Store-Operated Calcium Entry Contributes to Cisplatin-Induced Cell Death in Non-Small Cell Lung Carcinoma

Cisplatin (CDDP) is one of the principal chemotherapeutic agents used for the first-line treatment of many malignancies, including non-small cell lung carcinoma (NSCLC). Despite its use for over 40 years, its mechanism of action is not yet fully understood. Store-operated calcium entry (SOCE), the main pathway allowing Ca2+ entry in non-excitable cells, is involved in tumorogenesis, cancer progression and chemoresistance. It has become an attractive target in cancer treatment. In this study, we showed that siRNA-mediated depletion of stromal interaction molecule 1 (STIM1) and transient receptor potential channel 1 (TRPC1), two players of the store-operated calcium entry, dramatically reduced CDDP cytotoxicity in NSCLC cells. This was associated with an inhibition of the DNA damage response (DDR) triggered by CDDP. Moreover, STIM1 depletion also reduced CDDP-dependent oxidative stress. In parallel, SOCE activation induced Ca2+ entry into the mitochondria, a major source of reactive oxygen species (ROS) within the cell. This effect was highly decreased in STIM1-depleted cells. We then conclude that mitochondrial Ca2+ peak associated to the SOCE contributes to CDDP-induced ROS production, DDR and subsequent apoptosis. To the best of our knowledge, this is the first time that it is shown that Ca2+ signalling constitutes an initial step in CDDP-induced apoptosis.


Introduction
The anticancer activity of cisplatin (cis-diamminedichloroplatinum(II), CDDP) was shown for the first time by Barnett Rosenberg in 1969 [1]. It was FDA-approved for treatment of testicular and ovarian cancers in 1978. However, despite its long-standing and large use in the treatment of many malignancies, including sarcomas, carcinomas of lung, ovaries, head and neck and bladder, its mechanism of action is not yet completely understood [2]. The "canonical" pathway underlying the anticancer effects of CDDP is linked to its ability to form intrastrand and interstrand DNA-adducts and therefore to trigger DNA damage response (DDR) pathway. This leads to p53 accumulation and mitochondrial membrane permeabilization (MMP), which is the rate-limiting step of the intrinsec pathway of apoptosis [3]. However, intriguing results showed that CDDP induced mitochondrial-dependent apoptosis in enucleated cells, demonstrating that DNA is dispensable for CDDP-induced cell death [4,5]. This may be related to direct prooxidant activity of CDDP and/or to a reduction of antioxidant capacity of the cell. Moreover, oxidative stress triggers DDR and reactive oxygen species (ROS) might then constitute an intermediate signal between CDDP and DNA [6]. Depletion of STIM1 with a pool of four different siRNAs (named siSTIM1) reduced CDDPinduced ΔΨm dissipation and plasma membrane permeabilization (measured by the vital dye propidium iodide, PI) (Figure 2A,B). We previously showed that the cation channel TRPC1 mediated SOCE in A549 cells [27]. In this study, we observed that siRNA-mediated depletion of TRPC1 decreased CDDP-induced apoptosis in a similar level to STIM1 depletion (Figure 2A,B). Depletion of STIM1 with a pool of four different siRNAs (named siSTIM1) reduced CDDPinduced ΔΨm dissipation and plasma membrane permeabilization (measured by the vital dye propidium iodide, PI) (Figure 2A,B). We previously showed that the cation channel TRPC1 mediated SOCE in A549 cells [27]. In this study, we observed that siRNA-mediated depletion of TRPC1 decreased CDDP-induced apoptosis in a similar level to STIM1 depletion (Figure 2A,B).

