1 Introduction

It has been established from studies of mutant mice deficient in IgM secretion that both preimmune “natural” IgM and antigen-induced “immune” IgM are important in responses to pathogens and self-antigens (Ehrenstein and Notley 2010). Effector proteins interacting with the Fc portion of IgM, such as complement and its receptors, have so far mainly been elucidated in the context of IgM-mediated immune protection and regulation (Heyman 2000) (see also Chap. 4 by Anna Sörman and Birgitta Heyman). The role of IgM Fc receptor (FcµR) in such effector functions has just begun to be explored, since the FCMR was identified in 2009 (Kubagawa et al. 2009). In this chapter, recent findings about the FcµR in both humans and mice are reviewed, along with emphasis on its significance and the discrepancies among different reports especially in murine studies. Several review articles on FcµR and its ligand IgM have already been published elsewhere, and the authors recommend them for further information on both of these topics (Baumgarth 2016; Ehrenstein and Notley 2010; Klimovich 2011; Kubagawa et al. 2014a, c; Panda and Ding 2015; Wang et al. 2016).

2 Lymphocyte-Restricted Expression of FcµR

The cellular distribution of FcµR in mice is still not completely resolved. In our studies, FcµR expression at the level of transcription is only detectable in B-lineage cells and not in other cell types. This was investigated by reverse transcriptase polymerase chain reaction (RT-PCR) using RNAs from the following tissue/cell samples. (i) Splenocytes or liver tissues of mice deficient for recombination-activating gene 1 (Rag1), which are devoid of B and T cells but contain abundant granulocytes and macrophages and (ii) CD19-positive B or CD19-negative non-B cells and Gr1- or Ly6G-positive myeloid cells enriched twice to high purity by fluorescence-activated cell sorting (FACS) from wild-type (WT) mouse bone marrow (BM) and spleen (Honjo et al. 2012a, 2013). By contrast, Choi et al. (2013) using a similar RT-PCR analysis of FACS sorted cells found clear expression of FcµR by the splenic Gr1-positive cell population in addition to B cells, but curiously, total BM cells, which should contain more abundant Gr1-positive myeloid cells than B cells, did not express significant levels of FcµR. The B cell-restricted expression pattern of FcµR is also documented in two large-scale expression databases of immune cells: Immunological Genome Project (https://www.immgen.org) and Reference Database of Immune Cells (http://refdic.rcai.riken.jp/welcome.cgi). Intriguingly, however, another study has recently reported that by single-cell RNA sequencing along with complex algorithmic assessments and its functional annotation, FcµR is suggested as one of the critical regulators of Th17 pathogenicity in myelin oligodendrocyte glycoprotein (MOG)-induced autoimmune encephalomyelitis (EAE) (Gaublomme et al. 2015). In our studies using the T cell transfer colitis model in Rag1 −/− mice, however, none of the sorted T cell subpopulations with the phenotype of IL-17+, IFNγ+, or IL-17+/IFNγ+ express FcµR transcripts as determined by gene array analysis (Zimmermann et al. unpublished). In this regard, transcriptome analysis of even homogenously isolated cell populations (e.g., Th17 cells) is a mixed snapshot of asynchronously propagated, metabolically heterogeneous cell populations. On the other hand, single-cell RNA sequence analysis along with multiple algorithmic assessments may distinguish distinct states of cells within such populations. Even based on single-cell RNA sequencing data, however, it is difficult to distinguish whether such a small subpopulation of Th17 cells may indeed express full-length FcµR transcripts or may passively acquire them through exosomes or membrane vesicles from FcµR-bearing B cells via trogocytosis (Gyorgy et al. 2011). It will also be difficult to prove formally that such a minor population of Th17 cells expresses functional FcµR at low levels on their cell surface and plays a major regulatory role in the pathogenesis of MOG-induced EAE. The experimental data leading to this idea were based on the ex vivo data of Th17-polarizing cells from Fcmr-deficient (KO) and WT mice. This issue will be further discussed in the section of Fcmr KO mice (Sect. 4).

