A Comparison of Signaling Requirements for Apoptosis of Human B Lymphocytes Induced by the B Cell Receptor and CD95/Fas¹

Programmed cell death, or apoptosis, is an active process fundamental to the development and homeostasis of multicellular organisms (1). Apoptosis of B lymphocytes is an essential mechanism for both eliminating self-reactive B cells during development and regulating clonal B cell populations during the immune response (reviewed in 2 . Signaling through the B cell antigen receptor (BCR)³ complex by multivalent self Ags, such as those on cell surfaces, can eliminate immature self-reactive B cells (3, 4, 5). During B cell lymphopoiesis, death induced by self Ag most likely occurs at the stage when the BCR is first expressed (6). However, BCR ligation can lead to either clonal proliferation or deletion by apoptosis, depending on the developmental state of the B cell and signals derived from coreceptors (7; reviewed in 8).

The CD95/Fas receptor, a member of the TNF receptor family, also makes an important contribution to determining the lymphocyte life span (9, 10). Cross-linking Fas with agonistic Abs or Fas ligand results in rapid apoptosis of many cell types (reviewed in 11 . Fas-induced apoptosis plays an important role in activation-induced cell death and elimination of autoreactive B cells (reviewed in 2 . The intracellular domain of CD95/Fas contains a death domain that recruits other death domain-containing proteins, including the adapter protein FADD/MORT-1. FADD/MORT-1, in turn, recruits a member of the caspase family of cysteine proteases termed caspase 8 (FLICE/MACH) (12, 13, 14). Caspase 8 is then thought to undergo autoproteolytic activation, initiating a cascade of proteolytic events involving downstream caspases such as caspase 1 (ICE) and caspase 3 (CPP32/Yama) (15, 16). This caspase cascade, by cleaving critical targets, plays a central role in the induction of apoptosis by a wide variety of stimuli (reviewed in 17 . However, the mechanisms regulating the initiation of caspase activity by receptors outside the TNF receptor family and the identities of critical caspase substrates remain to be determined.

As with other cell fate decisions, it is likely that protein phosphorylation/dephosphorylation mechanisms play an important role in the initiation and progression of apoptosis. Recently, two families of mitogen-activated protein kinase (MAPK) have been implicated in apoptotic signaling (reviewed in 18 . These pathways, the stress-activated protein kinase (SAPK) or c-Jun amino-terminal kinase pathway and the p38 MAPK pathways, are strongly activated by stressful stimuli such as proinflammatory cytokines, UV light, and osmotic shock (19, 20, 21, 22). SAPK and p38 MAPK are also activated during apoptosis induced by nerve growth factor withdrawal of PC12 cells, UV irradiation, Fas cross-linking in human T lymphoma cells or neuroblastoma cells, and BCR ligation in human B lymphoma cells (23, 24, 25, 26, 27, 28). The ability of cysteine protease inhibitors such as ZVAD-fmk to block activation of SAPK and p38 MAPK by many apoptotic stimuli suggests that caspases may play an important upstream role in activating these pathways (25, 26).

Whether the SAPK and p38 MAPK pathways play an active role in apoptosis is controversial. Several studies have apparently dissociated SAPK and p38 MAPK activity from the induction of apoptosis (29, 30, 31). However, others have provided strong evidence to support an apoptotic role for the SAPK and p38 MAPK pathways (23, 32, 33, 34). For example, overexpression of MKK6b, an upstream activator of p38 MAPK, has been shown to induce apoptosis in Jurkat T cells (34). Furthermore, MKK6b activity was found to be required for Fas-induced apoptosis (34). Thus, depending on the stimuli and cell type, activation of the SAPK/p38 MAPK pathways may be sufficient to induce cell death. However, the precise cellular roles of these cascades in the cellular response to apoptotic stimuli are not clear. In this study we present data suggesting that p38 MAPK plays a differential role in BCR- and Fas-induced apoptosis of human B lymphoma cells and discuss a model to rationalize these findings.

