Rat C6 glioma cells were a gift from Dr. Nathalie Chevallier (INSERM U1198). They were maintained in DMEM supplemented by 10% fetal calf serum in the presence of 100 U/mL penicillin and 100 μg/mL streptomycin.

Changes in cellular content of GSH were measured after monobromobimane (mBBr) labeling as previously described (De Jesus Ferreira et al., 2005). For this, C6 rat glioma cells were seeded in 96-well culture plates. When reaching 80% of confluence, cells were treated with BSO (10 μM) or SFN (10 μM) for 24 h. The culture medium was then replaced by extracellular medium containing 50 μM mBrB. Extracellular medium comprised: 124 mM NaCl, 3.5 mM KCl, 25 mM NaHCO3, 1.25 mM NaH₂PO₄, 1 mM CaCl2, 2 mM MgSO4, 10 mM D-glucose, and 10 mM HEPES (pH: 7.4). Incubation with mBBr lasted for 30 min then cells were washed with extracellular medium to remove unbound mBBr. Fluorescence was then measured in cellulo with a plate reader (Tecan “Spark” 20M) at 527 nm after excitation at 380 nm. Background fluorescence was obtained from mBBr unlabeled cells. For data processing, background fluorescence was subtracted to all fluorimetric signals which were further normalized to mBBr fluorescence recorded in control cells. An experimental determination was performed eight times per experiment. Data are expressed as percentages of at least three distinct experiments performed on different cell cultures.

C6 rat glioma cells were seeded in 24-well culture plates. When reaching 80% of confluence cells were treated with either BSO (10 μM) or SFN (10 μM) or acivicin (100 μM) and returned to the incubator for 24 h. The oxidative stress was then induced by treating the cells with increasing concentrations of tert-butyl hydroperoxide (tBuOOH) for 24 h. When tested, the intracellular Ca²⁺ chelator 1,2-Bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetrakis(acetoxymethyl ester) (BAPTA-AM 1–10 μM) was added 1 h prior to the tBuOOH application. Cell viability was measured by assaying mitochondrial activity using MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bromide) transformation, as previously described (De Jesus Ferreira et al., 2005). Data are presented as percentages of MTT transformation normalized to basal obtained in cells without any treatment. They are means (±S.E.M.) of at least three individual determinations performed on different cell cultures; each determination being performed in triplicate per experiment.

Intracellular calcium concentration ([Ca²⁺]_i) was measured using the fluorescent indicator fura-2. For this purpose, cells grown on glass coverslips were loaded with fura-2 by a 30 min incubation at 37 °C with 1 μM fura-2-AM and 0.02% Pluronic in the extracellular solution described above. [Ca²⁺]_i was monitored by videomicroscopy. After rinsing, the glass coverslip was transferred to the recording chamber mounted on an inverted microscope (Leica, DMIRB), continuously superfused with extracellular medium. Fura-2 emission was obtained by exciting alternatively at 340 and 380 nm with a rotating filter wheel (Sutter Instruments) and by monitoring emissions (F340 and F380) at 510 nm. Fluorescent signals were collected with a CCD camera (Hamamatsu), digitized, and analyzed with image analysis software (Acquacosmos, Hamamatsu). The ratio of emissions at 510 nm (F340/F380) was recorded in cells every second. Experiments were carried out at room temperature and drug application was performed with a gravity-fed system. Data are expressed as averages (±SEM) of the ratio between the fura-2 fluorescence values of 340/380 nm excitation wavelengths ratios (F340/F380) normalized to the corresponding basal F340/F380 measured prior to any drug application. When required, F340/F380 ratio was calibrated into [Ca²⁺]_i (nM) according the following equation (Grynkiewicz et al., 1985):

With Kd = 224 nM, R = F340/F380, S2f = F380 without Ca²⁺ and S2b = F380 with saturating Ca²⁺.

R_max was obtained by applying the Ca²⁺ ionophore 4Br-A23187 (5 μM). R_min was then recorded by substituting the extracellular medium for Ca²⁺-free extracellular medium with no added Ca²⁺ and supplemented with EGTA (2 mM).

