To investigate the effect of Bcl-2 on Ca2+ signaling in T cells, we continuously monitored Ca2+ concentration in Bcl-2–positive and –negative clones of the WEHI7.2 T cell line after T cell receptor (TCR) activation by anti-CD3 antibody. In Bcl-2–negative cells, high concentrations of anti-CD3 antibody induced a transient Ca2+ elevation, triggering apoptosis. In contrast, low concentrations of anti-CD3 antibody induced Ca2+ oscillations, activating the nuclear factor of activated T cells (NFAT), a prosurvival transcription factor. Bcl-2 blocked the transient Ca2+ elevation induced by high anti-CD3, thereby inhibiting apoptosis, but did not inhibit Ca2+ oscillations and NFAT activation induced by low anti-CD3. Reduction in the level of all three inositol 1,4,5-trisphosphate (InsP3) receptor subtypes by small interfering RNA inhibited the Ca2+ elevation induced by high but not low anti-CD3, suggesting that Ca2+ responses to high and low anti-CD3 may have different requirements for the InsP3 receptor. Therefore, Bcl-2 selectively inhibits proapoptotic Ca2+ elevation induced by strong TCR activation without hindering prosurvival Ca2+ signals induced by weak TCR activation.
Ca2+ is a versatile second messenger that mediates a wide range of cellular processes, including cell division and apoptosis (Berridge et al., 2003). Under physiological conditions, cytoplasmic Ca2+ is maintained at a low level, and it is the elevation of cytoplasmic Ca2+ that generates Ca2+ signals. Elevated Ca2+ transmits information by activating Ca2+-sensitive effectors, including phosphatases and kinases. The Ca2+ elevation involved in signal transduction is often in the form of repetitive Ca2+ spikes or oscillations (Berridge, 1997b). The information-processing capability of Ca2+ signaling is enhanced by modulation of the frequency, amplitude, and spatial properties of Ca2+ elevations. This in part explains how a simple messenger such as Ca2+ can regulate diverse cellular processes.
In T cells, Ca2+ signals mediate a variety of responses to T cell receptor (TCR) activation, including cell proliferation and apoptosis (Winslow et al., 2003; for reviews see Berridge, 1997a; Lewis, 2001, 2003; Randriamampita and Trautmann, 2004). As in all nonexcitable cells, the T cell Ca2+ response begins with the release of Ca2+ from the ER through inositol 1,4,5-trisphosphate (InsP3)–dependent Ca2+ channels (InsP3 receptors). The resulting cytoplasmic Ca2+ elevation is amplified by Ca2+ entry through Ca2+-release–activated Ca2+ channels on the plasma membrane, producing either a transient Ca2+ elevation or Ca2+ oscillations (Donnadieu et al., 1992a,b; Hess et al., 1993; for review see Lewis, 2001). The Ca2+ signal is then transduced through Ca2+/calmodulin–mediated activation of the protein phosphatase calcineurin, which dephosphorylates and thereby activates the nuclear factor of activated T cells (NFAT; for review see Lewis, 2003; Winslow et al., 2003). NFAT is a transcription factor that activates the interleukin-2 promoter, increasing cell proliferation. Activation of calcineurin, and hence NFAT, is sustained more efficiently by Ca2+ oscillations than by a transient elevation of Ca2+, whereas other Ca2+ responses (e.g., nuclear factor kB and c-Jun NH2-terminal kinase activation) are preferentially activated by transient Ca2+ elevation (Dolmetsch et al., 1997, 1998). The importance of Ca2+ oscillations in T cell signaling is increasingly recognized, including evidence that Ca2+ oscillations regulate thymocyte motility during positive selection in the thymus (Bhakta et al., 2005).
We recently reported that the antiapoptotic protein Bcl-2 (Cory and Adams, 2002) interacts with InsP3 receptors on the ER and inhibits InsP3-mediated Ca2+ efflux (Chen et al., 2004). As a consequence, Bcl-2 dampens the cytoplasmic Ca2+ elevation induced by an antibody to the CD3 component of the TCR complex. These findings are intriguing in view of the known role of Ca2+ in signaling apoptosis (for reviews see Hajnoczky et al., 2003; Orrenius et al., 2003; Hanson et al., 2004), but an inhibitory effect of Bcl-2 on InsP3-mediated Ca2+ elevation would seem incompatible with the wide range of physiological processes governed by InsP3-mediated Ca2+ signals. Would not Bcl-2 interfere with Ca2+ signals that regulate physiological processes required for cell function and survival?
