptotic factors from the intermembrane and intercristae spaces such as cytochrome c, Smac/DIABLO, apoptosis-inducing factor (AIF) and endonuclease G (for review, see Green and Kroemer, 2004). The subsequent activation of caspase-dependent and caspase-independent signaling cascades then spreads the death program to other intracellular compartments. The major pathway resulting in the permeabilization of both membranes is the so-called permeability transition (PT). This process is thought to be mediated by the opening of the permeability transition pore complex (PTPC), a dynamic polyprotein complex, which spans both mitochondrial membranes at the contact sites (Zoratti and Szabo, 1994; Beutner et al., 1996; Marzo et al., 1998a, b; Crompton et al., 2002; Halestrap and Brenner, 2003). Lethal PTPC openings contribute to

Endoplasmic reticulum stress and mitochondrial membrane permeabilization A Deniaud et al

286

acute cell loss in toxic, hypoxic and reperfusion injury (Lemasters et al., 1998), Fas/CD95-induced hepatitis (Haouzi et al., 2001) and septic shock (Weaver et al., 2005). Although the exact composition of the PTPC is not known (and actually may depend on the experimental system and proapoptotic stimuli that are investigated), it is currently thought that the minimal constituents of the PTPC can include the voltagedependent anion channel (VDAC, in the outer membrane (OM)), the adenine nucleotide translocase (ANT, in the inner membrane (IM)) and cyclophilin D (CypD, in the matrix) (Zoratti and Szabo, 1994; Beutner et al., 1996; Marzo et al., 1998a). While knockout experiments performed in mice have confirmed the contribution of CypD to PT, data obtained with mitochondria from cell lacking some, but not all, ANT isoforms have questioned the importance of ANT for PT (Kokoszka et al., 2004; Baines et al., 2005; Dolce et al., 2005; Nakagawa et al., 2005). Irrespective of the exact nature of the protein(s) responsible for PT pore formation, PT has been demonstrated to participate in many scenarios of cell death, in particular those mediated by Ca2 þ and reactive oxygen species (ROS) (for review, see Le Bras et al., 2005). Previously, we showed a direct physical interaction between IP3R and PTPC suggesting a functional role of these protein–protein interactions (Verrier et al., 2004). However, fine mechanisms and key proteins mediating proapoptotic ER-induced MMP are still unclear. Based on these premises and incognita, we decided to further explore the functional link between ER, Ca2 þ and proapoptotic mitochondrial alterations, monitoring [Ca2 þ ]m either in cells driven into apoptosis and in a cell-free system containing purified mitochondria and ER vesicles. Our results delineate a scenario in which ER stress results in the IP3R-mediated release of Ca2 þ , which in turn causes MMP through a process, which involves VDAC and the PTPC.

Results ER stress induces Ca2 þ accumulation in the mitochondrial matrix To monitor Ca2 þ m, we loaded cervix (HeLa) and colon carcinoma (HCT116) cells with the fluorescent membrane permeable fluorochrome Rhod2-AM. This probe possesses a positive charge and is enriched in the mitochondrial matrix, driven by the DCm according to the Nernst equation (Figures 1a and b; Trollinger et al., 1997). The specificity of mitochondrial accumulation of the probe was checked by fluorescence microscopy, as demonstrated by a Rhod2-AM-punctuated fluorescence that colocalizes with the fluorescence of DIOC6(3), a prototypic mitochondrial probe (Figure 1a). Cell treatment with a Ca2 þ ionophore (A23187, A23) strongly increased the Rhod2-AM fluorescence, indicating a mitochondrial calcium accumulation (Figure 1b). Then, we quantified and analysed the kinetics of calcium accumulation in the mitochondrial compartment of HeLa and HCT116 cells in microtiter plates (Materials Oncogene

and methods). After stimulation of cells with A23 or an IP3-mobilizing agent (histamine (Hist) that only acts on HeLa, but not on HCT116 cells (Cianchi et al., 2005)), Ca2 þ m increased rapidly within seconds. In contrast, treatment with an inhibitor of N-glycosylation (tunicamycin (TN)) (McDowell and Schwarz, 1988) or an inhibitor of SERCA pumps (thapsigargin (TG)) (Inesi and Sagara, 1992) resulted in a slow sustained increase in Ca2 þ m, within several minutes. The level of Ca2 þ m remained high in both cell lines for A23, TG and TN treatments (Figures 1c and d). In contrast, Hist induced only a transient spike of Ca2 þ m increase (1–2 min). Brefeldin A (BFA), an inhibitor of the protein transport from the ER to the golgi apparatus (Misumi et al., 1986), was unable to elicit an increase in Ca2 þ m. A23, TG, TN and BFA, but not Hist, induced an enhanced expression of GRP78/BiP and CHOP/GADD153, two ER stress markers, both in HeLa cells (Figure 1e) and in HCT116 cells (not shown). Thus A23, TG, TN and BFA induce ER stress, although only A23, TG and TN trigger an elevation in Ca2 þ m. ER stress-induced proapoptotic mitochondrial alteration requires PT When HeLa cells were incubated with A23, TG or TN, a significant irreversible DCm dissipation was observed from 24 h to 48 h after treatment, as shown by staining with the DCm-sensitive fluorochrome DiOC(6)3 (Figures 2a and b). Moreover, A23, TG or TN, induced degradation of nuclear DNA leading to the accumulation of cells in the subdiploid compartment, as detectable by measurement of the DNA content (Figures 2a and b). For A23 treatment, the apoptotic population was inferior to the fraction of cells manifesting plasma membrane permeabilization as detected by staining with the vital dye propidium iodide (PI) (Figure 2a). The degree of DCm dissipation and apoptosis induced by TG or TN was similar to that triggered by lonidamine (LND) (Figure 2b), a chemotherapeutic agent that induces apoptosis via the induction of PT (Belzacq et al., 2001; Le Bras et al., 2006). Next, we used the calcein/cobalt method to monitor IM permeabilization (Nieminen et al., 1995; Petronilli et al., 1999; Poncet et al., 2003). This method relies on the loading of cells with the fluorescent probe calcein (MW ¼ 620 Da) and its quencher, cobalt (Co2 þ ). When loaded into cells in its acetoxymethyl ester form, calcein is liberated by the action of nonspecific esterases in all subcellular compartments including the mitochondrial matrix. In contrast, Co2 þ is excluded from the mitochondrial matrix due to the IM impermeability to this ion. As a consequence, when the barrier provided by IM is functional, cells are fluorescent due to the presence of non-quenched calcein in the mitochondrial matrix. Upon IM permeabilization, calcein is no longer entrapped in mitochondria and thus its fluorescence is quenched by Co2 þ . As expected, Cobalt alone had no effect on DCm, DNA fragmentation and membrane permeabilization (Figures 2a, b and d). TG, TN and A23 induced IM permeabilization concomitantly with


287

Figure 1 Ca2 þ m entry in cells and ER stress response. (a) HeLa cells were incubated with Hoechst, DIOC6(3) and Rhod2-AM to label nuclear and mitochondrial compartment, respectively. The merge panel shows clearly the labeling of mitochondria with the probe Rhod2-AM. (b) HeLa cells were stimulated by A23187 (A23) or not (control) and labeled with Rhod2-AM to visualize the mitochondrial accumulation of calcium. (c, n ¼ 7) HeLa cells were stimulated with 10 mM A23, 2.5 mg/ml BFA, 1 mM TG, 6.25 mg/ml TN, 100 mM Hist or solvent DMSO or H2O (noted Co). (d, n ¼ 6) HCT116 cells were stimulated with 10 mM A23, 2.5 mg/ml BFA, 1 mM TG, 6.25 mg/ml TN or 100 mM Hist. Ca2 þ m uptake was measured with 10 mM Rhod2-AM as described in Materials and methods. (e) Western blot analysis of the expression level of ER stress response-related proteins GRP78/BiP and CHOP/GADD153 in untreated cells (control) or after 24 h treatment with the indicated drugs. Tubulin is a loading control. A23, A23187; BFA, brefeldin A; DMSO, dimethylsulfoxide; ER, endoplasmic reticulum; Hist, histamine; TG, thapsigargin; TN, tunicamycin.

the loss of DCm. As a control, carbonylcyanide m-chlorophenylhydrazone (CCCP), a protonophore, uncoupled the mitochondria without promoting IM permeabilization (Figures 2a and b). A23, TN and TG induced the relocalization and activation of cytosolic Bax to the OM and the release of cytochrome c and AIF from the intermembrane space to the cytosol and the nucleus, respectively, as detected by immunofluorescence microscopy (Figure 2c, IF). These three events correlated with nuclear condensation and fragmentation, as shown by Hoechst staining (Figure 2c, Hoechst). Hist did not cause IM permeabilization, as measured by the calcein/cobalt method (Figure 2a) nor OM permeabilization (as measured by assessing the mitochondrial release of cytochome c and AIF (not shown)), correlating with the absence of signs of apoptosis, neither in HeLa cells (Figure 2a) nor in HCT116 cells (not shown).