Figure 2.
Effects of STIM1 and TRPC1 depletion on apoptosis-associated mitochondrial transmembrane potential (ΔΨ m) dissipation and plasma membrane permeabilization. (A) Following transient transfection (for 72 h) with either siRNA downregulating STIM1 (siSTIM1) or TRPC1 (siTRPC1) or with a control siRNA (siCTRL), A549 cells were left untreated or treated for additional 24 h with 25 µM CDDP, and labeled for the cytofluorometric assessment of ΔΨm (with the ΔΨmsensitive probe DiOC6(3)) and plasma membrane integrity (with propidium iodide, i.e., PI). (B) Quantification of data presented in A. White and grey columns represent the percentage of cells exhibiting ΔΨm loss alone (DiOC6(3) low ) or in association with plasma membrane breakdown (PI + ), respectively. Data are means of triplicate experiments ± S.D. and are representative of three independent experiments. Student's t test was employed to assess statistical significance. *** p < 0.001. (C) Immunoblot analysis of STIM1 expression 96 h after transfection of A549 cells with siCTRL, siSTIM1 or siTRPC1. β-actin was used as loading control. This confirms that the effect of STIM1 depletion on CDDP-induced apoptosis is not related to an off-target effect of the siRNA or to an unknown function of STIM1 but is actually due to SOCE inhibition. siRNA-mediated STIM1 depletion was confirmed by immunoblotting ( Figure 2C). We previously showed the efficiency of the same siRNAs targeted against TRPC1 in A549 cells [27].

SOCE Inhibition Reduced Expression of Specific Markers of Apoptosis Induced by CDDP
SOCE inhibition reduced biochemical hallmarks of the apoptosis activated by CDDP. As expected, 19 kDa and 17 kDa fragments from caspase-3, that is the consequence of its activation, were detected after CDDP treatment. STIM1 and TRPC1 depletion decreased CDDP-induced expression of active fragments of caspase-3, and this was correlated with a reduced degradation of PARP-1 ( Figure 3). This confirms that the effect of STIM1 depletion on CDDP-induced apoptosis is not related to an off-target effect of the siRNA or to an unknown function of STIM1 but is actually due to SOCE inhibition. siRNA-mediated STIM1 depletion was confirmed by immunoblotting ( Figure 2C). We previously showed the efficiency of the same siRNAs targeted against TRPC1 in A549 cells [27].

SOCE Inhibition Reduced Expression of Specific Markers of Apoptosis Induced by CDDP
SOCE inhibition reduced biochemical hallmarks of the apoptosis activated by CDDP. As expected, 19kDa and 17kDa fragments from caspase-3, that is the consequence of its activation, were detected after CDDP treatment. STIM1 and TRPC1 depletion decreased CDDP-induced expression of active fragments of caspase-3, and this was correlated with a reduced degradation of PARP-1 ( Figure 3).

SOCE Was Not Altered by CDDP
CDDP by itself was not able to induce Ca 2+ transients in A549 cells ( Figure 4A, insert). Moreover, CDDP did not modify the Ca 2+ concentration within the ER, as reflected by the absence of effect of CDDP on cytosolic Ca 2+ increase elicited by the SERCA pump inhibitor thapsigargin (Tg) in the absence of Ca 2+ in the external medium ( Figure 4A,B). To measure SOCE amplitude, we re-added Ca 2+ in the external medium after Tg-induced ER emptying. As expected, SOCE amplitude was highly decreased in siSTIM1 transfected cells ( Figure 4B). However, CDDP had no effect on SOCE amplitude.

SOCE Was Not Altered by CDDP
CDDP by itself was not able to induce Ca 2+ transients in A549 cells ( Figure 4A, insert). Moreover, CDDP did not modify the Ca 2+ concentration within the ER, as reflected by the absence of effect of CDDP on cytosolic Ca 2+ increase elicited by the SERCA pump inhibitor thapsigargin (Tg) in the absence of Ca 2+ in the external medium ( Figure 4A,B). To measure SOCE amplitude, we re-added Ca 2+ in the external medium after Tg-induced ER emptying. As expected, SOCE amplitude was highly decreased in siSTIM1 transfected cells ( Figure 4B). However, CDDP had no effect on SOCE amplitude. This confirms that the effect of STIM1 depletion on CDDP-induced apoptosis is not related to an off-target effect of the siRNA or to an unknown function of STIM1 but is actually due to SOCE inhibition. siRNA-mediated STIM1 depletion was confirmed by immunoblotting ( Figure 2C). We previously showed the efficiency of the same siRNAs targeted against TRPC1 in A549 cells [27].

