Published January 19, 2015, doi:10.4049/jimmunol.1401866 The Journal of Immunology

Unique Ligand-Binding Property of the Human IgM Fc

Kazuhito Honjo,* Yoshiki Kubagawa,* John F. Kearney,† and Hiromi Kubagawa*,1

The IgM (FcmR) is the newest FcR, and coligation of FcmR and Fas/CD95 on Jurkat cells with agonistic IgM anti-Fas mAb was shown to inhibit Fas-induced apoptosis. The ligand-binding activity of human FcmR was further examined. FcmR- mediated protection from apoptosis was partially blocked by addition of 104 molar excess of IgM or its soluble immune complexes, but it could be inhibited by addition of 10-fold excess of IgM anti-CD2 mAb. This suggests that FcmR binds more efficiently to the Fc portion of IgM reactive with plasma-membrane than to the Fc portion of IgM in solution. The former interaction occurred in cis on the same cell surface, but not in trans between neighboring cells. This cis engagement of FcmR resulted in modulation of Ca2+ mobilization via CD2 on Jurkat cells or BCRs on blood B cells upon cross-linkage with the corresponding IgM mAbs. Several functional changes were observed with FcmR mutants: 1) significant increase in IgM ligand binding in the cytoplasmic tail-deletion mutant, 2) enhanced cap formation in FcmR upon IgM binding at 4˚C with a point mutation of the transmembrane His to Phe, and 3) less protective activity of FcmR in IgM anti-Fas mAb-mediated apoptosis assays with a point mutation of the membrane-proximal Tyr to Phe. These findings show the importance of the cis engagement of FcmR and its critical role in receptor function. Hence, FcmR on B, T, and NK cells may modulate the function of surface proteins recognized by natural or immune IgM Abs on the shared membrane cell surface. The Journal of Immunology, 2015, 194: 000–000.

ntibodies have dual binding activity: to Ag via their response, has defied identification, despite extensive biochemical N-terminal variable regions, and to effector molecules evidence of IgM Fc–binding proteins accumulated over decades A such as FcRs via their C-terminal constant regions. FcRs (11–13). We previously successfully identified a cDNA encoding are expressed by many different cell types, and their interaction an authentic FcmR from cDNA libraries of human B-lineage cells with Abs can initiate a broad spectrum of effector functions es- using a functional cloning strategy (14). FCMR is a single-copy sential in host defense. These functions include phagocytosis gene located on 1q32.2, adjacent to two other IgM- of Ab-coated microbes, lysosomal degradation of endocytosed binding receptor genes, Fca/mR and polymeric Ig receptor. The immune complexes, Ab-dependent cell-mediated cytotoxicity, predicted FcmR is a transmembrane that consists of secretion of cytokines and chemokines, release of potent inflam- a single V-set Ig-like domain responsible for Fcm binding, an matory mediators, regulation of Ab production by B cells, survival additional extracellular region with no known domain structure, of plasma cells, and presentation of degradable as well as non- a transmembrane segment containing a charged His residue, and degradable Ags (1–7). These diverse functions depend on the Ab a relatively long cytoplasmic tail (118 aa) containing three con- isotype and the cell type expressing the FcR. Structurally and served Tyr and five conserved Ser residues. FcmR binds pentameric functionally diverse FcRs, namely FcR for IgG (FcgRI/CD64, IgM with a surprisingly high avidity of ∼10 nM as determined by FcgRII/CD32, FcgRIII/CD16, FcgRIV, FcRn), IgE (FcεRI, Scatchard plot analysis, with the assumption of a 1:1 stoichiometry FcεRII/CD23), IgA (FcaR/CD89), and both polymeric IgA and of FcmR to IgM ligand. Upon ligation of FcmR with IgM IgM (Fca/mR/CD351), have been characterized extensively at ligands, both Tyr and Ser residues in the cytoplasmic tail are both the protein and genetic levels (1–5, 8–10). phosphorylated (14), and receptors are rapidly internalized into It has long been a puzzle why an FcR for IgM (FcmR), the first Ig lysosomal compartments (15). Unlike other FcRs, the expres- isotype to appear during phylogeny, ontogeny, and the immune sion of FcmR is restricted to : B, T, and NK cells (14, 16). This suggests potentially distinct functions of FcmRas compared with other FcRs that are mainly expressed by myeloid *Division of Laboratory Medicine, Department of Pathology, University of Alabama cells. at Birmingham, Birmingham, AL 35294; and †Department of Microbiology, Univer- Alternatively, the FcmR was initially designated as Fas apo- sity of Alabama at Birmingham, Birmingham, AL 35294 ptotic inhibitory molecule 3 (FAIM3), because coligation of Fas 1 Current address: Deutsches Rheuma-Forschungszentrum Berlin, Berlin, Germany. and FcmR/FAIM3 with an agonistic IgM anti-Fas mAb prevented Received for publication July 22, 2014. Accepted for publication December 16, 2014. Fas-mediated apoptosis (17). Unlike the effect of IgM anti-Fas This work was supported in part by National Institutes of Health/National Institute mAb, however, ligation of Fas with an agonistic IgG mAb in- of and Infectious Diseases Grants AI4782-36 (to J.F.K.) and R21AI094625 (to H.K.). duced apoptosis irrespective of the expression of FcmR/FAIM3 (14, 16, 18). Notably, coligation of Fas and FcmR/FAIM3 with Address correspondence and reprint requests to Dr. Hiromi Kubagawa, Department of Pathology, University of Alabama at Birmingham, 506 Shelby, 1825 University the corresponding mouse IgG mAbs plus a secondary reagent Boulevard, Birmingham, AL 35294-2182. E-mail address: [email protected] [e.g., F(ab9)2 fragments of anti-mouse g Ab] had no demonstrable The online version of this article contains supplemental material. effects on the IgG anti-Fas mAb-induced apoptosis (14). This 2+ Abbreviations used in this article: 7-AAD, 7-aminoactinomycin D; [Ca ]i, intracel- suggests that the antiapoptotic activity of FcmR/FAIM3 depends 2+ lular Ca concentration; FAIM3, Fas apoptosis inhibitory molecule 3; NP, nitro- on usage of the IgM anti-Fas mAb, and not on physical proximity phenyl; WT, wild-type. of two receptors induced by artificial coligation. However, be- Copyright Ó 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00 cause the FcmR-mediated inhibition of the IgM anti-Fas mAb-