Next, we confirmed our findings in the HCT116 cell line (Figure 2d). We found that A23 induced a PTPCmediated MMP and an heterogeneous mode of cell death including signs of apoptosis (subdiploidy) and necrosis (incorporation of PI), whereas TN and TG induced IM permeabilization and subdiploidy in the absence of plasma membrane permeabilization (Figure 2d). Altogether, these data indicate that A23, TN and TG induce the permeabilization of both mitochondrial membranes when they induce cell death. ER stress-induced MMP is regulated by members of Bcl-2 family To determine the specific contribution of Bcl-2 family members to the MMP process, various cell lines, that have been genetically modified for the stable Oncogene


288

Figure 2 Measurement of MMP, necrosis and nuclear apoptosis in response to ER stress. Induction of IM permeabilization (Calcein-), DCm loss (Dioc-), nuclear apoptosis/hypoploidy (Sub-G1) and necrosis (PI þ ) were analysed by flow cytometry in response to 250 mM LND, 10 mM A23, 1 mM TG, 6.25 mg/ml TN, 20 mM CCCP and 100 mM Hist after the indicated times. (a, n ¼ 7) HeLa cells treated for 24 h or (b, n ¼ 7) 48h; (c) IF localization of Bax, cytochrome c and AIF in response to ER stress in HeLa cells (for details see Materials and methods). Apoptotic cells were identified by Hoechst staining as bright condensed or fragmented nuclei. (d, n ¼ 4) HCT116 treated for 24 h. Inserts show the absence of effect of cobalt alone on DCm loss, nuclear apoptosis and necrosis in the different cell lines. The data represent mean7s.d. Paired Student’s t-test was performed to analyse the statistical significance of the results (*Po0.1; **Po0.05). DCm, mitochondrial transmembrane potential; A23, A23187; AIF, apoptosis-inducing factor; CCCP, carbonylcyanide m-chlorophenylhydrazone; ER, endoplasmic reticulum; His, histamine; IF, immunofluorescence; IM, inner membrane; LND, lonidamine; MMP, mitochondrial membrane permeabilization; PI, propidium iodide; TG, thapsigargin; TN, tunicamycin.

overexpression of Bcl-2 (HeLa Bcl-2) or the genetic invalidation of Bax and/or Bak (HCT116-Bax-KO, HCT116-Bak-KO and HCT116-Bax/Bak-KO) were treated with the same death inducers: LND and CCCP as positive and negative controls, and A23187, TG and TN as ER stress inducers and analysed for the transmembrane inner potential and IM permeabilization by flow cytometry at 24 and 48 h (Figure 3). Thus, Bcl-2 can prevent the MMP and the mitochondrial depolarization triggered by all inducers at the exception of A23. Bak invalidation did not influence the death signal pathway, whereas Bax invalidation decreased the MMP and the mitochondrial permeability transition induced by all drugs. Therefore, these results confirm the Oncogene

capacity of Bax to cooperate with proteins of the PT pore to mediate the proapoptotic MMP. Pharmacological modulation of mitochondrial Ca2 þ accumulation, PT and cell death To identify the mechanisms involved in the Ca2 þ m uptake, we pretreated HeLa cells before inducing ER stress, with a series of pharmacological inhibitors, namely 2-aminoethoxydiphenyl borate (2-APB, an IP3R inhibitor (Peppiatt et al.,. 2003)), 4,40 -diisothiocyanatostilbene-2,20 -disulfonate (DIDS, a VDAC inhibitor (Madesh and Hajnoczky, 2001)), CCCP (an uncoupler), cyclosporin A (CsA, an inhibitor of PT


289

Figure 3 Involvement of the members of Bcl-2 family in the MMP induced by ER stress. Induction of IM permeabilization (Calcein-) and DCm loss (Dioc-) were analysed by flow cytometry in response to 250 mM LND, 10 mM A23, 1 mM TG, 6.25 mg/ml TN, 20 mM CCCP after the indicated times of treatment. HeLa cells overexpressing Bcl-2 (HeLa Bcl-2) were treated for 24 or 48 h (a and b, n ¼ 4); HCT116 cells KO for Bax, Bak or both were treated for 24 h (c, n ¼ 5). The data represent mean7s.d. DCm, mitochondrial transmembrane potential; A23, A23187; CCCP, carbonylcyanide m-chlorophenylhydrazone; ER, endoplasmic reticulum; IM, inner membrane; LND, lonidamine; MMP, mitochondrial membrane permeabilization; TG, thapsigargin; TN, tunicamycin.

acting on CypD (Halestrap and Brenner, 2003)) and bongkrekic acid (BA, an inhibitor of PT acting on ANT (Lauquin and Vignais, 1976)). The IP3R inhibitor (2-APB) efficiently decreased the Ca2 þ m entry induced by TN (4579%) but less by A23 (1579%) and TG (13714%) (Figure 4a). The VDAC blocker DIDS decreased the Ca2 þ m uptake induced by TG (49713%) and TN (57710%) but less by A23 (10710%) (Figure 4a). The uncoupler CCCP, which depolarizes mitochondria and hence removes the driving force (DCm) of Ca2 þ uptake, abolished the TG- and TN-induced Ca2 þ m increase (Figure 3a). CsA inhibited slightly the TG- and TN-induced Ca2 þ m increase (Figure 4a), confirming that PTPC can be involved in the Ca2 þ fluxes via the so-called Ca2 þ -induced Ca2 þ release (Ichas et al., 1997). However, BA had no effects on the Ca2 þ m rise induced by A23, TG or TN. The [Ca2]m spike induced by

Hist was partially inhibited by 2-APB (20711%) or DIDS (30712%) and was completely blocked by CCCP (90710%). Similarly, we found that 2-APB, DIDS, CCCP and CsA (but not BA) decreased the Ca2 þ m rise induced by TG and TN in HCT116 cells (Figure 4b). These observations suggest that IP3R, VDAC and the PTPC (and in particular the CsA target CypD) mediate the Ca2 þ m rise induced by ER stress. IM permeabilization was also measured in HeLa cells 24 h after the treatment by ER-stress inducers using the calcein/cobalt labeling method (Figures 4c and d). We found that A23-induced IM permeabilization was refractory to most inhibitors and was reduced only by the cytosolic calcium chelator BAPTA-AM and by the pan-caspase inhibitor z-VAD-fmk (Figures 4c and d). In contrast, BAPTA-AM, DIDS, CsA and BA all protected from TG- and TN-induced IM permeabilization Oncogene