Graphs presenting time-courses of F340/F380 ratio changes have been obtained by averaging data from a population of cells recorded individually during one single representative experiment. When required, the Areas Under Curve (AUC) of the Ca²⁺ entries elicited by tBuOOH application or by store depletion (SOCE) were calculated by integrating F340/F380 ratio over 5 min starting at the time of Ca²⁺ reinstatement in the extracellular medium. In graphs of pooled data, “n” values represent the entire population of cells recorded from at least three independent cultures (N).

Membrane currents were recorded using whole-cell patch-clamp method. Cells grown on glass coverslips were transferred to a recording chamber of an upright microscope, continuously superfused with the extracellular medium. Experiments were performed at room temperature with glass pipettes (4–5 MΩ resistance) filled with the intracellular solution comprising 140 mM KGluconate, 4 mM NaCl, 1 mM MgCl₂, 1 mM EGTA, 5 mM HEPES, 2 mM MgATP, 0.6 mM NaGTP, pH = 7.4 (KOH). Cells were held at 0 mV and membrane currents were recorded by applying voltage ramps from −100 to 100 mV every 10 s. Access resistance was monitored by applying a 10 mV voltage step at the end of the voltage ramp. Currents were collected and amplified with an Axoclamp 200B amplifier (Molecular Devices) and digitized (Digidata 1322, Molecular Devices). They were analyzed with John Dempster’s software WCP. Drugs were applied at the desired concentration via a gravity-fed application system. Data are presented as means ± S.E.M. on graphs plotting pooled data.

Total RNA extraction was performed after cell collection using TRI Reagent^® (Molecular Research Center, Inc.) according to the manufacturer’s instructions. Final concentration of 1 ng/mL total RNA was reverse-transcribed to first-strand cDNA using oligo-dT primer and SuperScript™ II Reverse Transcriptase (Invitrogen™) as described by the manufacturer. Real-time PCR was then carried out using SensiFAST™ SYBR^® No-ROX Kit (Bioline) in LightCycler^® 480 Instrument II (Roche Molecular Systems, Inc.). The primers used for the amplification of target cDNAs are shown in Table 1. They were designed with IDTDNA Technologies software. The access numbers for the target genes were obtained from NCBI data bank. All reactions were performed in a 1.5 μL volume using the following PCR program: initial denaturation at 95°C for 2 min, followed by 45 tri-step cycles of amplification: 10 s at 95°C, 10 s at 65°C, and 10 s at 72°C. Negative controls were obtained by running PCR with H₂O instead of the template. Specificity was verified by melting curve analyses and agarose gel electrophoresis of the amplicon. Levels of specific transcripts were given by the threshold cycle (Ct) values and the data were normalized for the corresponding β-actin RNA contents.

Cells were washed with PBS and lysed using RIPA Lysis Buffer (Boster) containing 1% NP-40, 0.5% sodium deoxycholate and 0.1% sodium dodecyl sulfate (SDS). Samples were then boiled for 5 min in a Laemmli Sample Buffer (BioRad) and loaded on a 12% SDS –polyacrylamide gel (6–8 mg/mL). After electrophoresis, proteins were electro-transferred (100 mV, 1 h) to nitrocellulose membrane using a sandwich blotting system. Blots were first incubated with 5% BSA and 0.1% Tween in Tris-buffered saline (TBS) for 1 h at room temperature, and then with either polyclonal rabbit primary antibody against Orai1 (1:1,000, rabbit Sigma-Aldrich, O8264), STIM2 (1:2,000, rabbit Sigma-Aldrich, S8572) or monoclonal mouse actin (1:100, mouse Sigma-Aldrich, A2066) at 4°C overnight. The appropriate horseradish peroxidase-conjugated secondary antibody IgG was added at a dilution of 1:10,000 for 1 h at RT. The peroxidase was detected with ECL Western Blotting Reagent (Amersham).

Statistical analyses were performed using SigmaStat software (Systat software Inc.). Data were first tested for normality and equality of the variances. In case of one or the other test failure (p < 0.05) then non-parametric methods, mainly ANOVA on rank, were used for comparisons. If both tests passed, then parametric methods were used. Specific tests used for each experiment were described in figure legends. Data obeying normality were represented as mean ± SEM; in other case data were represented as median, deciles and quartiles.

