Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

γ-secretase directly sheds the survival receptor BCMA from plasma cells

γ-secretase directly sheds the survival receptor BCMA from plasma cells ARTICLE Received 27 Feb 2015 | Accepted 28 Apr 2015 | Published 11 Jun 2015 DOI: 10.1038/ncomms8333 OPEN g-secretase directly sheds the survival receptor BCMA from plasma cells 1 1, 2,3, 4 5 Sarah A. Laurent , Franziska S. Hoffmann *, Peer-Hendrik Kuhn *, Qingyu Cheng , Yuanyuan Chu , 5 6 1 1 1 Marc Schmidt-Supprian , Stefanie M. Hauck , Elisabeth Schuh , Markus Krumbholz , Heike Ru¨bsamen , 2,3 7 7 4 4 8 Johanna Wanngren , Mohsen Khademi , Tomas Olsson , Tobias Alexander , Falk Hiepe , Hans-Walter Pfister , 9 10 11,12 1,12 2,3,12 Frank Weber , Dieter Jenne , Hartmut Wekerle , Reinhard Hohlfeld , Stefan F. Lichtenthaler & Edgar Meinl Survival of plasma cells is regulated by B-cell maturation antigen (BCMA), a membrane-bound receptor activated by its agonist ligands BAFF and APRIL. Here we report that g-secretase directly cleaves BCMA, without prior truncation by another protease. This direct shedding is facilitated by the short length of BCMA’s extracellular domain. In vitro, g-secretase reduces BCMA-mediated NF-kB activation. In addition, g-secretase releases soluble BCMA (sBCMA) that acts as a decoy neutralizing APRIL. In vivo, inhibition of g-secretase enhances BCMA surface expression in plasma cells and increases their number in the bone marrow. Furthermore, in multiple sclerosis, sBCMA levels in spinal fluid are elevated and associated with intracerebral IgG production; in systemic lupus erythematosus, sBCMA levels in serum are elevated and correlate with disease activity. Together, shedding of BCMA by g-secretase controls plasma cells in the bone marrow and yields a potential biomarker for B-cell involvement in human autoimmune diseases. 1 2 Institute of Clinical Neuroimmunology, Ludwig Maximilian University Munich, 81377 Munich, Germany. Neuroproteomics, Klinikum rechts der Isar, and Institute of Advanced Study, Technische Universita¨tMu¨nchen, 81377 Munich, Germany. German Center for Neurodegenerative Diseases (DZNE), 81377 Munich, Germany. Department of Rheumatology and Clinical Immunology, Charite - Universita¨tsmedizin Berlin and Deutsches Rheuma-Forschungszentrum Berlin-a Leibniz Institute, 10117 Berlin, Germany. Department of Internal Medicine III, Klinikum Rechts der Isar, Technische Universita¨tMu¨nchen, 81675 Munich, Germany. Research Unit Protein Science, Helmholtz Zentrum Mu¨nchen (GmbH), German Research Center for Environmental Health, 85764 7 8 Neuherberg, Germany. Karolinska University Hospital, Division of Clinical Neuroscience, 17176 Stockholm, Sweden. Department of Neurology, Klinikum 9 10 Grosshadern, Ludwig Maximilian University Munich, 81377 Munich, Germany. Max-Planck-Institute of Psychiatry, 80804 Munich, Germany. CPC 11 12 Helmholtz Zentrum Munchen (GmbH), 81377 Munich, Germany. Max-Planck-Institute of Neurobiology, 82152 Martinsried, Germany. Munich Cluster for Systems Neurology (SyNergy), 81377 Munich, Germany. * These authors contributed equally to this work. Correspondence and requests for materials should be addressed to E.M. (email: edgar.meinl@med.uni-muenchen.de). NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications 1 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 cells are of pathogenic relevance and serve as a therapeutic 2.0 target both in generalized immunopathological diseases *** Bsuch as systemic lupus erythematosus (SLE), and in 1.5 1,2 *** organ-specific diseases such as multiple sclerosis (MS) . 20 1.0 Activation and survival of B cells is largely regulated via the BAFF–APRIL system that comprises three receptors (BAFF-R, 10 0.5 TACI and BCMA (B-cell maturation antigen)) and two ligands 3 0 0 (BAFF and APRIL) . Membrane-bound BCMA (mBCMA) is OND CIS/MS HC OND CIS/MS NB 4 5,6 expressed on some activated B cells and Ig-secreting cells ;it n=20 n=37 n=26 n=20 n=37 n=5 binds both BAFF and APRIL . BCMA is essential for the 7,8 maintenance of long-lived plasma cells , an effect mediated by 4 150 * 9,10 APRIL or BAFF . These plasma cells produce IgG that protect *** * not only against pathogens but are also critically involved in 11,12 autoimmune diseases . Further, mBCMA engagement on activated B cells induces MHC class II, enhancing their ability to present antigen . The BAFF–APRIL system is targeted for therapeutic interven- 0 0 tion. In SLE, where BAFF levels in the serum are elevated, an 0.0 0.5 1.0 1.5 2.0 HC SLE SLE 13,14 n=29 n=17 n=22 –1 antibody-binding BAFF is already approved . On the other sBCMA (ng ml ) hand, the recombinant soluble receptor atacicept, which targets both BAFF and APRIL, unexpectedly worsened MS , indicating that essential features of this system are not fully understood. Other clinical trials targeting the BAFF–APRIL system in 13,14 immunopathological disorders have been launched . We aimed at identifying features of humoral immunity that 10 50 might be altered in autoimmune diseases. Thereby, we found that a soluble form of BCMA (sBCMA) is regularly detectable in 0 0 0 50 100 150 0 50 100 150 human blood. We then analysed sBCMA in human autoimmune –1 –1 sBCMA (ng ml ) sBCMA (ng ml ) diseases. In MS, sBCMA was elevated in the cerebrospinal fluid (CSF) and linked to local IgG production inside the brain. Figure 1 | sBCMA as a biomarker. (a) sBCMA plasma concentrations In SLE, sBCMA was systemically elevated and associated with were determined using ELISA in healthy controls (HC), patients with a disease activity. We went on to uncover the underlying clinically isolated syndrome (CIS) or MS, or other neurological diseases biochemical mechanism and found that BCMA was directly shed (OND). (b) sBCMA in the CSF was determined in patients with OND, by g-secretase, a ubiquitous intramembranous protease. Direct CIS/MS or neuroborreliosis (NB) (***Po0.001, Kruskal–Wallis test shedding of a membrane protein without prior processing by followed by Dunn’s Multiple Comparison Test). (c) sBCMA in the CSF another protease is a novel function of g-secretase, which is best correlated strongly with the intrathecal IgG production represented by the known for processing of amyloid precursor protein (APP) in IgG Index. This correlation was evident when all analysed CSF samples were 16,17 . To analyse the functional Alzheimer’s disease and Notch considered (Po0.0001, r ¼ 0.85) and in the MS/CIS group (Po0.0001, relevance of this shedding in vivo, we treated mice with a r ¼ 0.77, Spearman rank correlation, CIS/MS n¼ 36; OND n ¼ 20, g-secretase inhibitor; this enhanced surface BCMA on plasma NB n ¼ 5). (d) sBCMA in the serum was determined with ELISA in HC, cells and increased their number in the bone marrow. Thus, untreated (red) and treated (black) patients with SLE. sBCMA was elevated shedding of BCMA via g-secretase is an immunoregulatory in SLE patients and in the untreated SLE patients compared with the treated mechanism limiting plasma cells. patients (***Po0.001 and *Po0.05, Kruskal–Wallis test followed by Dunn’s Multiple Comparison Test). (e) sBCMA in the serum of SLE patients correlated strongly with disease activity quantified with SLE disease activity Results index (SLEDAI; Po0.001; r ¼ 0.54, Spearman correlation). (f) sBCMA in sBCMA shows local IgG production in MS and activity in SLE. the serum of SLE patients inversely correlated with the level of the We found sBCMA as a regular component of human blood from complement factor C3 (P ¼ 0.0374, r¼ 0.29, Spearman correlation). healthy subjects (Fig. 1a). In MS, sBCMA was elevated in the CSF, Bars represent means. but not in the blood (Fig. 1a,b). sBCMA levels in the plasma and serum were very similar as seen in 14 controls (Supplementary Fig. 1a). Levels of sBCMA in the CSF correlated strongly with sBCMA in the CSF (Supplementary Fig. 1b). High-dose steroids local IgG production in MS patients (Fig. 1c). This strong for treatment of acute relapses reduced serum sBCMA levels correlation was confirmed in a second cohort of 25 MS patients transiently (Supplementary Fig.1c). (Po0.0001, r ¼ 0.80, Spearman rank correlation). Also in In SLE, serum levels of sBCMA were elevated as seen in an neuroborreliosis, an infectious neuroinflammatory disease char- untreated and in a treated cohort (Fig. 1d). Immunosuppressive acterized by chronic IgG production within the CNS, sBCMA treatment of SLE patients reduced sBCMA levels (Fig. 1d). We levels were increased and correlated with local IgG production noted a strong correlation of serum sBCMA and disease activity (Fig. 1b,c). In CSF from MS patients, there was a slight inverse (Fig. 1e) and an inverse correlation with the paraclinical marker correlation of sBCMA with its high-affinity ligand APRIL complement factor 3 (Fig. 1f). There was a trend (P¼ 0.0767, (r¼ 0.35, P ¼ 0.033, Spearman rank correlation), but r¼ 0.23, Spearman rank correlation) for correlation between no association to the plasma level of sBCMA (r ¼ 0.07, anti-dsDNA titre and sBCMA in these SLE patients. Serum BAFF P ¼ 0.667, Spearman rank correlation). We analysed effects of levels were elevated in our SLE cohort (mean¼ 0.61 0.13 ng ml immunosuppressive treatment on sBCMA levels longitudinally in in healthy controls; mean¼ 2.11 4.32 ng ml in SLE patients, two cohorts of MS patients. Natalizumab, which blocks entry of P¼ 0.0004, Wilcoxon–Mann–Whithney test), and correlated with 18 2 lymphocytes into the CNS and local IgG production , reduced sBCMA levels (P¼ 0.0013, r¼ 0.47, Spearman rank correlation). 2 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. –1 SLEDAI IgG index sBCMA (ng ml ) –1 –1 –1 C3 (mg dl ) sBCMA (ng ml ) sBCMA (ng ml ) NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 ARTICLE BCMA is shed during differentiation to Ig-secreting cells. Since Again, differentiation towards IgG-secreting cells was mBCMA is known to be expressed on some activated B cells and accompanied by the appearance of sBCMA (Fig. 2b). 5,6 Ig-secreting cells , we analysed the release of sBCMA by After activation with CD40L þ IL-21, two B-cell populations, þ  þ þ primary human B cells. We applied two different protocols to CD19 CD38 and CD19 CD38 , cells could be distin- activate human primary B cells in order to differentiate them guished. We directly compared mBCMA and release of sBCMA towards Ig-secreting cells. First, purified blood-derived human by these different B-cell subsets. While mBCMA was absent on þ  þ B cells were activated via CD40L and further differentiated to unstimulated B cells, CD19 CD38 cells weakly and CD19 Ig-secreting cells by adding interleukin (IL)-21 (ref. 5). While CD38 cells strongly expressed mBCMA (Fig. 2c). We sorted human B cells activated via CD40L alone released low, but CD38 and CD38 cells, cultured them for another 24 h detectable, levels of sBCMA, this was strongly enhanced by without further stimulation and determined the amount of shed addition of IL-21 (Fig. 2a). In the second primary B-cell culture sBCMA (Fig. 2d). This revealed a close correlation between system we activated peripheral blood mononuclear cells (PBMC) released sBCMA and surface expression of mBCMA (Fig. 2e). with the TLR7 þ 8 ligand R848 and IL-2, which induces The transcription of BCMA in these B-cell subsets was further 19 þ differentiation of human memory B cells to Ig-secreting cells . substantiated with qPCR. BCMA transcript levels in the CD38 6 14 7 3.5 IgG IgG 12 6 sBCMA 3.0 sBCMA 5 2.5 8 4 2.0 6 3 1.5 4 2 1.0 2 1 0.5 0 0 0 0.0 Vehicle +CD40L +CD40L Vehicle R848 IL-21 IL-2 No stimulation CD40L+IL-21 stimulation + + – + + CD19 CD19 CD38 CD19 CD38 100 100 100 6 80 80 80 60 60 40 40 20 20 0 0 0 2 3 4 5 2 3 4 5 2 3 4 5 + – + 010 10 10 10 010 10 10 10 010 10 10 10 CD19 CD38 CD38 BCMA No stimulation CD40L+IL-21 12 3 4 50 IgH IgL 4.6 sBCMA 3.5 kDa kDa 0 50 100 150 200 250 BCMA surface expression (MFI) Extracellular domain Transmembranous domain N -MLQMA GQCSQ NEYFD SLLHA CIPCQ LRCSS NTPPL TCQRY CNASV TNSVK GTNAI LWTCL GLSLI ISLAV FVLMF LLRKI… Figure 2 | sBCMA is released when B cells differentiate towards plasma cells and comprises the extracellular domain plus part of the transmembranous region of BCMA. (a) Human purified B cells were activated for 5 days as indicated; IgG and sBCMA in the supernatant were quantified using ELISA. Combined data of three independent experiments (mean s.e.m., P ¼ 0.0073, paired t-test). (b) PBMCs were stimulated with R848 þ IL-2 for 7 days. IgG and sBCMA in the supernatant were quantified using ELISA. Combined data of three independent experiments (mean s.e.m., P ¼ 0.0227, paired t-test). (c–e) Human purified B cells were stimulated with CD40L þ IL-21. (c) surface BCMA was measured using flow cytometry on unstimulated B þ  þ þ þ cells, CD19 CD38 cells and CD19 CD38 cells. (d) Sorted CD38 and CD38 cells were cultured for another 24 h and the amount of shed sBCMA was measured using ELISA, combined data of two independent experiments. (e) Correlation between sBCMA release and surface expression of BCMA for a single replicate. (f,g) sBCMA was immunoprecipitated from supernatant of plasmacytoma cells (f, lanes 1, 2), serum (f, lane 3) and from supernatant of human purified B cells cultured with CD40L plus IL-21 (f, lane 4) with anti-BCMA monoclonal antibodies (mAbs) A7D12.2 (f, lanes 1 and 4) or C12A3.2 (f, lane 2) or goat-anti-BCMA (f, lane 3). Western blot analysis for BCMA (f) and silver staining of sBCMA immunoprecipitated from plasmacytoma supernatant (g) was performed. (h) The band at 6 kDa (from g) and sBCMA obtained using immunoprecipitation were analysed with mass spectrometry. The aa sequences of BCMA and peptides identified after tryptic (blue) or chymotryptic (red) digestion are shown. No peptide was detected with a C-terminal aa that was not a site for either tryptic or chymotryptic cleavage, indicating that the precise cleavage site of g-secretase within the membrane needs to be identified. NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications 3 & 2015 Macmillan Publishers Limited. All rights reserved. % of Max –1 –1 sBCMA (ng ml ) IgG (μg ml ) % of Max –1 sBCMA (ng ml ) –1 IgG (μg ml ) % of Max sBCMA (ng/ml) –1 sBCMA (ng ml ) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 cells reached 15.9 5.2% peptidyl-prolyl isomerase A (PPIA), while BCMA expression in the CD38 cells subset accounted for 80 80 ± ± 2.5 1.3% PPIA (mean s.e.m. of three independent replicates). Thus, in these B-cell subsets BCMA surface expression reflected BCMA transcription. 40 40 We also analysed BCMA shedding in tumour cell lines and 20 20 transfectants. The human plasmacytoma cell line JK-6L sponta- 0 0 Vehicle 0.01 0.03 0.1 0.3 1 Vehicle 1 3 10 25 50 neously shed sBCMA (Supplementary Fig. 2a). In BCMA- DAPT (μM) TAPI-1 (μM) transfected HeLa cells surface expression of mBCMA was accompanied by release of sBCMA (157 6ngml ) without Vehicle Vehicle requiring any further stimulus. HeLa cells did not secrete TAPI-1 DAPT Isotype control Isotype control detectable amounts of APRIL or BAFF, neither spontaneously 10 100 nor after transfection with BCMA or an empty vector. Together, our observations with primary human B-cell cultures, plasmacy- 60 60 toma cells and BCMA-transfected cells indicate that release of sBCMA is a direct consequence of surface expression of mBCMA; 20 20 it does not require additional stimulation or ligand binding. 