Apoptosis constitutes the primary mechanism by which noncycling plasma cells are eliminated after the secretion of Ag-specific Abs in a humoral immune response. The underlying mechanism is not known. Here, we demonstrate that the expression of both TRAIL/Apo-2 ligand and the death receptors (DR) DR5 and DR4, but not Fas, are sustained in IL-6-differentiated Ig-secreting human plasma cells as well as primary mouse plasma cells generated in a T-dependent immune response. Plasma cell apoptosis is induced by both endogenous and exogenous TRAIL ex vivo, suggesting that TRAIL-mediated killing may, in part, be plasma cell autonomous. By contrast, resting and activated B cells are resistant to TRAIL killing despite comparable expression of TRAIL and DRs. The preferential killing of plasma cells by TRAIL correlates with decreased expression of CD40 and inactivation of NF-κB. These results provide the first evidence that primary plasma cells synthesize TRAIL and are direct targets of TRAIL-mediated apoptosis, which may relate to the inactivation of the NF-κB survival pathway.
During B cell terminal differentiation, Ag-activated B cells are either eliminated by apoptosis, due to low Ag affinity or self-reactivity, or differentiated to affinity-matured memory B cells or Ab-secreting plasma cells (1). Plasma cells are permanently withdrawn from the cell cycle and memory cells cycle infrequently, if at all. With the exception of long-lived plasma cells found in the bone marrow and the lamina propria of the intestine, most plasma cells are rapidly eliminated by cell death after the synthesis and secretion of large amounts of Abs (2, 3, 4). Cell death thus represents the primary mechanism that controls plasma cell homeostasis; however, the mechanisms of plasma cell death are not known.
The expansion of activated B cells is tightly regulated by survival and apoptotic signals mediated by CD40 and Fas, two members of the TNFR family. Likewise, the proliferation and survival of germinal center B cells in a T-dependent immune response critically depends on CD40 signaling through TNFR-associated factor (TRAF)3 and NF-κB (5, 6). TRAIL/Apo-2 ligand (Apo2L) is a proapoptotic member of the TNF family (7, 8), which appears to predominantly induce apoptosis of tumor cells (9, 10), including lymphoma and myeloma (malignant plasma) cell lines (7, 8, 10, 11, 12, 13). Whether TRAIL has a role in the control of B lineage cells has not been defined.
TRAIL functions to either accelerate or attenuate apoptosis depending on its interaction with five distinct receptors (DRs). After ligand binding and receptor trimerization, DR4 (TRAIL-R1) and DR5 (TRAIL-R2) recruit Fas-associated death domain-containing protein (FADD) to their cytoplasmic domains (14, 15, 16, 17). This newly formed complex initiates an apoptotic cascade through the recruitment and activation of caspase-8 (15, 16, 17) or by a caspase-independent mechanism involving the receptor-interacting protein (RIP) (18). The other known TRAIL receptors are decoy receptor (DcR)1 (TRAIL-R3) and DcR2 (TRAIL-R4) (14), which lack intact cytoplasmic death domains, and the soluble osteoprotegerin (19). These receptors sequester TRAIL from DR4 and DR5, thereby antagonizing TRAIL-mediated apoptosis (14).
The humoral immune response requires IL-6, given that IgG and IgA responses are defective in the absence of IL-6 (20, 21). In vitro, stimulation of EBV-immortalized, IgG-bearing human lymphoblastoid cells with IL-6 recapitulates all major hallmarks of B cell terminal differentiation (22, 23, 24, 25). These include prominent increases in Ig synthesis and secretion, extinction of surface MHC class II expression, and cell cycle arrest (22, 23, 24, 25). Most importantly, the IL-6-differentiated plasma cells no longer exhibit a transformed phenotype due to loss of EBV-transforming gene expression and rapidly undergo apoptosis (22), thereby mimicking short-lived plasma cells. The human lymphoblastoid cells resemble CD40-activated B cells in that the TRAF signaling pathway is constitutively activated by the EBV-encoded latent membrane protein-1 (LMP1) (26, 27). The loss of LMP1 expression in IL-6-differentiated plasma cells suggests that apoptosis of plasma cells may relate to the loss of NF-κB activation and an altered balance between survival and death signals.
Here, we show that the IL-6-differentiated human plasma cells retain the expression of TRAIL, DR4, and DR5 and rapidly undergo apoptosis mediated by endogenous and exogenous TRAIL. Induction of apoptosis by TRAIL extends to primary mouse plasma cells but not resting or activated B cells. This preferential TRAIL-mediated killing correlates with loss of CD40 expression and NF-κB activation in plasma cells. Thus, plasma cells synthesize TRAIL and are subject to TRAIL-mediated apoptosis, which may relate to the inactivation of the CD40-NF-κB pathway.