Goat F(ab′)₂ anti-μ and anti-γ sera were obtained from Jackson ImmunoResearch (West Grove, PA) and Southern Biotechnology (Birmingham, AL). The anti-Fas Ab IPO-4 (35) was a gift from Dr. Svetlana Sidorenko (Kavetsky Institute, Kiev, Ukraine). SB203580 and PD98059 were purchased from Calbiochem (San Diego, CA). Rabbit polyclonal antiserum raised against MAPKAP kinase-2 was a gift from Dr. Philip Cohen (University of Dundee, Dundee, U.K.). Rabbit polyclonal antiserum raised against SAPK (C17) was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit polyclonal Ab against p38 MAPK was provided by Dr. Jeremy Saklatvala (Kennedy Institute of Rheumatology, London, U.K.). ZVAD-fmk was purchased from Kamiya Biomedical (Tukwila, WA). The CPP32 substrate peptide was obtained from Peptides International (Louisville, KY). The B104 human B lymphoma line, provided by Dr. Mitsufumi Mayumi (Kyoto University Hospital, Kyoto, Japan), was grown in culture as previously described (36). The BJAB human B lymphoma line, provided by Dr. Vishva Dixit (Genentech, South San Francisco, CA), was maintained as previously described (37).

After incubation with the indicated stimuli, 5 to 10 × 10⁶ cells were lysed for 15 min on ice in 1 ml of lysis buffer (20 mM HEPES, pH 7.4; 2 mM EGTA; 50 mM β-glycerophosphate; 1% Triton X-100; 10% glycerol; 1 mM DTT; 1 mM PMSF; 25 μg/ml leupeptin; 25 μg/ml aprotinin; 2 mM Na₂VO₄; and 10 mM NaF). Cell debris was removed by centrifugation at 10,000 × g for 10 min at 4°C. SAPK, p38 MAPK, or MAPKAPK-2 was then immunoprecipitated by addition of 2 μg of antiserum for 3 h followed by 50 μl of protein-A Sepharose (1/1) slurry for the final hour. The beads were pelleted by centrifugation and then washed three times each in lysis buffer, wash buffer (500 mM LiCl; 100 mM Tris-Cl, pH 7.6; 0.1% Triton X-100; and 1 mM DTT), and assay buffer (20 mM 3-[N-morpholino]-2-hydroxypropanesulfonic acid (MOPS) pH 7.2; 2 mM EGTA; 10 mM MgCl₂; 0.1% Triton X-100; and 1 mM DTT). The beads were left as a 1/1 suspension in assay buffer, and 20 μl of the appropriate substrate was added. For SAPK and p38 MAPK assays, 0.3 mg/ml of either GST-Jun (for SAPK) or GST-ATF2 (for p38 MAPK) was added. Kinase reactions were initiated by the addition of 15 μl of ³²P-labeled Mg-ATP solution (50 mM MgCl₂, 500 μM ATP, and 10 μCi of [γ-³²P]ATP) and were conducted at 30°C for 20 min. Reactions were stopped by the addition of 25 μl of 6× Laemmli sample buffer and boiling for 3 min. Samples were separated on a 10% SDS-PAGE gel and, after drying, were subjected to autoradiography. Quantitation by densitometry was performed with an imaging densitometer (Bio-Rad, Hercules, CA). For MAPKAPK-2 assays, 0.3 mg/ml substrate peptide (KKLNRTLSVA) was added. Kinase reactions were initiated by the addition of 15 μl of ³²P-labeled MgATP solution (50 mM MgCl₂, 500 μM ATP, and 10 μCi of [γ-³²P]ATP) and were conducted at 30°C for 20 min. Reactions were stopped, and ³²P incorporation was quantified by spotting 50 μl of the assay mixture onto p81 paper, washing in 0.1% phosphoric acid, and counting as previously described (38).

Dead cells were quantified with the use of annexin V flow cytometry as previously described (39). The annexin V assay exploits the fact that an early event during apoptosis of many cells is a loss of membrane lipid asymmetry, resulting in the exposure of phosphatidylserine in the outer leaflet of the plasma membrane. Briefly, cells were incubated with FITC-conjugated annexin V and counterstained with propidium iodide to exclude necrotic cells (Clontech, Palo Alto, CA). The cells were subsequently analyzed using a Becton Dickinson FACStar Plus flow cytometer (San Jose, CA). Other apoptosis assays employed included trypan blue dye exclusion and Hoechst 33342 7-amino-actinomycin D flow cytometry (27). These techniques yielded results similar to those obtained with annexin V (data not shown). Where indicated, cells were pretreated for 3 h with 10 μM ZVAD-fmk.