2 3 4 5 2 3 4 5 2 3 4 5 0 10 10 10 10 010 10 10 10 010 10 10 10 sBCMA comprises extracellular and intramembranous part. CD27 BCMA sBCMA was isolated by immunoprecipitation from the super- natant of primary Ig-secreting cells, plasmacytoma cells or serum; 120 120 in all these sources, sBCMA had a molecular weight (MW) of 100 100 B6 kDa as seen using western blot analysis (Fig. 2f). This size was 80 80 confirmed when silver staining was applied to detect material 60 60 obtained by immunoprecipitation with anti-BCMA from the 40 40 supernatant of plasmacytoma cells (Fig. 2g). This corresponds to the extracellular part of BCMA (54 amino acid (aa), calculated 20 20 MW 5.8 kDa). Unexpectedly, mass spectrometry revealed that 0 0 Vehicle 1 3 10 25 50 Vehicle 0.01 0.03 0.1 0.3 1 sBCMA comprised not only the complete extracellular domain TAPI-1 (μM) DAPT (μM) with an intact N terminus, but also part of the transmembrane region (Fig. 2h). This indicated that it was released by an intra- Vehicle Vehicle membranous protease. TAPI-1 DAPT Isotype control Isotype control c-secretase inhibitors block BCMA shedding from B cells. Since 10 mBCMA is a type-I oriented transmembrane protein with an 4 60 60 extracellular N terminus, g-secretase was a candidate for its intramembranous cleavage. We applied the g-secretase inhibitor 10 20 20 DAPT and compared it with the metalloprotease inhibitor TAPI- 2 3 4 5 2 3 4 5 1, which reduces the shedding of other TNFR-SF members. We 2 3 4 5 010 10 10 10 010 10 10 10 010 10 10 10 activated human B cells either via CD40L þ IL-21 (Fig. 3a,b) or CD19 BCMA via R848 þ IL-2 (Fig. 3c,d), and used both fluorescence-activated Figure 3 | c-secretase inhibitor DAPT reduces release of sBCMA and cell sorting (FACS) and enzyme-linked immunosorbent assay enhances surface expression of BCMA on activated human B cell. (ELISA) as read-out systems to quantify mBCMA and sBCMA. (a,b) Human B cells were differentiated into Ig-secreting cells via DAPT blocked the release of sBCMA even at low concentrations, CD40Lþ IL-21. (a) Release of sBCMA on treatment with DAPT or TAPI-1 while TAPI-1 had little or no effect (Fig. 3a,c). After CD40L þ IL- was measured using ELISA. sBCMA release was normalized to the amount 21 application, a high surface expression of mBCMA was noted in þþ þ of sBCMA shed under vehicle conditions, which was set as 100%. the CD27 CD38 subset (Fig. 3b), previously classified as late 20 ± Combined data of three independent experiments (mean s.e.m.). plasmablasts . DAPT enhanced surface expression of mBCMA (b) These activated primary human B cells were subgrouped on the basis of in these cells, while TAPI-I had little or no effect (Fig. 3b). When expression of CD27 and CD38. A high surface expression of BCMA was human PBMCs were activated with R848 þ IL-2, B20% of the þþ þ þ þ seen on the CD27 CD38 subset. Surface expression of BCMA was cells were CD19 CD38 after 7 days (Fig. 3d). These cells enhanced using DAPT treatment (1 mM), while TAPI-I (50 mM) had little strongly expressed mBCMA on their surface and this was greatly effect. (c,d) Human PBMCs were stimulated with R848 þ IL-2 for 7 days. enhanced by the g-secretase inhibitor DAPT; again, TAPI-I had (c) Release of sBCMA on treatment with DAPT or TAPI-1 was measured little effect (Fig. 3d). Similar to primary human B cells, we using ELISA. sBCMA release was normalized to the amount of sBCMA shed observed differential effects of DAPT and TAPI-1 on the release under vehicle conditions, which was set as 100%. Combined data of three of sBCMA and surface expression of mBCMA on human independent experiments (mean s.e.m.). (d) High surface expression of plasmacytoma cells (Supplementary Fig. 2a,b). þ þ BCMA was seen on the CD19 CD38 subset; this was further enhanced Further, we compared the effect of transition (LY-411575-I and by DAPT (1 mM), while TAPI-I (50 mM) had little effect. LY685,458) and non-transition state (DAPT and RO4929097) inhibitors of the g-secretase on BCMA shedding from human B cells. Human PBMCs were first stimulated with R848 þ IL2 for 7 days, and then CD19 B cells were positively selected and c-secretase directly sheds BCMA. Presenilin (PS)1 or PS2 is the 14,15 cultured overnight in the absence of these g-secretase inhibitors. catalytical component of the g-secretase complex . To finally We found that RO4929097, LY-411575-I and LY685,458 had prove that sBCMA is released by g-secretase, we switched from B similar effects as DAPT on the shedding of mBCMA as seen with cells to mouse embryonic fibroblasts (MEF) deficient for both PS1 both read-out systems, FACS and ELISA (Supplementary Fig. 3). and PS2 (PS  /  ) . These MEF cells were transduced with 4 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. sBCMA release (%) sBCMA release (%) CD38 CD38 % Of max % Of max sBCMA release (%) sBCMA release (%) % Of max % Of max NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 ARTICLE full-length human BCMA plus either wt PS1, or its catalytically short extracellular domain of BCMA facilitates its direct cleavage inactive mutant D385A (ref. 22). Cleavage of mBCMA and by g-secretase. release of sBCMA occurred only in the presence of an active g-secretase complex as seen with FACS, ELISA and western blot c-secretase regulates NF-jB-mediated activation via BCMA. analysis (Fig. 4a–d). To date, g-secretase is known to cleave only The reduced surface expression of mBCMA mediated by 16,17 membrane proteins after a previous cut by other proteases . g-secretase was linked to a reduced binding of APRIL This was in conflict with our mass spectrometry analysis (Supplementary Fig. 2c). To analyse whether the DAPT-increased indicating that sBCMA had an intact N terminus (Fig. 2e). mBCMA expression was accompanied by enhanced Therefore, we used an additional experimental approach to assure cellular responsiveness, we analysed NF-kB activation in that g-secretase cleaved mBCMA without prior N-terminal BCMA-transfected HEK cells. DAPT increased BCMA-mediated trimming. We transfected a cDNA construct coding for BCMA NF-kB activation induced via APRIL (Fig. 6a) and also via tagged with an N-terminal FLAG. Using an ELISA with anti- BAFF (Fig. 6b). FLAG for coating and anti-BCMA for detection, we found the release of sBCMA with N-terminal FLAG, which was blocked by sBCMA acts as a decoy inhibiting APRIL in vitro.We found the g-secretase inhibitor DAPT (Fig. 5a,b). This provides further that human sBCMA bound APRIL, but not BAFF (Fig. 7a). This is evidence that g-secretase sheds BCMA without prior N-terminal 3,23 different to mBCMA, which binds both APRIL and BAFF .We trimming by another protease. found that the naturally occurring sBCMA differed from To analyse the importance of the length of the short recombinant BCMA-Fc, which bound APRIL and BAFF; extracellular domain of BCMA (54 aa) for its direct cleavage by however, binding to APRIL was stronger (Fig. 7a). Differences g-secretase, we doubled its extracellular part (variant BCMA– between sBCMA and BCMA-Fc were also seen in NF-kBreporter BCMA) with an extracellular part of 108 aa. We transfected wild- assays: in HEK cells transfected with BCMA, APRIL-mediated type BCMA and BCMA–BCMA in HEK293T cells. While the NF-kB activation was blocked by both sBCMA and BCMA-Fc, g-secretase inhibitor DAPT strongly enhanced the surface while BAFF-mediated activation via mBCMA was suppressed only expression of transfected wild-type mBCMA, DAPT had little by BCMA-Fc, but not by sBCMA (Fig. 7b). Since APRIL can signal or no effect on surface expression of BCMA–BCMA (Fig. 5c,d). via mBCMA and TACI , we transfected HEK cells also with TACI. Similarly, when we measured the released sBCMA, we found that These experiments showed that sBCMA inhibited APRIL, but not DAPT strongly decreased its release from wild-type BCMA, but BAFF-mediated signalling via TACI (Supplementary Fig. 4). had little effect on shedding of BCMA–BCMA (Fig. 5e). Thus, the Further, sBCMA limited APRIL-mediated survival of activated primary B cells (Fig. 7c). The effects of BCMA shedding by 100 g-secretase are summarized in a diagram (Fig. 7d). 6,000 PS –/– PS –/– BCMA 60 c-secretase regulates plasma cells and B cells in vivo. We tested 4,000 PS –/– BCMA whether cleavage of mBCMA by g-secretase occurred also in vivo PS1 D385 using two different mouse models. First, we immunized mice with PS –/– BCMA 2,000 PS1 ovalbumin (OVA) þ lipopolysaccharide (LPS) to induce a 2 3 4 5 0 10 10 10 10 T-dependent immune response and plasma cell differentiation. BCMA Treatment 9 days later with the g-secretase inhibitor LY-411575-I for 1 day enhanced surface expression of mBCMA in CD138 plasma cells in the spleen (Fig. 8a) and bone marrow (Supplementary Fig. 5). In the absence of a g-secretase inhibitor, PS1 holo mBCMA was hardly detectable at all on the surface of murine 30 plasma cells, in line with previous observations . Second, we used an SLE model in NZB/W mice, in which germinal centres develop PS1 CTF 24,25 spontaneously . In addition here, 1 day of treatment with the NCT mature g-secretase inhibitor LY-411575-I enhanced surface expression of NCT immature BCMA on plasma cells in the spleen and bone marrow (Fig. 8b). 20 mBCMA To distinguish between short-lived and long-lived plasma cells, β-actin mice were fed bromodeoxyuridine (BrdU). We found that kDa inhibition of g-secretase for 1 day enhanced mBCMA both on BrdU plasma cells (designated also as short-lived plasmablasts) and on BrdU plasma cells (designated also as long-lived plasma cells) in the bone marrow and spleen (Fig.8b). We then analysed the effects of extended application of the g-secretase inhibitor and treated NZB/W mice for 7 days. This Figure 4 | Release of sBCMA requires active presenilin. (a–d) Presenilin- resulted again in enhanced expression of mBCMA on BrdU deficient MEF cells (PS / ) were transduced with human BCMA and BrdU plasma cells in the bone marrow and spleen (Fig. 8c). (PS / BCMA) and then with wild-type PS1 (PS / BCMA PS1) or This prolonged treatment period enhanced the number of plasma with a catalytically inactive mutant (PS / BCMA PS1-D385A). BCMA cells in the bone marrow, but not in the spleen (Fig. 8d). In the surface expression (a,b) and sBCMA release (d) were determined. In a,a spleen, the absolute number of BrdU plasma cells per organ representative experiment is shown; in b,d, mean s.e.m. of four independent decreased, but not their relative number (Fig. 8d,e). In the bone experimentsisgiven (respectively, P¼ 0.0313 and P¼ 0.0033, paired t-test). marrow, treatment with the g-secretase inhibitor increased both (c) Cells used in a,b,d were analysed by immunoblotting for expression of the absolute and also the relative number of plasma cells full-length PS1 (PS1 ), for autoendoproteolysis of PS1 generating a (Fig. 8d,e). holo C-terminal fragment (PS1 ) reflecting an active state of the g-secretase, for The increase in plasma cell number in the bone marrow after CTF maturation of nicastrin (NCT) and for full-length BCMA (mBCMA). inhibition of g-secretase could be attributed to the regulation of NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications 5 & 2015 Macmillan Publishers Limited. All rights reserved. PS–/– BCMA PS–/– BCMA-D385A PS–/– BCMA-PS1 PS–/– PS–/– BCMA PS–/– BCMA-D385A PS–/– BCMA-PS1 % Of max BCMA (MFI) –1 sBCMA (ng ml ) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 Vehicle DAPT Isotype control Biotinylated anti-BCMA 80 sBCMA 40 40 20 20 Anti-BCMA 2 3 4 5 2 3 4 5 010 10 10 10 010 10 10 10 / / / BCMA wt BCMA-BCMA DAPT DAPT DAPT BCMA Empty vector BCMA BCMA-FLAG 0.5 Anti-FLAG Control-Ig 0.4 ** ** 0.3 Biotinylated anti-BCMA 0.2 sBCMA-FLAG 0.1 Anti-FLAG 0 0 // / DAPT DAPT DAPT Empty vector BCMA BCMA-FLAG Figure 5 | Release of sBCMA occurs without prior N-terminal trimming and is facilitated by the short extracellular domain of BCMA. (a,b) HeLa cells were transfected with plasmids coding for full-length human BCMA or BCMA with an N-terminal FLAG and then cultured with increasing amounts of the g-secretase inhibitor DAPT (0.02, 0.1 and 0.5 mM). Twenty-four hours after transfection supernatants were harvested and the released sBCMA was analysed using ELISA. In (a), ELISA wells were coated with anti-BCMA, and in (b) with anti-FLAG or a control IgG (anti-myelin oligodendrocyte glycoprotein (MOG) 8.18 C5), both were developed with anti-BCMA. Schemes of the ELISAs are shown on the right. Combined data of two independent experiments (mean s.e.m.). (c–e) Human BCMA wild type (wt) or BCMA–BCMA with a doubled extracellular domain of BCMA were transfected into HEK293Tcells. (c,d) Surface expression of BCMA was determined in the absence or presence of the g-secretase inhibitor DAPT (1 mM). (d) The combined data of three experiments (P ¼ 0.0049, **Po0.01, unpaired t-test). (e) The effect of DAPT on the released sBCMA after transfection with BCMA wt or BCMA–BCMA (mean s.e.m. of three experiments), P ¼ 0.0081, **Po0.01, unpaired t-test. activation in vitro (Supplementary Fig. 6c), we gave sBCMA-AT ab three times a week for 4 weeks to NZB/W mice, which had 20 40 P=0.05 Vehicle * Vehicle been immunized with OVA before. The applied amount of DAPT 1 μM DAPT 1 μM 15 30 sBCMA-AT corresponded to the dose previously used for 26,27 BCMA-Fc . Nevertheless, under these conditions sBCMA- 10 20 AT did not modulate the total number of plasma cells, the number of OVA-specific plasma cells or the amount of circulating IgG, IgA and IgM (Supplementary Fig. 6d–f). Notably, inhibition of g-secretase had additional effects on Plasmid Empty BCMA Plasmid Empty BCMA B cells not expressing BCMA. Short-term treatment with Ligand// APRIL APRIL Ligand// BAFF BAFF the g-secretase inhibitor reduced GC B cells (defined as int high CD38 Fas ) in the spleen as seen in our immunization model Figure 6 | c-secretase regulates BCMA-mediated NF-jB activation. (Supplementary Fig. 7a,b). Moreover, after 7 days of treatment, (a,b) HEK293T cells were transfected with full-length human BCMA or an NZB/W mice had fewer B cells in the spleen and bone marrow empty vector. DAPT, APRIL (a) or BAFF (b) were added, and NF-kB and showed a striking reduction of pre-B cells in their bone activation was determined. Combined data of three independent marrow (Supplementary Fig. 7c). experiments (mean s.e.m., *Po0.05, paired t-test). mBCMA on plasma cells (Fig. 8, Supplementary Fig. 4) or on Discussion decoy effects of sBCMA, which we observed in vitro (Fig. 7b,c). We report that g-secretase directly sheds mBCMA and regulates To analyse possible immunoregulatory effects of the shed sBCMA the number of plasma cells in the bone marrow. Further, the in vivo, we produced recombinant sBCMA and applied it released sBCMA reflects B-cell activation in human autoimmu- systemically. We prepared sBCMA as a fusion protein with nity, namely compartmentalized Ig production in MS and disease a1-antitrypsin (AT; Supplementary Fig. 6a) to increase its activity in SLE. molecular weight and therefore its half live in vivo. When B cells are activated and differentiated towards 5,6 The monomeric structure of sBCMA-AT was confirmed using Ig-secreting cells, they start to express mBCMA , which is gel filtration (Supplementary Fig. 6b). Having shown that then shed by g-secretase as we show here. Our experiments with recombinant sBCMA-AT blocked APRIL-mediated NF-kB activated primary human B cells, plasmacytoma cells and BCMA 6 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. BCMA wt BCMA–BCMA BCMA wt BCMA–BCMA NF-κB (fold induction) –1 O.D. 450 nm sBCMA(ng ml ) NF-κB (fold induction) % Of max BCMA expression (fold increase) on DAPT % Of max sBCMA release (fold decrease) on DAPT NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 ARTICLE 0.5 2.0 0.4 Anti-BCMA 1.5 0.3 sBCMA / BCMA-Fc 1.0 Ligand-FLAG 0.2 0.5 Anti-FLAG 0.1 0.0 0.0 sBCMA +– – +– – sBCMA BCMA-Fc BCMA-Fc sBCMA BCMA-Fc – +– – +– PBS –+ – – – + Ligand –– –– APRIL-FLAG BAFF-FLAG ** 9 16 w/o sBCMA w/o sBCMA sBCMA 14 sBCMA 7 BCMA-Fc BCMA-Fc Plasmid BCMA Plasmid BCMA Ligand / APRIL Ligand / BAFF sBCMA BCMA-Fc sBCMA BCMA-Fc sBCMA BCMA-Fc sBCMA BCMA-Fc GSI *** ** BCMA APRIL γ-secretase Membrane B-cell activation and survival B-cell activation and survival sBCMA Figure 7 | sBCMA is a decoy for APRIL in vitro. (a) A scheme of the ELISA is shown on the left; it detects BCMA–APRIL–FLAG (left panel) or BCMA–BAFF–FLAG (right panel) complexes, but neither BCMA nor APRIL–FLAG nor BAFF–FLAG alone alone (*Po0.05, paired t-test). sBCMA was derived from the supernatant of plasmacytoma cultured under serum-free conditions. Combined data of three independent experiments (mean s.e.m.). (b) HEK293T cells were transfected with human BCMA and activated with APRIL (left panel) or BAFF (right panel). sBCMA (50 and 200 ng ml ) was applied as indicated. sBCMA and control supernatant were obtained as mentioned above. BCMA-Fc (50 and 200 ng ml ) was used as a positive control. Combined data of three independent experiments (mean s.e.m., *Po0.05; **Po0.01 paired test). (c) Murine B cells were activated via anti-IgM and cultured for 2 days with APRIL in the presence or absence of sBCMA (200 and 400 ng ml ). APRIL-induced survival was calculated as described in the Methods section. sBCMA was obtained from supernatant from HEK293T cells transfected with full-length BCMA (black bars). Control supernatant was obtained after transfection with an empty vector (white bars). sBCMA significantly inhibited APRIL-mediated survival (***Po0.001 and **Po0.01, paired t-test). Combined data of six independent experiments (mean s.e.m.). (d) Illustration of the consequences of sBCMA shedding by g-secretase: left: an active g-secretase cleaves mBCMA. This reduces the number of membrane-bound BCMA molecules and releases sBCMA, which binds its ligand APRIL functioning as a decoy. Right: g-secretase inhibitors (GSIs) result in elevated mBCMA on the surface and increased APRIL-mediated activation and survival. transfectants indicate that BCMA shedding by g-secretase is a example, NOTCH and APP) that they become accessible to consequence of surface expression of mBCMA; it does not require cleavage only after their extracellular domain has been trimmed 16,17 additional activation or ligand binding; however, we cannot by another protease . We now show that g-secretase can exclude that this shedding can be further enhanced by yet cleave mBCMA directly and that prolonging its extracellular part 16,17 unknown mechanisms. g-secretase is ubiquitously expressed greatly reduced this cleavage. Thus, constitutive cleavage of and accessibility of its transmembranous substrates is largely mBCMA by g-secretase is facilitated by its short (54 aa) regulated by the length of their extracellular part. While extracellular domain. Further we found that g-secretase reduced membrane proteins with an extracellular domain exceeding 100 BAFF and APRIL-mediated NF-kB activation via BCMA in vitro. aa are not efficiently cleaved, extracellular domains of B50 aa To get insight into the in vivo relevance of mBCMA shedding 28,29 permit effective cleavage . g-secretase may remove protein by g-secretase, we used an immunization protocol and an SLE stubs for further degradation and was hence called ‘proteasome of model with spontaneous formation of germinal centres and 16 24,25 the membrane’ , although this notion is controversially plasma cells . We found that inhibition of g-secretase discussed . Probably due to these length restrictions, it has enhances mBCMA on plasma cells in the bone marrow and been a unifying feature of all substrates of g-secretase (for spleen. This was observed in both BrdU and BrdU plasma NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications 7 & 2015 Macmillan Publishers Limited. All rights reserved. APRIL-induced survival (%) NF-κB (fold induction) O.D. 450 nm NF-κB (fold induction) O.D. 450 nm ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 ** *** *** 200 *** *** ** 250 Vehicle LY-411575-I 100 100 40 40 100 20 20 2 3 4 5 2 3 4 5 0 10 10 10 10 0 10 10 10 10 + – + – BCMA Spleen PC BrdU BrdU BM PC BrdU BrdU *** 800 1,500 **** *** **** **** *** *** 1,000 Vehicle LY-411575-I + – + – Spleen B220 PC BrdU BrdU BM B220 PC BrdU BrdU 1.0 *** * 10 ** 2.5 * ** 0.8 1.5 0.6 1.0 4 0.4 0.5 0.2 0.0 0 0.0 + – + – Spleen PC BrdU BrdU BM PC BrdU BrdU Spleen BM Figure 8 | c-secretase regulates plasma cells in mice. (a) Immunized (OVA–LPS in alum) C57/BL6 mice were treated with the g-secretase inhibitor LY-411575-I or vehicle, and the surface display of BCMA in splenocytes was measured using flow cytometry 1 day later. BCMA expression on gated þ þ B220 CD138 cells is shown, a representative example (left) and