290

Figure 4 Inhibition of Ca2 þ m entry and PTPC opening. (a, n ¼ 7) HeLa cells or (b, n ¼ 6) HCT116 cells were incubated with medium (Co.), 10 mM 2-APB, 25 mM DIDS, 10 mM CCCP, 1 mM CsA or 50 mM BA before stimulation with 10 mM A23, 1 mM TG, 6.25 mg/ml TN or 100 mM Hist. Ca2 þ m uptake was measured with 10 mM Rhod2-AM and its inhibition was measured as described in Materials and methods. IM permeabilization and nuclear apoptosis were recorded by flow cytometry after calcein/cobalt or PI staining, respectively. HeLa cells were pretreated for 2 h with 10 mM BAPTA-AM, 10 mM DIDS, 1 mM CsA or 50 mM BA and then incubated for 22 h with the indicated ER stress inducer. IM permeabilization (c, n ¼ 4) and sub-G1 population (e, n ¼ 4) were analysed. Involvement of caspase activation and protein neosynthesis was investigated by pretreating cells 2 h with either 50 mM z-VAD-fmk or 0.1 mg/ml CHX before ER stress induction. IM permeabilization (d, n ¼ 4) and sub-G1 population (f, n ¼ 4) were then analysed. The data represent mean7s.d. Paired Student’s t-test was used to analyse the statistical significance of the results (*Po0.1; **Po0.05). 2-APB, 2-aminoethoxydiphenyl borate; A23, A23187; BA, bongkrekic acid; CCCP, carbonylcyanide m-chlorophenylhydrazone; CHX, cycloheximide; CsA, cyclosporin A; DIDS, 4,40 -diisothiocyanatostilbene-2,20 -disulfonate; ER, endoplasmic reticulum; His, histamine; IM, inner membrane; PI, propidium iodide; PTPC, permeability transition pore complex; TG, thapsigargin; TN, tunicamycin.

(Figure 4c). In addition, TG- and TN-induced IM permeabilization required the activation of z-VAD-fmkinhibitable proteases and cycloheximide (CHX)-inhibitable protein synthesis (Figure 4d). All these inhibitors (BAPTA-AM, DIDS, CsA, BA, z-VAD-fmk and CHX) prevented the nuclear chromatin condensation indicative of apoptotic cell death, again implicating IP3R, VDAC and PTPC (Figures 4e and f). Oncogene

Ca2 þ m transfer in a cell-free system To further characterize the transfer of Ca2 þ from ER to mitochondria, we confronted purified organelles, namely ER vesicles (microsomes (MM)) and mitochondria in vitro, in a cell-free system. MM were purified by differential centrifugation from mouse brain or HeLa cells, and pure ER vesicles were obtained from mouse liver by sucrose gradient centrifugation. Mitochondria


291

were purified by differential and Percoll gradient centrifugations from mouse liver (Figures 5a and b), and by differential centrifugation from HeLa cells (Figure 5c). We first determined the capacity of isolated organelles (MM/ER and mitochondria) to take up or to release Ca2 þ , separately, in a buffer that allows mitochondria to maintain a DCm (not shown). This first experimental set-up confirmed that 2-APB specifically blocked ER Ca2 þ release (Ca2 þ ER), yet had no effect on Ca2 þ m and that DIDS, NADH and CCCP, but not CsA, inhibited Ca2 þ m uptake, with no effect on Ca2 þ ER (not shown). When we mixed the two types of organelles, a spontaneous, initial Ca2 þ m uptake was monitored by measuring the mitochondrial Rhod2 fluorescence. After stabilization of the system for 3–10 min, the exogenous addition of IP3 rapidly increased Ca2 þ m. This was a robust finding that did not depend on the origin (brain, liver or HeLa cells) of the two organelles fractions that were co-incubated (Figures 5a and c). When added onto isolated mitochondria, IP3 was devoid of any direct effect (not shown). In addition, the IP3induced Ca2 þ m increase could be reduced by IP3R inhibition (with 2-APB), VDAC closure (with DIDS and NADH) and dissipation of the DCm (with CCCP), but not by inhibition of PTPC (with CsA). These results validated the general set-up of the cell-free system that assesses the transfer of Ca2 þ from ER to mitochondria. Next, we investigated the effect of TG and TN on the cell-free system composed by brain MM and liver mitochondria. As expected, TN had a more sustained effect (Figure 5e) than IP3 (Figure 5a). TG, the SERCA pump inhibitor, was a comparatively weak inducer of Ca2 þ m increase (Figure 5d). The TN- or TG-induced Ca2 þ m rise was interrupted by inhibition of IP3R (with 2-APB) and by dissipation of the DCm (with CCCP). Although both DIDS and NADH are supposed to act as VDAC inhibitors (Lee et al., 1994; Thinnes et al., 1994), only NADH (but not DIDS) inhibited the TG-induced Ca2 þ m rise, while only DIDS (but not NADH) inhibited the TN-induced Ca2 þ m increase. This points to subtle differences in the involvement of VDAC in TG- versus TN-induced inter-organelle Ca2 þ fluxes or suggests the involvement of distinct OM channels in the Ca2 þ uptake by mitochondria (Figure 5e). CsA inhibited the TGinduced Ca2 þ m rise, in line with the previous observation that high doses of TG can trigger PT in isolated mitochondria (Korge and Weiss, 1999). Altogether, these data recapitulate the idea, in a cell-free system that the ER stress-induced Ca2 þ transfer between ER and mitochondria depends on IP3R, and is likely to involve DIDS- or NADH-inhibitable OM channels. Ca2 þ -dependent electrochemical transmembrane potential dissipation in vitro After isolation from the liver, mitochondria exhibit a stable DCm (Figure 6, Co.), which can be measured by the Rhod123 dequenching assay (Figure 6a). The DCm can be dissipated in a few seconds by CCCP (Mito þ CCCP) and after a latency of several minutes (15–25 min) by Ca2 þ (Mito þ Ca) or by the addition of

MM (Figure 6a). We checked that IP3, TG or TN did not affect the DCm, when added to isolated mitochondria (Figure 6b), although they did reduce the DCm when added to a cell-free system containing both mitochondria and MM. We then determined whether a series of inhibitors of Ca2 þ signaling would delay the point of inflexion of the DCm curve. 2-APB, NADH and CsA delayed the DCm loss induced by IP3, TG or TN (Figures 6c–f). These results confirm the importance of ER to mitochondria Ca2 þ transfer and the implication of the CsA-inhibitable PT pore in the mitochondrial effects of ER stress. Ca2 þ -dependent mitochondrial swelling in vitro To characterize the consequences of the Ca2 þ m uptake in the reconstituted cell-free system, we monitored mitochondrial matrix swelling by measuring the absorbance at 540 nm. Depending on the amount of Ca2 þ accumulated into the mitochondria, the swelling can be mediated by PTPC opening or occur independently of PTPC by cardiolipin-mediated effects (Petrosillo et al., 2004). By definition, only CsA can inhibit the PTPCdependent swelling. As previously, a delay of the inflexion point of the swelling curve was considered as a significative inhibitory effect. Thus, we found that the addition of brain MM to liver mitochondria induced a mitochondrial matrix swelling (Figure 7a). This swelling can be significantly potentiated by IP3 stimulation (50%), TG (13%) and TN (32%). As expected, CCCP inhibited the mitochondrial swelling by all treatments (Figure 7a). 2-APB and NADH inhibited the swelling induced by IP3, TG and TN, in agreement with the involvement of IP3R to release Ca2 þ from ER and that of VDAC to mediate Ca2 þ m entry into mitochondria. CsA inhibited also the swelling induced by IP3 or TN (but less that induced by TG), supporting again an involvement of the PT pore. Ca2 þ -dependent permeabilization of the inner and outer mitochondrial membranes in vitro To measure IM permeabilization, we measured the release of calcein that has been entrapped into isolated mitochondria. The experimental conditions were designed so that IM permeabilization resulting in calcein release would lead to the quenching of the calceindependent fluorescent signal by the cobalt present in the incubation buffer (Risso et al., 2002). After 100 min of incubation of the mitochondria with the microsomal vesicles alone, they were able to induce an IM permeabilization (Figure 7b) and this latest process was stimulated by IP3 (Figure 7b), TN (Figure 7b) and TG (Figure 7b) and prevented by CsA and CCCP (Figures 7c–e). Moreover, we found that MM alone or in the presence of IP3, TG or TN induced OM permeabilization as shown by the release of cytochrome c and AIF, but not VDAC (as a negative control of a mitochondrial protein), into the supernatant (Figure 7f). These results indicate that the calcium released from MM can promote the permeabilization of both mitochondrial membranes in the cell-free system. Oncogene


