Activation of an Unfolded Protein Response during Differentiation of Antibody-secreting B Cells*

The unfolded proteinresponse pathway (UPR) is believed to detect and compensate for excessive protein accumulation in the endoplasmic reticulum (ER). The UPR can be induced by pharmacological agents that perturb ER functions, but may also occur during cellular developmental processes such as the transition of B-lymphocytes into antibody-secreting plasma cells. Here we show that major UPR components are activated in B cells stimulated to secrete antibody. Increased expression of UPR targets including the ER chaperones BiP and GRP94 and the transcription factor XBP-1 initiates early in the differentiation program prior to up-regulated synthesis of Ig chains. Furthermore, these same kinetics are observed during differentiation for cleavage of the ER-localized ATF6α protein and splicing of XBP-1 mRNA to generate p50ATF6α and p54XBP-1, the two known UPR transcriptional activators. All of these UPR events reach maximal levels once Ig synthesis and secretion are markedly induced. Interestingly, these events are not accompanied by expression of CHOP, a transcription factor induced by ER stress agents commonly used to investigate the UPR. These results suggest that a physiological UPR elicited during differentiation of B-lymphocytes into high-rate secretory cells may be distinct from the UPR defined by agents that disrupt protein maturation in the ER.

The unfolded protein response (UPR) 1 is mediated by a multifaceted intracellular signaling pathway triggered by inhibition of glycosylation, Ca 2ϩ depletion, and other stress conditions that interfere with protein folding in the endoplasmic reticulum (ER) (1,2). As unfolded proteins accumulate, the UPR coordinates a broad down-regulation of protein synthesis with increased expression of various gene products including ER resident molecular chaperones that promote protein folding (1). In addition, the UPR can lead to arrest of cell growth (3) and, under conditions of chronic ER stress, culminate in apoptosis (4). Thus, the UPR pathway appears to monitor the protein folding capacity of the ER and transmit that information to mechanisms that can modulate the ER environment, regulate various aspects of cellular metabolism, and even influence cell fate.
Elucidating the UPR has largely relied on the use of pharmacological agents that disrupt protein folding and assembly in the ER, such as tunicamycin, an inhibitor of N-linked glycosylation, and thapsigargin, an inhibitor of the ER Ca 2ϩ -ATPase (5). When the mammalian ER is pharmacologically stressed, a set of ER transmembrane proteins initiates the UPR. One of these is ATF6␣, a 90-kDa type II transmembrane protein that undergoes ER stress-induced proteolysis to liberate its 50-kDa cytosolic domain (p50ATF6␣), a basic leucine zipper transcription factor (6). p50ATF6␣ translocates to the nucleus and participates in transcriptional induction of genes including BiP and GRP94, two ER resident molecular chaperones; XBP-1 (X-box binding protein 1), a basic leucine zipper transcription factor; and CHOP (C/EBP homologous protein (7), also known as GADD153 (8)), a member of the CAAT/enhancer-binding protein family of transcription factors (9 -12). CHOP has been linked to apoptosis of cells challenged with chronic ER stress conditions (13,14). The promoters of these genes all contain one or more copies of the cis-acting ER stress response element (ERSE) through which p50ATF6␣ exerts its activity (9 -12).
In parallel with ATF6, ER stress-inducing agents activate the IRE1␣/␤ proteins (15,16). IRE1␣ is a ubiquitously expressed ER type I transmembrane protein containing both a serine/threonine kinase module and an endoribonuclease domain in its cytosolic region (15). Upon UPR activation, IRE1␣ executes site-specific cleavage of XBP-1 mRNA to remove a 26-nucleotide intron. Religation of the 5Ј and 3Ј fragments yields a spliced XBP-1 mRNA with an altered reading frame encoding a 54-kDa basic leucine zipper transcription factor, p54XBP-1 (17)(18)(19). p54XBP-1 is more potent as a transcriptional activator and more stable than the 30-kDa protein translated from unprocessed XBP-1 mRNA (18,19). Like p50ATF6␣, p54XBP-1 can bind ERSEs and may also act through a second cis-acting element to regulate promoter activity (18). Therefore, it appears that ATF6 and IRE1␣ work together to regulate expression of XBP-1, ultimately generating a transcriptional activator that amplifies the UPR.
While primarily studied as a stress response triggered by pharmacological agents, the UPR may also take place in normal cellular developmental processes that increase the demand on the protein folding capacity of the ER. A classic example of increasing demands on the exocytic pathway is observed within the B-lymphocyte as it differentiates into a plasma cell that secretes thousands of antibody molecules each second (20). This impressive output is dependent on the proper folding and assembly of copious amounts of secretory immunoglobulin (Ig) heavy and light chains in the ER. Recent studies have, in fact, uncovered a potential link between the UPR and plasma cell differentiation as XBP-1, a gene regulated within the UPR (17)(18)(19), was shown to be required for normal antibody production in vivo (21). In addition, p54XBP-1 was shown to be expressed by splenocytes stimulated in vitro with lipopolysaccharide (LPS) (19), a stimulant that activates both proliferation and differentiation of B cells (22). The synthesis of various ER proteins, including BiP and GRP94, has also been shown to be elevated in LPS-stimulated B cells (23,24). Whether the apparent UPR elicited in B cells responding to LPS is a common feature of terminal B cell differentiation, is related to increased Ig synthesis in differentiating B cells, or is equivalent to the UPR triggered by pharmacological agents all remain unclear. To address these questions, we investigated the activation of the known UPR signaling pathway during the synchronous differentiation of CH12 B cells into cells that morphologically and functionally resemble antibody-secreting plasma cells (23).