compiled data from all 17 analysed animals (mean, ***Po0.001, unpaired t-test; right). The black symbols on the right indicate the samples shown on the left. Closed histograms indicate isotype controls. (b) NZB/W mice pretreated with BrdU received the g-secretase inhibitor LY-411575-I (red) or vehicle (blue) for 1 day. Surface expression of BCMA on all CD138 plasma cells (PC) and the BrdU and BrdU PC subgroups in the spleen and bone marrow (BM) was determined using flow cytometry. (c–e) Seven-day treatment of NZB/W mice with LY-411575-I. (c) BCMA surface expression in the spleen and BM on B220, and BrdU and BrdU plasma cells was determined. (d) Absolute number of plasma cells, BrdU and BrdU plasma cells in the spleen and BM. (e) Frequency (% of all cells in the organ) of plasma cells in the spleen and BM. Compiled data from 10 analysed animals per group (mean; *Po0.05; **Po0.01; ***Po0.001; ****Po0.0001, unpaired t-test). cells indicating that both short-lived plasmablasts and long-lived described previously for human recombinant monomeric BCMA, plasma cells are regulated. Application of a g-secretase inhibitor which bound and blocked only APRIL, while recombinant in vivo for 7 days enhanced the number of plasma cells in the human dimeric BCMA-Fc bound both APRIL and BAFF . bone marrow, but not in the spleen. We assume that the Binding assays revealed an avidity contribution of dimeric versus enhanced number of plasma cells in the bone marrow after monomeric recombinant BCMA, which resulted in a41,000-fold g-secretase inhibition is at least partly caused by an enhanced increase in apparent affinity of BCMA binding to BAFF . To get presence of mBCMA, since mBCMA mediates the induction of insight into the immunoregulatory capacity of sBCMA in vivo,we 7,8 the survival protein Mcl-1 in bone marrow plasma cells . The produced a recombinant variant of sBCMA (sBCMA-AT). differential effect of g-secretase inhibition on the plasma cell sBCMA-AT blocked APRIL-mediated NF-kB activation number in the spleen and bone marrow might be explained by the in vitro, but did, in contrast to what was observed with BCMA- previous observation that BCMA induced high expression of Fc previously , not affect plasma cell numbers in the bone Mcl-1 in bone marrow but not in spleen plasma cells .In marrow when given systemically. Murine bone marrow plasma addition, BCMA  /  mice had reduced plasma cell numbers in cells receive survival signals either via BAFF or APRIL , which is the bone marrow, but not in the spleen . consistent with our finding that sBCMA, which blocks mainly Reduction of plasma cell number in the bone marrow by APRIL, has no effect on plasma cell numbers in the bone marrow. g-secretase could be based on reducing membrane-bound Together, we conclude that the effect of BCMA shedding on the mBCMA or, alternatively, on the shed decoy sBCMA. We tested plasma cell number in the bone marrow is largely based on the the decoy potential of sBCMA in vitro and in vivo. In vitro, reduction of surface mBCMA. This would not exclude a function sBCMA blocked APRIL, but had little or no effect on BAFF, while of sBCMA, especially in a mucosal environment as a decoy for BCMA-Fc inhibited both BAFF and APRIL similarly. The APRIL, since APRIL-deficient mice had a reduced IgA response functional features we observed for sBCMA are similar to those to antigens encountered via the mucosal route . 8 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. BCMA (MFI) % Of max PC number (×10 ) % Of max BCMA expression (MFI) BCMA expression (MFI) PC number (×10 ) Frequency of PC (%) BCMA expression (MFI) BCMA expression (MFI) NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 ARTICLE We further found that inhibition of g-secretase affected B-cell albumin)/(serum IgG/serum albumin)). We examined longitudinally the sera of 10 patients with MS who were treated with high doses of steroids (1,000 mg per day subsets in vivo that do not express mBCMA, such as germinal methylprednisolone intravenously for 3–5 days) because of a relapse. Samples were centre B cells. This might be based on a blockade of NOTCH obtained directly before treatment, 3 days and 4 weeks later. We analysed serum pathways, which are known to determine T–B-cell lineage samples from 17 untreated and 22 treated SLE patients. The treated group included 32 33 commitment , maintenance of marginal zone B cells , patients treated with glucocorticosteroid, hydroxychloroquine, azathioprine, cyclophosphamide and mycophenolate mofetyl. Our study included 26 plasma survival of GC B cells and other, not yet identified, substrates 35 samples and 29 serum samples of 34 healthy control donors. Detailed data of the of g-secretase, which also contribute to B-cell activation . patients included in this study can be found in Supplementary Table 1. This study Together, our data indicate that inhibition of g-secretase was approved by the Ethics Committee of the Ludwig Maximilian University of enhances the survival receptor BCMA on plasma cells, increases Munich. Informed consent was obtained according to the Declaration of Helsinki. the number of plasma cells in the bone marrow and disturbs the B-cell compartment in the spleen. Cytokines and stimuli. In all stimulation and binding assays involving APRIL, we used mouse MegaAPRIL (EnzoLifeSciences, Farmingdale, NY), which binds both Release of membrane receptors is a general regulatory human and mouse BCMA and TACI; it is fused to FLAG-tag and is referred in this mechanism of inflammatory responses . TNFR1 is a study as APRIL. BAFF stimulation and binding assays were performed using either prominent example, and the immunoregulatory features of recombinant human BAFF (R&D Systems, Minneapolis, MN) or recombinant TNFR1 shedding were previously explored in vivo with a human BAFF–FLAG (EnzoLifeSciences). For stimulation of PBMC and native B knock-in mouse expressing a mutated nonsheddable TNFR1, cells, human IL-21 (EBioscience, San Diego, CA), TLR7 þ 8 ligand R848 (Sigma- Aldrich, St Louis, MO) and human IL-2 (R&D Systems) were used. Mouse L cells which resulted in autoinflammatory features and enhanced stably transfected with human CD40L were used and expression was continuously experimental autoimmune encephalitis . Biochemically, the monitored. release of sBCMA by g-secretase is different from the shedding of TNFR1, which is performed by the metalloprotease ADAM17 Antibodies and cell lines. To detect human BCMA the following antibodies were (ref. 36), and of TACI that is shed by ADAM10 (ref. 38). used: monoclonal antibodies A7D12.2 (IgG2b) and C12A3.2 (IgG1) (kindly pro- vided by BiogenIdec), and the polyclonal Ab AF193 (R&D Systems). The specificity Does shedding of BCMA also occur in human autoimmune of the two monoclonal antibodies to BCMA was confirmed using flow cytometry of diseases? We measured sBCMA levels in a compartmentalized BCMA-transfected MEF cells (Fig. 4a) and HEK cells (Fig. 5c), immunoprecipi- organ-specific autoimmune disease, MS, and in a systemic disease tation (ip) and subsequent western blot (Fig. 2f,g) or mass spectrometry (Fig. 2h). characterized by general activation of the B-cell compartment, Furthermore, an ELISA using A7D12.2 or C12A3.2 for coating and the polyclonal SLE. A hallmark of MS is B-cell persistence inside the brain goat antibody (AF193; R&D Systems, Minneapolis, MN) for detection with 39,40 recombinant BCMA (R&D Systems) as standard detected sBCMA with a sensitivity compartment with local Ig production , supported by local 41 of 30 pg ml . The following monoclonal antibodies were used for surface production of BAFF by astrocytes . We report that sBCMA is expression analysis using flow cytometry: fluorescein isothiocyanate (FITC)-con- elevated in the CSF in MS and closely correlates with intrathecal jugated anti-CD40L (BD PharMingen, San Diego, CA), FITC-conjugated anti- IgG production. Thus, sBCMA in the CSF reflects the local CD138 (Diaclone, Besanc¸on, France), Cy7-conjugated anti-CD27 and eFluor 450- conjugated anti-CD38 (EBioscience), Cy7-conjugated anti-CD19 (EBioscience). presence of Ig-secreting cells, which are known to be present in 2,39,40,42 For triple staining of BCMA with CD27 and CD38, we used the monoclonal the MS brain . This view is also supported by our antibody A7D12.2 and a 647-conjugated goat-anti-mouse IgG2 Ab (Invitrogen Life observation that reducing inflammation with natalizumab Technologies, Carlsbad, CA) along with the directly labelled IgG1 antibodies to decreases also sBCMA in the CSF. SLE is characterized by a CD38 and CD27 mentioned above. For the detection of APP-derived C99, the monoclonal antibody 4G8, SIG-39220 (Covance, Emeryville, CA) was used. systemic hyperactivation of the B-cell compartment with Single-cell suspensions were prepared from the bone marrow (femur and tibia) elevated surface expression of mBCMA on circulating immune and spleen. Mouse BCMA was detected using flow cytometry with biotin- cells . In this disease, we found a systemic elevation of sBCMA, conjugated anti-mouse BCMA (BAF 593; R&D Systems) along with eFluor450- or which is linked to disease activity. Further studies are required to PE-Cy7-conjugated streptavidin. BrdU staining of plasma cells was performed using a BrdU Flow Kit (BD Biosciences, San Jose, CA) according to the characterize sBCMA as a potential biomarker of SLE activity. manufacturer’s protocol. The detection of plasma cells was carried out with anti- sBCMA shedding adds to the complexity of the BAFF–APRIL CD138-PE (clone 281-2; BD Biosciences) for surface staining and anti-kappa- system. This system is highly relevant in health and disease Pacific Orange (clone 187.1; DRFZ) for intracellular staining. B and T cells were 13,45,46 and serves as a drug target . Release and function of identified with the following anti-mouse antibodies: CD21 (clone 7E9, BioLegend, sBCMA are of direct relevance for clinical trials targeting BAFF, CA, USA), CD23 (clone B3B4, BioLegend), CD24 (clone M1/69, BD Biosciences), CD93 (clone AA4.1, BioLegend), CD95 (clone Jo2, BD Biosciences), CD117 (clone APRIL and their receptors, which are currently under way . This 2B8, BD Biosciences), IgM (clone RMM-1, BioLegend), B220 (clone RA3-6B2, system also affects certain haematological malignancies such as DRFZ), IgD (clone 11-26c, DRFZ), GL-7 (clone GL-7, DRFZ), CD4 (clone GK1.5, 3 47 plasmacytoma , where sBCMA may also serve as a biomarker . DRFZ) and CD8 (clone 53-6.7, BioLegend). Identification of B-cell subsets in the high low/    þ þ Further, since g-secretase is involved not only in Alzheimer’s spleen: B1: IgM CD21 CD23 CD93 . Follicular B cells: IgM CD21 þ  high þ   þ CD23 CD93 . Marginal zone B cells: IgM CD21 CD23 CD93 . GL-7 : disease, but also in autoimmune diseases (including MS), where 48 GL-7 IgD . Identification of B-cell subsets in the bone marrow: pro-B cells: g-secretase inhibitors are being tested for therapeutic benefit , þ þ þ þ þ B220 CD93 CD117 . pre-B cells: B220 CD24 IgM IgD . Immature B our findings draw attention to potential side effects of g-secretase þ þ þ  þ low/ þ þ cells: B220 CD24 IgM IgD . Mature B cells: B220 CD24 IgM IgD . inhibitors related to the shedding of BCMA. Cytometric analysis was performed using a FACSCanto II cytometer (BD Biosciences) and data were analysed with the FlowJo software (Tree Star Inc.). Together, our study yields three main findings. First, Source and working concentration of antibodies used are listed in Supplementary g-secretase directly cleaves mBCMA. Regulation of surface Table 2. display of a ligand-binding receptor is a novel function of MEFs deficient for presenilin 1 (PS1) and presenilin 2 (PS2) (PS  /  ) were g-secretase. Second, g-secretase regulates the number of plasma kindly provided by Dr Bart De Strooper (Leuven, Belgium) ;PS  /  status was monitored using western blot analysis with the antibody PSEN1 (Epitomics, cells in the bone marrow. Third, the released sBCMA is a Burlingame, CA). The plasmacytoma cell line JK-6L (ref. 49) was kindly provided potential biomarker in human immunological diseases and could by Dr Silke Meister (Erlangen, Germany); expression of the surface markers CD138 be useful for therapeutic optimization. and BCMA was controlled. This cell line was cultured in RPMI supplemented with 10% FCS and in serum-free conditions (Hybridoma 6 direkt from Bio&SELL, Nu¨rnberg, Germany). Further, HeLa and HEK293T were applied in transfection Methods experiments. Clinical samples. In a first cohort we obtained plasma and the corresponding CSF from 37 untreated patients diagnosed with either clinical isolated syndrome (n ¼ 10) or MS (n ¼ 27) and from 20 untreated patients with other neurological Detection of BCMA and its ligands APRIL and BAFF. Surface expression of diseases; further five CSF samples from neuroborreliosis patients were obtained. In human BCMA was determined using flow cytometry on FACS Verse (BD Bios- a second cohort, we analysed CSF pairs from 25 additional MS patients before and ciences) using C12A3.2 or A7D12.2 and appropriate secondary antibodies. To about 1 year after continuous natalizumab therapy. The IgG production within the measure human sBCMA in the plasma, CSF or cell culture supernatants, a brain compartment (intrathecal) was calculated as the IgG index ((CSF IgG/CSF sandwich ELISA with polyclonal goat antibodies was used (BCMA/TNFRSF17 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications 9 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 Using lentiviral gene transduction, MEF deficient for PS1 and PS2 (PS  /  ) ELISA Duoset; R&D Systems). To detect sBCMA with an N-terminal FLAG, the were stably transduced with full-length human BCMA and then stably transduced monoclonal antibody anti-FLAG M2 antibody (Sigma-Aldrich) was used for with either wt PS1 or a catalytically inactive D385A mutant of PS1 ( ). coating and the biotinylated polyclonal goat-anti-BCMA antibody (R&D Systems) for detection. As a control the anti-myelin oligodendrocyte glycoprotein mono- clonal antibody 8.18C5 was used for coating. To measure APRIL in CSF, a sand- Enzyme inhibitors. We used the following g-secretase inhibitors: DAPT wich ELISA was used (Bender MedSystem, Vienna, Austria). BAFF levels in the (Calbiochem Merck, Darmstadt, Germany), L685,458 (R&D Systems), RO49290 serum were measured using the Quantokine human BAFF Elisa kit (R&D Systems). (Selleckchem, Houston, TX, USA) and the steroisomer SSR of LY-411575, referred To assess binding of sBCMA or BCMA-Fc (kindly provided by Biogen Idec) to as LY-411575-I (Sigma-Aldrich). This steroisomer of the g-secretase inhibitor APRIL and BAFF, an ELISA was used with the anti-FLAG M2 monoclonal LY-411575 is also a g-secretase inhibitor as observed with the APP fragment C99 antibody (Sigma-Aldrich) for coating to bind FLAG-tagged APRIL or and BCMA as substrates (Supplementary Figures 3 and 8). Metalloproteases were BAFF–FLAG. Then, sBCMA from supernatant of plasmacytoma cells, which was inhibited with TAPI-1 (EMD Chemicals/Calbiochem, Inc. Gibbstown US). concentrated with Amicon Ultra 3 K devices (Merck Millipore Ltd., Ireland), or Corresponding concentrations of DMSO (Sigma-Aldrich) were used as vehicle BCMA-Fc was added and incubated for 2 h at room temperature. BCMA was controls. Reduced BCMA shedding was observed at a vehicle concentration of 0.5% detected as described above. DMSO, which corresponds to 50 mM TAPI-1. Immunoprecipitation and western blot analysis. Immunoprecipitation of NF-jB reporter assay. To measure NF-kB activation, HEK293T cells were BCMA was performed with the monoclonal antibody A7D12.2, the monoclonal transiently transfected with a plasmid containing a firefly luciferase reporter gene antibody C12A3.2 or a polyclonal goat antibody (AF193; R&D Systems). These under the control of an NF-kB transcriptional response element, a plasmid with a antibodies were coupled to Dynabeads Protein G (Life Technologies, AS, Oslo) and Renilla reniformis luciferase reporter gene for normalization and the indicated cross-linked with bis-sulfosuccinimidyl-suberate (Pierce Chemical Co., Rockford, amounts of expression or control plasmids using Lipofectamine 2000 (Invitrogen IL). After successive incubation with either supernatant of plasmacytoma cells or Life Technologies). We used human BCMA or human TACI for transfection. To serum, we eluted with glycine or SDS loading buffer (NuPAGE LDS Sample Buffer, determine the effect of g-secretase inhibition on NF-kB activation, cells transfected Life Technologies). Cells were lysed at 4 C for 1 h in NP-40 lysis buffer (150 mM with BCMA were treated with DAPT or a solvent control and 6 h later stimulated NaCl, 50 mM Tris pH 7.5, 1% Nonidet P-40) containing complete protease with APRIL or BAFF. To analyse a possible decoy function, DAPT along with inhibitor cocktail (Roche Applied Science, Penzberg, Germany). BCMA was BAFF or APRIL were added to BCMA-Fc or supernatants generated by HEK293T detected by western blot analysis with the monoclonal antibody C12A3.2. Blots cells that had been transfected with full-length BCMA or an empty control vector were developed using the mouse true blot goat anti-mouse IgG-HRP system as described above. These supernatants were incubated at 37 C for 30 min and (EBioscience) and enhanced chemiluminescence (ECL). Expression and then added to BCMA or TACI-transfected and DAPT-treated cells used for the endoproteolysis of PS1 was analysed by immunoblotting of cell lysates with the reporter assay. 