292

Figure 5 Ca2 þ m entry measurement and pharmacological inhibition in an in vitro cell-free system. One micromolar of Rhod2-AMpreloaded mitochondria were pre-incubated in energizing buffer for 15 min before addition of ER/MM at t0. After stabilization of the fluorescence signal, 2.5 mM IP3, 1 mM TG or 6.25 mg/ml TN were added (arrow). For inhibition measurement, 2-APB (5 mM when not indicated), DIDS (25 mM when not indicated), NADH (1 mM when not indicated), CCCP (10 mM when not indicated) or CsA (1 mM) were added just before the MM, and inhibitions were calculated as the ratio between the amplitude spike with and without inhibitor as described in Materials and methods. (a, n ¼ 7) Liver mitochondria mixed with brain MM measurements. (b, n ¼ 3) Liver mitochondria mixed with liver ER measurements. (c, n ¼ 3) HeLa mitochondria mixed with HeLa MM measurements. (d and e, n ¼ 4) Liver mitochondria mixed with brain MM measurements. The data represent representative curves or mean7s.d. Paired Student’s t-test was performed to analyse the statistical significance of the results (*Po0.1; **Po0.05). 2-APB, 2-aminoethoxydiphenyl borate; CCCP, carbonylcyanide m-chlorophenylhydrazone; CsA, cyclosporin A; DIDS, 4,40 -diisothiocyanatostilbene-2,20 -disulfonate; ER, endoplasmic reticulum; IP3, inositol-1,4,5-triphosphate; MM, microsomes; TG, thapsigargin; TN, tunicamycin.

Oncogene


293

Figure 6 Mitochondrial depolarization in cell-free system. Liver mitochondria were incubated with 1 mM Rhod123 in energizing buffer 10 min before the fluorescence measurement. When fluorescence signal was stable, brain MM were added, first arrow on the graph except for (b) and (d). After the first depolarization spike due to non-stimulated Ca2 þ m entry, 2.5 mM IP3, 1 mM TG, 6.25 mg/ml TN or 1 mM CCCP were added as indicated by the second arrow (first for (b) and (d)). For inhibition measurement, 5 mM 2-APB, 1 mM NADH or 1 mM CsA were added before MM. In all graphs MM corresponds to liver mitochondria plus brain MM and Co. corresponds to mitochondria alone without treatment. (a, n ¼ 3) Mito corresponds to mitochondria alone. (b, n ¼ 3) Absence of direct effect of inducers on the mitochondrial potential of mitochondria alone. (c–h, n ¼ 3) All experiments except Co. correspond to mitochondria plus MM. The data are representative curves. 2-APB, 2-aminoethoxydiphenyl borate; CCCP, carbonylcyanide m-chlorophenylhydrazone; CsA, cyclosporin A; IP3, inositol-1,4,5-triphosphate; MM, microsomes; TG, thapsigargin; TN, tunicamycin.

Discussion In this study, we show that ER stress agents (that is, A23, TG and TN) and physiological stimuli (that is Hist and IP3) generate various fluxes of Ca2 þ to the mitochondrial compartment that can lead to MMP upon certain conditions. Thus, some compounds can (1)

induce a transient Ca2 þ m increase without cell death, for example Hist, (2) give a high, immediate and stable Ca2 þ m increase and an heterogeneous mode of cell death, for example A23, (3) generate a slower, but more sustained accumulation that induces MMP and leads to apoptosis, for example TG and TN, or (4) fail to increase Ca2 þ m, and induce apoptosis (Scorrano et al., Oncogene


294

Figure 7 Matrix swelling, calcein assay, cytochrome c and AIF release from isolated mitochondria. (a, n ¼ 3) Matrix swelling measurement was carried out with liver mitochondria plus brain MM. MM corresponds to mitochondria plus MM without any compounds. When presents, 5 mM 2-APB, 1 mM NADH, 1 mM CCCP or 1 mM CsA were first added to the mitochondria suspension just followed by the MM addition. Then, after 5 min, 2.5 mM IP3, 1 mM TG or 6.25 mg/ml TN were added. (b–e, n ¼ 2) Calcein release from mitochondria was measured by the quenching of fluorescence by Co2 þ . The concentration of each compound added and the order of addition were the same as the order used for swelling measurement. MM corresponds to mitochondria plus MM without any compounds and mitochondria stimulated by IP3, TG or TN were indicated by MM þ IP3 MM þ TG or MM þ TN, respectively, for Figure 6b, and noted IP3, TG or TN for Figure 6c–e. When added, CCCP or CsA is noted. (f, n ¼ 3) Untreated mitochondria (Mito), and mitochondria treated with Ca2 þ (Mito þ Ca), or with MM (Co.), MM plus 2.5 mM IP3 (MM þ IP3), MM plus 1 mM TG (MM þ TG), or MM plus 6.25 mg/ml TN (MM þ TN) for 120 min were centrifugated. The presence of AIF, cytochrome c and VDAC in the mitochondrial SN and P was revealed by western blotting. The data represent representative curves (b–e) or representative western blot (f). For (a) the data represent mean7s.d. 2-APB, 2-aminoethoxydiphenyl borate; AIF, apoptosis-inducing factor; CCCP, carbonylcyanide m-chlorophenylhydrazone; CsA, cyclosporin A; IP3, inositol-1,4,5-triphosphate; MM, micros; P, pellets; SN, supernatants; TG, thapsigargin; TN, tunicamycin; VDAC, voltage-dependent anion channel.

2003), for example BFA (Figure 1). Our results highlight a difference in the mode of Ca2 þ m accumulation promoted by physiological Ca2 þ mobilizers and ER stress agents and correlate with previous works indicating the capacity of some of these compounds to promote cell death via the Ca2 þ -induced mitochondrial pathway (for review see Rizzuto and Pozzan, 2006). In addition, our results suggest various underlying Ca2 þ -dependent molecular mechanisms. Despite its capacity to release Ca2 þ from the ER, Hist preserves cell viability in HeLa cells. This is presumably Oncogene

due to the transient nature of the Ca2 þ m uptake at a sub-lethal dose. In contrast, in two cancer cell lines, HeLa and HCT116 cells, A23, TG and TN triggered a DCm loss and a Bax, cytochrome c and AIF redistribution in mitochondrial, cytosolic and nuclear compartment, respectively (Figure 2). These effects were followed by a nuclear condensation and fragmentation (Figure 2c). Concomitantly with DCm loss, all compounds triggered PT as observed for other proapoptotic stimuli such as LND (Figures 2a, b and d), etoposide, as well as staurosporine (not shown). Moreover, the death