Here we report that synthesis of p54XBP-1 occurs in CH12 B cells differentiating into antibody-secreting cells in response to multiple stimuli. Interestingly, increased expression of the ER chaperones BiP and GRP94 and the induction of the transcription factor p54XBP-1 initiate prior to increased synthesis of nascent Ig heavy and light chains. We further demonstrate that the p50ATF6␣ transcription factor is generated in differentiating B cells and is also observed prior to induction of Ig synthesis. By contrast, these UPR events in differentiating B cells are not accompanied by induction of CHOP. Thus, the UPR elicited during differentiation of antibody-secreting B cells may be distinct from the UPR triggered by agents that grossly perturb normal ER functions. We propose that certain components of the known UPR signaling pathway play a critical role in mediating ER homeostasis as B cells terminally differentiate into dedicated secretory cells.

EXPERIMENTAL PROCEDURES
Cell Culture-The CH12 B cell lymphoma cell line was provided by Dr. Troy Randall (Trudeau Institute, Saranac Lake, NY) and maintained by weekly passage as an ascites tumor in B10.A mice (Jackson Laboratory, Bar Harbor, ME) (25). Cells were obtained by peritoneal lavage and immediately placed in culture in RPMI 1640 supplemented with minimal essential medium vitamins, minimal essential and nonessential amino acids, glucose, glutamine, penicillin, streptomycin, amphotericin, gentamycin (Mediatech, Herndon, VA), 50 M 2-mercaptoethanol (Invitrogen), and 10% fetal bovine serum (Atlanta Biologicals, Atlanta, GA). To induce differentiation, cells (10 5 cells/ml) were stimulated with 25 g/ml LPS (Escherichia coli 055:B5), 25 ng/ml IL-6, or 1 ng/ml IL-5 (all from Sigma). To induce ER stress, cells were treated with 1 g/ml tunicamycin (Sigma). HeLa cells were maintained in Dulbecco's modified Eagle's medium supplemented with the above antibiotics and 10% fetal bovine serum (Mediatech).
Secretory Rates and Enzyme-linked Immunosorbent Assay-Cells were washed twice in media and counted using trypan blue dye exclusion to determine viability. Cells were then re-plated at 1 or 2 ϫ 10 5 cells/ml and cultured for 1 or 4 h as indicated in the figure legends. IgM concentration in supernatants was determined by enzyme-linked immunosorbent assay using microtiter plates (BD Biosciences, Franklin Lakes, NJ) coated with goat anti-mouse (Southern Biotechnology Associates, Birmingham, AL). Goat anti-mouse -HRP (Southern Biotechnology Associates) was used as a secondary reagent.
SDS-PAGE and Immunoblotting-Samples were resolved on polyacrylamide gels under denaturing and reducing conditions and then transferred to supported nitrocellulose membranes (Schleicher & Schuell). Kaleidoscope (Bio-Rad) pre-stained molecular weight standards were used. Blots were blocked in Tris-buffered saline containing 0.1% Tween 20 (TBST), 5% milk, probed with antibodies diluted in TBST, 1% milk, and washed in TBST. Rabbit antisera against BiP, GRP94, and calnexin were generously provided by Dr. Linda Hendershot (St. Jude Children's Research Hospital, Memphis, TN). Rabbit anti-ATF6␣ sera with specificity for the cytosolic domain of ATF6␣ (recognizes both p90ATF6␣ and p50ATF6␣) was a kind gift from Dr. Kazutoshi Mori (Kyoto University, Kyoto, Japan). Rabbit anti-mouse and mouse anti-rabbit IgG were purchased from Jackson ImmunoResearch, West Grove, PA. Rabbit anti-mouse was purchased from ICN, Aurora, OH. Rabbit anti-mouse XBP-1 that recognizes both the 30-and 54-kDa versions of XBP-1 and rabbit anti-CHOP were purchased (Santa Cruz Biotechnology, Santa Cruz, CA). Protein A-HRP (EY Laboratories, San Mateo, CA), donkey anti-mouse IgG-HRP (Jackson Immu-noResearch), and ECL reagents (Amersham Biosciences) were used to develop the blots by chemiluminescence.