In order to modulate GSH levels, cells were pretreated for 24 h with either buthionine sulfoximine (BSO) or sulforaphane (SFN). The intracellular GSH concentration ([GSH]_i) was assayed in cellulo by labeling living cells with monobromobimane (mBBr). In cells exposed to 10 μM BSO, a γ-glutamyl-cysteine ligase (GCL) blocker, mBBr fluorescence was significantly reduced to 51 ± 4% (n = 11) as compared to untreated cells. Treatment with 20 μM BSO did not further reduce GSH content. On the opposite, mBBr fluorescence was increased to 135 ± 6% (n = 7) of control fluorescence in cells treated with 10 μM SFN, a strong activator of Nrf-2 pathway (Figure 1A).

We next verified whether [GSH]_i modulation effectively altered cell sensitivity toward oxidative stress. For this, the toxicity of tert-butyl hydroperoxide (tBuOOH) was measured on C6 cell pretreated or not with BSO or SFN (Figure 1B). Under control conditions, a 100 μM tBuOOH exposure led to little or no cell damage (93 ± 5% of cell survival, n = 12). By contrast, when applied at 500 μM, a strong cell death was observed (18 ± 1% of cell survival, n = 12). Treatment with 300 μM tBuOOH, resulted in moderate cell death survival (45 ± 5% n = 12 of cell survival). Thus, analyzing changes in cell survival profile obtained after tBuOOH treatments at 100, 300, and 500 μM enabled to measure both sensitization and resistance to oxidative stress.

Neither SFN, nor BSO pretreatment affected cell survival under basal conditions. Cell survival was 101 ± 3% and 102 ± 2% (n = 8) in SFN and BSO pretreated cells, respectively (Figure 1B). Increasing [GSH]_i with SFN pretreatment protected cells against peroxide toxicity: upon 500 μM tBuOOH, cell survival was 49 ± 12% (n = 8) in cells treated with SFN, as compared to 18 ± 3% obtained in control cells (Figure 1B). Other agents triggering Nrf2 pathway as dimethyl fumarate and tertio butyl hydroquinone exhibited similar effects as SFN (data not shown). In contrast, decreasing [GSH]i with BSO sensitized cells to peroxide toxicity. Indeed, in the presence of 100 μM tBuOOH, a concentration which did not induce any significant cell death in control cells, cell survival was reduced to 16 ± 2% after BSO pretreatment (Figure 1B).

To establish the putative link between oxidative stress and calcium homeostasis disruption in C6 glioma cells, intracellular Ca²⁺ variations were buffered using the cell permeant calcium chelator, BAPTA-AM. The treatment with 1–10 μM BAPTA-AM of C6 cells prevented cell death induced by the application of tBuOOH (100–500 μM) (Figure 1B). Cell survival was 94 ± 5% (n = 5) after exposing 10 μM BAPTA-loaded cells to 500 μM tBuOOH. This strengthened the requirement of intracellular calcium increases for the expression of peroxide-mediated toxicity in C6 cells.

We next evaluated how [GSH]_i manipulation impacted oxidative stress-mediated [Ca²⁺]_i changes. With this aim, the effect of tBuOOH was first examined on control cells. This peroxide induced a concentration-dependent increase in [Ca²⁺]_i in C6 cells which was linear for the first 10 min after applying it (Figures 2A,B). Cytoplasmic Ca²⁺ variation may result from influxes from the extracellular medium to cytoplasm or/and from Ca²⁺ release from internal stores including endoplasmic reticulum (ER). In extracellular Ca²⁺-depleted medium (Figure 2E) or in the presence of 100 μM lanthanum ions (La³⁺), a non-selective calcium channel blocker, tBuOOH-induced [Ca²⁺]_i accumulation was completely inhibited (not shown), suggesting the occurrence of Ca²⁺ entries under oxidative conditions.