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1401866 2 THE CIS ENGAGEMENT OF FcmR induced apoptosis is a reproducible and well-defined function, we FcmR mutants employed this antiapoptotic assay in the present study as one of FcmR cDNA constructs with point mutations (H253F, Y315F, Y366F, or the readouts to explore the mode of action of FcmR. In this study, Y385F) were generated by a QuickChange site-directed mutagenesis we have defined the unique ligand-binding properties of FcmR and (Stratagene) using nonmutated FcmR cDNA as a template DNA and a set of the key residues involved in its receptor function. primers: 1) 59-GAAGGCCAAGGATTTTTCATCCTGATCCCGACCAT-39 and 59-ATGGTCGGGATCAGGATGAAAAATCCTTGCCCTTC-39for H253F, 2) 59-CTCCCAAAACAACATCTTCAGCGCCTGCC-39 and 59-GGCA- Materials and Methods GGCGCTGAAGATGTTGTTTTGGGAG-39 for Y315F, 3) 59-GACCA- Cells GCTGTGAATTTGTGAGCCTCTACCACCAG-39 and 59-CTGGTGGTA- 9 9 The human leukemic (Jurkat) and mouse thymoma (BW5147) lines, GAGGCTCACAAATTCACAGCTGGTC-3 for Y366F, and 4) 5 -GG- 9 9 which stably expressed either only GFP as a control or both GFP and ACAGTGATTCAGATGACTTCATCAATGTTCCTG-3 and 5 -CAGGA- 9 human FcmR, were prepared by retrovirus-mediated transduction as ACATTGATGAAGTCATCTGAATCACTGTCC-3 for Y385F (underlining previously described (14) and were maintained in media containing indicates mutation sites), according to the manufacturer’s recommendation. For FcmR cytoplasmic tail deletion, an FcmR cDNA construct with a deletion puromycin at 1.5 and 0.75 mg/ml, respectively. BW5147 cells stably of most of the cytoplasmic tail (A281–A390; DCy) but with an inclusion of expressing human FcmR (without GFP) and wild-type (WT) Jurkat or BW5147 cells were maintained in media only as described (14). For Ca2+ the posttransmembrane basic aa-rich region (K273–K280) was generated by m mobilization, mononuclear cells isolated from healthy individuals by PCR using PrimeStar HS DNA polymerase (TaKaRa), Fc R cDNA as 9 GAATTC Ficoll-Hypaque density gradient centrifugation were enriched for B cells a template DNA, and a set of primers: 5 - TAGAAGGGACA- 9 9 GCGGCCGC 9 using a isolation kit II (Miltenyi Biotec) with MACS. The fre- ATGGACT-3 and 5 - CTATTTCCTCCTTTCAACGGCC-3 + Eco Not quency of CD19 B cells in the resultant enriched fractions was .95% (italics indicate RI and I sites, and underlining indicates translation start and stop sites). The amplified PCR products of the expected size were as determined by flow cytometry. This study involving human subjects gel purified and subcloned into the ZeroBlunt TOPO vector (Invitrogen). All was approved by the institutional review board of the University of m Alabama at Birmingham. plasmid DNAs encoding Fc R mutants were digested with appropriate re- striction enzymes, gel purified, and ligated into the pMX-PIE retroviral ex- Immunofluorescence analysis pression vector as described (14). After confirming the correct sequences with the expected mutation of the resultant cDNAs at both strands of DNA, Flow cytometric analysis of cell surface FcmR on transductants was per- they were transfected into 293T-A packaging cell line before transducing formed by using biotin-labeled, receptor-specific mAb (HM14, g1k iso- Jurkat cells as described (14). GFP+ transductants were enriched by FACS type) and isotype-matched control mAb, along with PE-labeled streptavidin before IgM binding, apoptosis, and Ca2+ mobilization analyses. as described (14). For IgM binding, cells were incubated with biotin-labeled mouse myeloma IgM (TEPC183) at 1 mg/ml or PBS, washed, and then Epifluorescence microscopic analysis incubated with PE-labeled streptavidin. Stained cells were examined by an Accuri C6 flow cytometer (BD Biosciences), and flow cytometric data were Viable cells were incubated for 20 min on ice with Alexa Fluor 555–labeled, analyzed with FlowJo software (Tree Star). affinity-purified mouse monoclonal IgM (MD4 anti-hen egg lysozyme mAb) at 10 mg/ml in PBS with or without 0.1% NaN3, washed in PBS/1% Apoptosis assay FCS/0.1% NaN3, and cytocentrifuged onto glass slides prior to being air- dried and fixed in 95% ethanol/5% glacial acetic acid at 220˚C. Some Apoptosis assays were performed essentially the same as described before aliquots of cells were resuspended in warmed RPMI 1640/10% FCS and m + (14). In brief, Fc R or control Jurkat cells were cultured for 15 h with or incubated for an additional 10 min at 37˚C before cytocentrifugation. The without agonistic IgM anti-Fas mAb (CH11, 10 ng/ml; Millipore), along stained cells were examined under a Leica/Leitz DMRB microscope. with various concentrations of inhibitors: 1) IgM mAbs specific for Images were acquired with a Hamamatsu C4742 video camera and pro- nitrophenyl (NP) hapten (B1-8) (19) or a1-3 dextran (clone 1-21) (20), 2) cessed with Openlab software. their soluble immune complexes with NP25-coupled BSA (molecular mass of ∼70 kDa; Biosearch Technologies) or a1-3 dextran (molecular mass of ∼1000 kDa), 3) FcmR-blocking (HM7) or nonblocking (HM3) mAbs (both Results g2bk), and 4) IgM mAb specific for the human CD2 (C373) or Jurkat TCR Interaction of FcmR with soluble IgM immune complexes (C305) (21). The latter two mAbs were provided by Dr. Arthur Weiss versus IgM bound to membrane components via the Fab region (Howard Hughes Medical Institute, University of California, San Fran- cisco) and were purified from ascites by Sephacryl S300 gel filtration, and In our previous studies apoptosis-prone human Jurkat cells stably the remainders were purified by Ag-coupled affinity columns. For soluble expressing FcmR were shown to be protected from Fas/CD95- immune complexes, IgM anti-NP and IgM anti-dextran mAbs and the mediated apoptosis when ligated with an agonistic IgM anti-Fas corresponding Ags were premixed to reach a final concentration in the cultures of 104 or 105 ng/ml for IgM mAbs with different molar ratios of mAb (CH11), but not when ligated with an agonistic IgG anti-Fas Ab/Ag (1:10 to 1:1022). Apoptotic cells were detected by staining