ASK1 promotes apoptosis of normal and malignant plasma cells

Plasma cells, the terminally differentiated antibody-secreting cells, are generated from mature B cells once they encounter antigens.¹  The generation of plasma cells can be aided by follicular helper T cells or independent of T cells.²  Plasma cells that arise in lymphoid organs either receive survival signals from the local microenvironment or home to bone marrow under the guidance of the bone marrow–derived chemoattractant, CXCL12, wherein they become competent for long-term survival; otherwise, they serve as short-lived plasma cells that eventually undergo apoptosis.¹  Plasma cell survival in bone marrow requires several external factors, such as IL-6, B cell–activating factor, and a proliferation-inducing ligand secreted by bone marrow stromal cells or osteoclasts.^3-5  In the lymphoid organs/tissues, plasma cells survive in the presence of B cell–activating factor or IL-10, each of which is secreted by epithelial cells⁶ ; in the presence of IL-6 secreted by dendritic cells; or in the presence of a proliferation-inducing ligand secreted by macrophages.⁷  Many of these factors are also important for the survival of the malignant counterparts of long-lived plasma cells,⁸  such as multiple myeloma (MM) cells.¹  However, the intracellular factors that regulate the survival of either short-lived or long-lived plasma cells need to be further investigated.

During an immune response, the mass production of immunoglobulins in plasma cells can invoke a cellular adaptive response, termed the endoplasmic reticulum (ER) stress response, to cope with the consequent accumulation of unfolded immunoglobulins in the ER lumen.⁹  Apoptosis can be triggered on prolonged ER stress.¹⁰  Apoptosis signal–regulating kinase 1 (ASK1), also known as mitogen-activated protein 3K5, plays a role in inducing apoptosis in response to cellular stress.¹¹  Mitogen-activated protein 3K5 is a member of MAP kinase kinase kinase family, and it selectively activates the JNK and p38 MAPK pathways on receiving various intracellular and extracellular stress signals, including ER stress and treatment with lipopolysaccharide (LPS).^12-14  Activation of ASK1 is tightly regulated by phosphorylation of a threonine (T) residue in ASK1 (T838 in human; T845 in mouse) that resides in the loop of the kinase domain.¹⁵  ASK1 can be activated by autophosphorylation.¹⁵  By contrast, protein phosphatase 5 (PP5) negatively regulates ASK1 by dephosphorylating T838.¹⁶  ASK1 is activated during stress responses, and the subsequent activation of JNK is required for stress-induced apoptosis. In fact, activation of ASK1 plays a role in a variety of diseases, such as those involving neuronal degeneration.¹³ 

B lymphocyte–induced maturation protein-1 (Blimp-1) is a crucial regulator of plasma cell development. B cell–specific deletion of Prdm1, the gene encoding Blimp-1, in mice showed defective production of antibody-secreting plasma cells generated after exposure to either thymus-dependent (TD) or thymus-independent (TI) antigens.¹⁷  Blimp-1 suppresses the transcription of genes responsible for germinal center B-cell or mature B-cell identity and for proliferation.¹⁸  In addition to its role in orchestrating plasma cell differentiation, Blimp-1 is also required for maintenance of long-lived plasma cells.^19,20  In our previous study, we reported that survival of MM cells or mouse plasmacytoma cells requires continuous Blimp-1 expression because depletion of Blimp-1 in these cells causes apoptosis.²⁰  In addition, MM cells treated with bortezomib, a therapeutic agent for MM relapse,²¹  had a reduced level of Blimp-1.²⁰  However, the detailed mechanism of how Blimp-1 ensures the survival of plasma cells remains elusive.

Because immunoglobulin overproduction in plasma cells can cause ER stress and ASK1 is involved in ER stress–induced apoptosis, we hypothesized that long-lived plasma cells possess a mechanism by which ER stress–induced apoptosis is prevented; thus, we investigated whether the regulation of ASK1 is relevant to plasma cell survival.

NCI-H929 (H929), U266, IM9 human MM cells, SKW6.4 lymphoblastoid cells, Raji mature B cells, and stable transfectants of WI-L2 cells were maintained as described.^18,20  Arachidonic acid (AA) and okadaic acid (OA) were purchased from Calbiochem. The small interfering RNA (siRNA) targeting human ASK1 (Stealth RNAi siRNA duplex) or negative-control siRNA duplexes were purchased from Invitrogen.