Materials and Methods
Terminal differentiation of CESS cells by IL-6 and induction of apoptosis by TRAIL
CESS is an EBV-immortalized, IgG-bearing human lymphoblastoid B cell line (22), Jurkat cells were obtained from the American Type Culture Collection, whereas the P11 human T cell line was provided by Dr. K. Elkon. CESS cells were terminally differentiated by treatment with baculovirally expressed recombinant human IL-6 and soluble IL-6 receptor (gp80) for 4 days as previously described (23). IL-6-differentiated (IgGhigh/surface MHC class IIlow) cells were separated from IL-6-refractory (IgGlow/surface MHC class IIhigh) cells using anti-MHC class II Abs conjugated to magnetic beads (Dynal, Lake Success, NY), also as previously described (23).
When indicated, cells were cultured with 20 μM benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone (ZVAD; Enzyme System Products, Livermore, CA) and 1 μg/ml recombinant soluble TRAIL/Apo2L (28), DR5-Fc (aa 52–180 of DR5 fused to the Fc portion of human IgG) (29), both generously provided by A. Ashkenazi (Genentech, South San Francisco, CA), or the Fc portion of human IgG (Rockland, Bridgeport, NJ). Recombinant TRAIL and DR5-Fc provided by Dr. Ashkenazi were used in Figs. 3 and 6 A. Otherwise, cells were incubated with recombinant soluble FLAG-tagged TRAIL/Apo2L (0.1 μg/ml) together with a FLAG enhancer Ab (1 μg/ml) and DR5-Fc (1 μg/ml) purchased from Alexis Biochemicals (San Diego, CA). Cell viability was determined by trypan blue exclusion.
In vitro terminal differentiation of primary mouse B cells
Resting splenic B cells were isolated from 5- to 8-wk-old BALB/c mice (The Jackson Laboratory, Bar Harbor, ME) and cultured with CD40 ligand (CD40L)-expressing L cells in the presence of human IL-6 (40 U/ml) and the soluble gp80 subunit of the IL-6R (40 U/ml) as previously described (30). From day 9 onward, the B cells were cocultured with osteoblastic MC3T3 cells (Deutsche Sammlungion Mikroorganismen und Zell Kulturen, purchased from the German Collection of Microorganisms and Cell Cultures, Braunschweig, Germany).
Isolation of plasma cells generated in the 4-hydroxy-3-nitrophenyl-chicken γ-globulin (NP-CGG) response and induction of apoptosis by TRAIL
Splenic B cells were isolated from B6 mice (8-wk of age) 10 days after s.c. injection with 75 μg of NP-CGG as previously described (30). To enrich for syndecan-1-positive plasma cells, viable cells were isolated by Ficoll gradient centrifugation and incubated at a concentration of 107 cells/ml with a biotinylated rat anti-mouse syndecan-1 Ab (1/500 dilution; BD Biosciences, Franklin Lake, NJ). This was followed by a second incubation with streptavidin conjugated to magnetic beads (Miltenyi Biotec, Auburn, CA). Syndecan-1-positive cells were either plated onto CD40L-expressing L cells in the absence or presence of DR5-Fc or Fc or cultured directly onto 2PK3 cells or a stable transfectant expressing membrane-bound murine TRAIL (mTRAIL) (31). Cell viability was determined by trypan blue exclusion after 5 h in culture. Resting and activated B cells present in the syndecan-1-negative fraction were separated by Percoll (Sigma-Aldrich, St. Louis, MO) gradient centrifugation, as previously described (30) and cultured onto mTRAIL or control 2PK3 cells for 24 h. Cell viability was determined by annexin V binding and trypan blue exclusion.
Immunofluorescence microscopy and flow cytometric analysis
Immunofluorescence microscopy was performed essentially as described (23). Unless indicated, all Abs were obtained from BD Biosciences. Fas expression in CESS cells was detected by incubation with an anti-Fas mouse mAb (1/100; PanVera, Madison, WI) followed by a second incubation with a FITC-conjugated rabbit anti-mouse Ab (1/200; DAKO, Carpinteria, CA). Intracellular IgG was detected with a Rhodamine-conjugated goat anti-human IgG Ab (1/400; Cappel, Bryan, OH). Cells were counterstained with 17 mM 4′,6′-diamino-2-phenylindole dihydrochloride and visualized by fluorescence microscopy.
Flow cytometric analysis was performed as previously described (30). CESS cells were stained with the above mentioned anti-Fas, or CD40 Abs (1/100), and were revealed by a secondary incubation with a FITC-conjugated rabbit anti-mouse Ab (1/200; DAKO). In vitro-differentiated primary mouse B cells were stained with a PE-conjugated hamster anti-mouse FasAb (1/100) or a FITC-conjugated rat anti-mouse CD40 Ab (1/100). For three-color staining shown in Fig. 7 A, syndecan-negative and -positive cells were stained with a PE-conjugated rat anti-mouse syndecan-1 Ab (1/200), a FITC-conjugated rat anti-mouse CD40 Ab (1/200) and a Cy-Chrome conjugated rat anti-mouse B220 Ab (BD PharMingen, San Diego, CA) (1/400). Cells were analyzed using a BD Biosciences FACSCalibur.