with mAb or (14, 18). Notably, coligation of FcmRand 7-aminoactinomycin D (7-AAD) and allophycocyanin-labeled annexin Fas with the corresponding IgG mAbs plus a common secondary + V (BD Biosciences) as described (14). For mixed cell cultures, FcmR or reagent could not prevent IgG anti-Fas mAb-induced apoptosis control Jurkat cells were cocultured with 10-fold excess number of FcmR+ BW5147 cells in the presence or absence of CH11 anti-Fas mAb, followed (14). To determine whether this apoptosis protection mediated by by cell surface staining with PE–anti-mouse CD3 mAb (145-2C11; BD FcmR is affected by IgM ligands, IgM mAbs specific for NP Biosciences). Apoptosis was assessed for Jurkat cell populations by ex- (clone B1-8) or a1-3 dextran (clone 1-21) and their immune + cluding CD3 live BW5147 cells. complexes were initially used as inhibitors in the apoptosis assays. Calcium mobilization As expected, addition of IgM anti-Fas mAb at 10 ng/ml induced + + + + robust apoptosis of control GFP cells, but not GFP /FcmR cells A mixture containing equal numbers of FcmR and WT Jurkat cells pre- + 2 2+ as determined by visualizing early (annexin V /7-AAD ) and late loaded with 5 mM Indo-1/AM (Molecular Probes), a Ca -sensitive dye, + + was stimulated by various concentrations of IgM mAbs specific for CD2 or (Annexin V /7-AAD ) apoptotic cells (Fig. 1A, upper second TCR or by 1 mM ionomycin as a positive control. The intracellular Ca2+ column). The addition of both types of IgM mAbs at 104 ng/ml 2+ concentrations ([Ca ]i) in both cell populations were simultaneously (upper third column; not shown for anti-dextran mAb) or their 2+ assessed by the ratio of emitted fluorescence of bound Ca at 405 nm and soluble immune complexes (upper fourth and lower first columns) free Ca2+ at 485 nm using an LSR II flow cytometer (BD Biosciences). In some experiments, an equal mixture of control GFP+ and WT Jurkat cells did not affect the FcmR-mediated protection, suggesting the in- was also assessed similarly as controls. For blood B cells, MACS-enriched ability of such IgM molecules to block the interaction of the Fc B cells (.95%) were loaded with Fluo-4/AM (0.25 mM; Molecular portion of IgM Fas mAb with FcmR. The failure to inhibit pro- Probes), rested, and then treated with a mixture of mouse IgM anti-human tection was observed with wide ranges of Ab/Ag molar ratios k mAb (END-5C1, 10 mg/ml) and mouse IgG2b anti-FcmR mAb (HM7 or (Fig. 1B, Supplemental Fig. 1). However, addition of both types of HM3, 100 mg/ml). Changes in Fluo-4 fluorescence in the labeled cells 5 following treatment were measured over real time with an Accuri C6 flow IgM at 10 ng/ml (Fig. 1B, Supplemental Fig. 1) and their immune cytometer and analyzed by FlowJo software as described (22). complexes with (Fig. 1A, lower second column, 1B, Supplemental The Journal of Immunology 3

FIGURE 1. Role of FcmR in IgM anti-Fas mAb–induced apoptosis. (A) Representative flow cytometric profiles. Jurkat cells stably expressing GFP only (GFP) as a control or both GFP and human FcmR (GFP/FcmR) were cultured with (+) or without (2) agonistic IgM anti-human Fas mAb (clone CH11, 10 ng/ml) in the presence or absence of the inhibitors with the indicated concentrations (ng/ml) in parentheses: IgM anti-NP mAb (B1-8) or its soluble complexes with NP25-BSA, IgM anti-dextran mAb (1-21, not shown) or its soluble complexes with dextran, IgM anti-CD2 mAb (C373), and IgG2b blocking (bl; HM7) or nonblocking (nb; HM3) anti-FcmR mAb. The Ab/Ag molar ratios were 1:1021 and 1:1022 in the upper fourth and fifth columns and 1:1 and 1:1021 in the lower first and second columns. After culture, cells were stained with allophycocyanin–annexin Vand 7-AAD before identification of early apoptotic (annexin V+/7-AAD2), late apoptotic (annexin V+/7-AAD+), and dead (annexin V2/7-AAD+) cells by flow cytometric analysis. Numbers in each quadrant indicate percentages of cells. (B) Effects of IgM and its immune complexes on FcmR-mediated antiapoptotic activity. GFP+ (open columns) or GFP+/FcmR+ (filled columns) cells were cultured with IgM anti-Fas mAb (10 ng/ml) in the presence of the indicated concentrations (ng/ml) of IgM anti-NP mAb alone or with

NP25-BSA as indicated for Ab/Ag molar ratios. Apoptotic (early and late) cells were similarly assessed and percentage relative apoptosis was determined as follows: 100 3 [(% anti-Fas–induced apoptotic cells with or without inhibitors in GFP+ or FcmR+/GFP+ cell line) 2 (% spontaneous apoptotic cells in the corresponding cell line)] 4 [(% anti-Fas mAb–induced apoptotic cells without inhibitors in GFP+ cell line) 2 (% spontaneous apoptotic cells in GFP+ cell line)]. Results are shown as means 6 1 SD from three to five independent experiments. The dotted line indicates percentage apoptosis of FcmR+/GFP+ cells in the presence of IgM anti-Fas mAb without inhibitors. (C) Efficient inhibition of IgM anti-CD2 mAb in FcmR-mediated antiapoptotic activity. GFP+ (open columns) or GFP+/FcmR+ (filled columns) cells were similarly cultured with the indicated doses (ng/ml) of IgM anti-CD2 mAb along with IgM anti-Fas mAb (10 ng/ml) before assessment of apoptotic cells. Results are shown as means 6 1 SD from three independent experiments. The dotted line indicates per- centage apoptosis of FcmR+/GFP+ cells in the presence of IgM anti-Fas mAb without IgM anti-CD2 mAb. Results using a Student t test comparison of anti- Fas mAb–induced apoptosis in the absence and presence of inhibitors are indicated as *p , 0.05, **p , 0.01, and ***p , 0.001. 4 THE CIS ENGAGEMENT OF FcmR