A possible clue to this dilemma was provided by earlier work indicating that Ca2+ responses after TCR activation vary according to the strength of TCR activation (Donnadieu et al., 1992a). Typically, strong signals induced by a high concentration of anti-CD3 antibody trigger a single transient elevation of cytoplasmic Ca2+, whereas weaker signals induced by a low concentration of anti-CD3 induce Ca2+ oscillations (Donnadieu et al., 1992a). Our previous experiments demonstrating an inhibitory effect of Bcl-2 on anti-CD3–induced Ca2+ elevation used a high concentration of anti-CD3 antibody that induced a transient Ca2+ elevation rather than Ca2+ oscillations. Therefore, in the present work, we investigated the effect of Bcl-2 on Ca2+ signals induced over a broad range of anti-CD3 concentrations. This led to the discovery that Bcl-2 differentially regulates Ca2+ signals according to the strength of TCR activation. Thus, Bcl-2 inhibited the transient Ca2+ elevation induced by a high concentration of anti-CD3 antibody, without interfering with Ca2+ oscillations induced by a low concentration of anti-CD3 antibody. Accordingly, Bcl-2 inhibited Ca2+-mediated apoptosis after strong TCR activation but did not inhibit NFAT activation after weak TCR activation. Therefore, by selectively regulating Ca2+ signals according to the strength of TCR activation, Bcl-2 discriminates between proapoptotic and prosurvival Ca2+ signals.
Bcl-2 inhibits Ca2+ elevation induced by high but not low anti-CD3 antibody
The WEHI7.2 T cell line corresponds to an immature double-positive stage of T cell differentiation, as WEHI7.2 cells express both CD4 and -8 antigens and are sensitive to glucocorticosteroid-induced apoptosis. Consistent with this stage of development, Bcl-2 is virtually undetectable in WEHI7.2 cells. In earlier work, Bcl-2–positive and –negative clones were derived by stably transfecting WEHI7.2 cells with an expression vector encoding full-length human Bcl-2 or an empty vector, respectively. The full characterization of the clones used in this work has been reported previously (Chen et al., 2004). All findings reported here are based on comparison of three Bcl-2–positive and three Bcl-2–negative clones. Findings were consistent among the different clones; therefore, data from individual clones have been pooled, unless otherwise noted.
Throughout this paper, cytoplasmic Ca2+ was measured at a single-cell level by digital imaging. An initial series of Ca2+ measurements was performed to determine the dose–response relationship between anti-CD3 concentration and cytoplasmic Ca2+ response patterns in a Bcl-2–negative clone (Fig. 1 A). In this experiment, a transient elevation of Ca2+ was defined as only one or two Ca2+ elevations reaching at least twice the basal level of Ca2+, whereas sustained Ca2+ oscillations were defined as three or more Ca2+ spikes at least twice the basal level of Ca2+ and separated by at least a 30-s interval. The percentage of cells responding with a transient Ca2+ elevation was maximal at 20 μg/ml anti-CD3 antibody and declined progressively with increasing antibody dilution (Fig. 1 A). Conversely, the percentage of cells developing Ca2+ oscillations increased progressively as anti-CD3 antibody concentration was reduced. Thus, there is a reciprocal dose–response relationship for transient elevations of Ca2+ versus Ca2+ oscillations after TCR activation. Based on these initial findings, subsequent studies used 20 μg/ml as representative of a high concentration of anti-CD3 antibody and 2, 0.75, and 0.33 μg/ml as representative of low concentrations of anti-CD3 antibody.
Representative Ca2+ traces comparing Bcl-2–negative and –positive cells at a high anti-CD3 concentration (20 μg/ml) are shown in Fig. 2. Bcl-2–negative cells responded to high anti-CD3 with an almost synchronous elevation of cytoplasmic Ca2+ as shown in Fig. 2 A, where multiple single-cell Ca2+ traces collected in a single experiment (i.e., a single microscope coverslip) are plotted together. In contrast, only a small proportion of Bcl-2–positive cells responded with a detectable Ca2+ elevation (Fig. 2 E). The inhibitory effect of Bcl-2 on anti-CD3–induced Ca2+ elevation in these experiments is also illustrated by averaging the Ca2+ response of multiple cells (Fig. 2, B and F). Characteristics of Ca2+ responses at a single-cell level are illustrated by the Ca2+ traces in Fig. 2 (C, D, G, and H). Though the number of responding cells and the amplitude of Ca2+ elevations differed markedly between Bcl-2–negative and –positive cells, the shape of the Ca2+ elevation was similar. In a small percentage of cells, secondary Ca2+ elevations were observed (e.g., Fig. 2 H), but as noted in Fig. 1 B, sustained oscillations were infrequent.