The contribution of intracellular calcium store release to the [Ca²⁺]_i increase elicited by tBuOOH and the potential induction of store-operated capacitive calcium entries (SOCE) as a source of Ca²⁺ influx was further evaluated. First, the occurrence of SOCE was examined in C6 cells thanks to their sensitivity to the selective CRAC blocker, GSK-7975A. SOCE were triggered by depleting Ca²⁺ from the ER with cyclopiazonic acid (10 μM, CPA), a reversible SERCA (Sarcoplasmic and Endoplasmic Reticulum Ca²⁺ ATPase) pump blocker, in Ca²⁺-free medium, and revealed by reinstating extracellular calcium concentration at 2 mM (Figure 2C). As expected, the addition of CPA induced a transient increase in Ca²⁺ in the absence of extracellular calcium, reflecting the release from ER; then reinstating extracellular calcium at 2 mM induced a large increase in intracellular Ca2+, reflecting the SOCE (Figure 2C). Consistently, this Ca²⁺ increase was blocked by 5 μM GSK-7975A or by 100 μM La³⁺ ions (Figure 2D). The AUC for the SOCE plotted in the Figure 2D were 0.67 ± 0.02 in control and 0.21 ± 0.06 in the presence of GK-7975A (n = 30 each). Then, the potential induction of SOCE by tBuOOH was evaluated. For this, tBuOOH was transiently applied in the absence of extracellular Ca²⁺. Under these conditions, no Ca²⁺ rise could be evidenced during the application of tBuOOH, but a large Ca²⁺ increase appeared when extracellular calcium was restored. This increase was partially blocked by GSK-7975A (Figures 2E, 3C). Thus, in C6 cells under control conditions tBuOOH did not induce any measurable intracellular calcium release, but evoked an entry of calcium upon extracellular calcium reinstatement partially sensitive to GSK-7975A. To confirm this entry was not a SOCE, intracellular Ca²⁺ stores were first depleted by CPA prior to the tBuOOH application. Under these conditions, the Ca²⁺ entry was still observed and retained its partial sensitivity to GSK-7975A (Figure 2F). The AUC for the Ca²⁺ entry plotted in the Figure 2F were 1.64 ± 0.34 in control and 1.04 ± 0.30 in the presence of GK-7975A (n = 20 each). The La³⁺-sensitive Ca²⁺ entry triggered by tBuOOH is thus largely independent of SOCE mechanism under control conditions.

We next evaluated how changing GSH content altered [Ca²⁺]_i changes elicited by tBuOOH. First of all, resting [Ca²⁺]_i was measured in control, BSO- and SFN-pretreated cells. It appears that [GSH]_i manipulations did not significantly alter basal [Ca²⁺]_i in these three groups of cells: resting [Ca²⁺]_i was 65 ± 2 nM, 66 ± 2 nM, 72 ± 5 nM, respectively (see Table 2). In SFN-pretreated cells, tBuOOH exposure elicited a Ca²⁺ entry lower than in control cells. By contrast, in GSH-depleted cells with BSO, a significant potentiation of the tBuOOH-evoked [Ca²⁺]_i increase was observed (Figures 3A,C). Moreover this potentiation [Ca²⁺]_i was almost completely inhibited by CRAC blocker (Figures 3B,C). Thus, GSH-depletion apparently sensitized CRAC to induce Ca²⁺ entry.

The sensitization by GSH depletion of CRAC observed under oxidative stress suggested that SOCE elicited by Ca²⁺ mobilization from the ER should be modulated by GSH content manipulation (Figure 3). We have thus tested whether the activation of the IP₃/Ca²⁺ mobilization pathway exhibited sensitivity to the GSH concentration. The stimulation of G_q-linked metabotropic P₂Y receptors with ADP elicited intracellular mobilization from IP₃-sensitive internal calcium stores in C6 cells (Figures 4A,B). Pretreating cells with BSO or SFN significantly impacted IP₃-mediated transient Ca²⁺ increase as shown by the comparisons of ADP concentration-response curves. As compared to control cells, the EC50 of ADP on calcium transient was slightly but non-significantly reduced in BSO-treated cells. By contrast, the EC50 of ADP effect on calcium transient was significantly increased in SFN-treated cells (1 ± 0.2 μM), while Hill number remained unchanged. Moreover, the maximal effect of ADP seemed to be similar in the different conditions (Table 2).

We have then evaluated whether store-operated calcium entry (SOCE) following P₂Y receptor activation was altered by [GSH]_i manipulations. Following extracellular calcium reinstatement ADP triggered a significant calcium influx in BSO-pretreated cells (Figures 4B,D). This influx was blocked by GSK-7975A (Figures 4C,D). SOCE was barely detected in either control or SFN-treated cells. Thus, in C6 glioma cells [GSH]_i depletion greatly potentiated SOCE triggered by IP₃-mediated responses.