Metabolic Labeling and Immunoprecipitation-At various intervals of LPS stimulation, CH12 cells were washed twice with warm phosphate-buffered saline and then cultured at 2.5 ϫ 10 6 cells/ml in warm media lacking methionine and cysteine (Invitrogen) for 20 min. Cells were then labeled for 15 min with [ 35 S]methionine and -cysteine using 100 Ci/ml Tran 35 S-label (ICN), harvested by centrifugation, and washed twice with cold phosphate-buffered saline. Cells were solubilized on ice in lysing buffer (0.5% Nonidet P-40, 0.5% sodium deoxycholate, 150 mM NaCl, 50 mM Tris-HCl, pH 7.5) containing protease inhibitors as described above. Postnuclear supernatants were prepared and incubated first with protein A-Sepharose beads (Sigma) precoated with normal rabbit sera and then with protein A-Sepharose beads precoated with antibodies specific for mouse , mouse , BiP, or GRP94 as described above. In each case, the beads were then washed 4 times with cold washing buffer (0.5% Nonidet P-40, 0.5% sodium deoxycholate, 400 mM NaCl, 50 mM Tris-HCl, pH 7.5), resuspended in reducing SDS-PAGE sample buffer and boiled for 5 min. Samples were resolved on 10.5% polyacrylamide gels under reducing SDS-PAGE conditions. Gels were fixed and stained (9% acetic acid, 45% methanol, 0.2% Coomassie Brilliant Blue (Bio-Rad)), de-stained (7% acetic acid, 15% methanol), and dried onto Whatman paper. Signals were visualized and quantitated using a Typhoon PhosphorImager and Image-Quant software (Amersham Biosciences).
Northern Blotting-Total RNA was isolated using the RNeasy kit (Qiagen, Valencia, CA). Equivalent amounts of RNA were resolved on 1.5% agarose-formaldehyde gels, transferred to Duralose-UV membranes (Stratagene, La Jolla, CA), and hybridized overnight at 42°C in a buffer containing 50% formamide, 6ϫ SSPE, 5ϫ Denhardt's solution, 0.5% SDS, and 100 g/ml salmon sperm DNA (Invitrogen). Probes were prepared by random primer labeling using the Prime-It II kit (Stratagene) and [␣-32 P]dCTP (Amersham Biosciences). The mouse C probe, a ϳ700-bp BglII-PstI insert from p107; the mouse probe, a 500-bp EcoRI insert from pSC33; and the CHO-B probe, a 600-bp EcoRI/ BamHI insert from the CHO-B plasmid were provided by Dr. Ronald Corley (Boston University, Boston, MA). The BiP probe was a 1.5-kb EcoRI-PstI fragment from a hamster cDNA clone provided by Dr. Amy Lee (University of Southern California, Los Angeles, CA). The probe for GRP94 was a 1.0-kb SalI-HindIII fragment from a mouse GRP94 cDNA provided by Dr. Michael Green (St. Louis University, St. Louis, MO). The mouse XBP-1 probe was a 687-bp cDNA fragment (nucleotides 421-1108) amplified by reverse transcriptase-PCR of CH12 B cell RNA. The mouse CHOP probe was a 600-bp EcoRI-XbaI fragment from a mouse cDNA provided by Dr. David Ron (New York University, New York). Blots were washed at 65°C in 2ϫ SSC, 0.1% SDS. Hybridization signals were analyzed by autoradiography and phosphorimaging (Amersham Biosciences).

Expression of p54XBP-1 in Differentiating B cells-Expres-
sion of the spliced form of the XBP-1 transcription factor, p54XBP-1, occurs in LPS-stimulated splenocytes (19). Therefore, we first determined whether synthesis of p54XBP-1 is unique to LPS-stimulated splenocytes or represents an event common to the differentiation of antibody-secreting B cells. We chose to utilize the murine B cell lymphoma, CH12, for our studies as these cells differentiate in vitro into high-rate IgMsecreting cells in response to IL-5, IL-6, or LPS (26 -28). Because IL-5 and LPS both augment proliferation of CH12 B cells, whereas IL-6 does not ( Fig. 1A and Ref. 28), use of these stimuli allows one to potentially distinguish events related to proliferation rather than differentiation. Immunoblot analysis using a rabbit antibody that recognizes both the p30 and p54 forms of XBP-1 was performed on CH12 cells harvested at 24-and 48-h intervals of culture under various conditions. As a control, a separate culture of LPS-stimulated cells was treated for 4 h with tunicamycin to induce the UPR and as expected, we observed p54XBP-1 expression in tunicamycin-treated cells (Fig.  1B). In cells stimulated with IL-5, IL-6, or LPS, we found that p54XBP-1 was expressed by 24 h and was increased at 48 h ( Fig. 1B). Each of these stimuli induced a similar level of IgM secretion (Fig. 1A). A much lower level of p54 XBP-1 was detected by 48 h in cells cultured in media alone that differentiated poorly in comparison to stimulated cells (Fig. 1, A and B). Importantly, cells cultured in media alone proliferated equivalently to cells stimulated with IL-6, yet did not express a similar level of p54XBP-1 (Fig. 1, A and B). This comparison suggests that generation of p54XBP-1 correlates with optimal differentiation rather than proliferation. A weaker signal corresponding to p30XBP-1 was also detected in the different culture conditions (Fig. 1B). These data indicate that UPRmediated post-transcriptional processing of XBP-1 mRNA and the apparent predominant expression of p54XBP-1 protein can be induced by multiple stimuli that are known to trigger the differentiation of antibody-secreting B cells. Therefore, UPR activation is not unique to LPS-stimulated B cells, but rather represents a general feature of terminal B cell differentiation.