Representative Ca2+ traces comparing Bcl-2–negative and –positive cells at lower concentrations of anti-CD3 antibody (2, 0.75, and 0.33 μg/ml) are shown in Fig. 3. In contrast to the broad elevation of Ca2+ triggered at 20 μg/ml anti-CD3, the Ca2+ responses at low concentrations of anti-CD3 antibody were typically in the form of narrow spikes. Also, in contrast to the inhibitory effect of Bcl-2 on Ca2+ responses at high anti-CD3, Bcl-2 did not inhibit the Ca2+ spikes evoked by low concentrations of anti-CD3. This is illustrated by the representative traces shown in Fig. 3. In Fig. 3, B–D and F–H show repetitive Ca2+ spikes, or oscillations, detected in both Bcl-2–negative and –positive cells over the range of anti-CD3 concentrations tested, whereas A and E show cells where only one or two Ca2+ spikes were detected. Note that the Ca2+ oscillations induced by low anti-CD3 were irregular in terms of both amplitude and frequency. This is characteristic of Ca2+ oscillations in T cells, as reported previously, and is in contrast to the more uniform pattern of Ca2+ oscillations observed in nonlymphoid cells (for reviews see Lewis, 2001; Randriamampita and Trautmann, 2004). Also, spontaneous Ca2+ oscillations were not detected in control experiments where anti-CD3 was not added to cells.
The responses of cells to either high (20 μg/ml) or low (2, 0.75, or 0.33 μg/ml) anti-CD3 in a large number of experiments are summarized in Fig. 1 (B and C). These data indicate that oscillations are much more frequent at the low than at the high concentration of anti-CD3 (Fig. 1 B). Moreover, these data confirm that Bcl-2 markedly inhibits Ca2+ responses to strong TCR activation (20 μg/ml anti-CD3 antibody), based on a significant (P ≤ 0.01) reduction in both the percentage of cells that respond (Fig. 1 B) and the amplitude of Ca2+ elevations in those cells that do respond (Fig. 1 C). In contrast, Bcl-2 did not inhibit Ca2+ responses to weak TCR activation (2, 0.75, or 0.33 μg/ml anti-CD3 antibody), based on the percentage of cells that respond (Fig. 1 B) and the mean amplitude of Ca2+ spikes (Fig. 1 C). Interestingly, at low anti-CD3 in Bcl-2–positive cells, there was a small but insignificant (P > 0.10) reduction in the percentage of cells developing a transient Ca2+ elevation (i.e., one or two Ca2+ spikes; Fig. 1 B, left) and a small but insignificant (P > 0.10) increase in the percentage of cells developing sustained Ca2+ oscillations (i.e., three or more Ca2+ spikes; Fig. 1 B, right). Moreover, the mean amplitude of Ca2+ spikes at the lowest anti-CD3 concentration (.33 μg/ml) was higher in Bcl-2–positive than in Bcl-2–negative cells (Fig. 1 C), although this difference was not statistically significant (P > 0.10).
As noted in Fig. 1 D, a major difference between the Ca2+ elevations induced at high versus low anti-CD3 antibody concentrations was in the duration (i.e., width) of the individual Ca2+ peaks. This is illustrated by the representative Ca2+ traces in Figs. 2 and 3 and is also documented quantitatively in Fig. 1 D. The mean peak width was 4 min at 20 μg/ml anti-CD3 antibody but <1 min at 2 μg/ml anti-CD3 antibody. Moreover, Bcl-2 did not alter the width of individual Ca2+ elevations (either transient induced by high anti-CD3 or transient and oscillatory at low anti-CD3; Fig. 1 D).
Detailed comparisons of Ca2+ oscillations induced by low concentrations of anti-CD3 antibody in Bcl-2–negative and –positive cells are summarized in Fig. 4. To quantitatively compare the oscillatory frequency in Bcl-2–positive and –negative cells, Ca2+ traces from numerous experiments were separated into successive 5-min time periods and the number of Ca2+ spikes during each period was logged, a method that has been described previously (Bird and Putney, 2005; Fig. 4, A and B). Overall, the frequency of Ca2+ oscillations induced by 2 μg/ml anti-CD3 antibody appeared higher in Bcl-2–positive than in Bcl-2–negative cells, although differences reached statistical significance only during the 15-min time interval (P = 0.01) and borderline significance during the 10-min time interval (P = 0.057; Fig. 4 A). The frequency of oscillations also appeared to be higher in Bcl-2–positive cells at 0.75 μg/ml anti-CD3 antibody, but differences were not statistically significant at any of the time intervals (Fig. 4 B). To analyze oscillatory frequency by a different method, the time interval between Ca2+ spikes was measured in multiple Ca2+ traces at 2 μg/ml anti-CD3 antibody and, based on these data, mean and mode frequencies were calculated. The mean frequency of Ca2+ oscillations in Bcl-2–negative cells was 5.2 ± 0.6 mHz, whereas the mode frequency was 7.3 ± 0.55 mHz. Mean frequency was lower than mode frequency because of the presence of low-frequency spiking detected in a small proportion of the Ca2+ traces. The mean frequency was not significantly different in Bcl-2–negative and –positive cells (Fig. 4 C), but the mode frequency was significantly higher in Bcl-2–positive cells (Fig. 4 D). Thus, consistent with the analysis in Fig. 4 A, this analysis suggests an increased frequency of Ca2+ oscillations in Bcl-2–positive compared to Bcl-2–negative cells. The total duration of Ca2+ oscillatory runs (i.e., time duration from initial to final Ca2+ spikes) appeared longer in Bcl-2–positive than in Bcl-2–negative cells, although this difference was of borderline significance (P = 0.05; Fig. 4 E). WEHI7.2 cells adhered loosely to coverslips, limiting the rate at which anti-CD3 antibody could be perfused onto cells. Therefore, to estimate latency period, the initial Ca2+ response to 2 μg/ml anti-CD3 antibody was recorded in cell suspensions fluorometrically (Fig. 4 F). Based on these data, the latency period was on the order of 2 min and was the same in Bcl-2–negative and –positive cells. Although the preceding analyses suggest that Bcl-2 may slightly increase the frequency and duration of Ca2+ oscillations, the major conclusion from these data is that Bcl-2 does not inhibit Ca2+ oscillations induced by low concentrations of anti-CD3 antibody. Consistent with this finding, Bcl-2 did not inhibit NFAT activation (Fig. 4 G).