295

elicited by these compounds required caspase activation and protein synthesis (Figure 4). Cell death induced by A23 was identified as apoptosis and necrosis independently of the cell line studied (Figure 2). A23 allows Ca2 þ ions to cross unspecifically cell membranes, and thus, stimulates a strong Ca2 þ signal, whose consequences are an heterogeneous mode of cell death that cannot be prevented by any inhibitor tested. Therefore, we focused on IP3, TG and TN for the mechanistic study and found that, in cells, as well as in vitro, IP3 induced a transient Ca2 þ m accumulation, and TG and TN a sustained Ca2 þ m uptake, correlating with the difference in cell viability in response to these two molecules. Thus, we show that physiological and ER stress agents induce quantitative and qualitative different mechanisms of mitochondrial Ca2 þ m uptake, which crucially determine their subsequent effects on mitochondria function and influence cell response as well. Ca2 þ effects in whole-cells versus in cell-free system The Ca2 þ -stimulated MMP process occurs in almost 24 h after cell treatment, independently of the cell line used (Figure 2). Surprisingly, IP3, TG and TN stimulation in the cell-free system induced the hallmarks of the mitochondrial phase of apoptosis in only 30–120 min in vitro, again independently of the origin of the organelles used. This includes the DCm loss, matrix swelling, PTPC opening and cytochrome c/AIF release (Figures 6 and 7). First, these differential kinetics could be due to an increased sensitivity of isolated mitochondria in comparison with intracellular mitochondria, which may be preserved by the cellular machinery and the dynamic of the mitochondrial network itself. Second, tumoral cells, that is HeLa or HCT116 cells, due to their higher resistance to cell death than normal tissues, may require a stronger apoptotic stimulation and the synergy of multiple amplification loops, such as the synthesis of new proteins (Figures 4d and f) to allow an ER-mitochondrial physical coupling as described recently by Csordas et al. (2006) and the activation of different effectors such as caspases (Figures 4d and f), to depolarize the mitochondria, open the PTPC and then lead to a complete apoptosis phenotype (Figure 4). These assumptions are coherent with the results published by Madesh et al. (2005), who measured simultaneously a ROS-induced Ca2 þ transfer between the ER and the mitochondria, and the loss of DCm. In their model, DCm dissipated only several minutes after the Ca2 þ increase and a full apoptosis phenotype was detected 5 h later in primary endothelial cells. Furthermore, their ROS stimulation implied the IP3/IP3R pathway, which proved the possibility to elicit an IP3dependent mechanism, which was upstimulated, and able to induce apoptosis. In other terms, it is conceivable that the IP3 stimulation in our cell-free system simulated an upstimulated IP3R pathway in cells. Irrespective of the kinetics difference of the in vitro and cellular processes, we show that, in conditions of purified organelles, in the absence of cytosolic Bax and activated caspases, we can recapitulate the Ca2 þ -induced signaling

cascade leading to MMP, from the release of ER Ca2 þ to MMP. IP3R and VDAC: two actors of the calcium signaling cascade Recently, many studies suggested a role for IP3R in the course of apoptosis. Thus, a caspase 3 cleavage of IP3R or an interaction between IP3R and cytochrome c generate an unregulated Ca2 þ -release state, which potentiates the cell death (Boehning et al., 2003; Bhanumathy et al., 2006). In our work, 2-APB modulated the IP3R. This compound is an antagonist of the IP3R or of other calcium channels, such as those implied in capacitative calcium entry (Iwasaki et al., 2001). Due to this dual effect, 2-APB use in whole cell is controversial. In our conditions, high doses of 2-APB (50 or 75 mM; Madesh et al., 2005) stimulated a direct important Ca2 þ m entry in our model. For this reason, we used a lower dose of 2-APB (10 mM), which led to a partial inhibition of Ca2 þ m entry (20–50%, Figures 4a and b), but supported clearly a specific IP3R implication in TG-, TN- and IP3-stimulated calcium fluxes in cells (Figure 4). As we cannot exclude multiple intracellular targets of 2-APB, we extended our approach in a cellfree system of purified organelles and confirmed in vitro that 2-APB inhibits specifically the univocal IP3Rinduced Ca2 þ release and its mitochondrial uptake, thus demonstrating the important role of IP3R for mitochondrial transfer of ER Ca2 þ (Figure 5). Furthermore, 2-APB delayed significantly the IP3-, TN- and TG-induced DCm loss and matrix swelling (Figures 6 and 7). So, our subcellular approach confirmed the functional implication of IP3R in the apoptotic ER/ mitochondria Ca2 þ cascade that leads ultimately to MMP. In our cell-free system, we also used different inhibitors of VDAC to block the Ca2 þ m entry through the OMM, as described earlier by Gincel et al. (2001). The IP3 stimulation experiments underscores an IP3R/ VDAC coupling for the ER to mitochondria Ca2 þ transfer, as shown by an inhibition of 50% (at least) with VDAC blockers, DIDS and NADH (Figure 5). Although the two VDAC inhibitors had a contradictory effect on TG- and TN-induced Ca2 þ transfer, our results suggest the involvement of VDAC in the TG- and TNinduced Ca2 þ movements. In fact, the regulation of OM permeability and OM channeling appears more complex than anticipated, and our data and results from Bathori et al. (2006) suggest that VDAC, being itself modulable by calcium is important, but that an other channel would also contribute to the Ca2 þ m uptake, as proposed by Csordas et al. (2002). Thus, in the future, it will be critical to characterize the OM Ca2 þ fluxes and to isolate and functionally characterize mitochondrial complexes containing OM and IM Ca2 þ transporter/carrier proteins, which might connect the two membranes in the so-called contact sites (Brdiczka, 1991). PTPC role in the signaling cascade In cells as well as in vitro, the measurement of permeability to calcein and pharmacological inhibition Oncogene


296

with CsA argue for an IM permeabilization mediated by PTPC opening in response to a sustained Ca2 þ m uptake. In cells, IM permeabilization coincides with mitochondrial depolarization and Bax translocation, whereas complete IM permeabilization in vitro occurred at the end of the mitochondrial matrix swelling or DCm loss (Figures 6 and 7). As indicated by CsA and BA inhibition of IM permeabilization and hypoploidy (Figures 4c and e), PTPC opening was required for ER stress-induced apoptosis. In agreement with current hypothesis (Szabadkai and Rizzuto, 2004), our results reveal the contribution of PTPC as an important actor of the Ca2 þ -induced MMP by ER stress agents. New insights in the mode of action of TN in ER stress TN is an inhibitor of the N-glycosylation of proteins in the golgi apparatus, but this compound is also able to alter the channel properties of the acetylcholine receptor. In this work, we demonstrate for the first time that TN generates a high and sustained Ca2 þ flow between ER and mitochondria (Figures 1c and d, and Figure 5). This effect was first observed in cells and was similar to those obtained with TG. Furthermore, this Ca2 þ transfer was confirmed in a minimal cell-free system, and appeared responsible for the hallmarks of mitochondrial apoptosis in this system. So, we propose that TN affect the N-glycosylation at low dose and both the N-glycosylation and Ca2 þ homeostasis at higher dose. Thus, TN is a highly potent proapoptotic drug that can kill cancer cells, plausibly by combining several ER stresses. To conclude, we characterized the mode of Ca2 þ accumulation induced by physiological and proapoptotic ER stress agents, identified the molecular actors as IP3R and VDAC channel as mediators of this Ca2 þ transfer. We demonstrated that this death signaling pathway and the MMP induction can be regulated by some members of the Bcl-2 family members, such as Bax and Bcl-2, but not by Bak. Moreover, this study underscores the involvement of PTPC and the permeabilization of both mitochondrial membranes during the release of cytochrome c and AIF into the cytosol. The development of a cell-free system in vitro corroborates these observations leading to the conclusion that the whole process can occur independently of any other component from other subcellular compartment (for example, plasma membrane, cytosol, nucleus or lysosomes). This cell-free system now opens the door to the molecular study of the pathophysiological modulation of the Ca2 þ -induced MMP by other oncogenes and tumor suppressors and to the screening of molecules or genes capable to modulate the Ca2 þ release from ER for treatment of human diseases involving an alteration of the ER structure or function.