ELISA and ELISPOT
Detection of human IgG secreted by IL-6-differentiated CESS cells by ELISA was performed in 96-well plates coated with 1 μg of a rabbit anti-human IgG polyclonal Ab (Cappel) (30). To detect NP-specific IgG secreted by primary plasma cells, wells were coated with 2.5 μg 4-hydroxy-3-iodo-5-nitrophenylacetyl succinimide ester conjugated to BSA. Bound Ab was revealed by incubation with biotinylated goat anti-human IgG (1/20,000; Jackson ImmunoResearch Laboratories, West Grove, PA), or biotinylated goat anti-mouse IgG (1/30,000; Sigma-Aldrich) followed by a second incubation with HRP-conjugated streptavidin (1/2000; Vector, Burlingame, CA).
A secondary immune response was elicited 5 wk after the primary response by injecting tail veins with 15 μg NP-CGG in PBS. Splenic B cells were isolated 6 days later, cultured for 4 h in medium alone, with parental 2PK3 cells or mTRAIL-expressing 2PK3 cells at a 10:1 ratio. An ELISPOT was then performed in 96-well plates coated with 1 μg of a goat anti-mouse κ Ab (Southern Biotechnology Associates, Birmingham, AL) or 2.5 μg 4-hydroxy-3-iodo-5-nitrophenylacetyl succinimide ester conjugated to BSA (30).
EMSA was performed to analyze the NF-κB DNA binding activity using 2.6 μg of whole cell lysate and a [32P]dATP-labeled H2K site as previously described (30), or to analyze the Oct-1 DNA binding activity using a [32P]dATP-labeled Oct-1 probe (H2B) (5′-GATCCCAACTCTTCACCTTATTTGCATAAGCGATTCTATAG). In competition assay, an unlabeled double-stranded oligonucleotide containing the Oct site of the Eμ enhancer (5′-AATTCACCCTGTCTCATGAATATGCAAATCAGGTGAGTCTATG-3′) was used at a 280-fold molar excess.
Whole cell lysates were prepared by incubating cells in lysis buffer (250 mM NaCl, 50 mM HEPES, pH 7; 0.1% Nonidet P-40) on ice for 10 min, supernatants were clarified by centrifugation, and 10–30 μg of total protein was resolved on SDS-PAGE and transferred onto polyvinylidine difluoride membranes. Membranes were blocked in TBS-T (10 mM Tris (pH 8), 150 mM NaCl, 0.1% Tween 20) containing 5% powdered milk for 1 h followed by a 3-h incubation with one of the following Abs: mouse anti-human caspase-7 (1-1-11; 1/1000), mouse anti-human caspase-8 (1-1-40; (1/1000), rabbit anti-human caspase-3 (585R; 1/1000) (32), all provided by Dr. Y. Lazbenik; rabbit anti-human DR4 (66901N; BD Transduction Laboratories, Lexington, KY; 1/1000), rabbit anti-human DR5 (AAP-430) (1/1000; Stressgen, Vancouver, Canada), rabbit anti-human A1 (FL-175; 1/1000; Santa Cruz Biotechnology, Santa Cruz, CA), hamster anti-mouse Bcl-2 (554218; 1/1000), or mouse anti-chick α-tubulin (T9026) (1/5000; Sigma-Aldrich). Membranes were rinsed in TBS-T and incubated with biotinylated donkey anti-mouse, donkey anti-hamster, or donkey anti-rabbit Abs (1/20,000; Jackson ImmunoResearch Laboratories) for 45 min followed by a second 30-min incubation in streptavidin-HRP (1/20,000; Jackson ImmunoResearch Laboratories). To detect IgG, membranes were probed directly with a biotinylated goat anti-human IgG Ab (1/20,000) or a biotinylated goat anti-mouse IgG Ab (1/20,000), followed by a streptavidin-HRP incubation. The immunoreactive proteins were detected by chemiluminescence (ECL; Amersham, Arlington Heights, IL). For sequential blotting, membranes were stripped in 2% SDS, 62.5 mM Tris, pH 6.8, at 50°C for 20 min.
RT-PCR and RNase protection assay
CESS cell RNA was isolated by the guanidine isothiocyanate procedure as previously described (25). Total RNA from primary B lineage cells were isolated using Trizol (Life Technologies, Gaithersburg, MD). For RT-PCR analysis, reverse transcription was performed on total RNA from 3 × 105 CESS cells or 1 μg of primary cell RNA using 1 μg oligo(dT) primer and Superscript II reverse transcriptase (Life Technologies). For the PCR analysis, 1 μl of cDNA was used with 10 pmol of each of the following primers: Fas ligand (FasL), 5′-TCAGCTCTTCCACCTACAGAA-3′ and 5′-TACAACATTCTCGGTGCCTG-3′; GAPDH, 5′-CCACCCATGGCAAATTCCATGGCA-3′ and 5′-TCTAGACGGCAGGTCAGGTCCACC-3′; mTRAIL, 5′-GGTCTCAAAGGACAAGGTG-3′ and 5′-TTAGTTAATTAAAAAGGCTCC-3′; mouse DR5, 5′-GTCAAAGCCGAAACACTGG-3′ and 5′-TCAAACGCACTGAGATCC-3′; mouse actin, 5′-AAGATCCTGACCGAGCGTGGC-3′ and 5′-CTGGAAGGTGGACAGTGAGGC-3′. DNA was amplified using the following PCR conditions: FasL, 35 cycles (30 s at 94°C; 30 s at 55°C; 45 s at 72°C); GAPDH, 20 cycles (1 min at 94°C; 1 min at 62°C; 1 min at 72°C), mTRAIL: 40 cycles (1 min at 94°C; 1 min at 52°C; 2 min at 72°C); mDR5, 40 cycles (1 min at 94°C; 1 min at 52°C; 2 min at 72°C) and mouse actin, 30 cycles (1 min at 94°C; 1 min at 68°C; 2 min at 72°C) The PCR products were electrophoresed on 2% agarose gels and stained with ethidium bromide.