Bcl-2 inhibits anti-CD3–induced apoptosis by inhibiting anti-CD3–induced Ca2+ elevation
The mean amplitudes of Ca2+ elevations induced by high and low concentrations of anti-CD3 antibody were compared in Fig. 1 C. Although the mean peak amplitude of the Ca2+ elevation induced by 20 μg/ml anti-CD3 was similar to the mean peak amplitude of Ca2+ spikes induced by 2 μg/ml anti-CD3, the net effect of amplitude and width of Ca2+ peak is a much more substantial Ca2+ elevation after high anti-CD3 than after low anti-CD3. Furthermore, analysis of the peak Ca2+ elevation induced by 20 μg/ml anti-CD3 on an individual cell basis illustrates a broad range of Ca2+ elevations detected in Bcl-2–negative cells (Fig. 5) . 27% of the Ca2+ spikes in Bcl-2–negative cells were >200 nM, an arbitrarily chosen threshold level, whereas only 7% of Ca2+ spikes in Bcl-2–positive cells were higher than this level after treatment with 20 μg/ml anti-CD3 antibody. Therefore, Bcl-2 not only reduces the percentage of cells that elevate Ca2+ in response to high anti-CD3 (Fig. 1 B) but also dampens Ca2+ elevations in the cells that do respond by preventing very high peak Ca2+ elevations. In contrast to the wide distribution of Ca2+ elevations observed at 20 μg/ml anti-CD3 antibody in Bcl-2–negative cells, only 6 and 1% of Ca2+ spikes were >200 nM at 2 and 0.75 μg/ml anti-CD3 antibody in Bcl-2–negative cells (Fig. 5). Even fewer cells elevated their Ca2+ to >200 nM at these low concentrations of anti-CD3 antibody in Bcl-2–positive cells (Fig. 5 C). Thus, although Bcl-2 did not reduce the percentage of cells that responded to low concentrations of anti-CD3 antibody by developing sustained Ca2+ oscillations (Fig. 1 B) and did not reduce the mean amplitude of these Ca2+ elevations (Fig. 1 C), it does appear to have set a threshold level above which the Ca2+ does not elevate even in response to weak TCR activation.
To investigate the contribution of high Ca2+ elevations to anti-CD3–induced apoptosis, cells were treated with 20 μg/ml anti-CD3 antibody and sorted by flow cytometry into two different populations based on relative levels of cytoplasmic Ca2+ (Fig. 6 A). The cells were then placed in culture, and the percentage of apoptotic cells was measured 24 h later. A significantly higher percentage of cells in the high Ca2+ population underwent apoptosis, compared to cells in the low Ca2+ population (Fig. 6 B). Conversely, reducing extracellular Ca2+ concentration, a condition that partially prevents Ca2+ elevation after treatment with high anti-CD3 (Fig. 7 A), inhibited apoptosis (Fig. 7 B). Thus, the induction of apoptosis in Bcl-2–negative cells is Ca2+ dependent, and the percentage of cells undergoing apoptosis is proportional to the peak amplitude of Ca2+ elevation after treatment with a high concentration of anti-CD3 antibody. Consistent with these findings, Bcl-2 inhibited apoptosis induction by high anti-CD3 (Fig. 7, C and D), in accordance with the dampening effect of Bcl-2 on anti-CD3–induced Ca2+ elevation described in preceding experiments (Fig. 5). Treatment with low anti-CD3 did not induce apoptosis (Fig. 7 C), consistent with the evidence that high Ca2+ elevation (≥200 nM) was much less common after treatment with low anti-CD3 than it was after treatment with high anti-CD3 (Fig. 5). Interestingly, the percentage of apoptotic cells was lower after treatment with low anti-CD3 antibody than it was in untreated cells (Fig. 7 C). This suggests that treatment with low anti-CD3 may have a prosurvival action, in contrast to the proapoptotic action of high anti-CD3.