Materials and methods Cell lines, cell culture and drugs HeLa, HeLa Bcl-2 (human cervix carcinoma, generously given by V Goldmacher, ImmunoGen, Cambridge, MA, USA), and Oncogene

HCT116, HCT116-Bax-KO, HCT116-Bak-KO and HCT116Bax/Bak-KO (human colorectal cancer, generously given by B Vogelstein, Johns Hopkins University School of Medecine, Baltimore, MD, USA) cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM:F12) supplemented with 100 mg/ml penicillin, 100 U/ml streptomycin, 1% glutamax and 10% fetal bovine serum (all from Invitrogen, Cergy Pontoise, France), under 5% CO2/95% air. The caspase inhibitor z-VAD-fmk was obtained from Bachem GmbH (Well am Rhein, Germany). Bongkrekik acid was from Calbiochem (Fontenay Sous Bois Cedex, France). All others compounds were purchased from Sigma (St Quentin, Falavier, France). Flow cytometry analysis For flow cytometry analysis of mitochondrial transmembrane potential (DCm), HeLa and HCT116 cells were stained with 100 nM DiOC(6)3 (Molecular Probes, Cergy Pontoise, France) for 15 min at 371C. Necrosis was estimated by adding 10 mg/ml of PI just before analysis. PTPC opening was assessed as described previously (Poncet et al., 2003). Briefly, cells were preincubated for 15 min at 371C with 1 mM calcein-AM (C-3100; Invitrogen) and 1 mM CoCl2 in Hanks’ balanced salt solution (without phenol red and without sodium bicarbonate; Invitrogen) supplemented with 1 mM HEPES, pH 7.3. Hanks’ solution was later replaced by complete culture medium during apoptosis induction. For the detection of apoptotic DNA loss (sub-G1 population), cells were harvested, fixed with cold 70% ethanol, washed three times with phosphate-buffered saline (PBS) and stained with 50 mg/ml PI in the presence of 250 mg/ml of RNase A. Immunofluorescence For the determination of Bax, cytochrome c and AIF redistribution, cells were washed with PBS, fixed with 3.7%. paraformaldehyde 10 min at room temperature (RT) and permeabilized 3 min at 201C with cold acetone. After two washes with PBS, cells were saturated with PBS supplemented with bovine serum albumin (BSA, 3%) for 30 min. Cells were incubated for 1 h with either anti-Bax (rabbit polyclonal, N20, Santa Cruz Biotechnology, Santa Cruz, CA, USA), or anti-cytochrome c (mAb 6H2.B4, Pharmingen, BD Biosciences, San Jose, CA, USA) or anti-AIF (rabbit polyclonal, Chemicon, Hampshire, UK) in PBS-BSA 1% at RT. After two washes, the second antibodies (goat anti-mouse IgG conjugated with Alexa fluor 350 or goat anti-rabbit IgG Alexa fluor 488; Invitrogen) was added in PBS-BSA 1% containing 2.5 mg/ ml Hoechst 33348 (Molecular Probes) for 45 min at RT. Cells were analysed with a fluorescence microscope (DMRH type, Leica, Rueil-Malmaison, France). Western blot analysis Attached and floating cells were recovered, rinsed in cold PBS and lysed in lysis buffer (50 mM Hepes-KOH pH 7.4, 250 mM NaCl, 2% NP-40, 5 mM EDTA, 0.5 mM DTT, and a cocktail of protease inhibitors (Roche/R-Biopharm, St-Didier au Mont d’Or, France)) for 30 min at 41C. Proteins (50 mg) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (12% acrylamide/0.2% bisacrylamide) for CHOP/GADD153 (R-20, Santa Cruz Biotechnology) detection or 10% acrylamide/0.1% bisacrylamide for GRP78/BIP (N-20, Santa Cruz Biotechnology) and tubulin (MCA78S, Serotec, Cergy Saint Christophe, France) detection and transferred to PVDF membranes (Millipore, Hampshire, UK). After overnight incubation at 41C with either antiGRP78/BiP (N20, Santa Cruz Biotechnology) or anti-CHOP/


297 GADD153 (R20, Santa Cruz Biotechnology) or anti-tubulin (YOL1/34, Serotec), proteins were detected by using the enhanced chemiluminescence method according to the manufacturer’s instructions (Amersham Pharmacia Biotech, Rockford, IL, USA). [Ca2 þ ]m determination with Rhod2-AM Cells were trypsinized and resuspended in DMEM without phenol red containing 10 mM Rhod2-AM (Invitrogen) for 30 min (Trollinger et al.,. 1997). Then, cells were washed two times with DMEM without phenol red and resuspended in the same medium. The variations of fluorescence of the Rhod2AM dye (lexc ¼ 540 nm lem ¼ 595 nm) were measured in microplate reader (TECAN Genios, Lyon, France). The cells were dispatched in 96-well white plates at 1.105 cells per well and the volume of each well was adjusted to 200 ml with DMEM without phenol red. This procedure allowed us to increase by a factor 10 the number of cells/well in comparison with adherent cells and accordingly to increase by a factor 10 the fluorescence signal. Cell treatment of trypsinized cells or adherent of cells gave qualitatively similar results. When indicated, inhibitors were added first, followed by the addition of the inducer after stabilization of the fluorescence baseline. The fluorescence was then recorded during 30 min. For inhibition determination, the amplitude of the fluorescence peaks induced by an inducer was normalized to 100%, and then the percent of inhibition was evaluated as the ratio between the amplitude of the fluorescence peak induced by one inducer and the amplitude of corresponding peak with the inhibitor þ the inducer 100. In all cases, the background of the solvent or the influence of the inhibitor alone was deduced from the measure of each peak of fluorescence. To control Rhod-AM localization, cells were labeled with 2 mM Hoechst 33348, 100 nM DIOC6(3) and 10 mM Rhod2-AM for 30 min at 371C and analysed by fluorescence microscopy. Isolation of organites Mouse liver mitochondria Mitochondria were isolated from mouse liver (Swiss females, 6-week old, CEC, Bray LU, France) by differential centrifugations and purified onto Percoll gradient according to Belzacq et al. (2003). Mouse liver ER ER was prepared according to Paulik et al. (1988). In brief, the mitochondrial supernatant was centrifuged at 41C on a discontinuous sucrose gradient (1.3 M/1.5 M/2 M sucrose) during 90 min at 85 000 gav. The material at the 1.3 M/ 1.5 M interface was collected and harvested for 90 min at 85 000 gav and 41C. The ER pellet was resuspended in buffer A (sucrose 0.3 M, TES 5 mM, EGTA 0.2 mM pH 7.2).

HeLa microsomes The mitochondrial supernatant was centrifuged 60 min at 100 000 gmax and 41C, and the microsomal pellet was resuspended in buffer A. Functional measurements in cell-free system All the in vitro measurements were made in buffer B (10 mM Tris-MOPS, pH 7.4, 5 mM succinate, 0.3 M sucrose, 1 mM Pi, 10 mM EGTA, 2 mM rotenone) at 231C. For Ca2 þ transfer determination, mitochondria and MM or ER were first tested alone. EGTA was added at a minimal concentration that countered the spontaneous release of calcium from MM but was too low to chelate the Ca2 þ released in response to IP3. It is noteworthy that, above a certain threshold, an excessive EGTA dose can hamper Ca2 þ m accumulation and its effects on mitochondria. Then, mitochondria (0.4 mg/ml) and MM/ ER (0.6 mg/ml) were mixed in the buffer B plus the optimized EGTA concentration. The last addition of MM/ER started the recording of Rhod2-AM fluorescence. When the fluorescence stabilized (5–10 min), the Ca2 þ transfer inducer was added and the measurement lasted until completion of the Ca2 þ transfer. When used, inhibitors were added just before the MM/ER. The fluorescence increase induced by the compound (IP3, TG and TN) was normalized to 100% and, to determine the percent of inhibition of the Ca2 þ uptake in mitochondria, the following ratio was calculated: (fluorescence peak with inhibitor/fluorescence peak without inhibitor) 100. For swelling, depolarization, calcein and cytochrome c/AIF/ VDAC release measurements, the incubation conditions were the same as those for the Ca2 þ transfer. The mitochondrial swelling was measured by the decrease in absorbance at 540 nm (Costantini et al., 2000). The depolarization of the mitochondria was measured by the Rhodamine 123 (1 mM) fluorescence dequenching (lexc ¼ 485 nm, lem ¼ 535 nm; Molecular Probes) (Vieira et al., 2002). For calcein measurements, mitochondria were loaded with calcein as described in Risso et al. (2002). The calcein release was measured by the quenching of the calcein fluorescence (lexc ¼ 485 nm, lem ¼ 535 nm; Fluka, St Quentin, Falavier, France) by 10 mM Co2 þ ions added to the buffer. Cytochrome c AIF and VDAC were detected by immunoblotting the supernatants and pellets of mitochondria and MM/ER, that have been incubated with various inducers for 2 h, with anti-cytochrome c (BD Biosciences, Le Pont de Clay, France), anti-AIF (Chemicon) or anti-VDAC antibodies (Eurogentec, Seraing, Belgium). Statistical analysis Data were analysed using Student’s t-test for all pairwise comparisons of mean responses among the different treatments tested (SigmaStat software). Results are presented as the mean7s.d. for replicates experiments. Abbreviations

Mouse brain microsomes MM were prepared by differential centrifugation. The brain was homogenized with a Thomas potter. The homogenate was centrifuged 10 min at 2000 g and 41C, and the supernatant was centrifuged 10 min at 12 000 g and 41C. This last supernatant was then centrifuged 60 min at 100 000 gmax and 41C. The microsomal pellet was resuspended in buffer A. HeLa mitochondria Freshly collected HeLa cells were resuspended in buffer A and were lysed in a dounce homogenizer (200 strokes). The lysate was first centrifuged 10 min at 1000 g. The supernatant was centrifuged 10 min at 10 000 g and 41C, and the mitochondrial pellet was resuspended in buffer A.