Following stimulation, cells were washed once in PBS and resuspended at 2 × 10⁸/ml in hypotonic lysis buffer (50 mM NaCl; 40 mM β-glycerophosphate; 10 mM HEPES, pH 7.0; 5 mM EGTA; and 2 mM MgCl₂). The lysate was then subjected to four freeze-thaw cycles before centrifugation at 10,000 × g for 10 min each (100 mM HEPES, pH 7.5; 10% sucrose; 0.1% 3-[(3-cholamidopropyl)dimethyl ammonio]-1-propanesulfonic acid; 10 mM DTT; and 0.1 mg/ml OVA). Protein concentrations were determined, and 25 μg of cell extract was incubated for 1 h at 37°C with either 50 μM Ac-DEVD-AMC or YVAD-AMC. Protease activity was determined by monitoring the release of 7-amino-4-trifluoromethyl coumarin at an excitation wavelength of 400 nm and an emission wavelength of 510 nm with the use of a CytoFluor II 96-well plate spectrofluorometer (PerSeptive Biosystems, Framingham, MA). Relative caspase activity was determined by dividing the activity observed at each time point by the values detected at time zero.

All experiments shown are representative of between three and five repetitions.

To gain insight into the mechanism of BCR-induced apoptosis we compared apoptotic signaling events stimulated by ligation of either surface IgM or CD95/Fas in two human B lymphoma cell lines. BJAB B cells undergo apoptosis in response to cross-linking CD95/Fas with anti-Fas mAbs, whereas B104 cells undergo apoptosis in response to ligation of the BCR with anti-IgM (27, 37, 40). Anti-IgM treatment induced an increase in annexin V binding to B104 cells (Fig. 1,A). This increased annexin V binding was blocked by pretreatment with ZVAD-fmk, a cell-permeable peptide inhibitor of caspases (Fig. 1,A). Treatment of BJAB cells with anti-Fas Abs also induced a ZVAD-fmk-sensitive increase in annexin V binding (Fig. 1 B). Thus, both BCR- and Fas-induced apoptosis, as measured by loss of membrane assymetry, are dependent upon caspase activity.

An analysis of the kinetics of apoptosis revealed that, whereas an increase in annexin V binding was not detectable until 2 to 4 h after BCR ligation of B104 cells, increased annexin V binding was observed within 1 h of treatment of BJAB cells with anti-Fas (Fig. 2, A and B). One possible explanation for the relative delay in apoptosis in B104 cells is that initiation of the cell death process requires synthesis of new genes and proteins. Consistent with this hypothesis, BCR-induced apoptosis was almost completely blocked by preincubation with the RNA synthesis inhibitor actinomycin D (Fig. 2,C). In contrast, actinomycin D did not inhibit anti-Fas-induced apoptosis (Fig. 2 D). Similar results were obtained with the protein synthesis inhibitor cycloheximide (data not shown). These data show that BCR-induced apoptosis, in contrast to Fas-induced apoptosis, is dependent on transcriptional and translational events.

BCR-induced apoptosis in B104 cells is accompanied by a delayed and sustained activation of both SAPK and p38 MAPK (27) (Fig. 3,A). To gain further insight into the roles of these kinase pathways in BCR- and Fas-induced apoptosis, we compared the effects of these agonists on SAPK and p38 MAPK activity in B104 and BJAB cells. Both anti-IgM and anti-Fas induced a slow and sustained increase in SAPK and p38 MAPK activity that peaked about 8 h after treatment (Fig. 3, A and B). However, the kinetics of SAPK and p38 MAPK activation in response to BCR ligation were delayed relative to those observed upon Fas cross-linking. In this respect, the increase in SAPK and p38 MAPK activities stimulated by these treatments parallels the kinetics observed for increased annexin V binding. Activation of SAPK and p38 MAPK are among the earliest measurable events associated with apoptosis in these cells.