Effect of InsP3 receptor knockdown on Ca2+ responses to strong and weak TCR activation
To address the question of how Bcl-2 differentially regulates Ca2+ signals induced by high versus low concentrations of anti-CD3 antibody, we used small interfering RNA (siRNA) to reduce levels of all three InsP3 receptor subtypes in WEHI7.2 cells (Fig. 8) . The siRNA olignucleotide pools were introduced into cells by electroporation with a transfection efficiency of 78 ± 2% (mean ± SEM). The siRNA-mediated reduction in InsP3 receptor levels was documented by Western blotting (Fig. 8 A). Substantial reduction in all three InsP3 receptor subtypes was reproducibly achieved (Fig. 8 B). This reduction in InsP3 receptors significantly inhibited the Ca2+ elevation induced by 20 μg/ml anti-CD3 antibody (Fig. 8, C and D). But the siRNA-mediated reduction in InsP3 receptor levels did not inhibit Ca2+ responses to 2 μg/ml anti-CD3 antibody. This is documented by representative Ca2+ traces (Fig. 8, E and F), by analysis of the percentage of responding cells (Fig. 8 G), and by amplitude analysis (Fig. 8 H). Thus, consistent with the inhibitory effect of Bcl-2 on Ca2+ responses to strong but not to weak TCR activation, the former appears to be more dependent on InsP3 receptor expression levels than the latter.
The principal finding in this study is that Bcl-2 differentially regulates Ca2+ signaling in T cells according to the strength of TCR activation. Bcl-2 was found to inhibit the cytoplasmic Ca2+ elevation induced by a high concentration of anti-CD3 antibody but did not inhibit Ca2+ elevation induced by low concentrations of anti-CD3 antibody. This finding evolved as a natural extension of our earlier investigation into the effect of Bcl-2 on InsP3 receptor–mediated Ca2+ release from the ER (Chen et al., 2004). Those studies initially used fluorometric measurements of Ca2+ elevation in response to a relatively high concentration of anti-CD3 antibody. It was this Ca2+ elevation that we found to be inhibited by Bcl-2. Although digital imaging was eventually used in addition to fluorometry in these earlier studies, a high concentration of anti-CD3 antibody was adhered to throughout. Thus, only the effect of Bcl-2 on the transient Ca2+ elevation induced by a relatively high concentration of anti-CD3 antibody was investigated. The present study was undertaken based on the prediction that Ca2+ oscillations would be detected if a lower concentration of anti-CD3 antibody were used. This prediction was based on evidence that high concentrations of cell surface receptor agonists generally induce transient elevations of Ca2+, whereas low concentrations of agonist are more likely to induce sustained oscillations (Berridge, 1990). Consistent with this paradigm, it was previously reported that strong TCR activation induces primarily a transient Ca2+ elevation, whereas weaker TCR activation induces primarily repetitive Ca2+ spikes, or oscillations (Donnadieu et al., 1992a).
As anticipated, the Ca2+ response pattern in WEHI7.2 cells underwent a transition from transient Ca2+ elevation to sustained oscillations as anti-CD3 antibody concentration was decreased (Fig. 1 A). The Ca2+ oscillations induced by anti-CD3 antibody were irregular in their amplitude and frequency (Fig. 3). This is characteristic of Ca2+ oscillations in T cells, as reported previously by others, and is in contrast to the more uniform pattern of Ca2+ oscillations observed in nonlymphoid cells (for reviews see Lewis, 2001; Randriamampita and Trautmann, 2004). The irregularity of anti-CD3–induced Ca2+ oscillations necessitated a large number of experiments to objectively compare oscillatory responses in Bcl-2–negative and –positive cells (Fig. 4). The only significant differences were an increase in oscillatory frequency (Fig. 4, A and D) and duration of oscillations (Fig. 4 E) in Bcl-2–positive cells. The oscillatory patterns induced in Bcl-2–negative and –positive cells were indistinguishable in all other respects, including the percentage of cells that developed oscillations (Fig. 1 B), amplitude (Fig. 1 C), width of Ca2+ spikes (Fig. 1 D), and latency period (Fig. 4 F). It has been reported that NFAT is optimally activated by Ca2+ oscillations (Dolmetsch et al., 1998; Tomida et al., 2003; for reviews see Lewis, 2003; Winslow and Crabtree, 2005). Therefore, NFAT activation was measured as a convenient “readout” of the Ca2+ oscillations induced by anti-CD3 in WEHI7.2 cells. Consistent with the finding that Bcl-2 did not inhibit anti-CD3–induced Ca2+ oscillations, Bcl-2 did not inhibit NFAT activation (Fig. 4 G).