With respect to the cell surface expression of FcµR using receptor-specific monoclonal antibodies (mAbs) and IgM-ligand binding, we and others have found that the expression of FcµR is restricted to adaptive immune lymphocytes: both B and T cells and, to a lesser extent, natural killer (NK) cells in humans (Kubagawa et al. 2009; Murakami et al. 2012) and only B cells in mice (Honjo et al. 2012a, 2013; Lapke et al. 2015; Ouchida et al. 2012; Shima et al. 2010). [NK cells are the only known non-adaptive immune cell in humans to express FcµR, but this cell is now thought to have both adaptive and innate immune cell features (Vivier et al. 2011)]. This lymphocyte-restricted expression suggests a distinct function of FcµR compared to FcRs for switched Ig isotypes (e.g., FcγRs, FcεRI, and FcαR), which are also expressed by various innate immune cells. The species difference in cellular distribution of FcµR indicates that the results from murine FcµR studies may not necessarily reflect the human situation. Given the fact that IgM is the first Ig isotype appearing during phylogeny, ontogeny, and immune responses and that IgM is considered as a first line of defense against infection, the lymphocyte-restricted expression pattern is somewhat unexpected and may have some functional significance. In this regard, it is also noteworthy that unlike the phylogenetically broad distribution of IgM from jawed vertebrates (i.e., cartilaginous fish) onward, analysis of currently existing genomic sequence databases indicates that the IgM FcR first appears in early reptiles and is found in all three major living (extant) groups of mammals (i.e., egg laying, marsupial and placental mammals) (Akula et al. 2014) (also see Chap. 1 by Srinivas Akula and Lars Hellman).

Contrary to the B cell-restricted expression pattern in mice described in the above-mentioned studies, other groups using a rat mAb (B68 clone) against mouse Toso, an original name of FcµR (Hitoshi et al. 1998; Kubagawa et al. 2015), reported the weak but “functional relevant expression” of FcµR by myeloid cells, dendritic cells (DCs), and T cells (Brenner et al. 2014; Lang et al. 2013a, b; Nguyen et al. 2011). Strangely, in their analyses, Ly6G-positive BM myeloid cells were weakly positive for B68 mAb staining, whereas Ly6G-negative cells, which should contain abundant B-lineage cells, were completely negative, suggesting a non-optimal assessment with this mAb (Lang et al. 2013a, b). In fact, Lapke et al. (2015) have recently demonstrated that the expression of Toso/FcµR in mice is restricted to B cells using the same B68 and additional A96 mAbs, consistent with the results from our analysis using a panel of five different mAbs (MM1, MM2, MM3, MM4, and MM6 clones) (Honjo et al. 2012a, 2013). Notably, most studies dealing with the predicted function of FcµR in granulocytes, monocyte/macrophages, and DCs were based on comparative analysis in adoptive transfer experiments of WT versus Fcmr KO BM cells (Brenner et al. 2014; Lang et al. 2013a, b, 2015) and not on any actual convincing data of the cell surface FcµR expression by non-B cells, hence the use of the phrase “functional relevant expression.” Apart from this, there is another concern regarding the specificity or cross-reactivity of several commercially available polyclonal or monoclonal reagents raised against synthetic peptides of both human and mouse FcµRs, which might account for the reported expression of Toso/FcµR by non-hematopoietic cell types such as pancreatic ß cells (Dharmadhikari et al. 2012).

The lymphocyte-restricted expression pattern of FcµR is thus distinct from the expression of other IgM-binding receptors. Polymeric Ig receptor (pIgR) is predominantly expressed by mucosal epithelial cells (Kaetzel 2005), and FcR for IgA and IgM (Fcα/μR) is expressed by follicular dendritic cells and other cell types including Paneth cells in small intestinal crypts, the proximal tubular epithelial cells in kidneys, and the serous acini and small epithelial cells of salivary glands (Kikuno et al. 2007).