Fig. 1) partially but significantly blocked the FcmR-mediated plus a 10-fold excess of mouse BW5147 cells stably expressing protection. All three IgM preparations (CH11, B1-8, and 1-21) human FcmR. When the Fc portion of IgM anti-Fas mAb binds were found to bind equally to FcmR+ cells (not shown), ruling out FcmRintrans between the excess of cocultured FcmR+ BW5147 the possibility that the CH11 IgM has a unique posttranslational cells, then FcmR+ Jurkat cells will become apoptotic. Alterna- modification to bind more efficiently the FcmR compared with tively, when the Fc portion of IgM anti-Fas mAb must bind FcmR others. In contrast, addition of FcmR-blocking IgG2b mAb (HM7), in cis on the same cell surface to perform this protective function, but not FcmR nonblocking IgG2b mAb (HM3), at 104 ng/ml FcmR+ Jurkat cells will still be protected against IgM anti-Fas specifically inhibited the interaction of the Fc portion of IgM mAb–mediated apoptosis even in the presence of a 10-fold excess Fas mAb with FcmR, thus permitting the FcmR+ cells to become number of FcmR+ BW5147 cells. As shown in Fig. 2B, addition of apoptotic (Fig. 1A, lower fourth and fifth columns). These findings excessive FcmR+ BW5147 cells did not affect the FcmR-mediated indicate that soluble IgM immune complexes are not good com- protection from apoptosis (second versus fourth column), sug- petitors in the interaction of FcmR with the Fc portion of IgM anti- gesting predominance of the protective mechanism involving a cis Fas mAb, although FcmR is shown by Scatchard plot analysis to interaction of the Fc portion of IgM anti-Fas mAb with FcmRon bind pentameric IgM with a surprisingly high avidity of ∼10 nM the same cell surface. Addition of excessive FcmR+ BW5147 cells (14). This also suggests the possibility that FcmR may bind more did not spontaneously induce apoptosis of either of FcmR+ or efficiently to the Fc portion of IgM, which is attached to plasma control Jurkat cells in the absence of anti-Fas mAb (third column). membranes via its Fab region, than to the Fc portion of free IgM Collectively, these findings show that although FcmR binds or its immune complexes in solution. soluble IgM pentamers at a high avidity (14), FcmRbindsmore To test this hypothesis, we employed additional IgM mAbs efficiently to the Fc portion of IgM Ab when it reacts with a reactive with CD2 or TCR on the surface of Jurkat cells as potential membrane component on the same cell surface. This finding competitors for the interaction of IgM Fas mAb with FcmR. Ad- implies that FcmR can modulate the functional activity of dition of IgM anti-CD2 mAb (C373) at 103 ng/ml efficiently cell surface receptors or molecules recognized by inhibited FcmR-mediated protection, thereby permitting the either natural or immune IgM Abs. FcmR+ cells to undergo apoptosis (see Fig. 1A, lower third col- Calcium mobilization by coligation of FcmR with CD2 or BCR umn, 1C). The anti-CD2 mAb alone did not induce apoptosis of by IgM mAbs either of cell types (not shown). Similar results were also obtained 2+ with an IgM anti-TCR mAb (C305, not shown). Thus, these To explore this possibility, we compared Ca mobilization in cells findings suggest that FcmR binds more efficiently to IgM mAbs, with a single ligation of CD2 or TCR versus the coligation of both + which attach to cell surface components via their Fab regions, than FcmR and CD2 or TCR. An equal mixture of WT and FcmR 2+ to IgM immune complexes in solution. Jurkat cells was preloaded with the Ca dye Indo-1/AM and si- multaneously stimulated by various concentrations of IgM mAbs m Cis engagement of Fc R with IgM anti-Fas mAb on the same specific for CD2 or TCR. Although it has been well documented cell surface that induction of CD2-mediated Ca2+ mobilization in T cells To determine whether the engagement of FcmR with IgM attached requires ligation of CD2 with two different IgG mAbs (23–25), to plasma membrane occurs in trans or cis (see Fig. 2A), we the IgM anti-CD2 mAb (C373) alone clearly induced Ca2+ mo- + 2+ performed mixed cell cultures of FcmR (or control) Jurkat cells bilization (Fig. 3A). Notably, the rise in [Ca ]i occurred signifi-

FIGURE 2. Cis engagement of FcmR. (A) Theoreti- cal FcmR-mediated protection in IgM anti-Fas mAb– induced apoptosis. In theory, human (three red vertical lines) on Jurkat cells ( yellow-filled circles) is recognized by CH11 IgM anti-human Fas mAb (purple Y; x5 indicates pentamer) and its Fc portion binds FcmR (blue ovals) in cis on the same cells or in trans between neighboring cells, thereby inhibit- ing Fas-induced apoptosis. Addition of 10-fold excess of mouse thymoma line (BW5147) expressing human FcmR (white circles) to the apoptosis assay of Jurkat cells is predicted to inhibit the trans (two-headed solid- line arrow), but not cis (two-headed broken-line arrow), interaction of FcmR with the CH11 IgM anti-Fas mAb. (B) Mixed cell cultures. FcmR+ (lower) or control (upper) Jurkat cells were cocultured with 10-fold ex- cess number of human FcmR+ BW5147 cells (of mouse origin) in the presence or absence of CH11 anti-Fas mAb. Cells were stained with PE-labeled anti-mouse CD3 mAb prior to assessment of apoptosis in Jurkat cell populations by excluding CD3+ live BW5147 cells. The surface levels of FcmR on BW5147 cells were higher than those on Jurkat cells. Result representative of three independent experiments. The Journal of Immunology 5 cantly faster in FcmR+ cells than in WT cells when ligated with To determine which part of the FcmR molecule is responsible for the anti-CD2 mAb at three different concentrations (3, 10, and these potential functions, we made FcmR cDNA constructs with 30 mg/ml, data at 3 and 30 mg/ml are not shown). A similar differ- point mutations (H253F, Y315F, Y366F, or Y385F) or a deletion ence, but to a much less extent, was also observed with the cross- of most of the cytoplasmic tail (A281–A390; DCy) and expressed linkage of TCR with IgM mAb (C305) (not shown). In contrast, them in Jurkat T cells. After enriching GFP+ cells by FACS, the 2+ the ionomycin-induced [Ca ]i rise occurred at the same time in surface expression of FcmR as determined by reactivity with both cell types. These findings suggest that the coligation of FcmR receptor-specific mAbs was comparable among the nonmutant and 2+ with CD2 or TCR induces a more rapid increase in [Ca ]i than mutant FcmR transductants (Fig. 4B), suggesting that any intro- does the ligation of CD2 or TCR alone. duced mutation does not affect the surface expression of FcmRon Because we found that freshly prepared blood T cells, unlike Jurkat cells. Remarkably, however, the IgM ligand-binding ac- Jurkat cells, were not activated by the C373 anti-CD2 mAb (not tivity was significantly increased