Splenic B cells (purified with B220 microbeads; Miltenyi Biotec) from 8- to 12-week-old C57BL/6 mice (BioLASCO), Ask1 knockout (KO) mice were generated as described,²²  and their littermate control wild-type (WT) mice were stimulated with LPS (2 μg/mL; Sigma-Aldrich) or IL-21 (100 ng/mL; eBioscience) plus αCD40 (5 μg/mL; BD PharMingen) and αIgM (10 μg/mL; Jackson ImmunoResearch Laboratories). The procedures for immunization of mice are described in supplemental Methods (available on the Blood Web site; see the Supplemental Materials link at the top of the online article). Primary human CD138⁺ plasma cells (isolated by CD138 MicroBeads; Miltenyi Biotech) were obtained from patients with MM at the National Taiwan University Hospital. Primary CD19⁺ human B cells were isolated as described²³  and cultured in RPMI-1640 that contained 10% fetal bovine serum (Gibco), penicillin/streptomycin, and 50μM β-mercaptoethanol at an initial cell density of 1 × 10⁶/mL. All the human samples were obtained with the approval of both the Human Subject Research Ethics Committee of Academia Sinica and the National Taiwan University Hospital.

cDNAs encoding full-length human ASK1 and constitutively active (CA) ASK1 lacking the N-terminal residues 1-649 were amplified by RT-PCR from H929 cells. The lentiviral vector that produces the shRNA against Blimp-1 and the negative control were described previously.²⁰  Detailed information on the generation of the dominant-negative ASK1, various expression vectors encoding different forms of ASK1, ASK1 promoter–driven luciferase reporter construct, and the constructs for establishment of Tet-on-inducible transfectants are described in supplemental Methods.

H929 and IM9 cells stably expressing luciferase were generated by selecting cells that carry the integrated luciferase expression vector with puromycin (2 μg/mL; Sigma-Aldrich). See supplemental Methods for details.

The protocol for preparation of retroviral and lentiviral vectors and viral transduction has been described.^20,24  See supplemental Methods for details.

Cells were harvested and washed with phosphate-buffered saline once and then suspended in annexin V binding buffer (BD PharMingen) at a density of 10⁶/mL. The antibodies used in this study were purchased from BD PharMingen and are described in supplemental Methods. Cellular fluorescence intensity was analyzed by FACSCanto (Becton Dickinson) and FCS Express 3.0 software (De Novo Software).

Total cell lysates (30-50 μg) or nuclear extracts (10 μg), prepared as described,²⁵  were subjected to SDS-PAGE and immunoblotting with the use of primary antibodies against Blimp-1, actin, α-tubulin, FLAG, Bim, Bcl-2, Bcl-x_L, and Mcl-1 as described.²⁰  Otherwise, the other antibodies are described in supplemental Methods. The immunoreactive proteins were detected as described.²⁰ 

Isolation of total RNA, cDNA synthesis, and subsequent quantitative PCR (QPCR) analysis in an ABI Prism 7300 sequence detection system (Applied Biosystems) were performed as described.²³  RT-QPCR analysis of cDNA was performed with the use of primers for the SYBR green method, and the primer sequences used are described in supplemental Methods.

Chromatin immunoprecipitation was performed essentially as described.²³  See supplemental Methods for details.

siRNA (1μM) against human ASK1 or negative-control siRNA was delivered into 2 × 10⁶ H929 cells by Nucleofector II (Lonza) according to the supplier's suggested protocol. The procedure for the luciferase reporter assay is described in supplemental Methods.

Luciferase-expressing ASK1-inducible and control Tet response element (TRE) H929 or IM9 transfectants (1 × 10⁷ cells) were intravenously injected into sublethally irradiated female SCID mice (BioLASCO) via the tail vein. At 6 days after engraftment, normal drinking water was replaced by 5% sucrose lacking doxycycline (Dox) or supplemented with Dox (0.2 mg/mL). Meanwhile, mice that had received a xenograft were orally administrated with or without Dox by gavage (1 mg in 1 mL per mice) twice a week. Bioluminescence was measured after intraperitoneal injection with D-luciferin (150 μg/g of body weight; Caliper), followed by acquiring bioluminescent images with the use of the Xenogen IVIS system.