16 h after stimulation cells were harvested and cell lysates were antibody PSEN1 (Epitomics); maturation of nicastrin (NCT) was evaluated by prepared using passive lysis buffer (Promega, Madison, WI, USA) and the reporter immunoblotting with the antibody N1660 (Sigma-Aldrich). gene activity was measured using firefly luciferase substrate (Biozym, Hameln, Germany) and renilla luciferase substrate (Promega) respectively. Mass spectrometry and sample preparation. sBCMA was purified by immu- noprecipitation and obtained by acidic elution. The eluate was then desalted and Flow cytometry sorting and quantitative RT-PCR. Human B cells were positively concentrated using StageTips, C18, microcolumns (Thermo Scientific, Bremen, selected from PBMC with CD19 MACS beads (Miltenyi Biotec) and differentiated Germany). Then two approaches were followed. (A) The material was digested in into IgG-secreting cells by coculture with CD40L-transfected mouse L cells and solution by trypsin or chymotrypsin and analysed using mass spectrometry (LTQ þþ þ recombinant human IL-21 as described above. CD27 CD38 and CD38 cells Orbitrap XL) as described before . (B) The immunoprecipitated material was were sorted using the BD FACS Aria cell sorting system (Becton Dickinson, separated via an SDS gel, silver-stained and the band corresponding to BCMA as Heidelberg, Germany). For quantitative PCR analysis, RNA was isolated using the detected using western blot analysis was excised and analysed using mass RNeasy Micro Kit (Qiagen, Hilden, Germany) and cDNA was generated using the spectrometry (LTQ Orbitrap XL). High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Darmstadt, Germany). To detect BCMA transcripts, primer and probes were used as published . As housekeeping gene Cyclophilin A (PPIA) was used and detected Cell culture and gene transfer. Two experimental systems were used to differ- with the primer and probes 43263416E from Applied Biosystems. Reactions were entiate primary human B cells to Ig-secreting cells. First, B cells were positively carried out in duplicate using TaqMan assays in combination with the TaqMan selected from PBMC with CD19 MACS beads (Miltenyi Biotec, Bergisch Gladbach, PCR Core Reagent Kit (Applied Biosystems). Samples were run in MicroAmp Germany) and differentiated into IgG-secreting cells by coculture with CD40L- Optical 96-well reaction plates in a 7900HT Fast Real-Time PCR System (Applied transfected mouse L cells and recombinant human IL-21 (50 ng ml ). Further Biosystems). Data were analysed using the SDSv2.3 software (Applied Biosystems). þ þ þ separation of CD19 CD38 and CD19 CD38 cells was performed using CD38 MACS beads. Second, PBMC were stimulated with the TLR7 þ 8 ligand 1  1 19 R848 (2.5 mgml ) and IL-2 (1,000 IU ml ) as described . Secreted IgG was sBCMA-AT fusion protein. The cDNA of the extracellular part of mouse BCMA quantified by ELISA (Mabtech, Nacks Strand, Sweden). To analyse the effects of (sBCMA) was synthesized using the Integrated DNA Technology with a BspEI and DAPT and TAPI-1, these Ig-secreting cells were washed after 7 days and cultured 0 0 0 an Age I site at the 5 and 3 ends, respectively, and fused to the 3 end of human for another 16 h with either inhibitor. AT in the expression vector pTT5 (ref. 52). The resulting fusion between AT and To measure APRIL-induced survival, mouse B cells (magnetically isolated from sBCMA (extracellular residues from position 2 to 49 lacking the first methionine) is spleen using the EasySep Mouse B Cell Isolation Kit (Stemcell Technologie, separated by a short flexible peptide linker TGSGSGA and terminates with Vancouver, Canada)) were cultured in 96-well microtiter plates, which were hexahistidine tag to facilitate purification. The free cysteine residue on the surface 1  1 precoated with anti-IgM (5mgml ). APRIL (100 ng ml ) was added for 48 h and of wild-type AT was replaced by a serine residue to prevent dimerization in cross-linked with anti-FLAG monoclonal antibody (Sigma-Aldrich). Supernatants oxidizing environments. The fusion protein was expressed in HEK293-EBNA cells generated by HEK293T cells that had been transfected with full-length BCMA 52 as described previously . MW was determined with standard Coomassie gel and (OriGene Technologies, Inc., Rockville) or an empty control vector and therefore the size was analysed using gel filtration with the Superose 12 HR 10/30 gel either contained sBCMA or did not, were added at a final sBCMA concentration of filtration column (Amersham/GE Healthcare Life Sciences, Freiburg, Germany) on 1  1 200 ng ml and 400 ng ml . Cell survival was quantified by flow cytometry using a fast protein liquid chromatography column (FPLC). Function of sBCMA-AT was TO-PRO -3 Iodide viability dye (Invitrogen Life Technologies) and APRIL-induced tested using NF-kB reporter assays. survival was calculated as followed: 100 x [(cell survival in presence of APRIL—cell survival without APRIL)/cell survival without APRIL]. To obtain BCMA with an N-terminal FLAG-tag, full-length human BCMA c-secretase inhibitor and recombinant sBCMA-AT in mice. Two different (h184) from a human BCMA cDNA clone (OriGene Technologies) was first mouse models were applied. First, 2-month-old C57/BL6 mice were immunized amplified by PCR. The PCR fragment was then digested and ligated into the intraperitoneally with 100 mg OVA and 10 mg LPS in alum (a protocol based on ref. pCMV6-AN-DDK vector (OriGene Technologies). This construct does not contain 53) and killed 10 days later. Immunized mice received an intraperitoneal dose 23  1 a signal peptide, since the native BCMA also does not contain one . To obtain a (10 mg kg ) of LY-411575-I on day 9, followed by another dose 6 h before being BCMA variant with a prolonged extracellular part, the extracellular part of human killed. Second, female 16- to 18-week NZB/W F1 mice were bred at the animal BCMA was amplified by PCR, digested and ligated at the N-terminal end of full- facility of the German Rheumatism Research Center Berlin (DRFZ) under defined, length human BCMA clone (OriGene Technologies). We thereby doubled the pathogen-free conditions. To distinguish between short-lived and long-lived extracellular part, and called the new construct BCMA–BCMA. The plasmid plasma cells, we fed the mice BrdU (Sigma-Aldrich, 1 mg ml ) with 1% glucose in peak12-SP-C99 expresses the C-terminal 99 amino acids of APP together with the drinking water for 2 weeks. One week after the start of BrdU feeding, mice were APP signal peptide. C99 is a direct substrate for g-secretase . Peak12-SP-C99 was treated daily with the g-secretase inhibitor LY-411575-I intraperitoneally at a dose 51  1 obtained by cloning SP-C99 from the pCEP4 vector into the peak12 plasmid . of 10 mg kg for 7 days. Finally, 2-month-old NZB/W mice were immunized HEK or HeLa cells were transfected with 200 ng of respective expression plasmids intraperitoneally with OVA (100 mg per mouse and alum (Thermo) as adjuvant) as using lipofectamine 2,000 (Invitrogen Life Technologies). described and boosted at 3 months. At 4 months, mice were treated for four 10 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 ARTICLE consecutive weeks intraperitoneally with recombinant sBCMA-AT fusion protein 29. Hemming, M. L., Elias, J. E., Gygi, S. P. & Selkoe, D. J. Proteomic profiling of (200 mg three times a week). The concentration of this fusion protein in the serum gamma-secretase substrates and mapping of substrate requirements. PLoS Biol. of the injected mice was determined with an AT ELISA, which is specific for 6, e257 (2008). human AT (ab189579 from Abcam, Cambridge, UK). Single-cell suspensions were 30. Day, E. S. et al. Selectivity of BAFF/BLyS and APRIL for binding to the TNF prepared from the spleen and bone marrow (femur and tibia) as described before. family receptors BAFFR/BR3 and BCMA. Biochemistry (Mosc.) 44, 1919–1931 Cells were first incubated with anti-CD138-PE and then fixed with Cytofix/ (2005). Cytoperm buffer (BD Bioscience) for 30 min. Intracellular Ig kappa light chain was 31. Castigli, E. et al. Impaired IgA class switching in APRIL-deficient mice. Proc. detected by anti-kappa-Pacific Orange and OVA-specific antibodies were detected Natl Acad. Sci. USA 101, 3903–3908 (2004). by intracellular OVA conjugated with FITC. 32. Radtke, F., MacDonald, H. R. & Tacchini-Cottier, F. Regulation of innate and adaptive immunity by Notch. Nat. Rev. Immunol. 13, 427–437 (2013). 33. Simonetti, G. et al. IRF4 controls the positioning of mature B cells in the References lymphoid microenvironments by regulating NOTCH2 expression and activity. 1. Do¨rner, T., Radbruch, A. & Burmester, G. R. B-cell-directed therapies for J. Exp. Med. 210, 2887–2902 (2013). autoimmune disease. Nat. Rev. Rheumatol. 5, 433–441 (2009). 34. Yoon, S. O., Zhang, X., Berner, P., Blom, B. & Choi, Y. S. Notch ligands 2. Krumbholz, M., Derfuss, T., Hohlfeld, R. & Meinl, E. B cells and antibodies in expressed by follicular dendritic cells protect germinal center B cells from multiple sclerosis pathogenesis and therapy. Nat. Rev. Neurol. 8, 613–623 apoptosis. J. Immunol. 183, 352–358 (2009). (2012). 35. Yagi, T. et al. Defective signal transduction in B lymphocytes lacking presenilin 3. Mackay, F. & Schneider, P. Cracking the BAFF code. Nat. Rev. Immunol. 9, proteins. Proc. Natl Acad. Sci. USA 105, 979–984 (2008). 491–502 (2009). 36. Levine, S. J. Molecular mechanisms of soluble cytokine receptor generation. 4. Yang, M. et al. B cell maturation antigen, the receptor for a proliferation- J. Biol. Chem. 283, 14177–14181 (2008). inducing ligand and B cell-activating factor of the TNF family, induces antigen 37. Xanthoulea, S. et al. Tumor necrosis factor (TNF) receptor shedding controls presentation in B cells. J. Immunol. 175, 2814–2824 (2005). thresholds of innate immune activation that balance opposing TNF functions in 5. Darce, J. R., Arendt, B. K., Wu, X. & Jelinek, D. F. Regulated expression of infectious and inflammatory diseases. J. Exp. Med. 200, 367–376 (2004). BAFF-binding receptors during human B cell differentiation. J. Immunol. 179, 38. Hoffmann, F. S. et al. The Immunoregulator Soluble TACI Is Released by 7276–7286 (2007). ADAM10 and Reflects B Cell Activation in Autoimmunity. J. Immunol. 194, 6. Avery, D. T. et al. BAFF selectively enhances the survival of plasmablasts 542–552 (2015). generated from human memory B cells. J. Clin. Invest. 112, 286–297 (2003). 39. von Budingen, H. C., Bar-Or, A. & Zamvil, S. S. B cells in multiple sclerosis: 7. O’Connor, B. P. et al. BCMA is essential for the survival of long-lived bone connecting the dots. Curr. Opin. Immunol. 23, 713–720 (2011). marrow plasma cells. J. Exp. Med. 199, 91–98 (2004). 40. Nylander, A. & Hafler, D. A. Multiple sclerosis. J. Clin. Invest. 122, 1180–1188 8. Peperzak, V. et al. Mcl-1 is essential for the survival of plasma cells. Nat. (2012). Immunol. 14, 290–297 (2013). 41. Krumbholz, M. et al. BAFF is produced by astrocytes and up-regulated 9. Belnoue, E. et al. APRIL is critical for plasmablast survival in the bone in multiple sclerosis lesions and primary central nervous system lymphoma. marrow and poorly expressed by early-life bone marrow stromal cells. Blood J. Exp. Med. 201, 195–200 (2005). 111, 2755–2764 (2008). 42. Cepok, S. et al. Short-lived plasma blasts are the main B cell effector subset 10. Benson, M. J. et al. Cutting edge: the dependence of plasma cells and during the course of multiple sclerosis. Brain 128, 1667–1676 (2005). independence of memory B cells on BAFF and APRIL. J. Immunol. 180, 43. Stohl, W. et al. B lymphocyte stimulator overexpression in patients with 3655–3659 (2008). systemic lupus erythematosus: longitudinal observations. Arthritis Rheum. 48, 11. Amanna, I. J. & Slifka, M. K. Mechanisms that determine plasma cell lifespan 3475–3486 (2003). and the duration of humoral immunity. Immunol. Rev. 236, 125–138 (2010). 44. Kim, J., Gross, J. A., Dillon, S. R., Min, J. K. & Elkon, K. B. Increased 12. Hiepe, F. et al. Long-lived autoreactive plasma cells drive persistent BCMA expression in lupus marks activated B cells, and BCMA receptor autoimmune inflammation. Nat. Rev. Rheumatol. 7, 170–178 (2011). engagement enhances the response to TLR9 stimulation. Autoimmunity 44, 69– 13. Croft, M., Benedict, C. A. & Ware, C. F. Clinical targeting of the TNF and 81 (2011). TNFR superfamilies. Nat. Rev. Drug Discov. 12, 147–168 (2013). 45. Vincent, F. B., Saulep-Easton, D., Figgett, W. A., Fairfax, K. A. & Mackay, F. 14. Stohl, W., Scholz, J. L. & Cancro, M. P. Targeting BLyS in rheumatic disease: The BAFF/APRIL system: emerging functions beyond B cell biology and the sometimes-bumpy road from bench to bedside. Curr. Opin. Rheumatol. 23, autoimmunity. Cytokine Growth Factor Rev. 24, 203–215 (2013). 305–310 (2011). 46. Osorio, C. et al. Selective regulation of axonal growth from developing 15. Kappos, L. et al. Atacicept in multiple sclerosis (ATAMS): a randomised, hippocampal neurons by tumor necrosis factor superfamily member APRIL. placebo-controlled, double-blind, phase 2 trial. Lancet Neurol. 13, 353–363 (2014). Mol. Cell Neurosci. 59C, 24–36 (2014). 16. Kopan, R. & Ilagan, M. X. Gamma-secretase: proteasome of the membrane? 47. Sanchez, E. et al. Serum B-cell maturation antigen is elevated in multiple Nat. Rev. Mol. Cell Biol. 5, 499–504 (2004). myeloma and correlates with disease status and survival. Br. J. Haematol. 158, 17. Selkoe, D. J. & Wolfe, M. S. Presenilin: running with scissors in the membrane. 727–738 (2012). Cell 131, 215–221 (2007). 48. Minter, L. M. et al. Inhibitors of gamma-secretase block in vivo and in vitro 18. Ransohoff, R. M. & Engelhardt, B. The anatomical and cellular basis of immune T helper type 1 polarization by preventing Notch upregulation of Tbx21. Nat. surveillance in the central nervous system. Nat. Rev. Immunol. 12, 623–635 Immunol. 6, 680–688 (2005). (2012). 49. Burger, R. et al. Interleukin-6 production in B-cell neoplasias and Castleman’s 19. Pinna, D., Corti, D., Jarrossay, D., Sallusto, F. & Lanzavecchia, A. Clonal disease: evidence for an additional paracrine loop. Ann. Hematol. 69, 25–31 dissection of the human memory B-cell repertoire following infection and (1994). vaccination. Eur. J. Immunol. 39, 1260–1270 (2009). 50. Hauck, S. M. et al. Deciphering membrane-associated molecular processes in 20. Avery, D. T. et al. Increased expression of CD27 on activated human memory B target tissue of autoimmune uveitis by label-free quantitative mass cells correlates with their commitment to the plasma cell lineage. J. Immunol. spectrometry. Mol. Cell Proteomics 9, 2292–2305 (2010). 174, 4034–4042 (2005). 51. Mitterreiter, S. et al. Bepridil and amiodarone simultaneously target the 21. Herreman, A. et al. Total inactivation of gamma-secretase activity in presenilin- Alzheimer’s disease beta- and gamma-secretase via distinct mechanisms. deficient embryonic stem cells. Nat. Cell Biol. 2, 461–462 (2000). J. Neurosci. 30, 8974–8983 (2010). 22. Yamasaki, A. et al. The GxGD motif of presenilin contributes to catalytic 52. Perera, N. C. et al. NSP4 is stored in azurophil granules and released function and substrate identification of gamma-secretase. J. Neurosci. 26, by activated neutrophils as active endoprotease with restricted specificity. 3821–3828 (2006). J. Immunol. 191, 2700–2707 (2013). 23. Bossen, C. & Schneider, P. BAFF, APRIL and their receptors: structure, 53. Pasare, C. & Medzhitov, R. Control of B-cell responses by Toll-like receptors. function and signaling. Semin. Immunol. 18, 263–275 (2006). Nature 438, 364–368 (2005). 24. Luzina, I. G. et al. Spontaneous formation of germinal centers in autoimmune 54. Hoyer, B. F. et al. Short-lived plasmablasts and long-lived plasma cells mice. J. Leukoc. Biol. 70, 578–584 (2001). contribute to chronic humoral autoimmunity in NZB/W mice. J. Exp. Med. 25. Cassese, G. et al. Inflamed kidneys of NZB/W mice are a major site for the 199, 1577–1584 (2004). homeostasis of plasma cells. Eur. J. Immunol. 31, 2726–2732 (2001). 26. Yu, G. et al. APRIL and TALL-I and receptors BCMA and TACI: system for regulating humoral immunity. Nat. Immunol. 1, 252–256 (2000). Acknowledgements 27. Rennert, P. et al. A soluble form of B cell maturation antigen, a receptor for the We thank Drs K. Berer, K. Dornmair and J. Ellwart for support, Drs S. Kalled and tumor necrosis factor family member APRIL, inhibits tumor cell growth. J. Exp. K. Otipoby for monoclonal antibodies to BCMA, K. Ogston and Drs C. Haass, H. Steiner, Med. 192, 1677–1684 (2000). N. Kawakami and G. Krishnamoorthy for comments on the manuscript. This work was 28. Struhl, G. & Adachi, A. Requirements for presenilin-dependent cleavage of supported by the DFG (SFB TRR 128, SFB TRR 130, SFB 650), the BMBF (‘Kran- notch and other transmembrane proteins. Mol. Cell 6, 625–636 (2000). kheitsbezogenes Kompetenznetz Multiple Sklerose’ and JPND RiMOD’), the Breuer NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications 11 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 Foundation Alzheimer Award, the Gemeinnu¨tzige-Hertie-Stiftung (P1130076 and HERTIE Competing financial interests: The authors declare no competing financial interest. Senior Professorship to HW), the Munich Cluster for Systems Neurology (ExC 1010 Reprints and permission information is available online at http://npg.nature.com/ SyNergy), the GIF and the Verein zur Therapieforschung fu¨r Multiple-Sklerose-Kranke. reprintsandpermissions/ Author contributions How to cite this article: Laurent, S. A. et al. g-secretase directly sheds the survival S.A.L., F.H., P.-H.K., Q.C., Y.C., M.S.-S., S.M.H., E.S., JW and H.R. performed experi- receptor BCMA from plasma cells. Nat. Commun. 6:7333 doi: 10.1038/ncomms8333 ments and analysed data; M.Kr. was involved in study design; F.W., T.A., F.H., M.Kh., (2015). T.O. and H.-W.P. provided clinical samples; D.J., H.W., R.H., S.F.L., S.A.L. and E.M. designed the study, analysed data and wrote the paper. All authors discussed results and This work is licensed under a Creative Commons Attribution 4.0 commented on the manuscript. International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, Additional information Supplementary Information accompanies this paper at http://www.nature.com/ users will need to obtain permission from the license holder to reproduce the material. naturecommunications To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ 12 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nature Communications Springer Journals

Loading next page...