Having established that SAPK and p38 MAPK were activated during both BCR- and Fas-induced apoptosis, we tested whether caspase activity was required for activating these kinase pathways using the caspase inhibitor ZVAD-fmk. Activation of both SAPK and p38 MAPK in response to either anti-IgM or anti-Fas was significantly inhibited by pretreatment of the cells with ZVAD-fmk (Fig. 3, A and B). These results indicates that the sustained activation of both SAPK and p38 MAPK by ligation of either BCR or CD95/Fas is dependent on the proteolytic activity of caspases.

We next sought to determine the role of p38 MAPK in BCR- and Fas-induced apoptosis using a p38 MAPK inhibitor. The pyridinyl imidazole p38 MAPK inhibitor, SB203580, has been shown to specifically block p38 MAPK activity by a variety of stimuli in many different cells (41, 42). SB203580 appears to be specific for the α and β isoforms of p38 MAPK (42, 43, 44, 45). The mechanism of SB203580 action involves reversible binding to p38 MAPK itself. Thus, to confirm inhibition of p38 MAPK in vivo, it is necessary to measure an event downstream of p38 MAPK. One such event is the activation of MAPKAPK-2 (38). Figure 4,A shows that anti-IgM treatment of B104 cells resulted in a late and sustained activation of MAPKAPK-2 activity. The kinetics of MAPKAPK-2 activation closely resemble those observed for p38 MAPK in these cells, consistent with its status as a downstream target for p38 MAPK. Preincubation of B104 cells with 10 μM SB203580 almost completely abolished the MAPKAPK-2 response to BCR ligation, indicating that p38 MAPK activity was profoundly inhibited. Under these conditions, BCR-induced apoptosis was also significantly inhibited (Fig. 4,B). For example, SB203580 reduced the percentage of apoptotic cells observed after 12 h of anti-IgM treatment from approximately 75 to 20%. Pretreatment of BJAB cells with SB203580 also completely blocked anti-Fas-induced activation of MAPKAPK-2 (Fig. 4,C). However, in contrast to the effects on BCR-induced apoptosis, SB203580 did not inhibit Fas-induced apoptosis (Fig. 4,D). The dose response for SB203580 inhibition of anti-IgM-induced MAPKAPK-2 activity closely matched that observed for inhibition of apoptosis, with a half-maximal concentration of approximately 2 μM. This provides further support for the hypothesis that p38 MAPK plays an important role in BCR-induced apoptosis of B104 cells (Fig. 5). Consistent with our previous results suggesting that the classical MAPK pathway does not play a determinant role in anti-IgM-induced apoptosis of B104 cells (27), the PD98059 inhibitor of MAP kinase kinase did not inhibit apoptosis in response to anti-IgM (data not shown).

Because induction of caspase activity is likely to represent a common commitment step during apoptosis, we measured caspase activity induced by anti-IgM and anti-Fas. Using a DEVD peptide, a substrate relatively selective for caspase 3 and related caspases, both BCR and Fas cross-linking stimulated an increase in caspase activity measured in cell extract (Fig. 6,A). Whereas increased DEVD-specific protease activity was detectable within 1 h after Fas ligation, anti-IgM-induced caspase activity was not detected until 3 to 4 h after BCR cross-linking. The kinetics of BCR-induced caspase activity paralleled the kinetics of apoptosis (Fig. 2) and SAPK/p38 MAPK activation (Fig. 3) and were delayed relative to Fas-induced caspase activation (Fig. 6 A). Little activity was detected against a YVAD peptide in response to either anti-IgM or anti-Fas (data not shown). This result is consistent with a low level of activity of caspase 1 and related caspases.

To determine whether BCR-induced caspase activity requires new gene expression and protein synthesis, we examined the ability of anti-IgM to induce DEVD-specific caspase activity in B104 cells pretreated with actinomycin D or cycloheximide. Actinomycin D almost completely blocked the increase in cysteine protease activity induced by anti-IgM (Fig. 6,B). Similar results were obtained with cycloheximide (data not shown), suggesting that activation of DEVD-specific caspases in response to BCR cross-linking is dependent on the induction of gene expression and protein synthesis. Because BCR-induced apoptosis also requires p38 MAPK activity, it is possible that p38 MAPK may function upstream of caspase activation. To test this hypothesis, the ability of anti-IgM to induce DEVD-specific caspase activity was examined in B104 cells pretreated with SB203580. SB203580 pretreatment significantly blocked BCR-induced caspase activity (Fig. 6 B). Neither actinomycin D nor SB203580 inhibited caspase activation induced by Fas cross-linking in BJAB cells (data not shown). Collectively, these findings establish a role for p38 MAPK in BCR-induced apoptosis both upstream and downstream of caspase activity.