Mice that had received a xenograft were killed 20 days after engraftment, and the lungs were fixed in 10% formaldehyde (Mallinckrodt Chemicals) at 4°C. Samples were dehydrated, followed by slicing paraffin-embedded sections at a thickness of 3 μm. Subsequent TUNEL staining was performed with an in situ cell death detection kit and TMR red (Roche) according to the manufacturer's suggestions Images, acquired with Leica DMG000B microscope equipped with a 20×/0.7 HC-Plan APO objective and Leica DFC490 camera, were analyzed with Leica FW4000 software (Leica Microsystems).

Statistical analysis was performed with a paired, 2-tailed Student t test. Log-rank tests were used to calculate the statistical significance in Kaplan-Meier survival curves. P < .05 was considered statistically significant.

We first examined if the expression level or activity of ASK1 is involved in generating short-lived plasma cells. Toward this goal, we generated short-lived plasma cells in vitro from murine splenic B cells that had been stimulated with LPS (which mimics a TI stimulus) or the combination of IL-21, αIgM, and αCD40 (which mimics a TD stimulus)²⁶ ; plasma cells typically begin to die at day 3 (TI-like stimulus) or day 5 (TD-like stimulus) after stimulation. As expected, Blimp-1 was induced after stimulation in both cultured cells as analyzed by immunoblotting, showing that these cells were committed to plasma cell fate (Figure 1A). The temporal expression and activation of ASK1 in these culture systems was next examined. The phosphorylation of ASK1 at T845 was increased in response to both types of stimulation, indicating that ASK1 was activated under both culture conditions; notably, the overall levels of ASK1 protein did not change substantially (Figure 1A). We next examined whether ASK1 contributed to stimulus-induced apoptosis during differentiation of short-lived plasma cells. WT or CA ASK1²⁷  was expressed from a bi-cistronic retroviral vector that coexpressed yellow fluorescent protein (yfp) in TD stimulant–treated splenic B cells, and apoptosis of plasma cells (B220^lowCD138⁺) was assessed by annexin V staining in the yfp⁺ gate. Ectopic expression of WT or CA ASK1 increased apoptosis compared with plasma cells transduced with the control vector (pGC; Figure 1B). We also examined apoptosis of short-lived plasma cells produced by TI or TD stimulant–treated Ask1 KO and WT splenic B cells; the B220^lowCD138⁺ plasma cells generated by KO B cells showed less apoptosis than WT plasma cells on day 3 (TI stimulation) and days 4-5 (TD stimulation; Figure 1C). Therefore, our data indicated that ASK1 was activated and played a role in enhancing apoptosis during the generation of short-lived plasma cells in vitro.

To further understand the physiologic relevance of ASK1 is regulating plasma cell survival in vivo, we examine the production of plasma cells after immunization. Seven days after immunization with TI antigen, (4-hydroxy-3-nitrophenyl) acetyl (NP)–Ficoll, KO mice produced significantly higher numbers of NP-specific IgM-secreting cells, which reflect short-lived plasma cells,¹⁷  in spleen compared with those produced by WT mice (Figure 1D), consistent with the notion that ASK1 promotes apoptosis of short-lived plasma cells. Along this line, we wondered if the numbers of long-lived plasma cells, generated by the TD antigen, NP–keyhole limpet hemocyanin, in bone marrow of mice was also affected by a lack of ASK1. Interestingly, at 6 weeks after immunization, the KO mice had more NP-specific IgG-secreting cells in bone marrow than did WT mice (Figure 1E), which coincided with the increased NP-specific IgG titers in the sera of KO mice 3 and 6 weeks after immunization (supplemental Figure 1). These results suggested that loss of ASK1 may prolong plasma cell survival, leading to increased numbers of long-lived and short-lived plasma cells.