SOCE Inhibition Reduced Expression of Specific Markers of Apoptosis Induced by CDDP
SOCE inhibition reduced biochemical hallmarks of the apoptosis activated by CDDP. As expected, 19kDa and 17kDa fragments from caspase-3, that is the consequence of its activation, were detected after CDDP treatment. STIM1 and TRPC1 depletion decreased CDDP-induced expression of active fragments of caspase-3, and this was correlated with a reduced degradation of PARP-1 ( Figure 3).

SOCE Was Not Altered by CDDP
CDDP by itself was not able to induce Ca 2+ transients in A549 cells ( Figure 4A, insert). Moreover, CDDP did not modify the Ca 2+ concentration within the ER, as reflected by the absence of effect of CDDP on cytosolic Ca 2+ increase elicited by the SERCA pump inhibitor thapsigargin (Tg) in the absence of Ca 2+ in the external medium ( Figure 4A,B). To measure SOCE amplitude, we re-added Ca 2+ in the external medium after Tg-induced ER emptying. As expected, SOCE amplitude was highly decreased in siSTIM1 transfected cells ( Figure 4B). However, CDDP had no effect on SOCE amplitude.

STIM1 Depletion Inhibited CDDP-Dependent ERK Activation
As mentioned above, CDDP is known to induce ERK activation. We confirmed that, in A549 cells, 25 µM CDDP increased phosphorylation of ERK1/2. Depletion of STIM1 almost abolished this effect ( Figure 5A). This is in line with our previous data showing that siTRPC1 inhibited ERK1/2 activation triggered by EGF in A549 cells [27]. Since the role of CDDP-dependent activation of ERK1/2 in cell death is a matter of debate, we inhibited the ERK1/2 pathway with PD98059 to identify a potential role of this pathway in apoptosis induced by CDDP. CDDP-induced PARP-1 cleavage was not modified by PD98059 ( Figure 5B). This demonstrated that the effect of STIM1 depletion on CDDP-induced cell death might not be explained by an inhibition of the ERK1/2 pathway. In line with this observation, accumulation of p53 triggered by CDDP was not impaired in the presence of PD98059 ( Figure 5B).
was added in the external medium, allowing SOCE. Traces are representative of at least 5 independent experiments (minimum 10 cells analysed per experiment). The acute effect of 25 µM CDDP on cytosolic Ca 2+ is shown in the insert (same scale as the main graph). (B) Quantification of experiments presented in A. Results are expressed as means ± S.D. (n ≥ 5). Student's t test was employed to assess statistical significance. *** p < 0.001.

STIM1 Depletion Inhibited CDDP-Dependent ERK Activation
As mentioned above, CDDP is known to induce ERK activation. We confirmed that, in A549 cells, 25 µM CDDP increased phosphorylation of ERK1/2. Depletion of STIM1 almost abolished this effect ( Figure 5A). This is in line with our previous data showing that siTRPC1 inhibited ERK1/2 activation triggered by EGF in A549 cells [27]. Since the role of CDDP-dependent activation of ERK1/2 in cell death is a matter of debate, we inhibited the ERK1/2 pathway with PD98059 to identify a potential role of this pathway in apoptosis induced by CDDP. CDDP-induced PARP-1 cleavage was not modified by PD98059 ( Figure 5B). This demonstrated that the effect of STIM1 depletion on CDDPinduced cell death might not be explained by an inhibition of the ERK1/2 pathway. In line with this observation, accumulation of p53 triggered by CDDP was not impaired in the presence of PD98059 ( Figure 5B).

STIM1 Depletion Inhibited DDR Induced by CDDP
It is well known that CDDP induces DDR by forming DNA-adducts or by producing ROS that in turn alter DNA. To detect activation of the DDR pathway, we measured phosphorylation of ATM, which is provoked by DNA lesions. As expected, CDDP induced ATM phosphorylation ( Figure 6A). This effect was reduced after STIM1 depletion. In response to DDR activation, p53 accumulates within the cell and activates intrinsic pathway of apoptosis or cell growth arrest. CDDP-dependent p53 accumulation was dramatically decreased in STIM1-depleted cells ( Figure 6A,B). This means that SOCE is required to trigger DDR after CDDP treatment.

STIM1 Depletion Inhibited DDR Induced by CDDP
It is well known that CDDP induces DDR by forming DNA-adducts or by producing ROS that in turn alter DNA. To detect activation of the DDR pathway, we measured phosphorylation of ATM, which is provoked by DNA lesions. As expected, CDDP induced ATM phosphorylation ( Figure 6A). This effect was reduced after STIM1 depletion. In response to DDR activation, p53 accumulates within the cell and activates intrinsic pathway of apoptosis or cell growth arrest. CDDP-dependent p53 accumulation was dramatically decreased in STIM1-depleted cells ( Figure 6A,B). This means that SOCE is required to trigger DDR after CDDP treatment.

STIM1 Depletion Reduced CDDP-Induced ROS Production
Since CDDP-triggered DDR was inhibited after SOCE inhibition, this suggested that DNA damage was not only induced by a direct effect of CDDP on DNA. As mentioned in the introduction, CDDP is known to elicit the production of ROS that might in turn damage DNA. In line with those observations, we showed that CDDP induced production of ROS that were measured by the increase of fluorescence of the H2DCFDA probe ( Figure 7A,B). We observed that ROS production after CDDP exposure was abolished in STIM1-depleted cells ( Figure 7A,B), indicating that production of ROS after CDDP exposure depends on Ca 2+ entry through store-operated channels.

STIM1 Depletion Reduced CDDP-Induced ROS Production
Since CDDP-triggered DDR was inhibited after SOCE inhibition, this suggested that DNA damage was not only induced by a direct effect of CDDP on DNA. As mentioned in the introduction, CDDP is known to elicit the production of ROS that might in turn damage DNA. In line with those observations, we showed that CDDP induced production of ROS that were measured by the increase of fluorescence of the H2DCFDA probe ( Figure 7A,B). We observed that ROS production after CDDP exposure was abolished in STIM1-depleted cells ( Figure 7A,B), indicating that production of ROS after CDDP exposure depends on Ca 2+ entry through store-operated channels. Data are normalized to control conditions (i.e., siCTRL transfected cells without CDDP) and presented as the mean of three independent experiments ± S.D. ANOVA test was employed to assess statistical significance. * p < 0.05.

STIM1 Depletion Reduced SOCE-Dependent Entry of Ca 2+ into the Mitochondria
Since mitochondria constitute a major source of ROS and that mitochondrial Ca 2+ is involved is mitochondrial ROS production, we measured mitochondrial Ca 2+ after activation of SOCE. To achieve that, we used the cell-permeant fluorescent Ca 2+ indicator Rhod-2 AM. Fluorescent imaging of A549 cells stained with this dye exhibited a typical mitochondrial pattern ( Figure 8A). Tg-induced release of Ca 2+ from the ER induced a peak of Ca 2+ within the mitochondrial matrix ( Figure 8B,C). This effect  . Student's t test was employed to assess statistical significance. * p < 0.05; *** p < 0.001.

STIM1 Depletion Reduced CDDP-Induced ROS Production
Since CDDP-triggered DDR was inhibited after SOCE inhibition, this suggested that DNA damage was not only induced by a direct effect of CDDP on DNA. As mentioned in the introduction, CDDP is known to elicit the production of ROS that might in turn damage DNA. In line with those observations, we showed that CDDP induced production of ROS that were measured by the increase of fluorescence of the H2DCFDA probe ( Figure 7A,B). We observed that ROS production after CDDP exposure was abolished in STIM1-depleted cells ( Figure 7A,B), indicating that production of ROS after CDDP exposure depends on Ca 2+ entry through store-operated channels. Data are normalized to control conditions (i.e., siCTRL transfected cells without CDDP) and presented as the mean of three independent experiments ± S.D. ANOVA test was employed to assess statistical significance. * p < 0.05.