In contrast to CD95/Fas, the components of the apoptotic pathway and the mechanism of activation of caspases in response to BCR ligation are not well understood. Our previous results established that activation of the SAPK and p38 MAPK pathways correlated with BCR-induced apoptosis in B104 cells (27). In this study we attempted to gain further insight into the mechanism of BCR-induced apoptosis by comparing BCR-induced apoptosis to apoptosis induced by CD95/Fas. Apoptosis stimulated by either the BCR in the human B lymphoma B104 or cross-linking Fas in the human B lymphoma BJAB is accompanied by induction of caspase activity increased exposure of phosphatidylserine in the outer leaflet of the plasma membrane and sustained activation of both SAPK and p38 MAPK. Our findings with respect to the role of caspases in Fas-induced activation of SAPK and p38 MAPK are in agreement with those previously reported (25, 26). These results support the hypothesis that caspases are critical effectors of the apoptotic response induced by both the BCR and CD95/Fas.

The precise caspase targets responsible for activation of the SAPK and p38 MAPK pathways are not known. However, it is reasonable to speculate that there might be upstream components of these protein kinase cascades that are targets for caspase-mediated proteolytic cleavage and activation. In this respect, known caspase substrates include D4-GDI, a GDP dissociation inhibitor for Rho family GTPases, DNA-dependent protein kinase, and the δ, θ, and ζ isoforms of protein kinase C (46, 47, 48, 49, 50). More recently, several protein kinases that are thought to function in mammalian MAPK pathways have been identified as caspase substrates. For example, p21-activated kinase 2 and MEKK1, have been shown to be activated by caspase-mediated proteolysis (51, 52). Additional studies will be required to determine the contributions of these proteins and other caspase targets to caspase-dependent SAPK and p38 MAPK activation during BCR- and Fas-induced apoptosis.

Our kinetic comparisons suggest that BCR-induced caspase activation is delayed relative to Fas-induced caspase activation. This delay may reflect a requirement for macromolecular synthesis to occur before apoptosis. The ability of actinomycin D or cycloheximide to inhibit BCR-induced caspase activation and apoptosis supports this hypothesis. Actinomycin D and cycloheximide promote apoptosis in many cells. For example, actinomycin D sensitizes some cells to TNF-α-mediated apoptosis by blocking the ability of the TNF receptor to induce an NFκB-mediated cell survival pathway (29, 53). However, actinomycin D has also been found to inhibit apoptosis induced by nerve growth factor withdrawal (54, 55), suggesting that the effects of actinomycin D may depend on cell type and stimulus. The ability of actinomycin D to block BCR-induced apoptosis could be mediated by either blocking the expression of factors that promote cell survival or inhibiting the induction of proapoptotic components. Candidates for such targets include transcription factors, components of the cellular protein synthesis machinery, and proteins that function in cellular apoptotic pathways. The sensitivity of immature B cells to apoptosis has recently been shown to correlate with increased expression levels of caspase 3/CPP32 (56, 57). Thus, up-regulation of the expression level of caspases themselves is one potential explanation for inhibition of BCR-induced apoptosis by actinomycin D.