Several types of stimulations or conditions have been shown to modulate cell surface expression of FcµR. Upon antigen receptor ligation with antibodies or phorbol myristate acetate stimulation, FcµR expression in humans is up-regulated on B cells but is down-modulated on T cells, suggesting differential regulation of FcµR expression during B cell and T cell activation (Kubagawa et al. 2009; Nakamura et al. 1993; Sanders et al. 1987). Stimulation of T and NK cells with IL-2 in vitro also down-regulates FcµR expression in a STAT5-dependent manner (Murakami et al. 2012). In diseases, the enhanced expression of FcµR is a hallmark of chronic lymphocytic leukemia (CLL) B cells, as first demonstrated many years ago by rosette formation with IgM-coated erythrocytes (Ferrarini et al. 1977; Pichler and Knapp 1977), followed by IgM-ligand binding (Ohno et al. 1990; Sanders et al. 1987), gene expression (Catera et al. 2008; Pallasch et al. 2008; Proto-Siqueira et al. 2008; Rosenwald et al. 2001; Wang et al. 2004), and receptor-specific mAbs (Li et al. 2011; Vire et al. 2011). Intriguingly, surface FcµR levels are also significantly elevated in the non-CLL B cells and T cells in CLL patients (Li et al. 2011). Immunotherapies targeting for the FcµR have been designed for CLL cells. One is an immunotoxin-coupled IgM Fc (Cµ2-Cµ4) (Vire et al. 2014), and another is chimeric antigen receptor-modified T cells using a single-chain fragment (scFv)-containing the variable regions of an anti-FcµR mAb (6B10) (Faitschuk et al. 2016). In both cases, patient CLL B cells appear to be selectively eliminated in vitro without affecting the non-leukemic B and T cells. In patients with selective IgM immunodeficiency, cell surface FcµR levels on a particular blood B cell subset with a marginal zone (MZ) phenotype (IgM+/IgD+/CD27+) are significantly diminished as compared to age-matched controls, but the molecular basis for this reduction remains unclear (Gupta et al. 2016).

3 Unique Ligand-Binding Activity

3.1 Fcµ-Specificity, Ligand-Binding Avidity, and Glycosylation

After identifying the FcµR cDNA from human B-lineage cell-derived cDNA libraries by a functional cloning strategy (i.e., IgM-ligand binding), cell lines stably expressing FcµR have mainly been used to investigate their ligand-binding specificity. The FcµR-bearing cells clearly bind IgM in a dose-dependent manner, but not other Ig isotypes (i.e., IgG1-4, IgA1-2, IgD, or IgE) (Kubagawa et al. 2009). The inability of FcµR to bind polymeric IgA clearly indicates that FcµR is distinct from pIgR and Fcα/µR, both of which bind IgM and polymeric IgA and are clustered within the FCMR locus on chromosome 1q32.2. [Recent domain swapping analysis has revealed that unlike pIgR, Fcα/µR can bind J chain-deficient IgM hexamers (Yoo et al. 2011); hence, all three IgM-binding receptors are different in terms of ligand-binding specificity]. Binding of IgM by FcµR is mediated by its Fc5µ fragments, consisting mostly of Cµ3/Cµ4 domains, but not by Fabµ fragments, thereby confirming its IgM Fc-binding specificity (Kubagawa et al. 2009; Murakami et al. 2012). Recent domain swapping analysis reveals the Cµ4 as the target of FcµR (Lloyd et al. 2017). By Scatchard plot analysis assuming a 1:1 stoichiometry of FcµR to IgM ligand, FcµR binds IgM pentamers with a strikingly high avidity of ~10 nM. This in turn suggests that serum IgM, the concentration of which is ~1 µM, constitutively binds FcµR on the surface of B, T, and NK cells and explains why detection of cell surface FcµR is enhanced by preculture of cells, especially T cells, in IgM-free media (Kubagawa et al. 2009; Nakamura et al. 1993). In this regard, Vire et al. (2011) found that FcµR on CLL B cells was rapidly internalized upon IgM binding and shuttled to the lysosomes for degradation. The configuration of IgM is also important for FcµR binding, as higher concentrations (>100-fold) are required for binding of monomeric IgM to the FcµR+ cells than pentameric IgM. In addition to IgM pentamer, J chain-deficient IgM hexamers are also present in normal sera, albeit at unknown concentrations, but the complement activation activity of the IgM hexamers is 50- to 100-fold higher than IgM pentamers (Randall et al. 1992; Wiersma et al. 1998). Intriguingly, our preliminary findings show that the dissociation constant (KD) of a recombinant IgM hexamer for FcµR is only twofold to threefold higher than that of IgM pentamers (Fig. 1).