STIM1 Depletion Reduced SOCE-Dependent Entry of Ca 2+ into the Mitochondria
Since mitochondria constitute a major source of ROS and that mitochondrial Ca 2+ is involved is mitochondrial ROS production, we measured mitochondrial Ca 2+ after activation of SOCE. To achieve that, we used the cell-permeant fluorescent Ca 2+ indicator Rhod-2 AM. Fluorescent imaging of A549 cells stained with this dye exhibited a typical mitochondrial pattern ( Figure 8A). Tg-induced release of Ca 2+ from the ER induced a peak of Ca 2+ within the mitochondrial matrix ( Figure 8B Data are normalized to control conditions (i.e., siCTRL transfected cells without CDDP) and presented as the mean of three independent experiments ± S.D. ANOVA test was employed to assess statistical significance. * p < 0.05.

STIM1 Depletion Reduced SOCE-Dependent Entry of Ca 2+ into the Mitochondria
Since mitochondria constitute a major source of ROS and that mitochondrial Ca 2+ is involved is mitochondrial ROS production, we measured mitochondrial Ca 2+ after activation of SOCE. To achieve that, we used the cell-permeant fluorescent Ca 2+ indicator Rhod-2 AM. Fluorescent imaging of A549 cells stained with this dye exhibited a typical mitochondrial pattern ( Figure 8A). Tg-induced release of Ca 2+ from the ER induced a peak of Ca 2+ within the mitochondrial matrix ( Figure 8B,C). This effect was similar in control cells and in siSTIM1-transfected cells. In contrast, STIM1 depletion dramatically reduced mitochondrial Ca 2+ peak after re-addition of Ca 2+ in the external medium ( Figure 8B,C). This is correlated with the decrease of SOCE-evoked cytosolic Ca 2+ peak and with the decrease of ROS production in STIM1-depleted cells. As in the cytosol, CDDP by itself did not affect mitochondrial Ca 2+ (Figure 8B, insert). was similar in control cells and in siSTIM1-transfected cells. In contrast, STIM1 depletion dramatically reduced mitochondrial Ca 2+ peak after re-addition of Ca 2+ in the external medium ( Figure 8B,C). This is correlated with the decrease of SOCE-evoked cytosolic Ca 2+ peak and with the decrease of ROS production in STIM1-depleted cells. As in the cytosol, CDDP by itself did not affect mitochondrial Ca 2+ (Figure 8B, insert). . Student's t test was employed to assess statistical significance. ** p < 0.01.

SOCE Is Also Involved in CDDP-Induced Cell Death in Cervix Carcinoma Cells
We aimed to investigate whether the results that we obtained in NSCLC were also true in another cancer routinely treated by CDDP. We observed that STIM1 depletion also decreased CDDPinduced cell death in cervix carcinoma cells HeLa, as demonstrated by the reduction of the percentage of cells with low ΔΨm and the reduction in PARP cleavage ( Figure 9A,B).

SOCE Is Also Involved in CDDP-Induced Cell Death in Cervix Carcinoma Cells
We aimed to investigate whether the results that we obtained in NSCLC were also true in another cancer routinely treated by CDDP. We observed that STIM1 depletion also decreased CDDP-induced cell death in cervix carcinoma cells HeLa, as demonstrated by the reduction of the percentage of cells with low ∆Ψm and the reduction in PARP cleavage ( Figure 9A,B).
This last result demonstrates that caspase activation in response to CDDP was inhibited in the absence of STIM1. As in A549 cells, ROS production in response to CDDP was also decreased in STIM1-depleted cells ( Figure 9C). This last result demonstrates that caspase activation in response to CDDP was inhibited in the absence of STIM1. As in A549 cells, ROS production in response to CDDP was also decreased in STIM1-depleted cells ( Figure 9C).