Given that lack of ASK1 increased long-lived plasma cell numbers and that Blimp-1 is important for the maintenance of long-lived plasma cells,^19,20  we speculated that Blimp-1 may repress ASK1 in these cells. We examined the association between the expression of Blimp-1 and of ASK1 in CD19⁺ human mature B cells isolated from peripheral blood of healthy persons and in CD138⁺ human plasma cells isolated from bone marrow aspirates of patients with MM. As expected, the level of Blimp-1 mRNA was ∼ 8-fold higher in plasma cells than in mature B cells (Figure 2A). Of note, the level of ASK1 mRNA was reduced ∼ 4-fold in plasma cells than in mature B cells (Figure 2A). Consistently, the expression of ASK1 and Blimp-1 protein in a human lymphoblastoid cell line, SKW6.4, and in a MM cell line, H929, was inverse (Figure 2B). The reduced expression of ASK1 appeared to be sensitive to the level of Blimp-1 because Blimp-1 depletion via a lentiviral vector expressing shRNA for Blimp-1, Blimp-1i,²⁰  in MM cell lines (including IM9, U266, and H929) increased ASK1 mRNA levels compared with cells transduced with a control vector expressing a scrambled shRNA sequence (ctrli; Figure 2C). In a reverse trend, the expression of ASK1 mRNA and protein was reduced on induction of Blimp-1 in an established WI–IL-2 B-cell line that expressed inducible FLAG-tagged Blimp-1 fused with estrogen ligand binding domain¹⁸  (Figure 2D left panel and right panel, respectively). Together, these opposite expression profiles suggested that Blimp-1 suppresses ASK1 expression.

We next verified whether ASK1 is directly repressed by Blimp-1. A genomic region that spanned 1.5 kb upstream of the transcriptional start site of human ASK1 was reported to be necessary for basal transcription from ASK1²⁸  and was thus used to analyze the potential Blimp-1 binding sites according to putative Blimp-1 binding sequences.²⁹  We identified 2 putative Blimp-1 binding sites, which are located at approximately −1309 to −1299 and −262 to −253 relative to the transcription start site of ASK1. Chromatin immunoprecipitation was used to examine whether Blimp-1 binds these 2 putative sites in the ASK1 promoter; compared with samples immunoprecipitated with the rabbit IgG control antibody, Blimp-1 bound to CIITA, a known direct target of Blimp-1 as a positive control,³⁰  and to the −262 to −253 region of ASK1 but not to the −1309 to −1299 region or the sequence encoding the 3′ untranslated region (Figure 2E). Likewise, in the mouse Ask1 5′ regulatory region, 2 potential Blimp-1 binding sites were found, and both appeared to be occupied by Blimp-1 in B220^lowCD138⁺ plasma cells sorted from LPS-stimulated splenic B-cell cultures (supplemental Figure 2). To determine whether the putative Blimp-1 binding site in the −262 to −253 region of human ASK1 is critical for Blimp-1 repression, we generated a reporter construct containing the −1.3-kb region of ASK1 fused to luciferase cDNA. A previously generated luciferase reporter driven by CIITA promoter III was used as a positive control for Blimp-1 repression.³⁰  We found that, indeed, Blimp-1 suppressed the WT ASK1 promoter as well as the CIITA promoter activity in a dose-dependent manner (Figure 2F), whereas the Blimp-1–mediated repression was significantly diminished when the Blimp-1 binding site was mutated (Figure 2F). Taken together, these data indicated that Blimp-1 directly repressed ASK1 transcription.

Because Blimp-1 suppresses ASK1, we examined whether ectopic expression of ASK1 is sufficient to induce the death of malignant plasma cells, even in the presence of endogenous Blimp-1. A retroviral vector that expressed FLAG-tagged WT or CA ASK1, or yfp alone (pGC), was transduced into the human MM cell lines, H929 and U266 cells, and the retrovirally transduced (yfp⁺) cells were sorted 1 day later and maintained in culture for various periods to determine the extent of cell death. The expression of the various FLAG-tagged ASK1 proteins was confirmed by immunoblotting with anti-FLAG (Figure 3A). Compared with pGC-transduced cells, expression of WT or CA ASK1 resulted in more pronounced cell death in both the H929 and U266 lines at most time points (Figure 3A). In addition, when H929 cells were transduced with a dominant-negative ASK1, which is a loss-of-kinase-function mutant that carries a lysine-to-arginine point mutation at 709 site (K709R),¹²  cell death was comparable with that of pGC-transduced cells (Figure 3A left panel), indicating that the kinase activity of ASK1 controls the death of plasma cells. In line with this, Tet-on–based, Dox-dependent, inducible H929 and IM9 stable transfectants that express FLAG-tagged ASK1 were generated to examine cell death after induction of ASK1. Dox induced ASK1 on various days, as confirmed by immunoblotting with anti-FLAG (Figure 3B). As expected, induction of ASK1 expression in these 2 inducible transfectants triggered increased cell death in a time-dependent manner compared with control transfectants that carried only the TRE (Figure 3B). The increased cell death by enforced expression of ASK1 resulted from apoptosis because H929 cells retrovirally transduced with CA or WT ASK1 had a greater proportion of annexin V–positive cells compared with pGC-transduced cells (Figure 3C). Together, these data indicated that ectopic expression of ASK1 triggered apoptotic cell death in malignant plasma cells.