2-APB, 2-aminoethoxydiphenyl borate; AIF, apoptosisinducing factor; ANT, adenine nucleotide translocase; BA, bongkrekic acid; BFA, brefeldin A; CHX, cycloheximide; CsA, cyclosporin A; DIDS, 4,40 -diisothiocyanatostilbene-2,20 disulfonate; DCm, mitochondrial transmembrane potential; ER, endoplasmic reticulum; Hist, histamine; IP3, inositol1,4,5-triphosphate; IP3R, inositol-1,4,5-trisphosphate receptor; CCCP, carbonylcyanide m-chlorophenylhydrazone; MMP, mitochondrial membrane permeabilization; IM, inner membrane; OM, outer membrane; PI, propidium iodide; ROS, reactive oxygen species; PT, permeability transition; PTPC, permeability transition pore complex; TG, thapsigargin; TN, tunicamycin; VDAC, voltage-dependent anion channel. Oncogene


298 Acknowledgements This work is supported by grants funded by ARC, CNRS and by Agence Nationale pour la Valorisation de la Recherche

(ANVAR) to CB, n1A0505001, INCa (n10610–3D1616–123/ PL-2006) as well by a grant from European Union (RIGHT to GK). A Deniaud receives a PhD fellowship of Ligue Nationale Contre le Cancer.

References Baines CP, Kaiser RA, Purcell NH, Blair NS, Osinska H, Hambleton MA et al. (2005). Loss of cyclophilin D reveals a critical role for mitochondrial permeability transition in cell death. Nature 434: 658–662. Bathori G, Csordas G, Garcia-Perez C, Davies E, Hajnoczky G. (2006). Ca2+-dependent control of the permeability properties of the mitochondrial outer membrane and voltage-dependent anion-selective channel (VDAC). J Biol Chem 281: 17347–17358. Belzacq A, Vieira H, Verrier F, Vandecasteele G, Cohen I, Prevost M et al. (2003). Bcl-2 and bax modulate adenine nucleotide translocase activity. Cancer Res 63: 541–546. Belzacq AS, El Hamel C, Vieira HL, Cohen I, Haouzi D, Metivier D et al. (2001). Adenine nucleotide translocator mediates the mitochondrial membrane permeabilization induced by lonidamine, arsenite and CD437. Oncogene 20: 7579–7587. Beutner G, Ruck A, Riede B, Welte W, Brdiczka D. (1996). Complexes between kinases, mitochondrial porin and adenylate translocator in rat brain resemble the permeability transition pore. FEBS Lett 396: 189–195. Bhanumathy C, Nakao S, Joseph S. (2006). Mechanism of proteasomal degradation of inositol trisphosphate receptors in CHO-K1 cells. J Biol Chem 281: 3722–3730. Boehning D, Patterson RL, Sedaghat L, Glebova NO, Kurosaki T, Snyder SH. (2003). Cytochrome c binds to inositol (1,4,5) trisphosphate receptors, amplifying calciumdependent apoptosis. Nat Cell Biol 5: 1051–1061. Brdiczka D. (1991). Contact sites between mitochondrial envelope membranes. Structure and function in energyand protein-transfer. Biochim Biophys Acta 1071: 291–312. Breckenridge D, Germain M, Mathai J, Nguyen M, Shore G. (2003). Regulation of apoptosis by endoplasmic reticulum pathways. Oncogene 22: 8608–8618. Cianchi F, Cortesini C, Schiavone N, Perna F, Magnelli L, Fanti E et al. (2005). The role of cyclooxygenase-2 in mediating the effects of histamine on cell proliferation and vascular endothelial growth factor production in colorectal cancer. Clin Cancer Res 11: 6807–6815. Costantini P, Belzacq AS, Vieira HL, Larochette N, de Pablo MA, Zamzami N et al. (2000). Oxidation of a critical thiol residue of the adenine nucleotide translocator enforces Bcl2-independent permeability transition pore opening and apoptosis. Oncogene 19: 307–314. Crompton M, Barksby E, Johnson N, Capano M. (2002). Mitochondrial intermembrane junctional complexes and their involvement in cell death. Biochimie 84: 143–152. Csordas G, Madesh M, Antonsson B, Hajnoczky G. (2002). tcBid promotes Ca(2+) signal propagation to the mitochondria: control of Ca(2+) permeation through the outer mitochondrial membrane. EMBO J 21: 2198–2206. Csordas G, Renken C, Varnai P, Walter L, Weaver D, Buttle K et al. (2006). Structural and functional features and significance of the physical linkage between ER and mitochondria. J Cell Biol 174: 915–921. Dolce V, Scarcia P, Iacopetta D, Palmieri F. (2005). A fourth ADP/ATP carrier isoform in man: identification, bacterial expression, functional characterization and tissue distribution. FEBS Lett 579: 633–637. Oncogene

Gincel D, Zaid H, Shoshan-Barmatz V. (2001). Calcium binding and translocation by the voltage-dependent anion channel: a possible regulatory role in mitochondrial function. Biochem J 351: 147–155. Green DR, Kroemer G. (2004). The pathophysiology of mitochondrial cell death. Science 305: 626–629. Hajnoczky G, Csordas G, Krishnamurthy R, Szalai G. (2000a). Mitochondrial calcium signaling driven by the IP3 receptor. J Bioenerg Biomembr 32: 15–25. Hajnoczky G, Csordas G, Madesh M, Pacher P. (2000b). Control of apoptosis by IP(3) and ryanodine receptor driven calcium signals. Cell Calcium 28: 349–363. Halestrap AP, Brenner C. (2003). The adenine nucleotide translocase: a central component of the mitochondrial permeability transition pore and key player in cell death. Curr Med Chem 10: 1507–1525. Haouzi D, Lekehal M, Tinel M, Vadrot N, Caussanel L, Letteron P et al. (2001). Prolonged, but not acute, glutathione depletion promotes Fas-mediated mitochondrial permeability transition and apoptosis in mice. Hepatology 33: 1181–1188. Ichas F, Jouaville L, Mazat J. (1997). Mitochondria are excitable organelles capable of generating and conveying electrical and calcium signals. Cell 89: 1145–1153. Inesi G, Sagara Y. (1992). Thapsigargin, a high affinity and global inhibitor of intracellular Ca2+ transport ATPases. Arch Biochem Biophys 298: 313–317. Iwasaki H, Mori Y, Hara Y, Uchida K, Zhou H, Mikoshiba K. (2001). 2-Aminoethoxydiphenyl borate (2-APB) inhibits capacitative calcium entry independently of the function of inositol 1,4,5-trisphosphate receptors. Receptors Channels 7: 429–439. Kokoszka JE, Waymire KG, Levy SE, Sligh JE, Cai J, Jones DP et al. (2004). The ADP/ATP translocator is not essential for the mitochondrial permeability transition pore. Nature 427: 461–465. Korge P, Weiss JN. (1999). Thapsigargin directly induces the mitochondrial permeability transition. Eur J Biochem 265: 273–280. Lauquin GJ, Vignais PV. (1976). Interaction of (3H) bongkrekic acid with the mitochondrial adenine nucleotide translocator. Biochemistry 15: 2316–2322. Le Bras M, Borgne-Sanchez A, Touat Z, Deniaud A, Maillier E, Lecellier G et al. (2006). Chemosensitization by knockdown of adenine nucleotide translocase-2. Cancer Res 66: 9143–9152. Le Bras M, Clement MV, Pervaiz S, Brenner C. (2005). Reactive oxygen species and the mitochondrial signaling pathway of cell death. Histol Histopathol 20: 205–220. Lee A, Zizi M, Colombini M. (1994). Beta-NADH decreases the permeability of the mitochondrial outer membrane to ADP by a factor of 6. J Biol Chem 269: 30974–30980. Lemasters JJ, Nieminen AL, Qian T, Trost LC, Elmore SP, Nishimura Y et al. (1998). The mitochondrial permeability transition in cell death: a common mechanism in necrosis, apoptosis and autophagy. Biochim Biophys Acta 1366: 177–196. Luo X, He Q, Huang Y, Sheikh M. (2005). Transcriptional upregulation of PUMA modulates endoplasmic reticulum