by the DCy mutant (Fig. 4C, shown), we switched to using a mitogenic IgM anti-k mAb (END- 4D), suggesting the influence of the FcmR cytoplasmic tail on its 5C1) and examined Ca2+ mobilization induced by BCR cross- ectodomain-mediated ligand binding. linkage in the presence of FcmR-blocking (HM7) or nonblocking Next, the fate of IgM ligands after binding to FcmR on these (HM3) mAbs. As shown in Fig. 2B, IgM anti-k mAb–induced Ca2+ transductants was examined by epifluorescence microscopy. IgM mobilization was indistinguishable in the absence or presence of binding at 4˚C in the presence of NaN3, an inhibitor of energy- FcmR nonblocking mAb (HM3). In contrast, the presence of dependent contraction of actin filament, was observed in a weak FcmR-blocking mAb (HM7) diminished the IgM anti-k mAb–in- patchy staining pattern at the peripheral rim of transductants re- duced Ca2+ mobilization, suggesting that FcmR provides a stimu- gardless of their mutations. However, when incubated at 4˚C latory signal upon BCR cross-linkage with IgM mAbs. without NaN3, distinct IgM binding, marked by a more intense and localized or capping pattern, was observed with the H253F Key residues in the FcmR transmembrane and cytoplasmic tail mutant (Fig. 4Eb) as compared with all others except for the DCy responsible for receptor functions mutant, which exhibited a more broadly localized staining pattern In FcRs and in other paired Ig-like receptors, it is generally ob- (Fig. 4Ea for nonmutated). This suggests an important role of the served that when the ligand-binding a-chain has a short cyto- His253 residue in the anchoring of FcmR in the transmembrane. plasmic tail with no signal-transmitting potential, then it contains After washing and incubating at 37˚C for 10 min without NaN3, a charged residue in the transmembrane segment that facilitates the IgM/FcmR complexes were internalized and localized in in- noncovalent association with another transmembrane protein– tracellular locations, presumably endosomal or lysosomal com- containing ITAMs. This association allows for the transmission of partments, in the FcmR nonmutated (Fig. 4Ec), H253F (Fig. 4Ed), activating signals to cells as seen in FcgRI, FcgRIII, FcaR, and and Y315F transductants (not shown), but not in other mutants FcεRI. Alternatively, when the ligand-binding a-chain has a con- (Y366F, Y385F, and DCy, not shown). These results are consistent ventional hydrophobic transmembrane segment, then it usually with previous findings reported by others that two C-terminal– has a long ITAM-containing cytoplasmic tail (26). This recruits conserved Tyr residues were involved in receptor-mediated en- the phosphatase SHIP upon phosphorylation to attenuate the sig- docytosis (15). nals as in FcgRIIB. However, FcmR is unusual in having both The FcmR-mediated protection from apoptosis was significantly features: a charged His residue (H253) in the transmembrane re- diminished in the FcmR Y315F and DCy mutants, as the fre- gion and a long cytoplasmic tail containing three conserved Tyr quencies of apoptotic cells in these mutants were indistinguishable (Y315, Y366, Y385) and five conserved Ser residues, features that from those in the control transductant (Fig. 4F). Other FcmR have been conserved in multiple species (see Fig. 4A) (14). These mutants (H253F, Y366F, and Y385F) still had significant protec- characteristics suggest a dual signaling ability of FcmR: one from tive activity compared with controls (GFP+), but the extent of their a potential adaptor protein noncovalently associating with the protective activity was significantly lower than that of the non- FcmR via the H253 residue, and the other from its own Tyr and/or mutated FcmR. These findings suggest that the membrane- Ser residues in the cytoplasmic tail. proximal Tyr residue (Y315) is largely responsible for, but other Tyr residues also contribute minimally to, the FcmR-mediated protection. Notably, the rapid rise in IgM anti-CD2 mAb– 2+ induced [Ca ]i observed in the FcmR nonmutant was abolished by all of the FcmR mutants, suggesting that both the cytoplasmic tail including all three Tyr residues and the transmembrane His residue are involved in signaling mediated by coligation of CD2 and FcmR.

Discussion FIGURE 3. Ca2+ mobilization by IgM mAbs against CD2 or Ig k.(A) The present study is focused on the IgM ligand-binding property of An equal mixture of WT (GFP2, blue lines) and FcmR+/GFP+ (red lines) FcmR, and several remarkable features of the FcmR have been Jurkat cells preloaded with Ca2+ dye Indo-1/AM were stimulated with IgM identified. First, although FcmR binds soluble pentameric IgM at anti-CD2 (C373) mAb at 10 mg/ml or by 1 mM ionomycin at the time point a relatively high avidity of ∼10 nM (14), FcmR actually interacts 2+ indicated by arrows. The [Ca ]i levels were assessed by the 405/485 nm more efficiently with the Fc portion of IgM Abs when bound to fluorescence ratio in each viable cell population using an LSR II flow plasma membranes via their Ag-binding Fab regions, and this cytometer. This is a representative result from four independent experi- interaction occurs in cis on the same cell surface. This conclusion ments. (B) Fluo-4–loaded blood B cells were treated with IgM anti-human k mAb (END-5C1, 10 mg/ml) in the absence (red line) or presence of is based on the functional assays rather than actual ligand-binding 2+ IgG2b, FcmR-blocking HM7 (blue line), or nonblocking HM3 (green line) assessments. Second, the Ca mobilization induced by cross- 2+ mAb (100 mg/ml) at the time point indicated by an arrow. The [Ca ]i linkage of CD2 or BCRs with IgM mAbs differs between the li- levels were assessed by fluorescence intensity during a 5-min period. gation of CD2 or BCR alone and the coligation of FcmR and CD2 Representative result from three independent experiments. or BCR. Third, several remarkable functional changes are observed 6 THE CIS ENGAGEMENT OF FcmR