For the RNase protection analysis, antisense RNA was generated from the human hAPO-3c multiprobe template set (BD PharMingen) as described (33). Total RNA (5 μg) was hybridized with 1 × 106 cpm of the radiolabeled probe, subjected to RNase protection analysis, and electrophoresed on 6% denaturing acrylamide gels containing urea (33).
Decreased Fas expression on IL-6-differentiated human plasma cells in vitro
IL-6-differentiated human IgG plasma cells rapidly undergo apoptosis as evidenced by annexin V binding and poly(ADP)ribose polymerase cleavage (22, 23). These plasma cells (IgGhigh), which no longer possess a transformed phenotype due to reversal of EBV immortalization (22), can be enriched to >80% purity by negative selection for the loss of MHC class II expression (23). To verify the involvement of caspases in plasma cell death, the IL-6-differentiated plasma cells and control CESS cells were cultured in the presence or absence of the general caspase inhibitor, ZVAD (Fig. 1 A). Although ZVAD did not alter the viability of control cells, it reduced the death of IL-6-differentiated plasma cells by 60% in 24 h. Plasma cell apoptosis is therefore primarily caspase dependent.
The role of Fas in plasma cell apoptosis was then addressed. Flow cytometric analysis showed that Fas was highly expressed on control CESS cells but substantially reduced, although not eliminated, on IL-6-differentiated plasma cells (Fig. 1,B). Confirming this observation, immunofluorescent staining revealed a prominent reduction of Fas on plasma cells expressing high levels of intracellular IgG (Fig. 1,C). Moreover, FasL mRNA was undetectable in CESS cells before or after IL-6 differentiation by either RT-PCR or RNase protection assays (Figs. 1 D and 2A). Thus, IL-6-differentiated plasma cells express reduced level of Fas and no FasL. The Fas pathway is unlikely to be instrumental in the death of these clonal plasma cells in vitro.
Expression of TRAIL, DR4 and DR5 on IL-6-differentiated human plasma cells
Apoptosis of IL-6-differentiated plasma cells in the absence of FasL prompted us to investigate the role of TRAIL in plasma cell death. RNase protection assays revealed that mRNAs encoding TRAIL, DR4, and DR5 were expressed in IL-6-differentiated plasma cells (IgGhigh), at levels comparable with those observed in IL-6-refractory (IgGlow) and untreated CESS cells (Fig. 2,A). The Fas mRNA levels were similarly maintained in IL-6-differentiated plasma cells (Fig. 2,A), implying that reduction of Fas protein (Fig. 1) might occur at the posttranscriptional level. The maintenance of DR4 and DR5 expression was confirmed at the protein level by immunoblot analysis (Fig. 2,B), where detection of DR5 but not DR4 in Jurkat T cells (15) served as a control for the Ab specificity. The purity of the plasma cell population was confirmed through the detection of high IgG levels, and tubulin expression controlled for protein loading (Fig. 2,B). There was no evidence for the presence of DR3 or TNFR p55 mRNAs in plasma cells (Fig. 2 A). The expression of TRAIL, DR4, and DR5 is therefore retained during IL-6 differentiation of human B lymphoblastoid cells to plasma cells. The reduction of surface Fas expression and the absence of FasL indicate that Fas is unlikely to be instrumental in the apoptosis of IL-6-differentiated plasma cells. The sustained expression of TRAIL and its receptors in plasma cells is consistent with a potential role for TRAIL in inducing plasma cell apoptosis in vitro.
Endogenous and exogenous TRAIL induces apoptosis of IL-6-differentiated human plasma cells
The function of TRAIL in plasma cell apoptosis was characterized, first in IL-6-differentiated human IgG-secreting plasma cells due to the relative ease in enriching this plasma cell population. To determine whether endogenous TRAIL induces plasma cell death, IL-6-differentiated plasma cells were incubated with DR5-Fc, an inhibitory chimeric soluble TRAIL receptor comprised of the extracellular domain of DR5 fused to the Fc portion of human IgG (29) (Fig. 3). Indeed, DR5-Fc inhibited plasma cell death by 40% within 24 h, as assessed by trypan blue staining (Fig. 3, top). This correlated with a 4-fold increase in the relative amounts of IgG secreted into the medium during the 24-h period, as measured by ELISA (Fig. 3, bottom). Inhibition of cell death by DR5-Fc was specific to the DR5 extracellular domain, because the Fc portion of human IgG had no effect (Fig. 3). Thus, endogenous TRAIL functions to induce plasma cell death. The addition of soluble, trimerized recombinant human TRAIL further enhanced plasma cell death, whereas the presence of both DR5-Fc and TRAIL led to a marked reduction of plasma cell death and a corresponding increase in IgG secretion (Fig. 3). These results suggest that plasma cell death is induced by endogenous TRAIL and augmented by exogenous TRAIL.