Kinetics of Ig Induction and Secretion during Differentiation of Ig-secreting B Cells-Our initial results concerning XBP-1 ( Fig. 1) prompted us to assess the kinetics of Ig induction and its potential relationship to the activation of UPR-related events during differentiation of antibody-secreting B cells. Because p54XBP-1 was induced by all factors that stimulate CH12 B cell differentiation (Fig. 1), we simplified our studies by focusing on LPS-mediated differentiation and began by characterizing the kinetics of increased Ig expression and secretion. We first assessed the abundance of transcripts for Ig heavy and light chains over the course of differentiation. Northern blot analyses revealed that increased amounts of and mRNAs were apparent by the 12-h intervals of LPS stimulation, and these continued to elevate thereafter ( Fig.  2A). The transcripts were resolved as a tight doublet, consistent with their identity being membrane-bound (2.7 kb) and secretory (2.4 kb) mRNAs as others have reported (29,30).
Next, we utilized metabolic labeling and immunoprecipitations to assess the translation of heavy and light chains at various intervals of LPS-induced CH12 differentiation. A slight

FIG. 2. Kinetics of increased Ig expression during CH12 B cell differentiation.
A, northern blot analyses of and expression. Cells were stimulated to differentiate with LPS and harvested at the indicated intervals. Equivalent amounts of total RNA prepared from each sample were resolved on a 1.5% agarose/formaldehyde gel. Northern blots were hybridized with radiolabeled probes specific for the indicated genes. CHO-B encodes a mitochondrial protein and served as a loading control. Data representative of 3 separate experiments are shown. B, analysis of and synthesis by metabolic labeling. Cells were stimulated with LPS for the indicated intervals. At each interval, cells were shifted to methionine-and cysteine-deficient media and labeled for 15 min with [ 35 S]methionine and -cysteine. Cell lysates were first precleared with protein A-Sepharose beads precoated with normal rabbit sera (NRS). and chains were then immunoprecipitated with rabbit anti-and rabbit anti-antibodies and protein A-Sepharose. Immunoprecipitates were resolved by 10.5% SDS-PAGE under reducing conditions. Signals were obtained by phosphorimaging analysis of the dried gels. C, quantitative data from 3 separate experiments as shown in B. Signal intensities for total and at each interval were quantitated by phosphorimaging. The mean Ϯ S.E. (n ϭ 3) for the -fold increase in synthesis of (q) and (f) at each interval relative to the value obtained for each at t ϭ 0 are plotted. decrease in the amount of [ 35 S]methionine and -cysteine incorporated into both nascent and chains was observed at the 4and 8-h intervals as compared with the 0-h starting point (Fig.  2, B and C). A decrease in metabolic activity during the initial few hours of in vitro culture for CH12 B cells has been previously reported (30). By the 12-h interval, the amount of label incorporated into both and chains had returned to the level observed at 0 h. Therefore, in total, the amount of label incorporated into both nascent and chains was strikingly similar in cells labeled at the 0-, 4-, 8-, and 12-h intervals (Fig. 2, B and C). By the 24-h interval, translation of both and chains was up-regulated and their synthesis continued to increase thereafter (Fig. 2, B and C). The chains were resolved as a tight doublet on reducing SDS-polyacrylamide gels (Fig. 2B), consistent with their identity being membrane-bound (upper band) and secretory (lower band) proteins as others have previously documented (30,31). This quantitative assessment of Ig translation in CH12 B cells indicates that the flow of nascent Ig polypeptides into the ER does not increase until after the 12-h interval of LPS-induced differentiation.
We then used immunoblotting to assess the steady-state level of heavy and light chains at various intervals of CH12 differentiation. Immunoblotting revealed significant increases in both and chains by the 24-h interval of LPS stimulation (Fig. 3A). The chains were again resolved as a doublet on reducing SDS-polyacrylamide gels (Fig. 3A), consistent with their identity being membrane-bound and secretory (30,31). The immunoblots also indicated that smaller increases in Ig chains occurred by 8 -12 h of stimulation (Fig. 3A). This may result from enhanced stability of these proteins early in the differentiation program prior to their elevated synthesis, a possibility that can be further explored using pulse-chase experiments. Finally, we found that CH12 B cells secreted very low, but measurable amounts of IgM by 12 h of LPS stimulation (Fig. 3B). The rate of secretion increased sharply between the 20-and 28-h intervals and reached maximal levels by 48 h (Fig.  3B).