SB203580, a pyridinyl imidazole inhibitor of p38 MAPK, has allowed the role of p38 MAPK in the regulation of cellular processes such as cytokine production, glucose transport, and transcriptional regulation to be determined (41, 58, 59, 60, 61). SB203580 appears to be highly specific for p38 MAPK, inhibiting the α and β isoforms of p38 MAPK, but not the more distantly related p38γ/SAPK3 or SAPK4 (42, 43, 44, 45). SB203580 significantly inhibited BCR-induced apoptosis in B104 cells. Consistent with previous reports of the effects of SB203580, the concentration required for 50% inhibition of apoptosis was ∼2 μM (41, 42). Since results obtained with SB203580 do not discriminate between events mediated directly by p38 MAPK and those mediated by downstream targets of p38 MAPK, it is possible that MAPKAPK-2 or related kinases perform critical functions during BCR-induced apoptosis. SB203580 has previously been shown to inhibit apoptosis of neural cells induced by either trophic factor withdrawal or glutamate treatment and sodium salicylate-induced apoptosis of fibroblasts (59, 60, 61). The ability of SB203580 to inhibit BCR-induced apoptosis in B104 cells, but not anti-Fas-induced apoptosis in BJAB cells, suggests that the p38 MAPK pathway plays a differential role in BCR- and Fas-induced signaling pathways (Ref. 26 and our results). In this respect, we cannot rule out the possibility that the differential effects of SB203580 are a result of cell type differences. It appears likely that the involvement of p38 MAPK in apoptosis may be dependent on both the cell type and the stimulus used. Our results, demonstrating an important role for p38 MAPK in BCR-induced apoptosis, raise a question regarding the role of SAPK. Since SAPK is insensitive to inhibition by SB203580 in vitro and in vivo (42), SAPK activation appears to be insufficient to induce apoptosis in these cells.

The identities of the substrates for p38 MAPK that function during apoptosis are unknown. However, in light of the differential sensitivity of BCR- and Fas-induced apoptosis to actinomycin D and cycloheximide, p38 MAPK substrates may function in pathways that regulate transcription and/or translation. Previous studies employing SB203580 have implicated p38 MAPK in the regulation of both transcriptional and translational events. For example, p38 MAPK has been shown to phosphorylate the transcription factors Elk-1 and ATF-2 that participate in regulation of the serum response element and activator protein-1 (62, 63, 64). The mechanism by which SB203580 blocks production of inflammatory cytokines appears to be at the level of translation rather than at that of transcription (41). The effects of SB203580 on BCR-induced apoptosis could reflect a role for p38 MAPK at either the transcriptional or the translational level. An alternative hypothesis is that components of the cellular apoptotic machinery are direct targets for kinases in the p38 MAPK pathway. Such targets could include caspases themselves, caspase regulators, or components of cell survival pathways such as members of the Bcl-2 family.

One interpretation of the effects of ZVAD-fmk and SB203580 is that p38 MAPK may function both upstream and downstream of caspases during BCR-induced apoptosis. A hypothesis to account for these data is that p38 MAPK may function as part of a positive feedback loop that serves to amplify the apoptotic response to BCR ligation (see Fig. 7). A similar model has been proposed to explain the role of MKK6b in Fas-induced apoptosis of Jurkat cells. Huang et al. (34) showed that expression of an activated mutant of MKK6b induced apoptosis and that a dominant negative MKK6b mutant inhibited Fas-induced apoptosis. However, SB203580 did not inhibit apoptosis induced by MKK6b. To account for these results, the authors suggested the existence of a positive feedback loop that comprises MKK6 but not p38 MAPK. Our data concerning BCR-mediated apoptosis differ from these findings in one important respect. The ability of ZVAD-fmk to block p38 MAPK activation and of SB203580 to inhibit the induction of caspase activity clearly implicate p38 MAPK as both a target and a positive effector of the BCR-induced apoptotic response. One explanation for these apparently paradoxical results is that MKK6 may function in both Fas- and BCR-mediated apoptotic pathways. However, in the case of Fas-induced apoptosis, MKK6 targets other than SB203580-inhibitable p38 MAPK may function to amplify the response. In contrast, BCR-induced apoptosis may depend on the activity of an MKK6 target, such as the α or β isoforms of p38 MAPK, that is sensitive to SB203580 (Fig. 7). Studies are underway to identify the signaling components that mediate caspase-dependent activation of SAPK and p38 MAPK and the critical p38 MAPK targets that function during BCR-induced apoptosis.

We thank Drs. Philip Cohen and Jeremy Saklatvala for generously providing antisera. We also thank Kate Elias for editorial assistance and our colleagues in the Krebs and Clark laboratories for helpful discussions.