Induction of cell death by TRAIL, however, did not extend to the control CESS lymphoblastoid cells (Fig. 3). They were intrinsically less apoptotic than IL-6-differentiated plasma cells, and were refractory to TRAIL-mediated killing at a concentration that enhanced plasma cell death (Fig. 3). TRAIL therefore preferentially induces the death of IL6-differentiated plasma cells.
TRAIL induces apoptosis of primary plasma cells generated in T-dependent immune responses
The susceptibility of IL-6-differentiated human plasma cells to TRAIL killing prompted us to address whether primary plasma cells are also direct targets of TRAIL-mediated apoptosis (Fig. 4). Primary mouse plasma cells were generated in vivo by immunization with NP-CGG, a T cell-dependent Ag and enriched to >70% homogeneity by selecting syndecan-1-positive cells from splenic B cells isolated on day 10 of immunization. As controls, resting and activated B cells present in the syndecan-1-negative fraction were subsequently separated by Percoll gradient centrifugation (30). TRAIL and DR5 (there is no DR4 in mice) mRNAs were expressed at comparable levels in primary resting, activated, and plasma cells, as indicated by RT-PCR analysis (Fig. 4 A).
Having verified the expression of TRAIL ligand and receptor in primary plasma cells, we then determined whether they were subject to TRAIL killing, by incubation with 2PK3 cells stably expressing the trimerized mTRAIL or the parental cell line as a negative control (31). Plasma cell death was enhanced 2-fold within 5 h of coincubation with mTRAIL (p = 0.03), but not with the control cells (Fig. 4,B, left panel). Conversely, it was reduced by 30% by incubation with DR5-Fc (p = 0.02), but not with Fc (Fig. 4,B, middle panel). The reduction of plasma cell death by DR5-Fc was corroborated by a 2-fold increase (p = 0.03) in the amount of NP-specific IgG Ab secreted in the medium (Fig. 4 B, right panel). Thus, endogenous and exogenous TRAIL cooperate to induce the death of primary plasma cells generated in a primary T cell-dependent response.
To further confirm that TRAIL induces apoptosis of primary plasma cells generated in vivo, the loss of plasma cells secreting NP-specific IgG in response to TRAIL was assayed functionally by ELISPOT (Fig. 4,C). NP-specific plasma cells were first amplified in vivo in a secondary response elicited 5 wk after primary NP-CGG immunization. Splenic B cells were isolated at the peak of the secondary Ab response (day 6) and directly cultured, without selection for syndecan-1-positive cells, for 4 h in medium, alone or together with mTRAIL or the control 2PK3 cells (Fig. 4 C). mTRAIL, but not the control cells, induced a 2-fold reduction of the number of NP-specific IgG plasma cells (Ab-forming cell) (p = 0.006) as well as the polyclonal IgGκ-secreting plasma cells (p = 0.03). Thus, TRAIL directly induces the death of Ag-specific and polyclonal, class-switched, primary plasma cells generated in a T-dependent immune response.
Incubation with mTRAIL-expressing cells ex vivo, however, did not enhance apoptosis of resting or activated splenic B cells isolated from a primary NP response, as determined by either the annexin V-binding activity of early apoptotic cells or trypan blue staining of dead cells (Fig. 4,D). Primary resting and activated mouse B cells are therefore refractory to TRAIL-mediated killing ex vivo, in agreement with our observation during IL-6 differentiation of human lymphoblastoid cells (Fig. 3). TRAIL therefore preferentially induces the death of primary plasma cells but not resting or activated B cells in a T-dependent immune response.
TRAIL-mediated apoptosis of IL-6-differentiated human plasma cells is primarily caspase dependent
TRAIL has been shown to initiate the caspase cascade by recruitment and activation of caspase-8 (15, 16, 17). Spontaneous and TRAIL-induced death of IL-6-differentiated plasma cells are inhibited by ZVAD (Figs. 1,A and 5A), suggesting that the TRAIL death signals are mediated by caspases. Consistent with this possibility, both isoforms of procaspase-8 (55 and 53 kDa) as well as procaspases-3 and -7 were abundantly expressed in IL-6-differentiated IgG plasma cells as determined by immunoblotting (Fig. 5,B). The initiator caspase-8 was activated within 1 h of TRAIL stimulation, as indicated by the emergence of its cleavage products. This was rapidly followed by the activation of the effector caspases-7 and -3, also in agreement with observations in other cell types (15, 16, 17, 34). Although the cleavage of all three caspases in response to TRAIL was inefficient in IL-6-differentiated plasma cells (Fig. 5,B), the levels of cleavage were comparable with those observed by others in a subset of cell lines including human B cell lines and colon and lung carcinomas (16, 17). Moreover, activation of these three caspases was also very modest after prolonged (7 h) Fas cross-linking, in striking contrast to their efficient activation in the control lymphoblastoid cells under the same Fas cross-linking conditions (Fig. 5 C and data not shown). Activation of caspases-8, -3, and -7 is therefore inefficient in plasma cells by either TRAIL or Fas.