FIGURE 4. Effects of FcmR mutations on receptor function. (A) Predicted protein structure of FcmR. The human FcmR cDNA encodes type I trans- membrane protein that consists of a single V-set Ig-like domain (oval shape), an additional extracellular region with no known structure, a transmembrane segment (between two thick lines) containing a charged His residue (d), and a relatively long cytoplasmic tail containing three conserved Tyr residues (s). Point mutations are indicated and the extent of the deletion of the cytoplasmic tail is shown by the black bracket. Hatch marks indicate exon boundariesin the FCMR gene. (B–E) Cell surface FcmR levels and IgM ligand binding. Jurkat cells stably expressing only GFP (Cont.) or both GFP and FcmRof nonmutated (NM) or indicated mutations (H253F, Y315F, Y366F, Y385F or DCy) were assessed for cell surface levels of FcmR determined by HM14 mAb (B) and for IgM ligand-binding activity (C). The mean fluorescence intensity (MFI) of FcmR levels was determined as MFI of anti-FcmR mAb minus MFI of control mA (B) and the MFI of IgM binding as MFI of IgM minus MFI of PBS (C). Results are shown as the MFI 6 1 SD from three to five independent experiments. *p , 0.05 when compared with nonmutated (NM) FcmR. (D) Representative flow cytometric profiles. FcmR NM (blue) and DCy (red) cells were stained with biotin-HM14 anti-FcmR (open) or control (shaded) mAb for cell surface expression of FcmR(left) and with biotin-IgM myeloma (open) or PBS (shaded) for ligand binding (right). Because profiles with isotype-matched mAb or PBS were the same between FcmR NM and DCy cells, only one shaded profile is shown. (E) Representative epifluorescence microscopic images. The FcmRNM(Ea and Ec) and H253F (Eb and Ed) cells were incubated with Alexa Fluor 555–IgM (without NaN3) on ice (Ea and Eb), washed, and then incubated for additional 10 min at 37˚C (Ec and Ed) before cytocen- trifugation. Fluorescence images were combined with phase-contrast cell images. Scale bars, 10 mm. (F)FcmR-mediated protection from apoptosis. FcmR NM or mutant and control Jurkat cells were incubated with CH11 anti-Fas mAb as described in the Fig. 1 legend, and the frequencies of early and late apoptotic cells are plotted as means 6 1 SD from three independent experiments. *p , 0.05, **p , 0.01, *** p , 0.001 when compared with FcmR-/GFP+ control cells. with FcmR mutants: 1) increased IgM ligand binding by the FcmR mune complexes, 103-fold excess of FcmR-blocking mAb (HM7, DCy mutant, 2) enhanced cap formation of the FcmR H253F g2bk isotype), or 10-fold excess of IgM anti-CD2 or anti-TCR mutant even at 4˚C upon IgM ligand binding, and 3) less protective mAb binding to the Jurkat cells. The interaction of FcmR with the activity of the FcmR Y315F mutant in IgM anti-Fas mAb–mediated Fc of IgM anti-Fas mAb occurred in cis on the same cell surface apoptosis. and not in trans between neighboring cells, as evidenced by the The preferential interaction of FcmR with IgM attached to fact that adding 10-fold excess of FcmR+ BW5147 cells (of mouse plasma membrane via its Fab region was initially notified in our origin) did not affect FcmR-mediated protection from IgM anti- previous reconcilement of antiapoptotic activity of Toso, the Fas mAb–induced apoptosis. The preferential cis engagement of original designation given to FcmR (14, 17). In the present study, FcmR is thus distinct from the trans engagement of FcgRIIb, an we have defined more quantitatively the ligand-binding property inhibitory FcgR, in death receptor–mediated apoptosis. The in- of FcmR. The interaction of FcmR with the Fc portion of IgM anti- teraction of agonistic IgG mAbs against death receptors, including Fas mAb in Jurkat cells was not inhibited by addition of 103-fold Fas/CD95, with FcgRIIb is essential for the death receptor–me- excess of IgM mAbs or their soluble immune complexes, but it diated apoptosis and occurs in trans, but not in cis (27–29). The was blocked by addition of 104-fold excess of IgM mAbs or im- cis engagement of FcmR is significant in that FcmR has the po- The Journal of Immunology 7 tential to modulate the functional activity of lymphocyte surface of interest, given the unique sequence around the Y315 residue: receptors on the same cell when they are recognized by natural or 308P·R·S/T·Q·N·N·I/V·Y·S/T·A·C·P·R·R·A·R323 (bold type indi- immune IgM Abs. The physiological relevance of this cis engage- cates conserved amino acids). This does not match any known ment may be related to unique features of the IgM ligand and its Tyr-based signaling motifs. The mechanism of Toso/FcmR- receptor. In idiotype network theory, BCRs (or their secreted mediated protection from apoptosis was suggested to result products) are proposed to be directed toward the variable regions from potentiation of the cellular FLIP, a master antiapoptotic of other BCRs or their Ig molecules. Two thirds of the newly regulator (17), or alternatively by prevention of the internaliza- generated B cells in bone marrow are now known to react with tion of Fas, an important step for apoptosis signaling (44, 45), self-antigens such as dsDNA, ssDNA, insulin, and LPS (30). owing to simultaneous cross-linkage of both Fas and FcmRwith According to our estimates, approximately a fourth of the IgM- IgM Fas mAb (16). With regard to the latter, it is noteworthy that containing culture supernatants from 96 different EBV-transformed the coligation of FcmR with CD2 or TCR induced a more rapid neonatal B cell lines react with lymphocyte surface components increase in intracellular Ca2+ levels than in the ligation of CD2 (H. Kubagawa, K. Honjo, and Y. Kubagawa, unpublished observa- or TCR alone. tions). IgM anti–lymphocyte Abs are often present in individuals In conclusion, FcmR, the newest member of the FcR family, with autoimmune diseases or chronic viral infections, such as HIV-1, interacts more efficiently with the Fc portion of the IgM Abs, and recognize many different surface Ags (e.g., CD45, CD175/Tn, which are bound to plasma membranes via their Ag-binding Fab CD3ε, CD4, chemokine receptors, sphingosine-1-phosphate re- region, as compared with the Fc portion of IgM Abs or their ceptor 1), and some of those Abs regulate T cell–mediated in- immune complexes in solution. The former interaction occurs in flammatory responses in vitro (31–41). Unlike FcRs for switched cis on the same cell