Since both ADP evoked calcium transients and SOCE could be modified following intracellular GSH content manipulation, we have then examined the impact of BSO or SFN pretreatments on the capacity of cells to accumulate Ca²⁺ within the ER and on SOCE occurrence using CPA (Figure 2C). The area under the curve (AUC) of both Ca²⁺ mobilization and Ca²⁺ entry did not exhibit significant differences whatever the treatment applied (Table 3). Therefore, GSH manipulation neither modified Ca²⁺ load within the ER nor SOCE elicited by SERCA inhibition. This suggested that GSH metabolism modulation predominantly affected IP₃ receptor-mediated SOCE.

The effects of GSH content manipulations on C6 glioma cells, were then studied on the expression of STIM and Orai, two interacting proteins responsible for SOCE after ER-Ca²⁺ store release.

First, mRNAs of NRF2, as well as of Heme Oxigenase 1 (HO-1) and GCL, two enzymes whose expression is known to be under the control of NRF2, were quantified by qRT-PCR (Figure 5A). Significant increases in HO-1, Nrf2, and GCLC (the catalytic subunit of GCL) mRNA expressions were evidenced in SFN treated cells. Treatments with BSO had no effect on NRF2 and CGLC mRNA expression, but slightly increased the expression of HO-1 mRNA.

The mRNA expressions of the oxidative stress-sensitive TRP channels were examined. The TRPM2 mRNA was not detectable in these cells (not shown). BSO pretreatment had no effect on the expression of TRPM7, TRPV1, and TRPC1 mRNAs, while SFN pretreatment significantly increased the TRPV1 mRNA expression, and to a lesser extent that of TRPC1 (Figure 5B).

Finally, mRNA expressions of the proteins involved in SOCE, i.e., STIM and Orai, were measured. SFN pretreatment induced a significant decrease in STIM2 mRNAs accompanied by significant increases in Orai1 and Orai3 mRNA expressions. BSO pretreatment significantly diminished STIM2 mRNA expression (Figure 5C).

SOCE protein expressions were further examined by western blot. We observed that SFN treatment led to a significant decrease in both STIM2, Orai1, and Orai3 expressions (Figures 6A,B). The decreased expression of STIM and Orai proteins could explain the diminished SOCE mediated by IP₃/Ca²⁺ mobilization observed in these cells after this treatment.

To further characterize membrane conductance changes elicited by [GSH]_i manipulations on P₂Y receptor signaling, electrophysiological recordings were carried out. The application of ADP resulted in the occurrence of a potassium current reversing around −77.0 ± 3.7 mV (n = 4) which exhibited an inward rectification at high voltages and a high sensitivity to 20 mM tetraethyl ammonium (TEA, Figure 7A). In cells pretreated with SFN, the current density measured at + 50 mV was not significantly decreased from 6.61 ± 1.94 pA/pF (n = 4) in control cells to 2.54 ± 0.66 pA/pF (n = 5). Under these conditions, the reversal potential was not significantly different from the one obtained in control cells (−69.6 ± 27.4 mV, n = 5, Figure 7B). By contrast, in cells pretreated with BSO, the ADP-elicited current was dramatically and significantly increased as current density reached 33.2 ± 5.0 pA/pF (n = 3) at +50 mV with shifted reversal potential to −48.0 ± 2.9 mV (n = 3, Figure 7B). These results agreed with the sensitization of the P₂Y receptor response as previously observed monitoring [Ca²⁺] changes (Figures 4A,B). Interestingly in BSO-treated cells, ADP-elicited currents were significantly reduced by 20 mM TEA as well as by 5 μM GSK-7975A. Indeed, in BSO-treated cells the current density at +50 mV was 7.7 ± 4.4 pA/pF (n = 3) and 3.9 ± 1.8 pA/pF (n = 3), in the presence of TEA and GSK-7975A, respectively. Under these conditions, both TEA and GSK shifted the reversal potential from −48.0 ± 2.9 mV (n = 3) to −35.94 ± 0.05 mV (n = 3) and −27.5 ± 1.2 mV (n = 3) respectively (Figure 7C).