The reduction of TRAIL-induced death by ZVAD and the activation of the caspase machinery in response to TRAIL demonstrate that TRAIL-mediated plasma cell apoptosis in vitro is primarily caspase dependent (Fig. 5). However, protection from apoptosis by ZVAD was incomplete, implying that a caspase-independent mechanism may function in concert with the caspase-dependent pathway to induce plasma cell apoptosis in response to TRAIL.
Decreased CD40 expression and inactivation of NF-κB in IL-6-differentiated human plasma cells
Our data point to preferential killing of plasma cells, but not B cells before terminal differentiation, by TRAIL despite comparable levels of expression of both the ligand and receptors. This raises the possibility that either functional TRAIL/DR complexes cannot form on activated (lymphoblastoid) cells, or that plasma cells have specifically lost the ability to respond to one or more survival signals. To address the latter possibility, we investigated the regulation of the TRAF/NF-κB survival pathway, which is constitutive in EBV-transformed lymphoblastoid cells (26, 27) but may no longer operate in IL-6-differentiated plasma cells (22) due to loss of LMP1 expression.
Electrophoretic mobility shift assays showed that NF-κB was highly activated in control and IL-6-refractory CESS cells, as evidenced by the formation of NF-κB-DNA complexes consisting of p50/p65, p50/RelB, and p50/p50 (Fig. 6,A, lanes 1 and 2). However, the NF-κB DNA-binding activity was drastically reduced in IL-6-differentiated plasma cells (Fig. 6,A, lane 3). Although TRAIL stimulation can activate NF-κB in other cells (35, 36, 37, 38), it did not enhance the NF-κB DNA-binding activity in either IL-6-differentiated plasma cells or refractory cells (Fig. 6,A, lanes 4 and 5). The loss of NF-κB activity could not be attributed to an overall reduction in DNA-binding activity of cellular transcription factors, because the Oct-1 DNA binding activity was comparable in each lysate (Fig. 6,B). Moreover, coordinated with the loss of LMP1 expression (22), the expression of CD40, a potent activator of NF-κB in B lineage cells (5, 6), was also profoundly reduced in IL-6-differentiated plasma cells (Fig. 6 C). This correlated with the failure of CD40L to delay or rescue apoptosis of these plasma cells (data not shown). Thus, terminal differentiation of lymphoblastoid cells by IL-6 correlated with coordinated loss of CD40 expression and NF-κB activation.
Reduction of CD40 and Fas expression in primary mouse plasma cells
Spontaneous apoptosis ex vivo of primary plasma cells generated in the NP response also could not be delayed or prevented by stimulation with CD40L (Fig. 4,B), suggesting that the expression of CD40 might be similarly reduced in primary plasma cells. Indeed, the expression of CD40 on primary plasma cells (Synhigh/B220low) was significantly lower than that on B cells (Synlow/B220high) from the same immune response (Fig. 7,A). To address the possibility that the expression of CD40 might be temporally regulated during B cell terminal differentiation, primary mouse plasma cells were generated in vitro by sequential coculture of resting splenic B cells with CD40L-expressing L cells and then MC3T3 osteoblastic cells in the presence of IL-6. Under these conditions, the majority of B cells (85%) entered the cell cycle by day 3 based on 5-bromo-2′-deoxyuridine uptake and continued to proliferate until around day 7, when cell cycle arrest began. From day 9 onward, coculture with MC3T3 cells facilitated terminal differentiation, leading to plasma cells that had lost B220 and MHC class II, expressed syndecan-1 and secreted IgM (W. Zhang and S. Chen-Kiang, unpublished observations). The expression of CD40 was maintained on activated B cells (day 4), reduced as cells withdrew from the cell cycle (day 7), and further decreased in plasma cells (day 11) to levels substantially below that of resting B cells (Fig. 7 B). Together, these in vivo and in vitro results demonstrate that CD40 expression is progressively reduced during terminal differentiation of primary B cells.
The reduction of CD40 expression suggests that the NF-κB survival pathway may be coordinately inactivated in primary plasma cells. To address this possibility, we determined the expression of A1 in activated B cells and plasma cells as a functional readout for NF-κB activity because A1 is a target for both NF-κB and CD40 (39, 40). The A1 protein level was markedly lower in plasma cells (Syn+) compared with activated B cells (Syn−) generated in the primary NP response, whereas Bcl-2 and tubulin levels remained unchanged (Fig. 7 C). These results confirmed that CD40 and a NF-κB target gene, A1, are selectively and coordinately reduced in primary plasma cells.