Discussion
Several studies suggest that SOCE components might be attractive targets in cancer treatment. In NSCLC, we showed that SOCE ablation reduced proliferation rate by disrupting EGFR-dependent signalling [27]. More recently, we discovered that disruption of the EGFR/ErbB2-dependent signalling by lapatinib and CP-724714, two inhibitors of the receptor tyrosine kinase (RTK), diminished the amplitude of the SOCE in breast cancer cells [28]. In the present study, we showed that siRNA-mediated SOCE inhibition dramatically reduced CDDP cytotoxicity in NSCLC. Interestingly, a similar effect was obtained in the cervix carcinoma cell line HeLa, suggesting that our observations are not restricted to a single type of cancer. Surprisingly, a previous report showed a slight increase in CDDP-induced apoptosis after STIM1 silencing [29]. We must admit that we do not have any convincing arguments to explain this discrepancy. The decreased cytoxicity after STIM1 depletion was correlated with a drop of oxidative stress triggered by CDDP and with a reduction of mitochondrial Ca 2+ peak following SOCE. This observation was in line with our previous results in skeletal muscles [30]. In contrast, CDDP was not able to induce Ca 2+ transient or to alter SOCE. A potential role of the MAPK pathway in CDDP-induced apoptosis has previously been described. We actually observed a major increase in the expression of pERK1/2 in A549 cells after CDDP exposure, effect that was considerably reduced in STIM1-depleted cells. However, pharmacological inhibition of ERK1/2 did not alter the potency of CDDP to induce cell death. We conclude that the MAPK pathway does not play a significant role in CDDP-induced apoptosis in A549 cells.
Production of ROS after CDDP treatment is well documented. This oxidative stress can contribute to the anti-tumor activity of CDDP [31][32][33]. Unfortunately, the well-known toxicity of CDDP, especially on kidney, is also related to ROS production [34]. The production of ROS by CDDP treatment mainly depends on the depletion of antioxidant molecule like glutathione and on injury in the mitochondrial respiratory chain [35].
Ca 2+ signalling can elicit the production of ROS, essentially via direct activation NADPH oxidase isoforms or by increasing the activity of Krebs cycle and electron transport chain enzymes that in turn drive the production of superoxide [36]. Reciprocally, ROS regulate Ca 2+ signalling at multiple levels

Discussion
Several studies suggest that SOCE components might be attractive targets in cancer treatment. In NSCLC, we showed that SOCE ablation reduced proliferation rate by disrupting EGFR-dependent signalling [27]. More recently, we discovered that disruption of the EGFR/ErbB2-dependent signalling by lapatinib and CP-724714, two inhibitors of the receptor tyrosine kinase (RTK), diminished the amplitude of the SOCE in breast cancer cells [28]. In the present study, we showed that siRNA-mediated SOCE inhibition dramatically reduced CDDP cytotoxicity in NSCLC. Interestingly, a similar effect was obtained in the cervix carcinoma cell line HeLa, suggesting that our observations are not restricted to a single type of cancer. Surprisingly, a previous report showed a slight increase in CDDP-induced apoptosis after STIM1 silencing [29]. We must admit that we do not have any convincing arguments to explain this discrepancy. The decreased cytoxicity after STIM1 depletion was correlated with a drop of oxidative stress triggered by CDDP and with a reduction of mitochondrial Ca 2+ peak following SOCE. This observation was in line with our previous results in skeletal muscles [30]. In contrast, CDDP was not able to induce Ca 2+ transient or to alter SOCE. A potential role of the MAPK pathway in CDDP-induced apoptosis has previously been described. We actually observed a major increase in the expression of pERK1/2 in A549 cells after CDDP exposure, effect that was considerably reduced in STIM1-depleted cells. However, pharmacological inhibition of ERK1/2 did not alter the potency of CDDP to induce cell death. We conclude that the MAPK pathway does not play a significant role in CDDP-induced apoptosis in A549 cells.
Production of ROS after CDDP treatment is well documented. This oxidative stress can contribute to the anti-tumor activity of CDDP [31][32][33]. Unfortunately, the well-known toxicity of CDDP, especially on kidney, is also related to ROS production [34]. The production of ROS by CDDP treatment mainly depends on the depletion of antioxidant molecule like glutathione and on injury in the mitochondrial respiratory chain [35].
Ca 2+ signalling can elicit the production of ROS, essentially via direct activation NADPH oxidase isoforms or by increasing the activity of Krebs cycle and electron transport chain enzymes that in turn drive the production of superoxide [36]. Reciprocally, ROS regulate Ca 2+ signalling at multiple levels and increasing evidences show that this crosstalk between Ca 2+ and ROS is of significant importance in tumorogenesis (for a recent and extensive review on this topic, see [37]).
More specifically, ROS are also known to regulate SOCE. Indeed, oxidation of cysteine 49 and 56 of STIM1 inhibits STIM1 oligomerization and SOCE [38]. Similarly, oxidation of ORAI1 locks the channel in a closed conformation [39]. However, much less studies have investigated the control of ROS production by SOCE. Here, we show for the first time that SOCE inhibition is able to reduce CDDP-induced ROS production. The mechanism underlying this effect is not clear. Mitochondria might constitute a possible target of STIM1-dependent Ca 2+ entry. Since Ca 2+ stimulates several enzymes of the Krebs cycle, the increased metabolic rate induced by Ca 2+ overload within the mitochondria might result in increased respiratory chain leakage and ROS levels [35]. Moreover, it has been shown that SOCE inhibition reduced oxidative stress in a cellular model of Parkinson's disease in which oxidative phosphorylation in mitochondria are injured by 1-methyl-4-phenylpyridinium (MPP+) [40].