The finding that Bcl-2 selectively inhibits the transient Ca2+ elevation induced by high anti-CD3 without interfering with Ca2+ oscillations induced by low anti-CD3 is relevant to the role of Bcl-2 in regulating apoptosis, as WEHI7.2 cells undergo apoptosis after treatment with high anti-CD3 but not when treated with low anti-CD3 (Fig. 7 C). Moreover, apoptosis induction by high anti-CD3 was Ca2+ mediated (Fig. 7, A and B), and the percentage of cells undergoing apoptosis was proportional to the level of Ca2+ elevation (Fig. 6). These findings are consistent with evidence that apoptosis induction after TCR activation is triggered by InsP3 receptor–mediated Ca2+ elevation (Nakayama et al., 1992; Jayaraman and Marks, 1997). Thus, by selectively repressing the Ca2+ elevation induced by strong TCR activation, Bcl-2 inhibits apoptosis without interfering with physiological Ca2+ signals induced by weak TCR activation.
The present findings are intriguing in light of the known role of Bcl-2 in T cell development. T cells located in the thymic cortex are TCR positive and both CD4+ and CD8+ (“double positive”). At this stage of T cell development, Bcl-2 expression is low and cortical thymocytes are highly susceptible to apoptosis induction after TCR activation by antigen or anti-CD3 antibody (Smith et al., 1989; Murphy et al., 1990; Shi et al., 1991; Nakayama et al., 1992). When T cells mature and migrate to the thymic medulla, they remain TCR positive but become either CD4+CD8− or CD4−CD8+ (“single positive”). Bcl-2 expression is increased at this stage of development, and as a consequence, single-positive thymocytes are less susceptible to apoptosis than are cortical thymocytes (Gratiot-Deans et al., 1993; Veis et al., 1993). A strong correlation has been demonstrated between Bcl-2 expression and susceptibility to Ca2+-induced apoptosis during T cell development (Andjelic et al., 1993). Thymocytes at the earliest stage of development (TCR−CD4−CD8−), the stage during which thymocytes relocate from bone marrow to thymus, express high levels of Bcl-2 and are resistant to Ca2+-mediated apoptosis, whereas thymocytes in the next stage of development (TCR+CD4+CD8+) are highly susceptible to Ca2+-mediated apoptosis. It is in this stage of development that thymocytes undergo either negative or positive selection. Strong ligation of the TCR (e.g., by self-peptide–major histocompatibility complex [MHC] complexes) induces negative selection, whereas weak ligation of the TCR (e.g., by foreign antigen–MHC complexes) induces positive selection (for review see Hogquist, 2001; Neilson et al., 2004). Double-positive thymocytes from Bcl-2–transgenic mice accumulate excessively because of reduced negative selection and are resistant to anti-CD3–induced apoptosis (for review see Cory, 1995). Therefore, the role of Bcl-2 expression during T cell development may be to regulate when and where negative selection occurs. Decreased Bcl-2 expression in double-positive cortical thymocytes, but not in earlier or later stages of T cell development, may limit negative selection to the cortical region of the thymus and to this stage of development. Elevated Bcl-2 expression at earlier (double negative) and later (single positive) stages of T cell development may dampen Ca2+ transients produced by strong TCR engagement while permitting repetitive Ca2+ oscillations that signal cell proliferation and survival.
The mechanism by which Bcl-2 differentially regulates Ca2+ elevation after strong but not weak TCR activation is not entirely understood. In our earlier work (Chen et al., 2004), the inhibitory effect of Bcl-2 on anti-CD3–induced Ca2+ elevation appeared to be mediated at the level of the InsP3 receptor rather than in upstream TCR signaling pathways. This conclusion was based on two experimental strategies in which upstream TCR signaling pathways were bypassed. In one strategy, we found that Bcl-2 inhibited Ca2+ elevation induced by a cell-permeant InsP3 ester. In the other strategy, we found that Bcl-2 inhibited ER Ca2+ release induced by adding InsP3 to cells in which the plasma membrane had been permeabilized by digitonin. In addition, Bcl-2 appeared to interact with InsP3 receptors, based on results of blue native gel electrophoresis and coimmunoprecipitation (Chen et al., 2004). Finally, purified Bcl-2 inhibited InsP3-gated single-channel opening when microsomal membrane fractions containing InsP3 receptors were incorporated into planar lipid bilayers (Chen et al., 2004). Therefore, the collective evidence that Bcl-2 interacts with InsP3 receptors and inhibits InsP3-mediated Ca2+ release from the ER raises the question of whether the induction of Ca2+ oscillations by low concentrations of anti-CD3 antibody is InsP3 receptor independent or at least requires far fewer functional InsP3 receptors than does the elevation of Ca2+ induced by a high concentration of anti-CD3 antibody.
To address this question, we used siRNA to reduce InsP3 receptor levels in WEHI7.2 cells (Fig. 8). This procedure inhibited Ca2+ elevation induced by strong TCR activation but did not inhibit the induction of Ca2+ oscillations by weak TCR activation. In contrast, Ca2+ responses evoked in HeLa cells by both high and low concentrations of ATP or histamine were repressed by InsP3 receptor type 1 knockdown (Hattori et al., 2004). Thus, mechanisms of Ca2+ oscillation formation after TCR activation and G protein–coupled receptor activation may differ. Our findings indicate that Ca2+ responses initiated by weak TCR activation are generated independent of InsP3 receptor–mediated Ca2+ release or that only a relatively small proportion of the full InsP3 receptor complement is required to initiate Ca2+ signals in response to weak TCR activation.