 
/lp/springer-journals/secretase-directly-sheds-the-survival-receptor-bcma-from-plasma-cells-03GrZhBym0

References (135)

Publisher
Springer Journals
Copyright
Copyright © 2015 by The Author(s)
Subject
Science, Humanities and Social Sciences, multidisciplinary; Science, Humanities and Social Sciences, multidisciplinary; Science, multidisciplinary
eISSN
2041-1723
DOI
10.1038/ncomms8333
Publisher site
See Article on Publisher Site

Abstract

ARTICLE Received 27 Feb 2015 | Accepted 28 Apr 2015 | Published 11 Jun 2015 DOI: 10.1038/ncomms8333 OPEN g-secretase directly sheds the survival receptor BCMA from plasma cells 1 1, 2,3, 4 5 Sarah A. Laurent , Franziska S. Hoffmann *, Peer-Hendrik Kuhn *, Qingyu Cheng , Yuanyuan Chu , 5 6 1 1 1 Marc Schmidt-Supprian , Stefanie M. Hauck , Elisabeth Schuh , Markus Krumbholz , Heike Ru¨bsamen , 2,3 7 7 4 4 8 Johanna Wanngren , Mohsen Khademi , Tomas Olsson , Tobias Alexander , Falk Hiepe , Hans-Walter Pfister , 9 10 11,12 1,12 2,3,12 Frank Weber , Dieter Jenne , Hartmut Wekerle , Reinhard Hohlfeld , Stefan F. Lichtenthaler & Edgar Meinl Survival of plasma cells is regulated by B-cell maturation antigen (BCMA), a membrane-bound receptor activated by its agonist ligands BAFF and APRIL. Here we report that g-secretase directly cleaves BCMA, without prior truncation by another protease. This direct shedding is facilitated by the short length of BCMA’s extracellular domain. In vitro, g-secretase reduces BCMA-mediated NF-kB activation. In addition, g-secretase releases soluble BCMA (sBCMA) that acts as a decoy neutralizing APRIL. In vivo, inhibition of g-secretase enhances BCMA surface expression in plasma cells and increases their number in the bone marrow. Furthermore, in multiple sclerosis, sBCMA levels in spinal fluid are elevated and associated with intracerebral IgG production; in systemic lupus erythematosus, sBCMA levels in serum are elevated and correlate with disease activity. Together, shedding of BCMA by g-secretase controls plasma cells in the bone marrow and yields a potential biomarker for B-cell involvement in human autoimmune diseases. 1 2 Institute of Clinical Neuroimmunology, Ludwig Maximilian University Munich, 81377 Munich, Germany. Neuroproteomics, Klinikum rechts der Isar, and Institute of Advanced Study, Technische Universita¨tMu¨nchen, 81377 Munich, Germany. German Center for Neurodegenerative Diseases (DZNE), 81377 Munich, Germany. Department of Rheumatology and Clinical Immunology, Charite - Universita¨tsmedizin Berlin and Deutsches Rheuma-Forschungszentrum Berlin-a Leibniz Institute, 10117 Berlin, Germany. Department of Internal Medicine III, Klinikum Rechts der Isar, Technische Universita¨tMu¨nchen, 81675 Munich, Germany. Research Unit Protein Science, Helmholtz Zentrum Mu¨nchen (GmbH), German Research Center for Environmental Health, 85764 7 8 Neuherberg, Germany. Karolinska University Hospital, Division of Clinical Neuroscience, 17176 Stockholm, Sweden. Department of Neurology, Klinikum 9 10 Grosshadern, Ludwig Maximilian University Munich, 81377 Munich, Germany. Max-Planck-Institute of Psychiatry, 80804 Munich, Germany. CPC 11 12 Helmholtz Zentrum Munchen (GmbH), 81377 Munich, Germany. Max-Planck-Institute of Neurobiology, 82152 Martinsried, Germany. Munich Cluster for Systems Neurology (SyNergy), 81377 Munich, Germany. * These authors contributed equally to this work. Correspondence and requests for materials should be addressed to E.M. (email: edgar.meinl@med.uni-muenchen.de). NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications 1 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 cells are of pathogenic relevance and serve as a therapeutic 2.0 target both in generalized immunopathological diseases *** Bsuch as systemic lupus erythematosus (SLE), and in 1.5 1,2 *** organ-specific diseases such as multiple sclerosis (MS) . 20 1.0 Activation and survival of B cells is largely regulated via the BAFF–APRIL system that comprises three receptors (BAFF-R, 10 0.5 TACI and BCMA (B-cell maturation antigen)) and two ligands 3 0 0 (BAFF and APRIL) . Membrane-bound BCMA (mBCMA) is OND CIS/MS HC OND CIS/MS NB 4 5,6 expressed on some activated B cells and Ig-secreting cells ;it n=20 n=37 n=26 n=20 n=37 n=5 binds both BAFF and APRIL . BCMA is essential for the 7,8 maintenance of long-lived plasma cells , an effect mediated by 4 150 * 9,10 APRIL or BAFF . These plasma cells produce IgG that protect *** * not only against pathogens but are also critically involved in 11,12 autoimmune diseases . Further, mBCMA engagement on activated B cells induces MHC class II, enhancing their ability to present antigen . The BAFF–APRIL system is targeted for therapeutic interven- 0 0 tion. In SLE, where BAFF levels in the serum are elevated, an 0.0 0.5 1.0 1.5 2.0 HC SLE SLE 13,14 n=29 n=17 n=22 –1 antibody-binding BAFF is already approved . On the other sBCMA (ng ml ) hand, the recombinant soluble receptor atacicept, which targets both BAFF and APRIL, unexpectedly worsened MS , indicating that essential features of this system are not fully understood. Other clinical trials targeting the BAFF–APRIL system in 13,14 immunopathological disorders have been launched . We aimed at identifying features of humoral immunity that 10 50 might be altered in autoimmune diseases. Thereby, we found that a soluble form of BCMA (sBCMA) is regularly detectable in 0 0 0 50 100 150 0 50 100 150 human blood. We then analysed sBCMA in human autoimmune –1 –1 sBCMA (ng ml ) sBCMA (ng ml ) diseases. In MS, sBCMA was elevated in the cerebrospinal fluid (CSF) and linked to local IgG production inside the brain. Figure 1 | sBCMA as a biomarker. (a) sBCMA plasma concentrations In SLE, sBCMA was systemically elevated and associated with were determined using ELISA in healthy controls (HC), patients with a disease activity. We went on to uncover the underlying clinically isolated syndrome (CIS) or MS, or other neurological diseases biochemical mechanism and found that BCMA was directly shed (OND). (b) sBCMA in the CSF was determined in patients with OND, by g-secretase, a ubiquitous intramembranous protease. Direct CIS/MS or neuroborreliosis (NB) (***Po0.001, Kruskal–Wallis test shedding of a membrane protein without prior processing by followed by Dunn’s Multiple Comparison Test). (c) sBCMA in the CSF another protease is a novel function of g-secretase, which is best correlated strongly with the intrathecal IgG production represented by the known for processing of amyloid precursor protein (APP) in IgG Index. This correlation was evident when all analysed CSF samples were 16,17 . To analyse the functional Alzheimer’s disease and Notch considered (Po0.0001, r ¼ 0.85) and in the MS/CIS group (Po0.0001, relevance of this shedding in vivo, we treated mice with a r ¼ 0.77, Spearman rank correlation, CIS/MS n¼ 36; OND n ¼ 20, g-secretase inhibitor; this enhanced surface BCMA on plasma NB n ¼ 5). (d) sBCMA in the serum was determined with ELISA in HC, cells and increased their number in the bone marrow. Thus, untreated (red) and treated (black) patients with SLE. sBCMA was elevated shedding of BCMA via g-secretase is an immunoregulatory in SLE patients and in the untreated SLE patients compared with the treated mechanism limiting plasma cells. patients (***Po0.001 and *Po0.05, Kruskal–Wallis test followed by Dunn’s Multiple Comparison Test). (e) sBCMA in the serum of SLE patients correlated strongly with disease activity quantified with SLE disease activity Results index (SLEDAI; Po0.001; r ¼ 0.54, Spearman correlation). (f) sBCMA in sBCMA shows local IgG production in MS and activity in SLE. the serum of SLE patients inversely correlated with the level of the We found sBCMA as a regular component of human blood from complement factor C3 (P ¼ 0.0374, r¼ 0.29, Spearman correlation). healthy subjects (Fig. 1a). In MS, sBCMA was elevated in the CSF, Bars represent means. but not in the blood (Fig. 1a,b). sBCMA levels in the plasma and serum were very similar as seen in 14 controls (Supplementary Fig. 1a). Levels of sBCMA in the CSF correlated strongly with sBCMA in the CSF (Supplementary Fig. 1b). High-dose steroids local IgG production in MS patients (Fig. 1c). This strong for treatment of acute relapses reduced serum sBCMA levels correlation was confirmed in a second cohort of 25 MS patients transiently (Supplementary Fig.1c). (Po0.0001, r ¼ 0.80, Spearman rank correlation). Also in In SLE, serum levels of sBCMA were elevated as seen in an neuroborreliosis, an infectious neuroinflammatory disease char- untreated and in a treated cohort (Fig. 1d). Immunosuppressive acterized by chronic IgG production within the CNS, sBCMA treatment of SLE patients reduced sBCMA levels (Fig. 1d). We levels were increased and correlated with local IgG production noted a strong correlation of serum sBCMA and disease activity (Fig. 1b,c). In CSF from MS patients, there was a slight inverse (Fig. 1e) and an inverse correlation with the paraclinical marker correlation of sBCMA with its high-affinity ligand APRIL complement factor 3 (Fig. 1f). There was a trend (P¼ 0.0767, (r¼ 0.35, P ¼ 0.033, Spearman rank correlation), but r¼ 0.23, Spearman rank correlation) for correlation between no association to the plasma level of sBCMA (r ¼ 0.07, anti-dsDNA titre and sBCMA in these SLE patients. Serum BAFF P ¼ 0.667, Spearman rank correlation). We analysed effects of levels were elevated in our SLE cohort (mean¼ 0.61 0.13 ng ml immunosuppressive treatment on sBCMA levels longitudinally in in healthy controls; mean¼ 2.11 4.32 ng ml in SLE patients, two cohorts of MS patients. Natalizumab, which blocks entry of P¼ 0.0004, Wilcoxon–Mann–Whithney test), and correlated with 18 2 lymphocytes into the CNS and local IgG production , reduced sBCMA levels (P¼ 0.0013, r¼ 0.47, Spearman rank correlation). 2 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. –1 SLEDAI IgG index sBCMA (ng ml ) –1 –1 –1 C3 (mg dl ) sBCMA (ng ml ) sBCMA (ng ml ) NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 ARTICLE BCMA is shed during differentiation to Ig-secreting cells. Since Again, differentiation towards IgG-secreting cells was mBCMA is known to be expressed on some activated B cells and accompanied by the appearance of sBCMA (Fig. 2b). 5,6 Ig-secreting cells , we analysed the release of sBCMA by After activation with CD40L þ IL-21, two B-cell populations, þ  þ þ primary human B cells. We applied two different protocols to CD19 CD38 and CD19 CD38 , cells could be distin- activate human primary B cells in order to differentiate them guished. We directly compared mBCMA and release of sBCMA towards Ig-secreting cells. First, purified blood-derived human by these different B-cell subsets. While mBCMA was absent on þ  þ B cells were activated via CD40L and further differentiated to unstimulated B cells, CD19 CD38 cells weakly and CD19 Ig-secreting cells by adding interleukin (IL)-21 (ref. 5). While CD38 cells strongly expressed mBCMA (Fig. 2c). We sorted human B cells activated via CD40L alone released low, but CD38 and CD38 cells, cultured them for another 24 h detectable, levels of sBCMA, this was strongly enhanced by without further stimulation and determined the amount of shed addition of IL-21 (Fig. 2a). In the second primary B-cell culture sBCMA (Fig. 2d). This revealed a close correlation between system we activated peripheral blood mononuclear cells (PBMC) released sBCMA and surface expression of mBCMA (Fig. 2e). with the TLR7 þ 8 ligand R848 and IL-2, which induces The transcription of BCMA in these B-cell subsets was further 19 þ differentiation of human memory B cells to Ig-secreting cells . substantiated with qPCR. BCMA transcript levels in the CD38 6 14 7 3.5 IgG IgG 12 6 sBCMA 3.0 sBCMA 5 2.5 8 4 2.0 6 3 1.5 4 2 1.0 2 1 0.5 0 0 0 0.0 Vehicle +CD40L +CD40L Vehicle R848 IL-21 IL-2 No stimulation CD40L+IL-21 stimulation + + – + + CD19 CD19 CD38 CD19 CD38 100 100 100 6 80 80 80 60 60 40 40 20 20 0 0 0 2 3 4 5 2 3 4 5 2 3 4 5 + – + 010 10 10 10 010 10 10 10 010 10 10 10 CD19 CD38 CD38 BCMA No stimulation CD40L+IL-21 12 3 4 50 IgH IgL 4.6 sBCMA 3.5 kDa kDa 0 50 100 150 200 250 BCMA surface expression (MFI) Extracellular domain Transmembranous domain N -MLQMA GQCSQ NEYFD SLLHA CIPCQ LRCSS NTPPL TCQRY CNASV TNSVK GTNAI LWTCL GLSLI ISLAV FVLMF LLRKI… Figure 2 | sBCMA is released when B cells differentiate towards plasma cells and comprises the extracellular domain plus part of the transmembranous region of BCMA. (a) Human purified B cells were activated for 5 days as indicated; IgG and sBCMA in the supernatant were quantified using ELISA. Combined data of three independent experiments (mean s.e.m., P ¼ 0.0073, paired t-test). (b) PBMCs were stimulated with R848 þ IL-2 for 7 days. IgG and sBCMA in the supernatant were quantified using ELISA. Combined data of three independent experiments (mean s.e.m., P ¼ 0.0227, paired t-test). (c–e) Human purified B cells were stimulated with CD40L þ IL-21. (c) surface BCMA was measured using flow cytometry on unstimulated B þ  þ þ þ cells, CD19 CD38 cells and CD19 CD38 cells. (d) Sorted CD38 and CD38 cells were cultured for another 24 h and the amount of shed sBCMA was measured using ELISA, combined data of two independent experiments. (e) Correlation between sBCMA release and surface expression of BCMA for a single replicate. (f,g) sBCMA was immunoprecipitated from supernatant of plasmacytoma cells (f, lanes 1, 2), serum (f, lane 3) and from supernatant of human purified B cells cultured with CD40L plus IL-21 (f, lane 4) with anti-BCMA monoclonal antibodies (mAbs) A7D12.2 (f, lanes 1 and 4) or C12A3.2 (f, lane 2) or goat-anti-BCMA (f, lane 3). Western blot analysis for BCMA (f) and silver staining of sBCMA immunoprecipitated from plasmacytoma supernatant (g) was performed. (h) The band at 6 kDa (from g) and sBCMA obtained using immunoprecipitation were analysed with mass spectrometry. The aa sequences of BCMA and peptides identified after tryptic (blue) or chymotryptic (red) digestion are shown. No peptide was detected with a C-terminal aa that was not a site for either tryptic or chymotryptic cleavage, indicating that the precise cleavage site of g-secretase within the membrane needs to be identified. NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications 3 & 2015 Macmillan Publishers Limited. All rights reserved. % of Max –1 –1 sBCMA (ng ml ) IgG (μg ml ) % of Max –1 sBCMA (ng ml ) –1 IgG (μg ml ) % of Max sBCMA (ng/ml) –1 sBCMA (ng ml ) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 cells reached 15.9 5.2% peptidyl-prolyl isomerase A (PPIA), while BCMA expression in the CD38 cells subset accounted for 80 80 ± ± 2.5 1.3% PPIA (mean s.e.m. of three independent replicates). Thus, in these B-cell subsets BCMA surface expression reflected BCMA transcription. 40 40 We also analysed BCMA shedding in tumour cell lines and 20 20 transfectants. The human plasmacytoma cell line JK-6L sponta- 0 0 Vehicle 0.01 0.03 0.1 0.3 1 Vehicle 1 3 10 25 50 neously shed sBCMA (Supplementary Fig. 2a). In BCMA- DAPT (μM) TAPI-1 (μM) transfected HeLa cells surface expression of mBCMA was accompanied by release of sBCMA (157 6ngml ) without Vehicle Vehicle requiring any further stimulus. HeLa cells did not secrete TAPI-1 DAPT Isotype control Isotype control detectable amounts of APRIL or BAFF, neither spontaneously 10 100 nor after transfection with BCMA or an empty vector. Together, our observations with primary human B-cell cultures, plasmacy- 60 60 toma cells and BCMA-transfected cells indicate that release of sBCMA is a direct consequence of surface expression of mBCMA; 20 20 it does not require additional stimulation or ligand binding. 