Cell Culture and Reagents
A549 and HeLa cell lines (American Type Culture Collection, Molsheim, France) were grown in DMEM-F12 and in DMEM, respectively, supplemented with 10% FCS at 37 • C in a humidified atmosphere of 5% CO 2 . Cells were cultured up to passage 30. CDDP, thapsigargin and PD98059 were purchased from Sigma-Aldrich (Overijse, Belgium).

Cytosolic and Mitochondrial Free Ca 2+ Measurements
A549 cells were plated on 22-mm round glass coverslips (2 × 10 5 cells/well, 6 well plates) 24 h after siRNA transfection. Ca 2+ measurement was performed 72 h later. For cytosolic Ca 2+ measurements, cells were incubated with 1 µM of Fura-2/AM (Invitrogen) in Krebs-HEPES buffer (11.5 mM HEPES, 120 mM NaCl, 6 mM KCl, 1.8 mM CaCl 2 , 1.2 mM MgCl 2 , 10 mM D-glucose, pH 7.6) for 1h15 at room temperature. Coverslips were then wash promptly and mounted in a heated (37 • C) microscope chamber. Cells were alternately excited at 340 and 380 nm using a Lambda DG-4 Ultra High Speed Wavelength Switcher (Sutter Instrument, Novato, CA, USA) coupled to an Axiovert 200M inverted microscope (Zeiss, Zaventem, Belgium). Images were acquired with a Zeiss Axiocam camera coupled to a 510 nm emission filter and analysed with Axiovision software (Zeiss). For mitochondrial Ca 2+ measurements, cells were incubated 30 min at room temperature in Krebs-HEPES buffer containing 5 µM Rhod-2/AM and rinsed 30 min in Krebs-HEPES buffer. Cells were excited at 545 nm and images were acquired at 590 nm.

Statistical Analysis
Data are presented as means ± SD. Student's t-test was used to determine statistical significance.

Conclusions
Taken together, our data revealed that SOCE triggers an increase in mitochondrial Ca 2+ concentration, which was dramatically reduced after STIM1 depletion. STIM1 depletion also decreased production of ROS evoked by CDDP treatment. This was accompanied by a diminution of the DDR and subsequent apoptosis. SOCE might therefore constitute a mechanism of resistance against chemotherapy and targeting SOCE mediators could then restore sensitivity to cytotoxic drugs such as CDDP.