299 calcium pool depletion-induced apoptosis via Bax activation. Cell Death Differ 12: 1310–1318. Madesh M, Hajnoczky G. (2001). VDAC-dependent permeabilization of the outer mitochondrial membrane by superoxide induces rapid and massive cytochrome c release. J Cell Biol 155: 1003–1015. Madesh M, Hawkins BJ, Milovanova T, Bhanumathy CD, Joseph SK, Ramachandrarao SP et al. (2005). Selective role for superoxide in InsP3 receptor-mediated mitochondrial dysfunction and endothelial apoptosis. J Cell Biol 170: 1079–1090. Marzo I, Brenner C, Zamzami N, Jurgensmeier JM, Susin SA, Vieira HL et al. (1998b). Bax and adenine nucleotide translocator cooperate in the mitochondrial control of apoptosis. Science 281: 2027–2031. Marzo I, Brenner C, Zamzami N, Susin SA, Beutner G, Brdiczka D et al. (1998a). The permeability transition pore complex: a target for apoptosis regulation by caspases and bcl-2-related proteins. J Exp Med 187: 1261–1271. McDowell W, Schwarz RT. (1988). Dissecting glycoprotein biosynthesis by the use of specific inhibitors. Biochimie 70: 1535–1549. Misumi Y, Miki K, Takatsuki A, Tamura G, Ikehara Y. (1986). Novel blockade by brefeldin A of intracellular transport of secretory proteins in cultured rat hepatocytes. J Biol Chem 261: 11398–11403. Nakagawa T, Shimizu S, Watanabe T, Yamaguchi O, Otsu K, Yamagata H et al. (2005). Cyclophilin D-dependent mitochondrial permeability transition regulates some necrotic but not apoptotic cell death. Nature 434: 652–658. Nieminen AL, Saylor AK, Tesfai SA, Herman B, Lemasters JJ. (1995). Contribution of the mitochondrial permeability transition to lethal injury after exposure of hepatocytes to t-butylhydroperoxide. Biochem J 307: 99–106. Oakes SA, Scorrano L, Opferman JT, Bassik MC, Nishino M, Pozzan T et al. (2005). Proapoptotic BAX and BAK regulate the type 1 inositol trisphosphate receptor and calcium leak from the endoplasmic reticulum. Proc Natl Acad Sci USA 102: 105–110. Paulik M, Nowack D, Morre D. (1988). Isolation of a vesicular intermediate in the cell-free transfer of membrane from transitional elements of the endoplasmic reticulum to Golgi apparatus cisternae of rat liver. J Biol Chem 263: 17738–17748. Peppiatt C, Collins T, Mackenzie L, Conway S, Holmes A, Bootman M et al. (2003). 2-Aminoethoxydiphenyl borate (2-APB) antagonises inositol 1,4,5-trisphosphate-induced calcium release, inhibits calcium pumps and has a usedependent and slowly reversible action on store-operated calcium entry channels. Cell Calcium 34: 97–108. Petronilli V, Miotto G, Canton M, Brini M, Colonna R, Bernardi P et al. (1999). Transient and long-lasting openings of the mitochondrial permeability transition pore can be monitored directly in intact cells by changes in mitochondrial calcein fluorescence. Biophys J 76: 725–734. Petrosillo G, Ruggiero FM, Pistolese M, Paradies G. (2004). Ca2+-induced reactive oxygen species production promotes cytochrome c release from rat liver mitochondria via

mitochondrial permeability transition (MPT)-dependent and MPT-independent mechanisms: role of cardiolipin. J Biol Chem 279: 53103–53108. Poncet D, Boya P, Metivier D, Zamzami N, Kroemer G. (2003). Cytofluorometric quantitation of apoptosis-driven inner mitochondrial membrane permeabilization. Apoptosis 8: 521–530. Reddy R, Mao C, Baumeister P, Austin R, Kaufman R, Lee A. (2003). Endoplasmic reticulum chaperone protein GRP78 protects cells from apoptosis induced by topoisomerase inhibitors: role of ATP binding site in suppression of caspase-7 activation. J Biol Chem 278: 20915–20924. Risso A, Braidot E, Sordano MC, Vianello A, Macri F, Skerlavaj B et al. (2002). BMAP-28, an antibiotic peptide of innate immunity, induces cell death through opening of the mitochondrial permeability transition pore. Mol Cell Biol 22: 1926–1935. Rizzuto R, Pozzan T. (2006). Microdomains of intracellular Ca2+: molecular determinants and functional consequences. Physiol Rev 86: 369–408. Schroder M, Kaufman RJ. (2005). The mammalian unfolded protein response. Annu Rev Biochem 74: 739–789. Scorrano L, Oakes SA, Opferman JT, Cheng EH, Sorcinelli MD, Pozzan T et al. (2003). BAX and BAK regulation of endoplasmic reticulum Ca2+: a control point for apoptosis. Science 300: 135–139. Szabadkai G, Rizzuto R. (2004). Participation of endoplasmic reticulum and mitochondrial calcium handling in apoptosis: more than just neighborhood? FEBS Lett 567: 111–115. Thinnes FP, Florke H, Winkelbach H, Stadtmuller U, Heiden M, Karabinos A et al. (1994). Channel active mammalian porin, purified from crude membrane fractions of human B lymphocytes or bovine skeletal muscle, reversibly binds the stilbene-disulfonate group of the chloride channel blocker DIDS. Biol Chem Hoppe Seyler 375: 315–322. Trollinger D, Cascio W, Lemasters J. (1997). Selective loading of Rhod 2 into mitochondria shows mitochondrial Ca2+ transients during the contractile cycle in adult rabbit cardiac myocytes. Biochem Biophys Res Commun 236: 738–742. Verrier F, Deniaud A, LeBras M, Metivier D, Kroemer G, Mignotte B et al. (2004). Dynamic evolution of the adenine nucleotide translocase interactome during chemotherapyinduced apoptosis. Oncogene 23: 8049–8064. Vieira HL, Boya P, Cohen I, El Hamel C, Haouzi D, Druillenec S et al. (2002). Cell permeable BH3-peptides overcome the cytoprotective effect of Bcl-2 and Bcl-X(L). Oncogene 21: 1963–1977. Weaver JGR, Tarze A, Moffat TC, Le Bras M, Deniaud A, Brenner C et al. (2005). Inhibition of adenine nucleotide translocator pore function and protection against apoptosis in vivo by an HIV protease inhibitor. J Clin Inves 115: 1828–1838. White C, Li C, Yang J, Petrenko N, Madesh M, Thompson C et al. (2005). The endoplasmic reticulum gateway to apoptosis by Bcl-X(L) modulation of the InsP(3)R. Nat Cell Biol 7: 1021–1028. Zoratti M, Szabo I. (1994). Electrophysiology of the inner mitochondrial membrane. J Bioenerg Biomembr 26: 543–553.

Oncogene