2 3 4 5 2 3 4 5 2 3 4 5 0 10 10 10 10 010 10 10 10 010 10 10 10 sBCMA comprises extracellular and intramembranous part. CD27 BCMA sBCMA was isolated by immunoprecipitation from the super- natant of primary Ig-secreting cells, plasmacytoma cells or serum; 120 120 in all these sources, sBCMA had a molecular weight (MW) of 100 100 B6 kDa as seen using western blot analysis (Fig. 2f). This size was 80 80 confirmed when silver staining was applied to detect material 60 60 obtained by immunoprecipitation with anti-BCMA from the 40 40 supernatant of plasmacytoma cells (Fig. 2g). This corresponds to the extracellular part of BCMA (54 amino acid (aa), calculated 20 20 MW 5.8 kDa). Unexpectedly, mass spectrometry revealed that 0 0 Vehicle 1 3 10 25 50 Vehicle 0.01 0.03 0.1 0.3 1 sBCMA comprised not only the complete extracellular domain TAPI-1 (μM) DAPT (μM) with an intact N terminus, but also part of the transmembrane region (Fig. 2h). This indicated that it was released by an intra- Vehicle Vehicle membranous protease. TAPI-1 DAPT Isotype control Isotype control c-secretase inhibitors block BCMA shedding from B cells. Since 10 mBCMA is a type-I oriented transmembrane protein with an 4 60 60 extracellular N terminus, g-secretase was a candidate for its intramembranous cleavage. We applied the g-secretase inhibitor 10 20 20 DAPT and compared it with the metalloprotease inhibitor TAPI- 2 3 4 5 2 3 4 5 1, which reduces the shedding of other TNFR-SF members. We 2 3 4 5 010 10 10 10 010 10 10 10 010 10 10 10 activated human B cells either via CD40L þ IL-21 (Fig. 3a,b) or CD19 BCMA via R848 þ IL-2 (Fig. 3c,d), and used both fluorescence-activated Figure 3 | c-secretase inhibitor DAPT reduces release of sBCMA and cell sorting (FACS) and enzyme-linked immunosorbent assay enhances surface expression of BCMA on activated human B cell. (ELISA) as read-out systems to quantify mBCMA and sBCMA. (a,b) Human B cells were differentiated into Ig-secreting cells via DAPT blocked the release of sBCMA even at low concentrations, CD40Lþ IL-21. (a) Release of sBCMA on treatment with DAPT or TAPI-1 while TAPI-1 had little or no effect (Fig. 3a,c). After CD40L þ IL- was measured using ELISA. sBCMA release was normalized to the amount 21 application, a high surface expression of mBCMA was noted in þþ þ of sBCMA shed under vehicle conditions, which was set as 100%. the CD27 CD38 subset (Fig. 3b), previously classified as late 20 ± Combined data of three independent experiments (mean s.e.m.). plasmablasts . DAPT enhanced surface expression of mBCMA (b) These activated primary human B cells were subgrouped on the basis of in these cells, while TAPI-I had little or no effect (Fig. 3b). When expression of CD27 and CD38. A high surface expression of BCMA was human PBMCs were activated with R848 þ IL-2, B20% of the þþ þ þ þ seen on the CD27 CD38 subset. Surface expression of BCMA was cells were CD19 CD38 after 7 days (Fig. 3d). These cells enhanced using DAPT treatment (1 mM), while TAPI-I (50 mM) had little strongly expressed mBCMA on their surface and this was greatly effect. (c,d) Human PBMCs were stimulated with R848 þ IL-2 for 7 days. enhanced by the g-secretase inhibitor DAPT; again, TAPI-I had (c) Release of sBCMA on treatment with DAPT or TAPI-1 was measured little effect (Fig. 3d). Similar to primary human B cells, we using ELISA. sBCMA release was normalized to the amount of sBCMA shed observed differential effects of DAPT and TAPI-1 on the release under vehicle conditions, which was set as 100%. Combined data of three of sBCMA and surface expression of mBCMA on human independent experiments (mean s.e.m.). (d) High surface expression of plasmacytoma cells (Supplementary Fig. 2a,b). þ þ BCMA was seen on the CD19 CD38 subset; this was further enhanced Further, we compared the effect of transition (LY-411575-I and by DAPT (1 mM), while TAPI-I (50 mM) had little effect. LY685,458) and non-transition state (DAPT and RO4929097) inhibitors of the g-secretase on BCMA shedding from human B cells. Human PBMCs were first stimulated with R848 þ IL2 for 7 days, and then CD19 B cells were positively selected and c-secretase directly sheds BCMA. Presenilin (PS)1 or PS2 is the 14,15 cultured overnight in the absence of these g-secretase inhibitors. catalytical component of the g-secretase complex . To finally We found that RO4929097, LY-411575-I and LY685,458 had prove that sBCMA is released by g-secretase, we switched from B similar effects as DAPT on the shedding of mBCMA as seen with cells to mouse embryonic fibroblasts (MEF) deficient for both PS1 both read-out systems, FACS and ELISA (Supplementary Fig. 3). and PS2 (PS  /  ) . These MEF cells were transduced with 4 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. sBCMA release (%) sBCMA release (%) CD38 CD38 % Of max % Of max sBCMA release (%) sBCMA release (%) % Of max % Of max NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 ARTICLE full-length human BCMA plus either wt PS1, or its catalytically short extracellular domain of BCMA facilitates its direct cleavage inactive mutant D385A (ref. 22). Cleavage of mBCMA and by g-secretase. release of sBCMA occurred only in the presence of an active g-secretase complex as seen with FACS, ELISA and western blot c-secretase regulates NF-jB-mediated activation via BCMA. analysis (Fig. 4a–d). To date, g-secretase is known to cleave only The reduced surface expression of mBCMA mediated by 16,17 membrane proteins after a previous cut by other proteases . g-secretase was linked to a reduced binding of APRIL This was in conflict with our mass spectrometry analysis (Supplementary Fig. 2c). To analyse whether the DAPT-increased indicating that sBCMA had an intact N terminus (Fig. 2e). mBCMA expression was accompanied by enhanced Therefore, we used an additional experimental approach to assure cellular responsiveness, we analysed NF-kB activation in that g-secretase cleaved mBCMA without prior N-terminal BCMA-transfected HEK cells. DAPT increased BCMA-mediated trimming. We transfected a cDNA construct coding for BCMA NF-kB activation induced via APRIL (Fig. 6a) and also via tagged with an N-terminal FLAG. Using an ELISA with anti- BAFF (Fig. 6b). FLAG for coating and anti-BCMA for detection, we found the release of sBCMA with N-terminal FLAG, which was blocked by sBCMA acts as a decoy inhibiting APRIL in vitro.We found the g-secretase inhibitor DAPT (Fig. 5a,b). This provides further that human sBCMA bound APRIL, but not BAFF (Fig. 7a). This is evidence that g-secretase sheds BCMA without prior N-terminal 3,23 different to mBCMA, which binds both APRIL and BAFF .We trimming by another protease. found that the naturally occurring sBCMA differed from To analyse the importance of the length of the short recombinant BCMA-Fc, which bound APRIL and BAFF; extracellular domain of BCMA (54 aa) for its direct cleavage by however, binding to APRIL was stronger (Fig. 7a). Differences g-secretase, we doubled its extracellular part (variant BCMA– between sBCMA and BCMA-Fc were also seen in NF-kBreporter BCMA) with an extracellular part of 108 aa. We transfected wild- assays: in HEK cells transfected with BCMA, APRIL-mediated type BCMA and BCMA–BCMA in HEK293T cells. While the NF-kB activation was blocked by both sBCMA and BCMA-Fc, g-secretase inhibitor DAPT strongly enhanced the surface while BAFF-mediated activation via mBCMA was suppressed only expression of transfected wild-type mBCMA, DAPT had little by BCMA-Fc, but not by sBCMA (Fig. 7b). Since APRIL can signal or no effect on surface expression of BCMA–BCMA (Fig. 5c,d). via mBCMA and TACI , we transfected HEK cells also with TACI. Similarly, when we measured the released sBCMA, we found that These experiments showed that sBCMA inhibited APRIL, but not DAPT strongly decreased its release from wild-type BCMA, but BAFF-mediated signalling via TACI (Supplementary Fig. 4). had little effect on shedding of BCMA–BCMA (Fig. 5e). Thus, the Further, sBCMA limited APRIL-mediated survival of activated primary B cells (Fig. 7c). The effects of BCMA shedding by 100 g-secretase are summarized in a diagram (Fig. 7d). 6,000 PS –/– PS –/– BCMA 60 c-secretase regulates plasma cells and B cells in vivo. We tested 4,000 PS –/– BCMA whether cleavage of mBCMA by g-secretase occurred also in vivo PS1 D385 using two different mouse models. First, we immunized mice with PS –/– BCMA 2,000 PS1 ovalbumin (OVA) þ lipopolysaccharide (LPS) to induce a 2 3 4 5 0 10 10 10 10 T-dependent immune response and plasma cell differentiation. BCMA Treatment 9 days later with the g-secretase inhibitor LY-411575-I for 1 day enhanced surface expression of mBCMA in CD138 plasma cells in the spleen (Fig. 8a) and bone marrow (Supplementary Fig. 5). In the absence of a g-secretase inhibitor, PS1 holo mBCMA was hardly detectable at all on the surface of murine 30 plasma cells, in line with previous observations . Second, we used an SLE model in NZB/W mice, in which germinal centres develop PS1 CTF 24,25 spontaneously . In addition here, 1 day of treatment with the NCT mature g-secretase inhibitor LY-411575-I enhanced surface expression of NCT immature BCMA on plasma cells in the spleen and bone marrow (Fig. 8b). 20 mBCMA To distinguish between short-lived and long-lived plasma cells, β-actin mice were fed bromodeoxyuridine (BrdU). We found that kDa inhibition of g-secretase for 1 day enhanced mBCMA both on BrdU plasma cells (designated also as short-lived plasmablasts) and on BrdU plasma cells (designated also as long-lived plasma cells) in the bone marrow and spleen (Fig.8b). We then analysed the effects of extended application of the g-secretase inhibitor and treated NZB/W mice for 7 days. This Figure 4 | Release of sBCMA requires active presenilin. (a–d) Presenilin- resulted again in enhanced expression of mBCMA on BrdU deficient MEF cells (PS / ) were transduced with human BCMA and BrdU plasma cells in the bone marrow and spleen (Fig. 8c). (PS / BCMA) and then with wild-type PS1 (PS / BCMA PS1) or This prolonged treatment period enhanced the number of plasma with a catalytically inactive mutant (PS / BCMA PS1-D385A). BCMA cells in the bone marrow, but not in the spleen (Fig. 8d). In the surface expression (a,b) and sBCMA release (d) were determined. In a,a spleen, the absolute number of BrdU plasma cells per organ representative experiment is shown; in b,d, mean s.e.m. of four independent decreased, but not their relative number (Fig. 8d,e). In the bone experimentsisgiven (respectively, P¼ 0.0313 and P¼ 0.0033, paired t-test). marrow, treatment with the g-secretase inhibitor increased both (c) Cells used in a,b,d were analysed by immunoblotting for expression of the absolute and also the relative number of plasma cells full-length PS1 (PS1 ), for autoendoproteolysis of PS1 generating a (Fig. 8d,e). holo C-terminal fragment (PS1 ) reflecting an active state of the g-secretase, for The increase in plasma cell number in the bone marrow after CTF maturation of nicastrin (NCT) and for full-length BCMA (mBCMA). inhibition of g-secretase could be attributed to the regulation of NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications 5 & 2015 Macmillan Publishers Limited. All rights reserved. PS–/– BCMA PS–/– BCMA-D385A PS–/– BCMA-PS1 PS–/– PS–/– BCMA PS–/– BCMA-D385A PS–/– BCMA-PS1 % Of max BCMA (MFI) –1 sBCMA (ng ml ) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 Vehicle DAPT Isotype control Biotinylated anti-BCMA 80 sBCMA 40 40 20 20 Anti-BCMA 2 3 4 5 2 3 4 5 010 10 10 10 010 10 10 10 / / / BCMA wt BCMA-BCMA DAPT DAPT DAPT BCMA Empty vector BCMA BCMA-FLAG 0.5 Anti-FLAG Control-Ig 0.4 ** ** 0.3 Biotinylated anti-BCMA 0.2 sBCMA-FLAG 0.1 Anti-FLAG 0 0 // / DAPT DAPT DAPT Empty vector BCMA BCMA-FLAG Figure 5 | Release of sBCMA occurs without prior N-terminal trimming and is facilitated by the short extracellular domain of BCMA. (a,b) HeLa cells were transfected with plasmids coding for full-length human BCMA or BCMA with an N-terminal FLAG and then cultured with increasing amounts of the g-secretase inhibitor DAPT (0.02, 0.1 and 0.5 mM). Twenty-four hours after transfection supernatants were harvested and the released sBCMA was analysed using ELISA. In (a), ELISA wells were coated with anti-BCMA, and in (b) with anti-FLAG or a control IgG (anti-myelin oligodendrocyte glycoprotein (MOG) 8.18 C5), both were developed with anti-BCMA. Schemes of the ELISAs are shown on the right. Combined data of two independent experiments (mean s.e.m.). (c–e) Human BCMA wild type (wt) or BCMA–BCMA with a doubled extracellular domain of BCMA were transfected into HEK293Tcells. (c,d) Surface expression of BCMA was determined in the absence or presence of the g-secretase inhibitor DAPT (1 mM). (d) The combined data of three experiments (P ¼ 0.0049, **Po0.01, unpaired t-test). (e) The effect of DAPT on the released sBCMA after transfection with BCMA wt or BCMA–BCMA (mean s.e.m. of three experiments), P ¼ 0.0081, **Po0.01, unpaired t-test. activation in vitro (Supplementary Fig. 6c), we gave sBCMA-AT ab three times a week for 4 weeks to NZB/W mice, which had 20 40 P=0.05 Vehicle * Vehicle been immunized with OVA before. The applied amount of DAPT 1 μM DAPT 1 μM 15 30 sBCMA-AT corresponded to the dose previously used for 26,27 BCMA-Fc . Nevertheless, under these conditions sBCMA- 10 20 AT did not modulate the total number of plasma cells, the number of OVA-specific plasma cells or the amount of circulating IgG, IgA and IgM (Supplementary Fig. 6d–f). Notably, inhibition of g-secretase had additional effects on Plasmid Empty BCMA Plasmid Empty BCMA B cells not expressing BCMA. Short-term treatment with Ligand// APRIL APRIL Ligand// BAFF BAFF the g-secretase inhibitor reduced GC B cells (defined as int high CD38 Fas ) in the spleen as seen in our immunization model Figure 6 | c-secretase regulates BCMA-mediated NF-jB activation. (Supplementary Fig. 7a,b). Moreover, after 7 days of treatment, (a,b) HEK293T cells were transfected with full-length human BCMA or an NZB/W mice had fewer B cells in the spleen and bone marrow empty vector. DAPT, APRIL (a) or BAFF (b) were added, and NF-kB and showed a striking reduction of pre-B cells in their bone activation was determined. Combined data of three independent marrow (Supplementary Fig. 7c). experiments (mean s.e.m., *Po0.05, paired t-test). mBCMA on plasma cells (Fig. 8, Supplementary Fig. 4) or on Discussion decoy effects of sBCMA, which we observed in vitro (Fig. 7b,c). We report that g-secretase directly sheds mBCMA and regulates To analyse possible immunoregulatory effects of the shed sBCMA the number of plasma cells in the bone marrow. Further, the in vivo, we produced recombinant sBCMA and applied it released sBCMA reflects B-cell activation in human autoimmu- systemically. We prepared sBCMA as a fusion protein with nity, namely compartmentalized Ig production in MS and disease a1-antitrypsin (AT; Supplementary Fig. 6a) to increase its activity in SLE. molecular weight and therefore its half live in vivo. When B cells are activated and differentiated towards 5,6 The monomeric structure of sBCMA-AT was confirmed using Ig-secreting cells, they start to express mBCMA , which is gel filtration (Supplementary Fig. 6b). Having shown that then shed by g-secretase as we show here. Our experiments with recombinant sBCMA-AT blocked APRIL-mediated NF-kB activated primary human B cells, plasmacytoma cells and BCMA 6 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. BCMA wt BCMA–BCMA BCMA wt BCMA–BCMA NF-κB (fold induction) –1 O.D. 450 nm sBCMA(ng ml ) NF-κB (fold induction) % Of max BCMA expression (fold increase) on DAPT % Of max sBCMA release (fold decrease) on DAPT NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 ARTICLE 0.5 2.0 0.4 Anti-BCMA 1.5 0.3 sBCMA / BCMA-Fc 1.0 Ligand-FLAG 0.2 0.5 Anti-FLAG 0.1 0.0 0.0 sBCMA +– – +– – sBCMA BCMA-Fc BCMA-Fc sBCMA BCMA-Fc – +– – +– PBS –+ – – – + Ligand –– –– APRIL-FLAG BAFF-FLAG ** 9 16 w/o sBCMA w/o sBCMA sBCMA 14 sBCMA 7 BCMA-Fc BCMA-Fc Plasmid BCMA Plasmid BCMA Ligand / APRIL Ligand / BAFF sBCMA BCMA-Fc sBCMA BCMA-Fc sBCMA BCMA-Fc sBCMA BCMA-Fc GSI *** ** BCMA APRIL γ-secretase Membrane B-cell activation and survival B-cell activation and survival sBCMA Figure 7 | sBCMA is a decoy for APRIL in vitro. (a) A scheme of the ELISA is shown on the left; it detects BCMA–APRIL–FLAG (left panel) or BCMA–BAFF–FLAG (right panel) complexes, but neither BCMA nor APRIL–FLAG nor BAFF–FLAG alone alone (*Po0.05, paired t-test). sBCMA was derived from the supernatant of plasmacytoma cultured under serum-free conditions. Combined data of three independent experiments (mean s.e.m.). (b) HEK293T cells were transfected with human BCMA and activated with APRIL (left panel) or BAFF (right panel). sBCMA (50 and 200 ng ml ) was applied as indicated. sBCMA and control supernatant were obtained as mentioned above. BCMA-Fc (50 and 200 ng ml ) was used as a positive control. Combined data of three independent experiments (mean s.e.m., *Po0.05; **Po0.01 paired test). (c) Murine B cells were activated via anti-IgM and cultured for 2 days with APRIL in the presence or absence of sBCMA (200 and 400 ng ml ). APRIL-induced survival was calculated as described in the Methods section. sBCMA was obtained from supernatant from HEK293T cells transfected with full-length BCMA (black bars). Control supernatant was obtained after transfection with an empty vector (white bars). sBCMA significantly inhibited APRIL-mediated survival (***Po0.001 and **Po0.01, paired t-test). Combined data of six independent experiments (mean s.e.m.). (d) Illustration of the consequences of sBCMA shedding by g-secretase: left: an active g-secretase cleaves mBCMA. This reduces the number of membrane-bound BCMA molecules and releases sBCMA, which binds its ligand APRIL functioning as a decoy. Right: g-secretase inhibitors (GSIs) result in elevated mBCMA on the surface and increased APRIL-mediated activation and survival. transfectants indicate that BCMA shedding by g-secretase is a example, NOTCH and APP) that they become accessible to consequence of surface expression of mBCMA; it does not require cleavage only after their extracellular domain has been trimmed 16,17 additional activation or ligand binding; however, we cannot by another protease . We now show that g-secretase can exclude that this shedding can be further enhanced by yet cleave mBCMA directly and that prolonging its extracellular part 16,17 unknown mechanisms. g-secretase is ubiquitously expressed greatly reduced this cleavage. Thus, constitutive cleavage of and accessibility of its transmembranous substrates is largely mBCMA by g-secretase is facilitated by its short (54 aa) regulated by the length of their extracellular part. While extracellular domain. Further we found that g-secretase reduced membrane proteins with an extracellular domain exceeding 100 BAFF and APRIL-mediated NF-kB activation via BCMA in vitro. aa are not efficiently cleaved, extracellular domains of B50 aa To get insight into the in vivo relevance of mBCMA shedding 28,29 permit effective cleavage . g-secretase may remove protein by g-secretase, we used an immunization protocol and an SLE stubs for further degradation and was hence called ‘proteasome of model with spontaneous formation of germinal centres and 16 24,25 the membrane’ , although this notion is controversially plasma cells . We found that inhibition of g-secretase discussed . Probably due to these length restrictions, it has enhances mBCMA on plasma cells in the bone marrow and been a unifying feature of all substrates of g-secretase (for spleen. This was observed in both BrdU and BrdU plasma NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications 7 & 2015 Macmillan Publishers Limited. All rights reserved. APRIL-induced survival (%) NF-κB (fold induction) O.D. 450 nm NF-κB (fold induction) O.D. 450 nm ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 ** *** *** 200 *** *** ** 250 Vehicle LY-411575-I 100 100 40 40 100 20 20 2 3 4 5 2 3 4 5 0 10 10 10 10 0 10 10 10 10 + – + – BCMA Spleen PC BrdU BrdU BM PC BrdU BrdU *** 800 1,500 **** *** **** **** *** *** 1,000 Vehicle LY-411575-I + – + – Spleen B220 PC BrdU BrdU BM B220 PC BrdU BrdU 1.0 *** * 10 ** 2.5 * ** 0.8 1.5 0.6 1.0 4 0.4 0.5 0.2 0.0 0 0.0 + – + – Spleen PC BrdU BrdU BM PC BrdU BrdU Spleen BM Figure 8 | c-secretase regulates plasma cells in mice. (a) Immunized (OVA–LPS in alum) C57/BL6 mice were treated with the g-secretase inhibitor LY-411575-I or vehicle, and the surface display of BCMA in splenocytes was measured using flow cytometry 1 day later. BCMA expression on gated þ þ B220 CD138 cells is shown, a representative example (left) and compiled data from all 17 analysed animals (mean, ***Po0.001, unpaired t-test; right). The black symbols on the right indicate the samples shown on the left. Closed histograms indicate isotype controls. (b) NZB/W mice pretreated with BrdU received the g-secretase inhibitor LY-411575-I (red) or vehicle (blue) for 1 day. Surface expression of BCMA on all CD138 plasma cells (PC) and the BrdU and BrdU PC subgroups in the spleen and bone marrow (BM) was determined using flow cytometry. (c–e) Seven-day treatment of NZB/W mice with LY-411575-I. (c) BCMA surface expression in the spleen and BM on B220, and BrdU and BrdU plasma cells was determined. (d) Absolute number of plasma cells, BrdU and BrdU plasma cells in the spleen and BM. (e) Frequency (% of all cells in the organ) of plasma cells in the spleen and BM. Compiled data from 10 analysed animals per group (mean; *Po0.05; **Po0.01; ***Po0.001; ****Po0.0001, unpaired t-test). cells indicating that both short-lived plasmablasts and long-lived described previously for human recombinant monomeric BCMA, plasma cells are regulated. Application of a g-secretase inhibitor which bound and blocked only APRIL, while recombinant in vivo for 7 days enhanced the number of plasma cells in the human dimeric BCMA-Fc bound both APRIL and BAFF . bone marrow, but not in the spleen. We assume that the Binding assays revealed an avidity contribution of dimeric versus enhanced number of plasma cells in the bone marrow after monomeric recombinant BCMA, which resulted in a41,000-fold g-secretase inhibition is at least partly caused by an enhanced increase in apparent affinity of BCMA binding to BAFF . To get presence of mBCMA, since mBCMA mediates the induction of insight into the immunoregulatory capacity of sBCMA in vivo,we 7,8 the survival protein Mcl-1 in bone marrow plasma cells . The produced a recombinant variant of sBCMA (sBCMA-AT). differential effect of g-secretase inhibition on the plasma cell sBCMA-AT blocked APRIL-mediated NF-kB activation number in the spleen and bone marrow might be explained by the in vitro, but did, in contrast to what was observed with BCMA- previous observation that BCMA induced high expression of Fc previously , not affect plasma cell numbers in the bone Mcl-1 in bone marrow but not in spleen plasma cells .In marrow when given systemically. Murine bone marrow plasma addition, BCMA  /  mice had reduced plasma cell numbers in cells receive survival signals either via BAFF or APRIL , which is the bone marrow, but not in the spleen . consistent with our finding that sBCMA, which blocks mainly Reduction of plasma cell number in the bone marrow by APRIL, has no effect on plasma cell numbers in the bone marrow. g-secretase could be based on reducing membrane-bound Together, we conclude that the effect of BCMA shedding on the mBCMA or, alternatively, on the shed decoy sBCMA. We tested plasma cell number in the bone marrow is largely based on the the decoy potential of sBCMA in vitro and in vivo. In vitro, reduction of surface mBCMA. This would not exclude a function sBCMA blocked APRIL, but had little or no effect on BAFF, while of sBCMA, especially in a mucosal environment as a decoy for BCMA-Fc inhibited both BAFF and APRIL similarly. The APRIL, since APRIL-deficient mice had a reduced IgA response functional features we observed for sBCMA are similar to those to antigens encountered via the mucosal route . 8 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. BCMA (MFI) % Of max PC number (×10 ) % Of max BCMA expression (MFI) BCMA expression (MFI) PC number (×10 ) Frequency of PC (%) BCMA expression (MFI) BCMA expression (MFI) NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 ARTICLE We further found that inhibition of g-secretase affected B-cell albumin)/(serum IgG/serum albumin)). We examined longitudinally the sera of 10 patients with MS who were treated with high doses of steroids (1,000 mg per day subsets in vivo that do not express mBCMA, such as germinal methylprednisolone intravenously for 3–5 days) because of a relapse. Samples were centre B cells. This might be based on a blockade of NOTCH obtained directly before treatment, 3 days and 4 weeks later. We analysed serum pathways, which are known to determine T–B-cell lineage samples from 17 untreated and 22 treated SLE patients. The treated group included 32 33 commitment , maintenance of marginal zone B cells , patients treated with glucocorticosteroid, hydroxychloroquine, azathioprine, cyclophosphamide and mycophenolate mofetyl. Our study included 26 plasma survival of GC B cells and other, not yet identified, substrates 35 samples and 29 serum samples of 34 healthy control donors. Detailed data of the of g-secretase, which also contribute to B-cell activation . patients included in this study can be found in Supplementary Table 1. This study Together, our data indicate that inhibition of g-secretase was approved by the Ethics Committee of the Ludwig Maximilian University of enhances the survival receptor BCMA on plasma cells, increases Munich. Informed consent was obtained according to the Declaration of Helsinki. the number of plasma cells in the bone marrow and disturbs the B-cell compartment in the spleen. Cytokines and stimuli. In all stimulation and binding assays involving APRIL, we used mouse MegaAPRIL (EnzoLifeSciences, Farmingdale, NY), which binds both Release of membrane receptors is a general regulatory human and mouse BCMA and TACI; it is fused to FLAG-tag and is referred in this mechanism of inflammatory responses . TNFR1 is a study as APRIL. BAFF stimulation and binding assays were performed using either prominent example, and the immunoregulatory features of recombinant human BAFF (R&D Systems, Minneapolis, MN) or recombinant TNFR1 shedding were previously explored in vivo with a human BAFF–FLAG (EnzoLifeSciences). For stimulation of PBMC and native B knock-in mouse expressing a mutated nonsheddable TNFR1, cells, human IL-21 (EBioscience, San Diego, CA), TLR7 þ 8 ligand R848 (Sigma- Aldrich, St Louis, MO) and human IL-2 (R&D Systems) were used. Mouse L cells which resulted in autoinflammatory features and enhanced stably transfected with human CD40L were used and expression was continuously experimental autoimmune encephalitis . Biochemically, the monitored. release of sBCMA by g-secretase is different from the shedding of TNFR1, which is performed by the metalloprotease ADAM17 Antibodies and cell lines. To detect human BCMA the following antibodies were (ref. 36), and of TACI that is shed by ADAM10 (ref. 38). used: monoclonal antibodies A7D12.2 (IgG2b) and C12A3.2 (IgG1) (kindly pro- vided by BiogenIdec), and the polyclonal Ab AF193 (R&D Systems). The specificity Does shedding of BCMA also occur in human autoimmune of the two monoclonal antibodies to BCMA was confirmed using flow cytometry of diseases? We measured sBCMA levels in a compartmentalized BCMA-transfected MEF cells (Fig. 4a) and HEK cells (Fig. 5c), immunoprecipi- organ-specific autoimmune disease, MS, and in a systemic disease tation (ip) and subsequent western blot (Fig. 2f,g) or mass spectrometry (Fig. 2h). characterized by general activation of the B-cell compartment, Furthermore, an ELISA using A7D12.2 or C12A3.2 for coating and the polyclonal SLE. A hallmark of MS is B-cell persistence inside the brain goat antibody (AF193; R&D Systems, Minneapolis, MN) for detection with 39,40 recombinant BCMA (R&D Systems) as standard detected sBCMA with a sensitivity compartment with local Ig production , supported by local 41 of 30 pg ml . The following monoclonal antibodies were used for surface production of BAFF by astrocytes . We report that sBCMA is expression analysis using flow cytometry: fluorescein isothiocyanate (FITC)-con- elevated in the CSF in MS and closely correlates with intrathecal jugated anti-CD40L (BD PharMingen, San Diego, CA), FITC-conjugated anti- IgG production. Thus, sBCMA in the CSF reflects the local CD138 (Diaclone, Besanc¸on, France), Cy7-conjugated anti-CD27 and eFluor 450- conjugated anti-CD38 (EBioscience), Cy7-conjugated anti-CD19 (EBioscience). presence of Ig-secreting cells, which are known to be present in 2,39,40,42 For triple staining of BCMA with CD27 and CD38, we used the monoclonal the MS brain . This view is also supported by our antibody A7D12.2 and a 647-conjugated goat-anti-mouse IgG2 Ab (Invitrogen Life observation that reducing inflammation with natalizumab Technologies, Carlsbad, CA) along with the directly labelled IgG1 antibodies to decreases also sBCMA in the CSF. SLE is characterized by a CD38 and CD27 mentioned above. For the detection of APP-derived C99, the monoclonal antibody 4G8, SIG-39220 (Covance, Emeryville, CA) was used. systemic hyperactivation of the B-cell compartment with Single-cell suspensions were prepared from the bone marrow (femur and tibia) elevated surface expression of mBCMA on circulating immune and spleen. Mouse BCMA was detected using flow cytometry with biotin- cells . In this disease, we found a systemic elevation of sBCMA, conjugated anti-mouse BCMA (BAF 593; R&D Systems) along with eFluor450- or which is linked to disease activity. Further studies are required to PE-Cy7-conjugated streptavidin. BrdU staining of plasma cells was performed using a BrdU Flow Kit (BD Biosciences, San Jose, CA) according to the characterize sBCMA as a potential biomarker of SLE activity. manufacturer’s protocol. The detection of plasma cells was carried out with anti- sBCMA shedding adds to the complexity of the BAFF–APRIL CD138-PE (clone 281-2; BD Biosciences) for surface staining and anti-kappa- system. This system is highly relevant in health and disease Pacific Orange (clone 187.1; DRFZ) for intracellular staining. B and T cells were 13,45,46 and serves as a drug target . Release and function of identified with the following anti-mouse antibodies: CD21 (clone 7E9, BioLegend, sBCMA are of direct relevance for clinical trials targeting BAFF, CA, USA), CD23 (clone B3B4, BioLegend), CD24 (clone M1/69, BD Biosciences), CD93 (clone AA4.1, BioLegend), CD95 (clone Jo2, BD Biosciences), CD117 (clone APRIL and their receptors, which are currently under way . This 2B8, BD Biosciences), IgM (clone RMM-1, BioLegend), B220 (clone RA3-6B2, system also affects certain haematological malignancies such as DRFZ), IgD (clone 11-26c, DRFZ), GL-7 (clone GL-7, DRFZ), CD4 (clone GK1.5, 3 47 plasmacytoma , where sBCMA may also serve as a biomarker . DRFZ) and CD8 (clone 53-6.7, BioLegend). Identification of B-cell subsets in the high low/    þ þ Further, since g-secretase is involved not only in Alzheimer’s spleen: B1: IgM CD21 CD23 CD93 . Follicular B cells: IgM CD21 þ  high þ   þ CD23 CD93 . Marginal zone B cells: IgM CD21 CD23 CD93 . GL-7 : disease, but also in autoimmune diseases (including MS), where 48 GL-7 IgD . Identification of B-cell subsets in the bone marrow: pro-B cells: g-secretase inhibitors are being tested for therapeutic benefit , þ þ þ þ þ B220 CD93 CD117 . pre-B cells: B220 CD24 IgM IgD . Immature B our findings draw attention to potential side effects of g-secretase þ þ þ  þ low/ þ þ cells: B220 CD24 IgM IgD . Mature B cells: B220 CD24 IgM IgD . inhibitors related to the shedding of BCMA. Cytometric analysis was performed using a FACSCanto II cytometer (BD Biosciences) and data were analysed with the FlowJo software (Tree Star Inc.). Together, our study yields three main findings. First, Source and working concentration of antibodies used are listed in Supplementary g-secretase directly cleaves mBCMA. Regulation of surface Table 2. display of a ligand-binding receptor is a novel function of MEFs deficient for presenilin 1 (PS1) and presenilin 2 (PS2) (PS  /  ) were g-secretase. Second, g-secretase regulates the number of plasma kindly provided by Dr Bart De Strooper (Leuven, Belgium) ;PS  /  status was monitored using western blot analysis with the antibody PSEN1 (Epitomics, cells in the bone marrow. Third, the released sBCMA is a Burlingame, CA). The plasmacytoma cell line JK-6L (ref. 49) was kindly provided potential biomarker in human immunological diseases and could by Dr Silke Meister (Erlangen, Germany); expression of the surface markers CD138 be useful for therapeutic optimization. and BCMA was controlled. This cell line was cultured in RPMI supplemented with 10% FCS and in serum-free conditions (Hybridoma 6 direkt from Bio&SELL, Nu¨rnberg, Germany). Further, HeLa and HEK293T were applied in transfection Methods experiments. Clinical samples. In a first cohort we obtained plasma and the corresponding CSF from 37 untreated patients diagnosed with either clinical isolated syndrome (n ¼ 10) or MS (n ¼ 27) and from 20 untreated patients with other neurological Detection of BCMA and its ligands APRIL and BAFF. Surface expression of diseases; further five CSF samples from neuroborreliosis patients were obtained. In human BCMA was determined using flow cytometry on FACS Verse (BD Bios- a second cohort, we analysed CSF pairs from 25 additional MS patients before and ciences) using C12A3.2 or A7D12.2 and appropriate secondary antibodies. To about 1 year after continuous natalizumab therapy. The IgG production within the measure human sBCMA in the plasma, CSF or cell culture supernatants, a brain compartment (intrathecal) was calculated as the IgG index ((CSF IgG/CSF sandwich ELISA with polyclonal goat antibodies was used (BCMA/TNFRSF17 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications 9 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 Using lentiviral gene transduction, MEF deficient for PS1 and PS2 (PS  /  ) ELISA Duoset; R&D Systems). To detect sBCMA with an N-terminal FLAG, the were stably transduced with full-length human BCMA and then stably transduced monoclonal antibody anti-FLAG M2 antibody (Sigma-Aldrich) was used for with either wt PS1 or a catalytically inactive D385A mutant of PS1 ( ). coating and the biotinylated polyclonal goat-anti-BCMA antibody (R&D Systems) for detection. As a control the anti-myelin oligodendrocyte glycoprotein mono- clonal antibody 8.18C5 was used for coating. To measure APRIL in CSF, a sand- Enzyme inhibitors. We used the following g-secretase inhibitors: DAPT wich ELISA was used (Bender MedSystem, Vienna, Austria). BAFF levels in the (Calbiochem Merck, Darmstadt, Germany), L685,458 (R&D Systems), RO49290 serum were measured using the Quantokine human BAFF Elisa kit (R&D Systems). (Selleckchem, Houston, TX, USA) and the steroisomer SSR of LY-411575, referred To assess binding of sBCMA or BCMA-Fc (kindly provided by Biogen Idec) to as LY-411575-I (Sigma-Aldrich). This steroisomer of the g-secretase inhibitor APRIL and BAFF, an ELISA was used with the anti-FLAG M2 monoclonal LY-411575 is also a g-secretase inhibitor as observed with the APP fragment C99 antibody (Sigma-Aldrich) for coating to bind FLAG-tagged APRIL or and BCMA as substrates (Supplementary Figures 3 and 8). Metalloproteases were BAFF–FLAG. Then, sBCMA from supernatant of plasmacytoma cells, which was inhibited with TAPI-1 (EMD Chemicals/Calbiochem, Inc. Gibbstown US). concentrated with Amicon Ultra 3 K devices (Merck Millipore Ltd., Ireland), or Corresponding concentrations of DMSO (Sigma-Aldrich) were used as vehicle BCMA-Fc was added and incubated for 2 h at room temperature. BCMA was controls. Reduced BCMA shedding was observed at a vehicle concentration of 0.5% detected as described above. DMSO, which corresponds to 50 mM TAPI-1. Immunoprecipitation and western blot analysis. Immunoprecipitation of NF-jB reporter assay. To measure NF-kB activation, HEK293T cells were BCMA was performed with the monoclonal antibody A7D12.2, the monoclonal transiently transfected with a plasmid containing a firefly luciferase reporter gene antibody C12A3.2 or a polyclonal goat antibody (AF193; R&D Systems). These under the control of an NF-kB transcriptional response element, a plasmid with a antibodies were coupled to Dynabeads Protein G (Life Technologies, AS, Oslo) and Renilla reniformis luciferase reporter gene for normalization and the indicated cross-linked with bis-sulfosuccinimidyl-suberate (Pierce Chemical Co., Rockford, amounts of expression or control plasmids using Lipofectamine 2000 (Invitrogen IL). After successive incubation with either supernatant of plasmacytoma cells or Life Technologies). We used human BCMA or human TACI for transfection. To serum, we eluted with glycine or SDS loading buffer (NuPAGE LDS Sample Buffer, determine the effect of g-secretase inhibition on NF-kB activation, cells transfected Life Technologies). Cells were lysed at 4 C for 1 h in NP-40 lysis buffer (150 mM with BCMA were treated with DAPT or a solvent control and 6 h later stimulated NaCl, 50 mM Tris pH 7.5, 1% Nonidet P-40) containing complete protease with APRIL or BAFF. To analyse a possible decoy function, DAPT along with inhibitor cocktail (Roche Applied Science, Penzberg, Germany). BCMA was BAFF or APRIL were added to BCMA-Fc or supernatants generated by HEK293T detected by western blot analysis with the monoclonal antibody C12A3.2. Blots cells that had been transfected with full-length BCMA or an empty control vector were developed using the mouse true blot goat anti-mouse IgG-HRP system as described above. These supernatants were incubated at 37 C for 30 min and (EBioscience) and enhanced chemiluminescence (ECL). Expression and then added to BCMA or TACI-transfected and DAPT-treated cells used for the endoproteolysis of PS1 was analysed by immunoblotting of cell lysates with the reporter assay. 