Because ASK1 overexpression could overcome the survival-promoting effect of Blimp-1 and could increase plasma cell death, we next examined whether ASK1 activity is crucial for this effect. ASK1 activity can be negatively regulated by PP5.¹⁶  The phosphatase activity of PP5 can be inhibited or activated in vitro by OA³¹  and AA,³²  respectively (Figure 4A). Indeed, the phosphorylation of ASK1 at T838 in the ASK1-inducible H929 cells treated with Dox was blocked by cotreatment with AA in a dose-dependent manner (Figure 4B). Furthermore, the ASK1-induced apoptosis could be prevented by cotreating ASK1-inducible H929 cells with Dox and AA (Figure 4C). Reciprocally, ASK1 in H929 cells could be phosphorylated at T838 3 days after OA treatment (Figure 4D), which dose and time dependently led to enhanced apoptosis of CD138⁺ human plasma cells isolated from bone marrow of patients with MM (Figure 4E). This effect largely relied on the induction of ASK1 activity by OA treatment because cells depleted of ASK1 showed reduced OA-mediated apoptosis compared with cells treated with ctrli (Figure 4F). Taken together, our data showed that regulation of ASK1 activity can modulate MM cell survival.

We next investigated whether induction of ASK1 decreases MM cell growth in vivo with the use of a xenograft model of human MM. We adopted a luciferase-based, noninvasive bioluminescent imaging system in a SCID mouse–human MM xenograft model in which the bioluminescence in each individual mouse could be quantified by real-time photon emission that correlates with MM cell number. ASK1-inducible H929 or IM9 transfectants stably expressing luciferase were intravenously xenografted into SCID mice that had received a sublethal dose of radiation. At 6 days after transplantation, mice that had received ASK1-inducible IM9 transfectants were administrated Dox, and in vivo bioluminescent imaging analysis showed that mice carrying ASK1-inducible transfectants or control TRE transfectants exhibited undistinguishable MM cell load before Dox treatment on day 6. The induction of ASK1 after Dox administration in vivo was validated by immunoblotting with the use of lysates from extramedullar solid tumors of mice that received a transplant (supplemental Figure 3A). Remarkably, mice bearing ASK1-inducible IM9 transfectants showed significantly reduced bioluminescence after Dox administration than did mice without Dox treatment on day 20. By contrast, mice engrafted with control TRE IM9 transfectants in the presence or absence of Dox administration had similar bioluminescence (Figure 5A). In addition to the diminished MM cell growth in vivo on ASK1 induction, overall survival was significantly prolonged in mice that received a transplant with ASK1-inducible IM9 transfectants and administrated Dox (Figure 5B) compared with mice that received control TRE IM9 transfectants treated with or without Dox or ASK1-inducible IM9 transfectants without Dox. Similar results, in terms of reduced bioluminescence in vivo (supplemental Figure 3B) and prolonged survival (Figure 5C) after Dox administration, were evident for mice that received a transplant with a stable ASK1-inducible H929 transfectant compared with control TRE H929 transfectants.

On the basis of our results that induction of ASK1 diminished MM cell growth in vivo, we wondered whether the reduced cell growth could reflect enhanced apoptosis. We used a TUNEL assay with lung tissues of engrafted mice because of their strong bioluminescence on day 20 (supplemental Figure 3C). Indeed, a pronounced increase in TUNEL⁺ signals was observed in lung tissues of mice bearing ASK1-inducible H929 transfectants on Dox administration compared with mice without Dox treatment and the control TRE transfectants (Figure 5D).