Surface Fas expression was prominently elevated on CD40L-activated B cells (Fig. 7,B), in agreement with previous reports (41). Of interest, Fas expression was also drastically reduced on plasma cells (day 11), to levels characteristic of resting B cells (Fig. 7,B). Fas and CD40 are therefore coordinately regulated during terminal differentiation of primary B cells initiated by CD40 signaling. Together with the reduction of Fas on IL-6-differentiated human plasma cells (Fig. 1 B), the loss of surface Fas expression appears to be common in plasma cells.
In this study, we demonstrate that primary plasma cells are susceptible to killing mediated by endogenous and exogenous TRAIL and that TRAIL-mediated apoptosis does not extend to resting or activated primary B cells despite comparable levels of ligand and death receptor expression. This sensitivity of plasma cells to TRAIL-mediated apoptosis may relate to reduced CD40 expression and NF-κB survival signals.
Induction of plasma cell apoptosis by TRAIL
The death of plasma cells is tightly coordinated with cell cycle arrest and cellular differentiation to ensure their rapid elimination at the end of a humoral immune response (2, 3, 22, 23). Here, we provide the first direct evidence that TRAIL mediates spontaneous and accelerated plasma cell death in two independent model systems: primary plasma cells generated in a T-dependent immune response; and human IgG plasma cells differentiated in vitro by IL-6. First, we showed that the expression of TRAIL and the DR death receptors were maintained in Ag-specific and polyclonal mouse plasma cells generated in the 4-hydroxy-3-nitrophenyl (NP) response (Fig. 4,A) as well as IL-6-differentiated human plasma cells (Fig. 2). Next, we demonstrated that both plasma cells were susceptible to killing mediated by endogenous and exogenous TRAIL ex vivo (Figs. 3 and 4). In light of the lack of understanding of TRAIL function in primary cells, lymphocytes in particular, these findings have significant implications for the mechanisms that control primary plasma cell apoptosis.
The susceptibility of plasma cells to TRAIL-mediated killing is in part determined by temporal changes in the composition of death receptors during B cell terminal differentiation. Contrasting the sustained expression of DR, Fas protein expression was drastically reduced in IL-6-differentiated human plasma cells, and in primary mouse plasma cells (Figs. 1 and 7,B). FasL expression was absent in the former (Figs. 1,D and 2A). Although cross-linking of Fas can be facilitated by an extracellular domain in the absence of ligand (42), the low level of Fas expressed on plasma cells would likely preclude efficient ligand-independent Fas oligomerization. Fas signaling is therefore not responsible for the rapid apoptosis of IL-6-differentiated plasma cells in vitro (Figs. 3 and 4) and unlikely to be the primary mechanism that controls the death of primary plasma cells.
The death of plasma cells, either spontaneous or induced by exogenous TRAIL, is caspase dependent based on its inhibition by ZVAD (Fig. 5). The extent to which TRAIL activates the caspase machinery has been shown to vary greatly according to cell types (15, 16, 17). In some cell lines, including B cell lines, the levels of caspase cleavage (15, 16, 17) are comparable with our findings in plasma cells (Fig. 5). Despite the inefficiency of caspase cleavage by TRAIL, which may be inherent to specific cell types, the ability of ZVAD to block the majority of TRAIL-induced killing suggests that plasma cells respond to TRAIL death signals primarily through caspase-dependent mechanisms. However, this does not preclude the involvement of a caspase-independent pathway that may function in concert with activation of the caspase-8 pathway to promote plasma cell death. This possibility is consistent with the finding that TRAIL can also activate a caspase-independent pathway through the RIP serine/threonine kinase to promote cell death (18). In addition, plasma death is unlikely to be induced exclusively by TRAIL, because neutralization of TRAIL killing by DR5-Fc significantly reduced, but did not block, plasma cell death (Figs. 3 and 4). Further studies are required to determine the relative contributions of TRAIL and additional death signals to plasma cell death and the intracellular pathways that mediate TRAIL-mediated plasma cell death.
TRAIL-mediated apoptosis is coincident with inactivation of the CD40-NF-kB signaling pathway
TRAIL induces apoptosis of terminally differentiated plasma cells, but not resting or activated primary mouse B cells or human lymphoblastoid cells (Figs. 3 and 4). Because the expression of key components of the TRAIL death pathway (TRAIL, DR4, and DR5) does not vary during B cell terminal differentiation (Figs. 2 and 4), this differential sensitivity of TRAIL-mediated killing must be determined by other factors. The most likely possibilities are regulated assembly of the TRAIL death-inducing signaling complex and altered balance between intracellular survival and apoptotic signals.