Induction of XBP-1 and ER Chaperone Transcripts during Differentiation of Ig-secreting B Cells-The XBP-1 transcription factor and the ER chaperones BiP and GRP94 are all transcriptionally up-regulated in the UPR (1,11) and are expressed at elevated levels in differentiating B cells (19,21,23,24). Therefore, we assessed the kinetics of induction of these UPR target genes in CH12 B cells differentiating in response to LPS. Northern blotting revealed that transcripts for BiP and GRP94 were elevated by 12 h of stimulation (Fig. 4). The significance of the initial increase in BiP transcripts observed at the 2-h interval is unclear, but may reflect changes in gene expression or mRNA turnover that might accompany the slowed metabolic activity of CH12 cells early during in vitro culture (30). More importantly, we reproducibly observed a coordinate induction of BiP and GRP94 mRNAs that began by the 12-h interval (Fig. 4). XBP-1 transcripts diminished during the first several hours of LPS stimulation (Fig. 4). The significance of this initial decline is unclear, but as described above, it may also be related to the decreased metabolic activity of CH12 B cells at early intervals of in vitro culture. However, a marked induction of XBP-1 mRNA was observed by 8 -12 h of stimulation, with a maximal level reached by 32 h (Fig. 4). Hence, XBP-1 transcripts were elevated with kinetics similar to that of BiP and GRP94. Therefore, the mRNAs for the UPR target genes BiP, GRP94, and XBP-1 are all elevated early in the LPS-induced differentiation program (12 h interval) prior to increased translation of Ig chains.
Synthesis of ER Chaperone and XBP-1 Proteins during Differentiation of Ig-secreting B Cells-We next determined whether the increase in transcripts for BiP, GRP94, and XBP-1 during the differentiation of CH12 B cells correlated with elevated synthesis of these proteins. The ER chaperones BiP and GRP94 are constitutively expressed, abundant proteins. Therefore, we first used metabolic labeling and immunoprecipitations to assess the translation of BiP and GRP94 at various intervals of LPS stimulation. Increased incorporation of [ 35 S]methionine and -cysteine into both BiP and GRP94 was clearly evident by the 24-h interval, and smaller increases were apparent by 12 h (Fig. 5A). By 48 h, the amount of label incorporated into these chaperones had modulated back to the level observed at the earlier time points (Fig. 5A). This intriguing observation will be further investigated as it suggests that

FIG. 3. Intracellular Ig levels and kinetics of Ig secretion during CH12 B cell differentiation.
Cells were stimulated to differentiate with LPS and harvested at the indicated intervals. A, whole cell lysates were prepared and equal cell equivalents were resolved by 10.5% SDS-PAGE under reducing conditions. Immunoblot analyses were performed using rabbit antibodies specific for the indicated proteins, protein A-HRP, and chemiluminescence. Data representative of 3 separate experiments are shown. B, the amount of IgM secreted per hour by 2 ϫ 10 5 cells was determined at each interval by enzyme-linked immunosorbent assay.

FIG. 4. Northern blot analyses of UPR target genes during differentiation of CH12 B cells.
Cells were stimulated to differentiate with LPS and harvested at the indicated intervals. Equivalent amounts of total RNA prepared from each sample were resolved on a 1.5% agarose/formaldehyde gel. Northern blots were hybridized with radiolabeled probes specific for the indicated genes. CHO-B encodes a mitochondrial protein and served as a loading control. Data representative of 3 separate experiments are shown. the synthesis of these ER lumenal proteins may be enhanced during ER expansion and then equilibrated to baseline levels at later stages of differentiation. In agreement with the biosynthetic labeling analysis, immunoblotting revealed a measurable increase in the levels of the long-lived BiP and GRP94 proteins that were apparent at the 24-and 48-h intervals (Fig.  5B).
In contrast to BiP and GRP94, there is little or no synthesis of XBP-1 in CH12 B cells in the absence of a differentiation stimuli (Fig. 1B). Therefore, we used immunoblotting to assess the synthesis of XBP-1 over the course of LPS-induced differentiation. Both the 30-and 54-kDa forms of XBP-1 protein were clearly present by 12 h and remained throughout the time course, with p54XBP-1 as the apparent predominant form at least by the 24-h interval (Fig. 6A, top and bottom panels). The presence of p30XBP-1 at all intervals of LPS stimulation indicates that a portion of the XBP-1 transcripts remained unprocessed throughout the differentiation program. In contrast, when LPS-stimulated cells were treated with tunicamycin for 2 h, only p54XBP-1 was detected (Fig. 6A), suggesting that complete processing of XBP-1 transcripts occurred under these conditions. IRE1␣, the ER transmembrane kinase/endoribonuclease that excises the 26-nucleotide cryptic intron from XBP-1 transcripts (19), and the complete XBP-1 mRNA splicing mechanism may be more robustly activated in response to tunicamycin than during differentiation. These data indicate that induction of XBP-1 synthesis, including the UPR-induced p54XBP-1, initiates prior to increased synthesis of nascent Ig chains as CH12 B cells terminally differentiate into antibodysecreting cells (Figs. 2, B and C, and 6A). Furthermore, these data suggest that IRE1␣-mediated splicing is closely coordinated with induction of XBP-1 mRNA and proceeds throughout the differentiation program.