In future studies, the mechanism of how Bcl-2 regulates InsP3 receptor function will be addressed in greater depth. In preliminary studies, we found that Bcl-2 overexpression decreases InsP3 receptor phosphorylation in WEHI7.2 cells. Moreover, it has recently been reported that Bcl-2 interacts with InsP3 receptors in a manner that is dependent on the Bcl-2 phosphorylation state and may regulate Ca2+ dynamics in the ER through regulation of InsP3 receptor phosphorylation (Bassik et al., 2004; Oakes et al., 2005). Others have reported that in neuronal cells Bcl-2 shuttles calcineurin to InsP3 receptors and regulates Ca2+ release from internal stores (Erin et al., 2003a,b; Erin and Billingsley, 2004). Therefore, one hypothesis is that strong TCR signals enhance InsP3 receptor phosphorylation, enhancing InsP3-induced Ca2+ release, and that Bcl-2 dampens the Ca2+ response to strong TCR activation by mediating dephosphorylation of InsP3 receptors. Although untested, this theory is consistent with evidence that phosphorylation regulates the Ca2+ channel activity of InsP3 receptors (Cameron et al., 1995; Jayaraman et al., 1996; deSouza et al., 2002; Straub et al., 2002; Cui et al., 2004).
In summary, we previously discovered that the known antiapoptotic protein Bcl-2 interacts with InsP3 receptors and inhibits InsP3-induced Ca2+ release from the ER in T cells. In this paper, we report that Bcl-2 selectively inhibits Ca2+ elevation induced by high but not low anti-CD3. As a consequence, Bcl-2 represses the transient elevation of Ca2+ associated with apoptosis induction after strong TCR activation but does not interfere with Ca2+ oscillations that activate NFAT after weak TCR activation. The capacity of Bcl-2 to differentially regulate Ca2+ signals induced by strong versus weak TCR activation allows Bcl-2 to selectively inhibit apoptotic Ca2+ signals without interfering with Ca2+ signals that mediate cell proliferation and survival.
Materials And Methods
EGTA and standard reagents were purchased from Sigma-Aldrich. Fura-2–AM and Hoechst 33342 were purchased from Invitrogen. Hamster anti–mouse CD3e ε chain monoclonal antibody (clone 145-2C11) and mouse anti–hamster IgG1 monoclonal antibody were obtained from BD Biosciences. Mouse monoclonal antibody NFATc2 was obtained from Santa Cruz Biotechnology, Inc.
WEHI7.2 cells were cultured in DME supplemented with 10% bovine calf serum, 2 mM l-glutamine, and 100 μM of nonessential amino acids. Transfection procedures, isolation of Bcl-2–positive and –negative clones, and the characterization of these clones were reported previously (Chen et al., 2004).
Methods of Ca2+ imaging, described in detail previously (Chen et al., 2004), were used here with only minor modifications. In brief, cells adhered to poly-l-lysine–coated coverslips (35-mm coverslip dishes; MatTek Corp.) were loaded with 1 μM fura-2–AM for 45 min at 25°C in extracellular buffer (ECB; 130 mM NaCl, 5 mM KCl, 1.5 mM CaCl2, 1 mM MgCl2, 25 mM Hepes, pH 7.5, 1 mg/ml BSA, and 5 mM glucose). The buffer was replaced with fresh ECB and the incubation was continued for 45 min at 25°C to permit deesterification. Culture dishes were mounted on the nonheated stage of an inverted microscope (CKX41; Olympus) equipped with a 20× fluor objective. Excitation light was alternated between 340 and 380 nm by a filter wheel (Sutter Instrument Co.), with 0.8- and 0.2-s exposure times, respectively, and emitted light was filtered at >510 nm and collected with an intensified charge-coupled device camera (12-bit VGA; Cooke). Anti-CD3 antibody was gently added to buffer overlaying the coverslip so as not to disturb cells loosely adherent to the coverslip. The video signal was digitized using InCyt Im2 software (Intracellular Imaging) and subsequently processed using Excel (Microsoft). To determine Rmin, cells were perfused with ECB deficient in Ca2+ and supplemented with 4 mM EGTA and 10 μM ionomycin. Rmax was obtained by perfusing cells with ECB supplemented with 4 mM CaCl2 and 10 μM ionomycin. Ca2+ concentration was calculated based on the published Kd for fura-2 of 220 nM, by the equation of Grynkiewicz et al. (1985).
Fluorometric Ca2+ measurements
The measurement of Ca2+ concentration in cell suspensions by fluorometry using fura-2–AM have been described in detail previously (Chen et al., 2004).