surface and not in trans between neighboring Ig isotypes, the expression of FcmR is restricted to lymphocytes, cells. This strongly implies that FcmR expressed on B, T, and NK namely, B, T, and NK cells (14). Therefore, it is highly probable cells may have a potential to modulate the function of target Ags that FcmR on B, T, and NK cells preferentially binds the Fc or receptors, which are recognized by natural or immune IgM Abs, portion of IgM Ab reactive with their cell surface Ags or recep- on the same cell surface. tors, thereby modulating the function of their target Ags or receptors. Supporting this idea, the CD2-mediated Ca2+ mobi- Acknowledgments lization after cross-linkage with IgM mAb was clearly faster in We thank Dr. Arthur Weiss for the gift of IgM mAbs specific for CD2 (C373) + 2+ FcmR than control Jurkat cells. The Ca mobilization induced or Jurkat TCR (C305) and for valuable suggestions, Dr. Peter Burrows for by cross-linkage of BCR on blood B cells with IgM anti-k mAb critical reading of the manuscript, Enid Keyser for FACS sorting, and was significantly different in the presence of an FcmR blocker Marion Spell for Ca2+ mobilization. (HM7) versus an FcmR nonblocker (HM3). The HM7 blocker diminished BCR-mediated Ca2+ mobilization, whereas the HM3 Disclosures nonblocker did not. Unlike the coligation of FcmR and Fas re- The authors have no financial conflicts of interest. ceptor, FcmR provides a stimulatory signal upon coligation with CD2 or BCR. Mutational analyses of FcmR revealed several remarkable References alterations in its function. Whereas nonmutated and mutant FcmR 1. Ravetch, J. V., and J. P. Kinet. 1991. Fc receptors. Annu. Rev. Immunol. 9: 457– 492. transductants expressed comparable levels of cell surface FcmRas 2. Dae¨ron, M. 1997. Fc receptor biology. Annu. Rev. Immunol. 15: 203–234. judged by receptor-specific mAbs, the IgM ligand-binding activity 3. Ravetch, J. V., and S. Bolland. 2001. IgG Fc receptors. Annu. Rev. Immunol. 19: 275–290. was significantly increased in the DCy mutant. This suggests ei- 4. Monteiro, R. C., and J. G. Van De Winkel. 2003. IgA Fc receptors. Annu. Rev. ther the formation of oligomeric FcmR due to its presumably Immunol. 21: 177–204. mobile nature within the plasma membrane or the inside–out 5. Nimmerjahn, F., and J. V. Ravetch. 2008. Fcg receptors as regulators of immune responses. Nat. Rev. Immunol. 8: 34–47. regulation of ligand-binding activity of FcmR by its cytoplasmic 6. Xiang, Z., A. J. Cutler, R. J. Brownlie, K. Fairfax, K. E. Lawlor, E. Severinson, tail. Given the precedent of inside–out signaling regulation in the E. U. Walker, R. A. Manz, D. M. Tarlinton, and K. G. C. Smith. 2007. FcgRIIb integrin family of cell surface adhesion molecules (42) and the controls bone marrow persistence and apoptosis. Nat. Immunol. 8: 419–429. FcaR (43), both possibilities are equally conceivable and remain 7. Bergtold, A., D. D. Desai, A. Gavhane, and R. Clynes. 2005. Cell surface to be elucidated. recycling of internalized antigen permits dendritic cell priming of B cells. Im- munity 23: 503–514. When the fate of IgM-bound FcmR was examined by immu- 8. Conrad, D. H. 1990. Fc ε RII/CD23: the low affinity receptor for IgE. Annu. Rev. nofluorescence microscopy, enhanced cap formation was clearly Immunol. 8: 623–645. observed with the H253F mutant even at 4˚C without NaN as 9. Roopenian, D. C., and S. Akilesh. 2007. FcRn: the comes of 3 age. Nat. Rev. Immunol. 7: 715–725. compared with FcmR nonmutated or other mutant transductants, 10. Shibuya, A., N. Sakamoto, Y. Shimizu, K. Shibuya, M. Osawa, T. Hiroyama, except the DCy mutant, which displayed a more broadly localized H. J. Eyre, G. R. Sutherland, Y. Endo, T. Fujita, et al. 2000. Fc a/m receptor staining pattern. This finding suggests a functional role of His253 mediates endocytosis of IgM-coated microbes. Nat. Immunol. 1: 441–446. 11. Sanders, S. K., H. Kubagawa, T. Suzuki, J. L. Butler, and M. D. Cooper. 1987. residue in the anchoring of FcmR in the transmembrane. Although IgM binding protein expressed by activated B cells. J. Immunol. 139: 188–193. we have not yet identified a membrane protein noncovalently 12. Ohno, T., H. Kubagawa, S. K. Sanders, and M. D. Cooper. 1990. Biochemical nature of an Fc m receptor on human B-lineage cells. J. Exp. Med. 172: 1165– associating with the ligand-binding chain of FcmR through its 1175. H253, there are several possible candidates, including the ITAM- 13. Nakamura, T., H. Kubagawa, T. Ohno, and M. D. Cooper. 1993. Characterization containing CD3z,FcεRIg, or DAP12, the PI3K-recruiting DAP10, of an IgM Fc-binding receptor on human T cells. J. Immunol. 151: 6933–6941. 14. Kubagawa, H., S. Oka, Y. Kubagawa, I. Torii, E. Takayama, D. W. Kang, or the unknown 40-kDa protein (p40), which was often copreci- G. L. Gartland, L. F. Bertoli, H. Mori, H. Takatsu, et al. 2009. Identity of the pitated with the 60-kDa FcmR in our previous analyses (14). elusive IgM Fc receptor (FcmR) in humans. J. Exp. Med. 206: 2779–2793. Notably, unlike other multichain FcRs, the FcmR H253F mutant 15. Vire, B., A. David, and A. Wiestner. 2011. TOSO, the Fcm receptor, is highly expressed on chronic lymphocytic leukemia B cells, internalizes upon IgM presented on the cell surface without a potentially associated binding, shuttles to the lysosome, and is downregulated in response to TLR membrane protein. activation. J. Immunol. 187: 4040–4050. 16. Murakami, Y., S. Narayanan, S. Su, R. Childs, K. Krzewski, F. Borrego, J. Weck, The findings that the Y315F and DCy mutants had no significant and J. E. Coligan. 2012. Toso, a functional IgM receptor, is regulated by IL-2 in protective activity against IgM anti-Fas mAb–induced apoptosis is T and NK cells. J. Immunol. 189: 587–597. 8 THE CIS ENGAGEMENT OF FcmR