We next examined the downstream signaling pathways regulated by ASK1-induced apoptosis in long-lived plasma cells. ASK1 activates JNK and p38 MAPKs.¹²  As expected, increased phosphorylation of JNK and p38 MAPK was observed in Dox-treated ASK1-inducible H929 transfectants but not in control TRE transfectants. Of note, ASK1 was phosphorylated at T838 after Dox treatment for various times (Figure 6A). Bcl-2 family proteins are important apoptotic regulators,³³  and the balance between Mcl-1 and Bim is especially crucial for MM cell survival.³⁴  We examined the expression of antiapoptotic and proapoptotic Bcl-2 family proteins after ASK1 induction. We found that total Bim, especially Bim_EL, which is encoded by an alternatively spliced form of Bim mRNA,³⁵  was significantly increased on the induction of ASK1, whereas the levels of Bcl-2, Bcl-xL, and Mcl-1 did not change (Figure 6B). Phosphorylation of Bim_EL at serine (S) 65 is required for JNK- or p38-mediated apoptosis,^36,37  and phosphorylation of Mcl-1 at S159 potentiates Mcl-1 degradation³⁸ ; we thus checked whether the increased Bim_EL was phosphorylated at S65 and if Mcl-1 could be phosphorylated at S159 on ASK1 induction. Indeed, the levels of S65-phosphorylated Bim_EL increased after ASK1 induction; however, Mcl-1 phosphorylation at S159 was not up-regulated (Figure 6B).

We previously showed that knockdown of Blimp-1 caused apoptosis of MM cells with a concomitant increase in Bim.²⁰  We thus speculated that ASK1 is required for Blimp-1 knockdown–induced apoptosis and Bim elevation. Indeed, Blimp-1 knockdown increased expression of Bim, which was diminished when ASK1 was simultaneously depleted (Figure 6C). Furthermore, knocking down ASK1 and Blimp-1 together efficiently blocked Blimp-1 knockdown–mediated apoptosis in H929 cells (Figure 6D), showing that ASK1 is required for Blimp-1 knockdown–induced Bim elevation and apoptosis.

In this study, we identified a new role for ASK1 in apoptotic regulation of plasma cells. ASK1 is activated during differentiation of murine short-lived plasma cells in vitro and contributes to apoptosis of these cells. We hypothesized that ASK1 is directly suppressed by Blimp-1 in long-lived plasma cells to avoid the possibility of activation of stress-induced apoptotic signaling pathways. The physiologic relevance of ASK1 in regulating plasma cell survival was further supported by our findings that Ask1 KO mice produced higher numbers of antigen-specific short-lived and long-lived plasma cells.

The limited lifespan of short-lived plasma cells is attributable to spontaneous apoptosis caused by activation of caspase-12, which is the human caspase-4 counterpart and is involved in ER stress–induced apoptosis.³⁹  Our results showed that ASK1 is phosphorylated, but its expression does not significantly change during differentiation of short-lived plasma cells after treatment with a TD or TI stimulant in vitro. Therefore, ASK1 function may be mainly regulated at the posttranslational level during the generation of short-lived plasma cells, thereby providing a means to rapidly respond to ER stress. Accordingly, short-lived plasma cells generated from B cells of Ask1 KO mice appeared to survive longer, and more antigen-specific short-lived plasma cells were found in Ask1 KO mice after immunization; this was not due, however, to abnormal B-cell development in those mice because Ask1 KO and control mice were comparable for the distribution of various splenic B-cell subsets (data not shown). In malignant counterparts of long-lived plasma cells, however, the regulation of ASK1 may be controlled at the transcriptional level by Blimp-1–mediated suppression. Enforced induction of ASK1 expression was sufficient to induce MM cell death in vitro and in vivo, and the latter led to reduced MM cell number and prolonged the survival of engrafted animals. However, residual ASK1 was still detectable in MM cells, suggesting that the chromatin environment of the ASK1 locus is not completely silenced in MM cells. Supporting this notion, Blimp-1 knockdown in MM cells could reestablish ASK1 expression. In addition, the residual ASK1 could be activated by treating cells with OA; as a consequence, apoptosis was triggered in the MM cell line or in MM cells from patients, indicating that ASK1 function is manipulatable in MM cells. However, why Blimp-1 induction during short-lived plasma cell differentiation was not able to repress Ask1 transcription remains to be mechanistically investigated. It could be that Blimp-1 binding to the Ask1 locus in short-lived plasma cells is not sufficient to repress transcription. Supporting this notion, we found that in short-lived plasma cells generated by stimulation of splenocytes with LPS, the region in mouse Ask1 that contained the Blimp-1 binding site (site 2) still contained high levels of acetylated histone (H)3, a marker for transcriptional activation, compared with those in day 0 culture (supplemental Figure 4).