Cells were treated with 2 μg/ml anti-CD3 antibody for various time periods at ambient temperature, after which they were placed on ice, pelleted, and resuspended in RIPA buffer (1% Triton X-100, 0.1% SDS, 50 mM Tris, pH 7.6, 150 mM NaCl, and 200 mM DTT) supplemented with Complete mini protease inhibitors (Roche) and Phosphatase inhibitor cocktails I and II (Sigma-Aldrich). Cell extracts were resolved by electrophoresis on 7% SDS–polyacrylamide gels under reducing conditions. The separated proteins were transferred to Immobilon-P PVDF membranes (Millipore) and incubated with anti-NFATc2 antibody at a dilution of 1:500, followed by incubation with horseradish peroxidase–conjugated goat anti–mouse IgG and visualized by the ECL Western blotting detection reagent (GE Healthcare).
InsP3 receptor Westerns
Western analysis for InsP3 receptors was performed as described previously (Chen et al., 2004). Protein samples were extracted as in the preceding method and resolved (60 μg/lane) through 4–20% gradient gels (Bio-Rad Laboratories). The antibodies for types 1 and 3 InsP3 receptors were purchased from EMD Biosciences and BD Biosciences, respectively. The antibody for type 2 InsP3 receptor was a gift from R. Wojcikiewicz (State University of New York Upstate Medical University, Syracuse, NY). The antibody for actin was obtained from Sigma-Aldrich. Secondary antibodies were obtained from GE Healthcare.
Cells were stained with Hoechst 33342 (final concentration 10 μg/ml), and typical apoptotic nuclear morphology was detected by epifluorescence microscopy using a microscope (Axiovert S100; Carl Zeiss MicroImaging, Inc.) equipped with a 63× oil/1.4 NA plan apochromat objective (Carl Zeiss MicroImaging, Inc.) and a filter cube (model XF23; Omega Optical; excitation = 485 nm, emission = 535 nm). Images were taken on a charge-coupled device camera (ORCA C4742-95-cooled; Hamamatsu) operating with Simple PCI software (Compix, Inc.).
Cells (1 million/ml) were loaded with 5 μM calcium green–AM (Invitrogen) for 45 min at 37°C in ECB. The cells were then pelleted and resuspended in ECB at the same concentration and incubated at room temperature for 30 min to allow dye deesterification. The cells were then pelleted and concentrated to 5 million/ml. The cells were then analyzed and sorted on a flow cytometer (Epics Elite; Beckman Coulter). Calcium green fluorescence was measured after dye excitation with a 488-nM argon laser, and emitted light collection was measured through a 525-nM band-pass filter. The cells were initially run through the flow cytometer for 1 min to assess basal cytosolic Ca2+, and 20 μg/ml anti-CD3 antibody was then added. The cells were gated and sorted into two populations: cells with a high level of Ca 2+ elevation and cells with a low level of Ca 2+ elevation. The sorted cells were pelleted and resuspended in fresh culture medium and 20 μg/ml anti-CD3 antibody was re-added, and 30 min later an equal concentration of anti–hamster IgG was added. Apoptosis was measured 24 h later, as described in the previous section.
The negative control, siCONTROL Non-Targeting siRNA Pool, and siGENOME SMARTpools for all three subtypes of InsP3 receptor were purchased from Dharmacon. After suspension in 1× siRNA buffer, SMARTpools were added at a concentration of 1 μM each to 0.2-cm cuvettes containing 5 million WEHI7.2 cells suspended in 200 μl Opti-MEM I (Invitrogen). Cuvettes were then subjected to a single 140V 10-ms2–wave pulse from a GenePulser Xcell (Bio-Rad Laboratories), and the contents of the cuvette were immediately added to fresh media. Cells were grown in culture after transfection for 48 h before use in experiments. Transfection efficiency was measured by transfecting siGLO Cyclophilin (Dharmacon) at a concentration of 1 μM. After 30 min, cells were pelleted and then resuspended in phosphate-buffered saline. Cells were visualized by fluorescence microscopy, with excitation at 546 nm, and at least 200 cells were counted in three separate experiments to determine the percentage of transfected cells.
Comparisons were made using the two-tailed t test for two samples, assuming equal variance.
The authors especially thank Michael Berridge for suggesting the use of low anti-CD3 concentrations to trigger calcium oscillations and for critical comments and suggestions during the course of this work. The authors also thank William Schilling, George Dubyak, Martin Bootman, Llewelyn Roderick, Gary Bird, and Doug Green for helpful discussions and advice. We thank Michael Sramkoski and Tammy Stefan for assistance with flow cytometry and Richard Wojcikiewicz for providing antibodies.
This work was supported by National Institutes of Health grants RO1 CA085804 (to C.W. Distelhorst) and T32 HL07147 (to M.C. Davis).
F. Zhong and M.C. Davis contributed equally to this paper.
Abbreviations used in this paper: InsP3, inositol 1,4,5-trisphosphate; NFAT, nuclear factor of activated T cells; siRNA, small interfering RNA; TCR, T cell receptor.