16 h after stimulation cells were harvested and cell lysates were antibody PSEN1 (Epitomics); maturation of nicastrin (NCT) was evaluated by prepared using passive lysis buffer (Promega, Madison, WI, USA) and the reporter immunoblotting with the antibody N1660 (Sigma-Aldrich). gene activity was measured using firefly luciferase substrate (Biozym, Hameln, Germany) and renilla luciferase substrate (Promega) respectively. Mass spectrometry and sample preparation. sBCMA was purified by immu- noprecipitation and obtained by acidic elution. The eluate was then desalted and Flow cytometry sorting and quantitative RT-PCR. Human B cells were positively concentrated using StageTips, C18, microcolumns (Thermo Scientific, Bremen, selected from PBMC with CD19 MACS beads (Miltenyi Biotec) and differentiated Germany). Then two approaches were followed. (A) The material was digested in into IgG-secreting cells by coculture with CD40L-transfected mouse L cells and solution by trypsin or chymotrypsin and analysed using mass spectrometry (LTQ þþ þ recombinant human IL-21 as described above. CD27 CD38 and CD38 cells Orbitrap XL) as described before . (B) The immunoprecipitated material was were sorted using the BD FACS Aria cell sorting system (Becton Dickinson, separated via an SDS gel, silver-stained and the band corresponding to BCMA as Heidelberg, Germany). For quantitative PCR analysis, RNA was isolated using the detected using western blot analysis was excised and analysed using mass RNeasy Micro Kit (Qiagen, Hilden, Germany) and cDNA was generated using the spectrometry (LTQ Orbitrap XL). High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Darmstadt, Germany). To detect BCMA transcripts, primer and probes were used as published . As housekeeping gene Cyclophilin A (PPIA) was used and detected Cell culture and gene transfer. Two experimental systems were used to differ- with the primer and probes 43263416E from Applied Biosystems. Reactions were entiate primary human B cells to Ig-secreting cells. First, B cells were positively carried out in duplicate using TaqMan assays in combination with the TaqMan selected from PBMC with CD19 MACS beads (Miltenyi Biotec, Bergisch Gladbach, PCR Core Reagent Kit (Applied Biosystems). Samples were run in MicroAmp Germany) and differentiated into IgG-secreting cells by coculture with CD40L- Optical 96-well reaction plates in a 7900HT Fast Real-Time PCR System (Applied transfected mouse L cells and recombinant human IL-21 (50 ng ml ). Further Biosystems). Data were analysed using the SDSv2.3 software (Applied Biosystems). þ þ þ separation of CD19 CD38 and CD19 CD38 cells was performed using CD38 MACS beads. Second, PBMC were stimulated with the TLR7 þ 8 ligand 1  1 19 R848 (2.5 mgml ) and IL-2 (1,000 IU ml ) as described . Secreted IgG was sBCMA-AT fusion protein. The cDNA of the extracellular part of mouse BCMA quantified by ELISA (Mabtech, Nacks Strand, Sweden). To analyse the effects of (sBCMA) was synthesized using the Integrated DNA Technology with a BspEI and DAPT and TAPI-1, these Ig-secreting cells were washed after 7 days and cultured 0 0 0 an Age I site at the 5 and 3 ends, respectively, and fused to the 3 end of human for another 16 h with either inhibitor. AT in the expression vector pTT5 (ref. 52). The resulting fusion between AT and To measure APRIL-induced survival, mouse B cells (magnetically isolated from sBCMA (extracellular residues from position 2 to 49 lacking the first methionine) is spleen using the EasySep Mouse B Cell Isolation Kit (Stemcell Technologie, separated by a short flexible peptide linker TGSGSGA and terminates with Vancouver, Canada)) were cultured in 96-well microtiter plates, which were hexahistidine tag to facilitate purification. The free cysteine residue on the surface 1  1 precoated with anti-IgM (5mgml ). APRIL (100 ng ml ) was added for 48 h and of wild-type AT was replaced by a serine residue to prevent dimerization in cross-linked with anti-FLAG monoclonal antibody (Sigma-Aldrich). Supernatants oxidizing environments. The fusion protein was expressed in HEK293-EBNA cells generated by HEK293T cells that had been transfected with full-length BCMA 52 as described previously . MW was determined with standard Coomassie gel and (OriGene Technologies, Inc., Rockville) or an empty control vector and therefore the size was analysed using gel filtration with the Superose 12 HR 10/30 gel either contained sBCMA or did not, were added at a final sBCMA concentration of filtration column (Amersham/GE Healthcare Life Sciences, Freiburg, Germany) on 1  1 200 ng ml and 400 ng ml . Cell survival was quantified by flow cytometry using a fast protein liquid chromatography column (FPLC). Function of sBCMA-AT was TO-PRO -3 Iodide viability dye (Invitrogen Life Technologies) and APRIL-induced tested using NF-kB reporter assays. survival was calculated as followed: 100 x [(cell survival in presence of APRIL—cell survival without APRIL)/cell survival without APRIL]. To obtain BCMA with an N-terminal FLAG-tag, full-length human BCMA c-secretase inhibitor and recombinant sBCMA-AT in mice. Two different (h184) from a human BCMA cDNA clone (OriGene Technologies) was first mouse models were applied. First, 2-month-old C57/BL6 mice were immunized amplified by PCR. The PCR fragment was then digested and ligated into the intraperitoneally with 100 mg OVA and 10 mg LPS in alum (a protocol based on ref. pCMV6-AN-DDK vector (OriGene Technologies). This construct does not contain 53) and killed 10 days later. Immunized mice received an intraperitoneal dose 23  1 a signal peptide, since the native BCMA also does not contain one . To obtain a (10 mg kg ) of LY-411575-I on day 9, followed by another dose 6 h before being BCMA variant with a prolonged extracellular part, the extracellular part of human killed. Second, female 16- to 18-week NZB/W F1 mice were bred at the animal BCMA was amplified by PCR, digested and ligated at the N-terminal end of full- facility of the German Rheumatism Research Center Berlin (DRFZ) under defined, length human BCMA clone (OriGene Technologies). We thereby doubled the pathogen-free conditions. To distinguish between short-lived and long-lived extracellular part, and called the new construct BCMA–BCMA. The plasmid plasma cells, we fed the mice BrdU (Sigma-Aldrich, 1 mg ml ) with 1% glucose in peak12-SP-C99 expresses the C-terminal 99 amino acids of APP together with the drinking water for 2 weeks. One week after the start of BrdU feeding, mice were APP signal peptide. C99 is a direct substrate for g-secretase . Peak12-SP-C99 was treated daily with the g-secretase inhibitor LY-411575-I intraperitoneally at a dose 51  1 obtained by cloning SP-C99 from the pCEP4 vector into the peak12 plasmid . of 10 mg kg for 7 days. Finally, 2-month-old NZB/W mice were immunized HEK or HeLa cells were transfected with 200 ng of respective expression plasmids intraperitoneally with OVA (100 mg per mouse and alum (Thermo) as adjuvant) as using lipofectamine 2,000 (Invitrogen Life Technologies). described and boosted at 3 months. At 4 months, mice were treated for four 10 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 ARTICLE consecutive weeks intraperitoneally with recombinant sBCMA-AT fusion protein 29. Hemming, M. L., Elias, J. E., Gygi, S. P. & Selkoe, D. J. Proteomic profiling of (200 mg three times a week). The concentration of this fusion protein in the serum gamma-secretase substrates and mapping of substrate requirements. PLoS Biol. of the injected mice was determined with an AT ELISA, which is specific for 6, e257 (2008). human AT (ab189579 from Abcam, Cambridge, UK). Single-cell suspensions were 30. Day, E. S. et al. Selectivity of BAFF/BLyS and APRIL for binding to the TNF prepared from the spleen and bone marrow (femur and tibia) as described before. family receptors BAFFR/BR3 and BCMA. Biochemistry (Mosc.) 44, 1919–1931 Cells were first incubated with anti-CD138-PE and then fixed with Cytofix/ (2005). Cytoperm buffer (BD Bioscience) for 30 min. Intracellular Ig kappa light chain was 31. Castigli, E. et al. Impaired IgA class switching in APRIL-deficient mice. Proc. detected by anti-kappa-Pacific Orange and OVA-specific antibodies were detected Natl Acad. Sci. USA 101, 3903–3908 (2004). by intracellular OVA conjugated with FITC. 32. Radtke, F., MacDonald, H. R. & Tacchini-Cottier, F. Regulation of innate and adaptive immunity by Notch. Nat. Rev. Immunol. 13, 427–437 (2013). 33. Simonetti, G. et al. IRF4 controls the positioning of mature B cells in the References lymphoid microenvironments by regulating NOTCH2 expression and activity. 1. Do¨rner, T., Radbruch, A. & Burmester, G. R. B-cell-directed therapies for J. Exp. Med. 210, 2887–2902 (2013). autoimmune disease. Nat. Rev. Rheumatol. 5, 433–441 (2009). 34. Yoon, S. O., Zhang, X., Berner, P., Blom, B. & Choi, Y. S. Notch ligands 2. Krumbholz, M., Derfuss, T., Hohlfeld, R. & Meinl, E. B cells and antibodies in expressed by follicular dendritic cells protect germinal center B cells from multiple sclerosis pathogenesis and therapy. Nat. Rev. Neurol. 8, 613–623 apoptosis. J. Immunol. 183, 352–358 (2009). (2012). 35. Yagi, T. et al. Defective signal transduction in B lymphocytes lacking presenilin 3. Mackay, F. & Schneider, P. Cracking the BAFF code. Nat. Rev. Immunol. 9, proteins. Proc. Natl Acad. Sci. USA 105, 979–984 (2008). 491–502 (2009). 36. Levine, S. J. Molecular mechanisms of soluble cytokine receptor generation. 4. Yang, M. et al. B cell maturation antigen, the receptor for a proliferation- J. Biol. Chem. 283, 14177–14181 (2008). inducing ligand and B cell-activating factor of the TNF family, induces antigen 37. Xanthoulea, S. et al. Tumor necrosis factor (TNF) receptor shedding controls presentation in B cells. J. Immunol. 175, 2814–2824 (2005). thresholds of innate immune activation that balance opposing TNF functions in 5. Darce, J. R., Arendt, B. K., Wu, X. & Jelinek, D. F. Regulated expression of infectious and inflammatory diseases. J. Exp. Med. 200, 367–376 (2004). BAFF-binding receptors during human B cell differentiation. J. Immunol. 179, 38. Hoffmann, F. S. et al. The Immunoregulator Soluble TACI Is Released by 7276–7286 (2007). ADAM10 and Reflects B Cell Activation in Autoimmunity. J. Immunol. 194, 6. Avery, D. T. et al. BAFF selectively enhances the survival of plasmablasts 542–552 (2015). generated from human memory B cells. J. Clin. Invest. 112, 286–297 (2003). 39. von Budingen, H. C., Bar-Or, A. & Zamvil, S. S. B cells in multiple sclerosis: 7. O’Connor, B. P. et al. BCMA is essential for the survival of long-lived bone connecting the dots. Curr. Opin. Immunol. 23, 713–720 (2011). marrow plasma cells. J. Exp. Med. 199, 91–98 (2004). 40. Nylander, A. & Hafler, D. A. Multiple sclerosis. J. Clin. Invest. 122, 1180–1188 8. Peperzak, V. et al. Mcl-1 is essential for the survival of plasma cells. Nat. (2012). Immunol. 14, 290–297 (2013). 41. Krumbholz, M. et al. BAFF is produced by astrocytes and up-regulated 9. Belnoue, E. et al. APRIL is critical for plasmablast survival in the bone in multiple sclerosis lesions and primary central nervous system lymphoma. marrow and poorly expressed by early-life bone marrow stromal cells. Blood J. Exp. Med. 201, 195–200 (2005). 111, 2755–2764 (2008). 42. Cepok, S. et al. Short-lived plasma blasts are the main B cell effector subset 10. Benson, M. J. et al. Cutting edge: the dependence of plasma cells and during the course of multiple sclerosis. Brain 128, 1667–1676 (2005). independence of memory B cells on BAFF and APRIL. J. Immunol. 180, 43. Stohl, W. et al. B lymphocyte stimulator overexpression in patients with 3655–3659 (2008). systemic lupus erythematosus: longitudinal observations. Arthritis Rheum. 48, 11. Amanna, I. J. & Slifka, M. K. Mechanisms that determine plasma cell lifespan 3475–3486 (2003). and the duration of humoral immunity. Immunol. Rev. 236, 125–138 (2010). 44. Kim, J., Gross, J. A., Dillon, S. R., Min, J. K. & Elkon, K. B. Increased 12. Hiepe, F. et al. Long-lived autoreactive plasma cells drive persistent BCMA expression in lupus marks activated B cells, and BCMA receptor autoimmune inflammation. Nat. Rev. Rheumatol. 7, 170–178 (2011). engagement enhances the response to TLR9 stimulation. Autoimmunity 44, 69– 13. Croft, M., Benedict, C. A. & Ware, C. F. Clinical targeting of the TNF and 81 (2011). TNFR superfamilies. Nat. Rev. Drug Discov. 12, 147–168 (2013). 45. Vincent, F. B., Saulep-Easton, D., Figgett, W. A., Fairfax, K. A. & Mackay, F. 14. Stohl, W., Scholz, J. L. & Cancro, M. P. Targeting BLyS in rheumatic disease: The BAFF/APRIL system: emerging functions beyond B cell biology and the sometimes-bumpy road from bench to bedside. Curr. Opin. Rheumatol. 23, autoimmunity. Cytokine Growth Factor Rev. 24, 203–215 (2013). 305–310 (2011). 46. Osorio, C. et al. Selective regulation of axonal growth from developing 15. Kappos, L. et al. Atacicept in multiple sclerosis (ATAMS): a randomised, hippocampal neurons by tumor necrosis factor superfamily member APRIL. placebo-controlled, double-blind, phase 2 trial. Lancet Neurol. 13, 353–363 (2014). Mol. Cell Neurosci. 59C, 24–36 (2014). 16. Kopan, R. & Ilagan, M. X. Gamma-secretase: proteasome of the membrane? 47. Sanchez, E. et al. Serum B-cell maturation antigen is elevated in multiple Nat. Rev. Mol. Cell Biol. 5, 499–504 (2004). myeloma and correlates with disease status and survival. Br. J. Haematol. 158, 17. Selkoe, D. J. & Wolfe, M. S. Presenilin: running with scissors in the membrane. 727–738 (2012). Cell 131, 215–221 (2007). 48. Minter, L. M. et al. Inhibitors of gamma-secretase block in vivo and in vitro 18. Ransohoff, R. M. & Engelhardt, B. The anatomical and cellular basis of immune T helper type 1 polarization by preventing Notch upregulation of Tbx21. Nat. surveillance in the central nervous system. Nat. Rev. Immunol. 12, 623–635 Immunol. 6, 680–688 (2005). (2012). 49. Burger, R. et al. Interleukin-6 production in B-cell neoplasias and Castleman’s 19. Pinna, D., Corti, D., Jarrossay, D., Sallusto, F. & Lanzavecchia, A. Clonal disease: evidence for an additional paracrine loop. Ann. Hematol. 69, 25–31 dissection of the human memory B-cell repertoire following infection and (1994). vaccination. Eur. J. Immunol. 39, 1260–1270 (2009). 50. Hauck, S. M. et al. Deciphering membrane-associated molecular processes in 20. Avery, D. T. et al. Increased expression of CD27 on activated human memory B target tissue of autoimmune uveitis by label-free quantitative mass cells correlates with their commitment to the plasma cell lineage. J. Immunol. spectrometry. Mol. Cell Proteomics 9, 2292–2305 (2010). 174, 4034–4042 (2005). 51. Mitterreiter, S. et al. Bepridil and amiodarone simultaneously target the 21. Herreman, A. et al. Total inactivation of gamma-secretase activity in presenilin- Alzheimer’s disease beta- and gamma-secretase via distinct mechanisms. deficient embryonic stem cells. Nat. Cell Biol. 2, 461–462 (2000). J. Neurosci. 30, 8974–8983 (2010). 22. Yamasaki, A. et al. The GxGD motif of presenilin contributes to catalytic 52. Perera, N. C. et al. NSP4 is stored in azurophil granules and released function and substrate identification of gamma-secretase. J. Neurosci. 26, by activated neutrophils as active endoprotease with restricted specificity. 3821–3828 (2006). J. Immunol. 191, 2700–2707 (2013). 23. Bossen, C. & Schneider, P. BAFF, APRIL and their receptors: structure, 53. Pasare, C. & Medzhitov, R. Control of B-cell responses by Toll-like receptors. function and signaling. Semin. Immunol. 18, 263–275 (2006). Nature 438, 364–368 (2005). 24. Luzina, I. G. et al. Spontaneous formation of germinal centers in autoimmune 54. Hoyer, B. F. et al. Short-lived plasmablasts and long-lived plasma cells mice. J. Leukoc. Biol. 70, 578–584 (2001). contribute to chronic humoral autoimmunity in NZB/W mice. J. Exp. Med. 25. Cassese, G. et al. Inflamed kidneys of NZB/W mice are a major site for the 199, 1577–1584 (2004). homeostasis of plasma cells. Eur. J. Immunol. 31, 2726–2732 (2001). 26. Yu, G. et al. APRIL and TALL-I and receptors BCMA and TACI: system for regulating humoral immunity. Nat. Immunol. 1, 252–256 (2000). Acknowledgements 27. Rennert, P. et al. A soluble form of B cell maturation antigen, a receptor for the We thank Drs K. Berer, K. Dornmair and J. Ellwart for support, Drs S. Kalled and tumor necrosis factor family member APRIL, inhibits tumor cell growth. J. Exp. K. Otipoby for monoclonal antibodies to BCMA, K. Ogston and Drs C. Haass, H. Steiner, Med. 192, 1677–1684 (2000). N. Kawakami and G. Krishnamoorthy for comments on the manuscript. This work was 28. Struhl, G. & Adachi, A. Requirements for presenilin-dependent cleavage of supported by the DFG (SFB TRR 128, SFB TRR 130, SFB 650), the BMBF (‘Kran- notch and other transmembrane proteins. Mol. Cell 6, 625–636 (2000). kheitsbezogenes Kompetenznetz Multiple Sklerose’ and JPND RiMOD’), the Breuer NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications 11 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8333 Foundation Alzheimer Award, the Gemeinnu¨tzige-Hertie-Stiftung (P1130076 and HERTIE Competing financial interests: The authors declare no competing financial interest. Senior Professorship to HW), the Munich Cluster for Systems Neurology (ExC 1010 Reprints and permission information is available online at http://npg.nature.com/ SyNergy), the GIF and the Verein zur Therapieforschung fu¨r Multiple-Sklerose-Kranke. reprintsandpermissions/ Author contributions How to cite this article: Laurent, S. A. et al. g-secretase directly sheds the survival S.A.L., F.H., P.-H.K., Q.C., Y.C., M.S.-S., S.M.H., E.S., JW and H.R. performed experi- receptor BCMA from plasma cells. Nat. Commun. 6:7333 doi: 10.1038/ncomms8333 ments and analysed data; M.Kr. was involved in study design; F.W., T.A., F.H., M.Kh., (2015). T.O. and H.-W.P. provided clinical samples; D.J., H.W., R.H., S.F.L., S.A.L. and E.M. designed the study, analysed data and wrote the paper. All authors discussed results and This work is licensed under a Creative Commons Attribution 4.0 commented on the manuscript. International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, Additional information Supplementary Information accompanies this paper at http://www.nature.com/ users will need to obtain permission from the license holder to reproduce the material. naturecommunications To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ 12 NATURE COMMUNICATIONS | 6:7333 | DOI: 10.1038/ncomms8333 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved.

Journal

Nature CommunicationsSpringer Journals

Published: Jun 11, 2015

There are no references for this article.