Aberrant activation of ASK1 is involved in various diseases, such as cardiac hypertrophy, diabetes, and neuronal degeneration.^11,40,41  ASK1 is also involved in tumor necrosis factor–induced generation of reactive oxygen species.⁴²  Thus, ASK1 is required for inflammation because bone marrow–derived dendritic cells from Ask1 KO mice exhibit an attenuated inflammatory response.²²  Here, we showed that suppression of ASK1 expression or reduction of its activity is required for the survival of short-lived plasma cells, long-lived plasma cells, and MM cells. Particularly, abnormal survival of short-lived or long-lived plasma cells correlates with autoimmune diseases in which autoantibodies are overproduced.⁴³  Therefore, regulation of ASK1 activity could be a potential therapeutic strategy for modulating immune responses, malignant plasma cell survival, or the above-mentioned diseases.

In our previous study, Blimp-1 knockdown–mediated apoptosis of MM cells was accompanied by increased levels of Bim and decreased Mcl-1.²⁰  In the present study, we show that the level of Bim, but not of Mcl-1, is altered in response to ASK1 induction in MM cells, suggesting that an elevated level of Bim after Blimp-1 knockdown has its effect downstream of ASK1 activation. Supporting this notion, we found that the increase in Bim level was reversed when Blimp-1 and ASK1 were simultaneously depleted in MM cells. However, Mcl-1 expression was not sensitive to ASK1 induction. In fact, the regulation of Mcl-1 expression is also controlled by ER stress.⁴⁴  The Mcl-1 level decreases because of cessation of mRNA translation through phosphorylation of eukaryotic initiation factor 2α,⁴⁴  whose activation is triggered by the RNA-dependent protein kinase–like ER protein kinase pathway on ER stress.⁴⁵  We previously showed that Blimp-1 knockdown–induced apoptosis could trigger caspase-4 activation,²⁰  a hallmark for ER stress–induced apoptosis.⁴⁶  Therefore, taken all together, we think that Blimp-1 knockdown–induced apoptosis may be influenced by 2 pathways in ER stress response: the protein kinase–like ER protein kinase pathway, which leads to a pause of Mcl-1 translation, and the inositol-requiring transmembrane kinase/endonuclease 1 (IRE1) pathway because ASK1 is one of its downstream effectors.⁴⁵  IRE1 is particularly important in the context of plasma cells because its activation triggers the splicing of X-box–binding protein-1 mRNA, which encodes an active transcription factor.⁴⁵  X-box–binding protein-1 is required for the production of immunoglobulins⁴⁷  and for MM cell survival in response to ER stress.⁴⁸  Accordingly, disruption of IRE1 in mice causes decreased levels of serum immunoglobulins produced by plasma cells.⁴⁹  In addition to up-regulation of Bim expression, phosphorylation of Bim at S65 was enhanced after forced expression of ASK1 in MM cells. It is known that phosphorylation of S65 in Bim is required for p38 MAPK– or JNK-mediated apoptosis^36,37  and for its release from binding to Mcl-1.⁵⁰  Taken together, the evidence suggests that ASK1 induction–mediated apoptosis of MM cells also is potentiated by the enhanced phosphorylation of Bim at S65 by JNK and p38 MAPK, which act downstream of ASK1; as a consequence, Bim can be released from binding to Mcl-1 to execute apoptosis.

In summary, we here show a new role for ASK1 in regulating apoptosis of normal and malignant plasma cells. We rationalized that long-lived or malignant plasma cells may require sustained suppression of ASK1 by Blimp-1 to overcome the activation of stress-induced apoptosis. Our study suggests the importance of regulating the ASK1 signaling axis for optimal control of plasma cell numbers.

The authors thank Dr W.-H. Lee for discussion and S.-Y. Yang for technical assistance.

This work was supported by Academia Sinica (AS-99-CDA-L12, K.-I.L.) and the National Science Council (100-2628-B-001-015-MY4, K.-I.L.), Taiwan.

Contribution: F.-R.L. and K.-I.L. designed the experiments, analyzed the data, and wrote the paper; F.-R.L., K.-H.H., S.-T.S., and C.-H.C. performed the experiments; and S.-Y.H., A.M., M.H., and H.I. provided important materials.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Kuo-I Lin, Genomics Research Center, Academia Sinica, Taipei, 115, Taiwan; e-mail: kuoilin@gate.sinica.edu.tw.