Activation of ATF6 during Differentiation of Ig-secreting B Cells-ATF6␣ functions as a proximal inducer of the UPR as p50ATF6␣ can positively regulate the promoters of genes including BiP, GRP94, and XBP-1 (9,11,12). Therefore, we next determined whether p50ATF6␣ is generated as B cells differentiate into antibody-secreting cells. Immunoblot analysis of CH12 cells harvested at various intervals of LPS stimulation was performed using a rabbit anti-human ATF6␣ antisera that recognizes both the full-length (p90) and cleaved (p50) forms of mouse and human ATF6␣. To serve as controls, CH12 cells were treated with tunicamycin for 2 h to ensure cleavage of ATF6␣. Also, HeLa cells, a human cell line in which ATF6␣ processing has been extensively characterized (6,9), were cultured in the absence or presence of tunicamycin for 2 h to provide specificity controls. p90ATF6␣ was readily detectable and its level increased over the time course of CH12 B cell differentiation (Fig. 6B, top panel). In contrast, the level of calnexin, an ER transmembrane protein that acts as a molec- FIG. 5. Kinetics of increased ER chaperone synthesis during CH12 B cell differentiation. Cells were stimulated with LPS for the indicated intervals. A, at each interval, cells were shifted to methionineand cysteine-deficient media and labeled for 15 min with [ 35 S]methionine and -cysteine. Cell lysates were precleared with protein A-Sepharose and normal rabbit sera (NRS). BiP and GRP94 were then immunoprecipitated with protein A-Sepharose and specific rabbit antisera and resolved by 10.5% SDS-PAGE under reducing conditions. Signals were obtained by phosphorimaging analysis of the dried gels. B, whole cell lysates were prepared and equal cell equivalents were resolved by 10.5% SDS-PAGE under reducing conditions. Immunoblot analyses were performed using rabbit antibodies specific for the indicated proteins, protein A-HRP, and chemiluminescence. Data representative of 3 separate experiments are shown.
FIG. 6. Analysis of XBP-1 synthesis and ATF6␣ activation during CH12 B cell differentiation. Cells were stimulated to differentiate with LPS and harvested at the indicated intervals. At 20 h, a portion of cells was treated with tunicamycin for 2 h to provide a positive control for synthesis of p54XBP-1 synthesis and cleavage of p90ATF6␣. Whole cell lysates were prepared and equal cell equivalents were resolved by 10% SDS-PAGE under reducing conditions. The same cell lysates were used for the immunoblots shown in A and B. A, immunoblot analysis of XBP-1. Rabbit antibodies specific for XBP-1, mouse anti-rabbit IgG as a secondary reagent, donkey anti-mouse IgG-HRP, and chemiluminescence were used. Top panel, 20 s exposure to film. Bottom panel, 5 min exposure to film. Data representative of 3 separate experiments are shown. B, immunoblot analysis of ATF6␣. HeLa cells cultured in the absence (Ϫ) or presence of tunicamycin for 2 h were included to provide specificity controls for identification of p90 and p50ATF6␣. Rabbit antibodies specific for ATF6␣, mouse anti-rabbit IgG as a secondary reagent, and donkey anti-mouse IgG-HRP were used to detect ATF6␣ proteins. Calnexin was detected using a rabbit anticalnexin antibody and protein A-HRP and included as a control for equal loading of cell equivalents for each interval of LPS stimulation. Signals were visualized by chemiluminescence. Top panel, 10 s exposure to film of ATF6 blot. Middle panel, 1 min exposure to film of ATF6 blot. Bottom panel, calnexin blot. Data representative of 3 separate experiments are shown. ular chaperone, exhibited a more modest increase over the course of CH12 differentiation and therefore provided a control for the loading of equal cell equivalents (Fig. 6B, bottom panel). The increase in p90ATF6␣ was a reproducible finding and will be studied further in regard to the regulation of the UPR proximal signal transducers during differentiation. Nonglycosylated p90ATF6␣ in the tunicamycin-treated cells exhibited enhanced electrophoretic mobility, providing a positive control for tunicamycin activity (Fig. 6B, top panel). A protein of ϳ50 kDa was also clearly detected by the rabbit anti-ATF6␣ antisera by the 12-h interval of LPS stimulation. This ϳ50-kDa protein migrated identically to a predominant ϳ50-kDa protein detected by the anti-ATF6␣ antisera in extracts from tunicamycin-treated CH12 cells. Comparison of this signal to the parallel HeLa cell control was consistent with its identity being p50ATF6␣ (Fig. 6B, top and middle panels). The level of p50ATF6␣ detected in differentiating B cells was less than the amount observed in tunicamycin-treated cells, again suggesting that UPR activation may be more robust in response to tunicamycin. The level of p50ATF6␣ was maximal by 24 h and was diminished by 48 h, the interval at which p90ATF6␣ was most abundant (Fig. 6B). This interesting result is currently under investigation and is important as it argues that the appearance of p50ATF6␣ over the course of CH12 B cell differentiation cannot be attributed solely to increased levels of p90ATF6␣. These data indicate that p90ATF6␣ is proteolytically cleaved to yield the p50ATF6␣ transcription factor in differentiating B cells. The generation of p50ATF6␣, like induction and processing of XBP-1 mRNA and increased expression of BiP and GRP94 transcripts, appears to initiate prior to increased synthesis of nascent Ig chains in terminally differentiating B cells.