Fig. 1
figure 1

Binding of hexameric and pentameric IgM to FcµR. An equal mixture of murine thymoma line BW5147 stably expressing both FcµR and green fluorescent protein (GFP) (•), and WT (FcµR−/GFP−) control cell line (□) was incubated with various concentrations (25 pM–30 nM) of recombinant hexameric (left IgM-αTp) or pentameric (right rIgM) human IgM ligands, the preparation and purity of which were described elsewhere (Yoo et al. 2011) and were confirmed prior to use in the present studies. The bound IgM was assessed by addition of phycoerythrin-labeled goat antibodies specific for human µ heavy chain, followed by flow cytometry. The mean fluorescent intensity (MFI) of IgM binding at each concentration was plotted, and the KD was calculated by nonlinear regression analysis using GraphPad Prism software. Two independent experiments yielded similar KD values (nM), and one of them is shown

Full size image

FcµR does not have N-linked glycosylation motifs (NxS/T; x indicates any amino acid) in the extracellular region, consistent with our previous biochemical characterization of the IgM-binding protein (Ohno et al. 1990). Since the core peptide of FcµR is predicted to have a M r of ~41 kDa and the FcµR expressed on B and T cells has a M r of ~60 kDa, one-third of the M r of the mature FcµR is thus made up of O-linked glycans. Potential glycosylation sites were determined by mutagenesis experiments: Thr residues at positions of 161, 164, 165, 181, 182, and 185 and Ser at 178 and 179 (Vire et al. 2011). Removal of sialic acids from FcµR+ cells by neuraminidase treatment slightly enhanced IgM-ligand binding, suggesting that desialylated FcµR has better ligand-binding activity (Kubagawa et al. 2009). Notably, Colucci et al. have recently shown that natural IgM, which is rich in terminal sialic acid residues, is internalized by T cells in humans and inhibits T cell responses such as anti-CD3/anti-CD28 mAbs- or PHA-mediated proliferation ex vivo and expression of pro-inflammatory cytokine genes. In contrast, desialylated natural IgM is not internalized and has poor inhibitory activity (Colucci et al. 2015). It remains to be elucidated how carbohydrate moieties of FcµR and its IgM ligands affect their interaction, as has been well documented in the case of FcγRs and IgG ligands (Pincetic et al. 2014; Schwab and Nimmerjahn 2013). In this regard, it has recently been reported that IgM binds FcµR and is internalized irrespective of its glycosylation (Lloyd et al. 2017).

3.2 Cis Engagement

In addition to the above findings of ligand binding in solution, a unique ligand-binding property of FcµR was found in the assay system using cell surface-attached IgM such as an agonistic IgM anti-Fas mAb (CH11 clone). Apoptosis-prone human Jurkat cells stably expressing FcµR were shown to be protected from Fas-/CD95-mediated apoptosis when ligated with the IgM anti-Fas mAb (Hitoshi et al. 1998), but not when ligated with an agonistic IgG anti-Fas mAb or Fas ligand (Honjo et al. 2012b; Kubagawa et al. 2009; Murakami et al. 2012). Notably, co-ligation of FcµR and Fas with the corresponding IgG mAbs plus a common secondary reagent [e.g., F(ab’)2 fragments of anti-mouse γ Ab] had no inhibitory effects on the IgG anti-Fas mAb-induced apoptosis (Kubagawa et al. 2009). This suggests that the anti-apoptotic activity of FcµR depends on usage of the IgM anti-Fas mAb, and not on physical proximity of two receptors by artificial co-ligation as observed in immunoreceptor tyrosine-based inhibitory motif (ITIM)-containing receptors such as FcγRIIb (Ravetch and Lanier 2000) and paired Ig-like receptor of inhibitory isoform (PIR-B) (Bléry et al. 1998).