Inactivation of the CD40-NF-κB pathway during B cell terminal differentiation may contribute to the selective killing of plasma cells by TRAIL (Figs. 6 and 7). Although CD40 expression is believed to be constitutive in mature B cells, our results revealed that it is in fact markedly reduced in primary mouse plasma cells generated in vivo and in vitro as well as IL-6 differentiated human plasma cells (Figs. 6 and 7). This correlates with the failure of CD40L to protect plasma cells from apoptosis mediated by endogenous and exogenous TRAIL (Fig. 4,B; J. Ursini-Siegel, W. Zhang and S. Chen-Kiang, unpublished observations). In the case of human lymphoblastoid cells, the extinction of LMP1 expression during terminal differentiation by IL-6 further ensures that the NF-κB pathway no longer functions in plasma cells (22). Confirming this prediction, the p50/p50, p50/RelB, and p50/p65 NF-κB DNA-binding activities were drastically reduced in IL-6-differentiated plasma cells and could not be restored by TRAIL (Fig. 6).
p50 and p65 are essential for the survival of primary lymphocytes (43, 44) and highly activated in freshly isolated primary resting and activated B cells (30). NF-κB may attenuate TRAIL-mediated killing through activation of specific Bcl-2 family proteins, based on the inverse correlation between sensitivity to TRAIL killing and expression of a specific NF-κB target gene A1 in primary plasma cells and activated B cells (Fig. 7). In addition to modulating the intracellular balance between survival and death signals, NF-κB may also impair the assembly and signaling of an active TRAIL/death-inducing signaling complex through induction of DcR1 and c-FLIP expression (45, 46, 47). It would be of interest to determine the expression of DcR1 and c-FLIP in plasma cells and their roles in TRAIL-mediated apoptosis.
In the context of a humoral immune response, CD40 has dual functions in B cell activation and survival, and NF-κB is essential for the transcription of Ig genes (48). Moreover, CD40 engagement of germinal center B cells induces memory B cell formation but inhibits their terminal differentiation into Ab-secreting plasma cells (1, 49, 50), and removal of CD40 signals allowed germinal center B cells to undergo proper terminal differentiation (49). Thus, reduction of CD40 expression and inactivation of NF-κB during B cell terminal differentiation must be exquisitely regulated. Consistent with this possibility, we show that reduction of CD40 expression on activated B cells is coincidental with the onset of cell cycle arrest and continues progressively during subsequent differentiation to plasma cells in vitro (Fig. 7). On this basis, we suggest that the reduction of CD40 expression promotes not only B cell terminal differentiation but also the decline of CD40-activated NF-κB activity, at a point when plasma cells have accumulated and secreted sufficient amounts of Ig, may in fact signal the end of their life span.
Conditional induction of apoptosis of tumor and primary cells by TRAIL
TRAIL is thought to preferentially kill transformed cells of various lineages in vitro, including B and T cell lymphomas and multiple myeloma cells (7, 8, 10, 11, 12). Induction of apoptosis by the administration of TRAIL in vivo has been shown to reduce tumor incidence in experimental mouse models of colon carcinoma (9) and mammary adenocarcinoma (10). Conversely, neutralization of endogenous TRAIL signals led to accelerated liver metastases (51). The protection of HL60 cells from TRAIL-mediated killing by p65 activation (52) and sensitization of myeloma cells to TRAIL killing by inactivation of NF-κB (13, 53), however, suggest that susceptibility of tumor cells to TRAIL-mediated apoptosis may also be determined by the balance between apoptosis and survival signals mediated by NF-κB.
With the exception of human astrocytes and prostate epithelial cells, most human primary cells are highly resistant to TRAIL-mediated apoptosis (9, 10, 54). However, they can be sensitized to TRAIL-mediated killing by CD3 ligation in the case of primary human thymocytes (55) and by inhibition of protein synthesis of primary human thyroid follicular cells and keratinocytes (56, 57). Our finding that primary resting and activated B cells are refractory to TRAIL killing does not preclude the possibility that additional signals may render them sensitive to TRAIL killing. Elucidating the pathways that mediate and modulate TRAIL-induced apoptosis in primary B lineage cells should help to better understand the control of TRAIL-mediated apoptosis of both normal and malignant plasma cells.
We thank Dr. Michelle Tourigny for contributing to the FACS analysis of CD40 expression in B cells in the NP response; Dr. Yuri Lazebnik for providing Abs to caspases and helpful suggestions; Dr. Avi Ashkenazi (Genentech) for providing DR-Fc and TRAIL used in experiments indicated; Drs. Beatrice Knudsen, Pengbo Zhou, Marcel van den Brink, and Cornelius Schmaltz for stimulating discussions; and Dr. Lee Kiang for critical reading of the manuscript.
This work was supported by a Postdoctoral Fellowship from the Lymphoma Research Foundation of America (to J.U.-S.), Cornell-Rockefeller University-Sloan-Kettering Institute Tri-Institutional National Institutes of Health Medical Scientist Training Program Grant GM07739 (to R.K.G.D.), research grants from the National Institute of Arthritis and Musculoskeletal and Skin Diseases (AR44580) and the National Cancer Institute (CA80204), and a Specialized Center of Research for Multiple Myeloma grant from the Leukemia and Lymphoma Society (to S.C.-K.).
Abbreviations used in this paper: TRAF, TNFR-associated factor; Apo2L, Apo-2 ligand; DR, death receptor; FADD, Fas-associated death domain-containing protein; DcR, decoy receptor; LMP1, latent membrane protein-1; CD40L, CD40 ligand; mTRAIL, membrane-bound murine TRAIL; NP, 4-hydroxy-3-nitrophenyl; NP-CGG, NP-chicken γ-globulin; RIP, receptor-interacting protein; ZVAD, benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone; FasL, Fas ligand.