Analysis of CHOP Expression during Differentiation of Igsecreting B Cells-Like BiP, GRP94, and XBP-1, transcription of CHOP is induced by pharmacological agents or culture conditions that hinder protein folding in the ER (32,33). The CHOP promoter contains a single ERSE through which p50ATF6␣ participates in its induction in the UPR (11). Our results demonstrating the generation of p50ATF6␣ and the corresponding induction of ERSE-regulated genes during differentiation of antibody-secreting CH12 B cells led us to determine whether CHOP is also induced under these conditions. We stimulated CH12 cells with LPS for various intervals and also treated a portion of cells at each interval with tunicamycin for an additional 4 h to generate appropriate positive controls. Northern blot analyses revealed the expected induction of and BiP transcripts over the course of LPS stimulation (Fig.  7A). In addition, BiP was further induced by tunicamycin treatment (Fig. 7A). In contrast, we found little or no evidence for induction of CHOP transcripts during differentiation, even at the 24-h interval when p50ATF6␣ was at maximal levels ( Fig.  6B) and XBP-1, BiP, and GRP94 transcripts were all up-regulated (Fig. 4). However, at all intervals of LPS stimulation, CHOP was robustly induced by tunicamycin (Fig. 7A), indicating that the arms of the UPR pathway that act upon the CHOP promoter can be pharmacologically activated throughout the course of CH12 B cell differentiation. In agreement with the northern blot data, we found no evidence for CHOP protein in differentiating CH12 B cells, yet CHOP was easily detected when cells were treated with tunicamycin (Fig. 7B). These data indicate that CHOP, unlike BiP, GRP94, and XBP-1, is not induced as a normal feature of CH12 B cell differentiation, suggesting that some, but not all, aspects of the known UPR pathway are triggered as B cells terminally differentiate into antibody-secreting cells. DISCUSSION We have provided the first analysis of the relationship of UPR activation to increased Ig synthesis during the differentiation of antibody-secreting B cells. It has previously been proposed that the enhanced flow of nascent Ig heavy and light chains into the ER in terminally differentiating B cells might elicit a feedback response, namely the UPR pathway, that acts to accommodate the increased demand on the protein folding capacity of the ER (19,23,34). In this plausible model, the UPR would up-regulate synthesis of ER chaperones and folding enzymes to ensure efficient antibody assembly and possibly also coordinate the enhanced membrane biosynthesis necessary for generation of the highly developed ER network that is characteristic of plasma cells. However, our data indicate that UPR events including induction of XBP-1, BiP, and GRP94 transcripts, induction of p54XBP-1 synthesis, and generation of p50ATF6␣ may all be initiated prior to increased translation of Ig chains during differentiation (Figs. 2, B and C, and 4 -6). This raises the possibility that another, as yet unknown, signal(s) provides the initial trigger for the UPR in this differentiation program.
Whereas our results strongly argue that certain elements of the UPR pathway are activated prior to increased synthesis of nascent Ig chains, the data do not rule out a potential role for Ig levels in eliciting these events. First, our immunoblot analyses of and levels over the course of differentiation suggest To ensure that the media was not depleted of any required nutrients, cells were switched to fresh media at 24 h. Expression of the indicated gene products was assessed by northern blotting as in Figs. 2 and 4. B, immunoblot analysis. CH12 cells were stimulated with LPS for the indicated time intervals. To generate a positive control for CHOP, a portion of cells was stimulated with LPS for 24 h and then treated with tunicamycin for an additional 4 h. Equivalent amounts of cellular protein were resolved by reducing SDS-PAGE and analyzed by immunoblotting with rabbit antibodies specific for , GRP94, and CHOP followed by protein A-HRP. Signals were visualized by chemiluminescence. The and GRP94 data are provided as controls for differentiation. * denotes nonglycosylated synthesized in the presence of tunicamycin. that the intracellular pools of these proteins may begin to increase prior to their elevated translation (Fig. 3). Also, it is noteworthy that all of the UPR events we assessed were optimal at the 24-h interval of LPS stimulation (Figs. 4 -6), a time at which Ig synthesis had sharply increased (Fig. 2, B and C). Thus, additional studies are necessary to more precisely define whether Ig levels in the ER contribute to UPR activation in differentiating B cells. Whatever the exact nature of the activating signal(s), we propose that UPR events initiate early in the differentiation program and then proceed, at least in part, concomitantly with increased Ig synthesis. This may allow a differentiating B cell to guard against significant deficiencies in the protein folding capacity of the ER even at the earliest stages of antibody secretion, a strategy that would optimize humoral immune responses.