17. Hitoshi, Y., J. Lorens, S. I. Kitada, J. Fisher, M. LaBarge, H. Z. Ring, U. Francke, 31. Warren, R. Q., E. A. Johnson, R. P. Donnelly, M. F. Lavia, and K. Y. Tsang. J. C. Reed, S. Kinoshita, and G. P. Nolan. 1998. Toso, a cell surface, specific 1988. Specificity of anti-lymphocyte in sera from patients with AIDS- regulator of Fas-induced apoptosis in T cells. Immunity 8: 461–471. related complex (ARC) and healthy homosexuals. Clin. Exp. Immunol. 73: 168– 18. Honjo, K., Y. Kubagawa, and H. Kubagawa. 2012. Is Toso an antiapoptotic 173. protein or an Fc receptor for IgM? Blood 119: 1789–1790. 32. Daniel, V., K. Schimpf, and G. Opelz. 1989. Lymphocyte autoantibodies and 19. Reth, M., G. J. Ha¨mmerling, and K. Rajewsky. 1978. Analysis of the repertoire alloantibodies in HIV-positive haemophilia patients. Clin. Exp. Immunol. 75: of anti-NP antibodies in C57BL/6 mice by cell fusion. I. Characterization of 178–183. families in the primary and hyperimmune response. Eur. J. Immunol. 8: 33. Silvestris, F., B. S. Edwards, O. M. Sadeghi, M. A. Frassanito, R. C. Williams, 393–400. Jr., and F. Dammacco. 1989. Isotype, distribution and target analysis 20. Foote, J. B., T. I. Mahmoud, A. M. Vale, and J. F. Kearney. 2012. Long-term of lymphocyte reactive antibodies in patients with human immunodeficiency maintenance of polysaccharide-specific antibodies by IgM-secreting cells. J. virus infection. Clin. Immunol. Immunopathol. 53: 329–340. Immunol. 188: 57–67. 34. Koren, E., M. W. Reichlin, M. Koscec, R. D. Fugate, and M. Reichlin. 1992. 21. Weiss, A., and J. D. Stobo. 1984. Requirement for the coexpression of T3 and the Autoantibodies to the ribosomal P proteins react with a plasma membrane- T cell antigen receptor on a malignant human T cell line. J. Exp. Med. 160: related target on human cells. J. Clin. Invest. 89: 1236–1241. 1284–1299. 35. Muller,€ C., S. Kukel, K. E. Schneweis, and R. Bauer. 1994. Anti-lymphocyte 22. Honjo, K., Y. Kubagawa, D. M. Jones, B. Dizon, Z. Zhu, H. Ohno, S. Izui, antibodies in plasma of HIV-1-infected patients preferentially react with MHC J. F. Kearney, and H. Kubagawa. 2012. Altered Ig levels and antibody responses class II-negative T cells and are linked to antibodies against gp41. Clin. Exp. in mice deficient for the Fc receptor for IgM (FcmR). Proc. Natl. Acad. Sci. USA Immunol. 97: 367–372. 109: 15882–15887. 36. Winfield, J. B., P. D. Fernsten, and J. K. Czyzyk. 1997. Anti-lymphocyte auto- 23. Meuer, S. C., R. E. Hussey, M. Fabbi, D. Fox, O. Acuto, K. A. Fitzgerald, antibodies in systemic lupus erythematosus. Trans. Am. Clin. Climatol. Assoc. J. C. Hodgdon, J. P. Protentis, S. F. Schlossman, and E. L. Reinherz. 1984. An 108: 127–135. alternative pathway of T-cell activation: a functional role for the 50 kd T11 sheep 37. Cappione, A. J., A. E. Pugh-Bernard, J. H. Anolik, and I. Sanz. 2004. Lupus IgG erythrocyte receptor protein. Cell 36: 897–906. VH4.34 antibodies bind to a 220-kDa glycoform of CD45/B220 on the surface of 24. Olive, D., M. Ragueneau, C. Cerdan, P. Dubreuil, M. Lopez, and C. Mawas. human B lymphocytes. J. Immunol. 172: 4298–4307. 1986. Anti-CD2 (sheep red blood cell receptor) monoclonal antibodies and 38. Lobo, P. I., K. H. Schlegel, C. E. Spencer, M. D. Okusa, C. Chisholm, T cell activation. I. Pairs of anti-T11.1 and T11.2 (CD2 subgroups) are strongly N. McHedlishvili, A. Park, C. Christ, and C. Burtner. 2008. Naturally occurring mitogenic for T cells in presence of 12-O-tetradecanoylphorbol 13-acetate. Eur. IgM anti-leukocyte autoantibodies (IgM-ALA) inhibit T cell activation and J. Immunol. 16: 1063–1068. chemotaxis. J. Immunol. 180: 1780–1791. 25. June, C. H., J. A. Ledbetter, P. S. Rabinovitch, P. J. Martin, P. G. Beatty, and 39. Liao, J. J., M. C. Huang, K. Fast, K. Gundling, M. Yadav, J. R. Van Brocklyn, J. A. Hansen. 1986. Distinct patterns of transmembrane calcium flux and in- M. R. Wabl, and E. J. Goetzl. 2009. Immunosuppressive human anti-lymphocyte tracellular calcium mobilization after differentiation antigen cluster 2 (E rosette autoantibodies specific for the type 1 sphingosine 1-phosphate receptor. FASEB receptor) or 3 (T3) stimulation of human lymphocytes. J. Clin. Invest. 77: 1224– J. 23: 1786–1796. 1232. 40. Lobo, P. I., K. H. Schlegal, J. Vengal, M. D. Okusa, and H. Pei. 2010. Naturally 26. Ve´ly, F., and E. Vivier. 1997. Conservation of structural features reveals the occurring IgM anti-leukocyte autoantibodies inhibit T-cell activation and che- existence of a large family of inhibitory cell surface receptors and noninhibitory/ motaxis. J. Clin. Immunol. 30(Suppl. 1): S31–S36. activatory counterparts. J. Immunol. 159: 2075–2077. 41. Liedtke, M., C. J. Twist, B. C. Medeiros, J. R. Gotlib, C. Berube, M. M. Bieber, 27. Bando, M., Y. Miyake, M. Shiina, M. Wachi, K. Nagai, and T. Kataoka. 2002. N. M. Bhat, N. N. Teng, and S. E. Coutre. 2012. Phase I trial of a novel human Actin cytoskeleton is required for early apoptosis signaling induced by anti-Fas monoclonal antibody mAb216 in patients with relapsed or refractory B-cell antibody but not Fas ligand in murine B A20 cells. Biochem. Biophys. acute lymphoblastic leukemia. Haematologica 97: 30–37. Res. Commun. 290: 268–274. 42. Kinashi, T. 2005. Intracellular signalling controlling integrin activation 28. Xu, Y., A. J. Szalai, T. Zhou, K. R. Zinn, T. R. Chaudhuri, X. Li, W. J. Koopman, in lymphocytes. Nat. Rev. Immunol. 5: 546–559. and R. P. Kimberly. 2003. FcgRs modulate cytotoxicity of anti-Fas antibodies: 43. Bakema, J. E., A. Bakker, S. de Haij, H. Honing, M. Bracke, L. Koenderman, implications for agonistic antibody-based therapeutics. J. Immunol. 171: 562– G. Vidarsson, J. G. van de Winkel, and J. H. Leusen. 2008. Inside-out regulation 568. of FcaRI (CD89) depends on PP2A. J. Immunol. 181: 4080–4088. 29. Li, F., and J. V. Ravetch. 2011. Inhibitory Fcg receptor engagement drives ad- 44. Vignaux, F., E. Vivier, B. Malissen, V. Depraetere, S. Nagata, and P. Golstein. juvant and anti-tumor activities of agonistic CD40 antibodies. Science 333: 1995. TCR/CD3 coupling to Fas-based cytotoxicity. J. Exp. Med. 181: 781–786. 1030–1034. 45. Yamauchi, A., K. Taga, H. S. Mostowski, and E. T. Bloom. 1996. Target 30. Wardemann, H., S. Yurasov, A. Schaefer, J. W. Young, E. Meffre, and cell-induced apoptosis of interleukin-2-activated human natural killer cells: M. C. Nussenzweig. 2003. Predominant autoantibody production by early human roles of cell surface molecules and intracellular events. Blood 87: 5127– B cell precursors. Science 301: 1374–1377. 5135.