Our data also provide the first evidence for activation of the ER membrane-localized transcription factor, ATF6, during the differentiation of B cells into antibody-secreting cells. It is noteworthy that while only a small amount of p90ATF6␣ appeared to be processed into the p50 form during differentiation, complete proteolytic cleavage of p90ATF6␣ is also not typically observed when cells are treated with pharmacological agents such as tunicamycin (6). It is intriguing that the abundance of p50ATF6␣ diminished at later time points of CH12 B differentiation (Fig. 6B). In contrast, the presence of p54XBP-1 throughout the 48 h of LPS stimulation (Fig. 6A) suggests that IRE1␣ activation was maintained. We emphasize that our current data for IRE1␣ activation are indirect. The IRE1 proteins are the only known mammalian proteins capable of mediating the cleavage of XBP-1 mRNA necessary for its alternative splicing (19); however, it will be important to directly assess the activation status of IRE1␣ in differentiating B cells.
The observations regarding ATF6 cleavage in differentiating B cells raise the possibility that ATF6 may participate in activating ERSE-regulated genes as part of the overall B cell terminal differentiation program. This is a testable hypothesis, although the issue is complicated by the fact that both ATF6 and XBP-1 can positively regulate ERSE-containing promoters (18). Therefore, determining whether ATF6 participates in regulating the XBP-1 promoter in differentiating B cells will be particularly important. The XBP-1 promoter is negatively regulated in B cells by the B cell lineage-specific activator protein encoded by Pax5 (35). B-lymphocyte-induced maturation protein (Blimp-1), a critical regulator of plasma cell differentiation (36,37), represses Pax5 (38). Recent evidence suggests that Blimp-1 is necessary, but not sufficient to induce XBP-1 (38). Taking our data into consideration, an intriguing and testable model is that p50ATF6␣ collaborates with Blimp-1 to optimally activate transcription of XBP-1 in differentiating B cells.
Our studies also reveal that some, but not all, aspects of the known UPR may be activated as B cells differentiate into antibody-secreting cells. The finding that CHOP is not induced, at least not to readily detectable levels, in differentiating CH12 B cells is striking (Fig. 7), given that all other ERSE-regulated genes that we have assessed (BiP, GRP94, and XBP-1) are up-regulated (Fig. 4). Interestingly, a previous study of IL-3dependent progenitor-myeloid cell lines revealed that re-stimulation of IL-3-deprived cells with IL-3 induced expression of BiP and GRP94, but not CHOP (39), providing precedence for the data reported here. Similar to the increase in protein traffic through the exocytic pathway that occurs in terminally differentiating B cells, re-stimulation of IL-3-deprived progenitor myeloid cells with IL-3 leads to elevated glycoprotein synthesis as protein translation is restored to normal levels (39). Thus, there may be some commonality in the demands placed upon the ER in these two distinct cellular systems.
What might be the molecular distinction between the "UPR" activated in differentiating B cells and the UPR elicited by pharmacological agents or conditions that severely disrupt the normal environment of the ER? The answer may involve PERK (PKR-like ER kinase), an ER transmembrane serine/threonine kinase that, like the ATF6 and IRE1 proteins, is activated by ER stress-inducing agents (40,41). Upon activation, PERK phosphorylates the ␣ subunit of elongation initiation factor-2 on serine 51, thereby preventing formation of translation initiation complexes and effectively inhibiting protein synthesis (42). Induction of certain UPR target genes, including CHOP, in response to ER stress agents such as tunicamycin and thapsigargin is dependent on PERK (43). Indeed, recent studies indicate that induction of the CHOP promoter in the UPR may require both p50ATF6␣ and ATF4 (44,45). Paradoxically, translation of ATF4 increases upon activation of PERK and other elongation initiation factor-2␣ kinases (43). In light of our B cell differentiation data and the previous work regarding mitogenic signaling (39), we speculate that PERK activation or the propagation of its downstream signals does not always accompany other aspects of the known UPR. We reason that PERK-mediated repression of protein synthesis, at least in a sustained manner, would be counterproductive in a differentiating B cell committed to high-level synthesis of Ig chains. Perhaps, mechanisms exist for differential utilization of distinct components of the UPR pathway to mediate ER homeostasis according to the needs of specific cell types and physiological situations. Testing these hypotheses will certainly require a direct and complete assessment of PERK activation and PERK-mediated signaling during B cell terminal differentiation.
We have demonstrated that activation of a UPR is characteristic of B cells differentiating into antibody-secreting cells. Delineating the relative importance of the various components of the UPR pathway to the differentiating B cell may yield insight into how this interorganelle signaling mechanism facilitates antibody secretion and, more broadly, meets the dynamic needs of the exocytic pathway in differentiating secretory cell types.