To determine how FcµR protects from apoptosis induced by IgM anti-Fas mAb, we added various concentrations of soluble IgM or its immune complexes as inhibitors into the apoptosis assays. The results are summarized in a cartoon fashion in Fig. 2. Addition of IgM anti-Fas mAb at 10 ng/ml induced robust apoptosis of FcµR-negative WT control, but not of FcµR-positive Jurkat cells. This FcµR-mediated protection was found to occur in cis, but not in trans, interactions of the Fc portion of IgM anti-Fas mAb with FcµR, because addition of excessive FcµR-positive/Fas-negative cells did not diminish the protection (Honjo et al. 2015). Addition of more than 10,000 molar excess of IgM Ab or its immune complexes (e.g., 1-21 IgM anti-α1-3 dextran mAb/α1-3 dextran) was required for partial, but significant, blockade of the cis interaction of the Fc portion of IgM anti-Fas mAb with FcµR. This suggests that the soluble IgM immune complexes are not potent competitors in the FcµR-mediated protection from apoptosis. When the IgM mAb reactive with CD2 (C373 clone) (Weiss and Stobo 1984) on the surface of Jurkat cells was employed as a potential competitor for the interaction of IgM Fas mAb with FcµR, a tenfold excess of IgM anti-CD2 mAb was sufficient to efficiently block the above cis interaction, thereby permitting the FcµR+ cells to undergo apoptosis. [In particular, same results were also obtained with an IgM anti-TCR mAb (C305 clone)]. Collectively, these findings show that although FcµR binds soluble IgM pentamers at a high avidity of ~10 nM, FcµR binds more efficiently to the Fc portion of IgM antibody when it is attached to a membrane component via its Fab region on the same cell surface. The preferential cis engagement of FcµR is thus distinct from the trans engagement of FcγRIIb, an inhibitory FcγR, in death receptor-mediated apoptosis. The interaction of agonistic IgG mAbs against death receptors, including Fas/CD95, with FcγRIIb is essential for the death receptor-mediated apoptosis and occurs in trans, but not in cis (Bando et al. 2002; Li and Ravetch 2011; Xu et al. 2003). The cis engagement of FcµR in turn implies that FcµR can modulate the functional activity of lymphocyte surface receptors or proteins recognized by either natural or immune IgM antibody.

Fig. 2
figure 2

Cis engagement of FcµR. Ligation of Fas/CD95 death receptor trimer (yellow) with agonistic IgM anti-Fas mAb (black broom shape) induces apoptosis in WT Jurkat cells (1st gray circle), but not in FcµR (blue tennis racket shape)-positive Jurkat cells (2nd). This FcµR-mediated protection is not blocked by addition of 104 molar excess of IgM or its immune complexes (pale blue broom) or tenfold excess of FcµR-positive, Fas-negative cells (pale purple small circles), suggesting an efficient cis interaction of IgM Fas mAb and FcµR on the same cell surface, but not a trans interaction between neighboring cells (3rd). Addition of tenfold excess of IgM mAb (orange broom) reactive with CD2 (orange peanut-shell shape) on Jurkat cells can efficiently block the interaction of IgM Fas mAb and FcµR, resulting in apoptosis (4th)

Full size image

3.3 Modulatory Effect of FcµR by Cis Engagement

The physiological relevance of the cis engagement of FcµR may be related to the unique features of IgM antibody, especially natural IgM antibody. Serum levels of IgM in mice raised in germfree conditions are similar to those in mice maintained under conventional or specific pathogen-free housing conditions (Hashimoto et al. 1978; Haury et al. 1997; Thurnheer et al. 2003). Two-thirds of the newly generated B cells in BM react with self-antigens such as double- or single-stranded DNA, insulin, and lipopolysaccharide (Wardemann et al. 2003). In our assessments, one-fourth of the IgM secreted from Epstein–Barr virus-transformed B cell lines derived from neonatal B cells reacts with lymphocyte surface components (unpublished observation). IgM anti-lymphocyte antibodies are often present in individuals with autoimmune diseases or chronic viral infections and recognize many different surface antigens (e.g., CD45, CD175/Tn, CD3ε, CD4, chemokine receptors, sphingo-sine-1-phosphate receptor 1), and some of those antibodies regulate T cell-mediated inflammatory responses in vitro (Cappione et al. 2004; Daniel et al. 1989; Koren et al. 1992; Liao et al. 2009; Lobo et al. 2008; Muller et al. 1994; Silvestris et al. 1989; Warren et al. 1988; Winfield et al. 1997) (see also Chap. 5 by Peter Lobo). It is thus quite possible that these IgM antibodies reactive with lymphocyte surface components engage FcµR in a cis interaction on the shared membrane surface, thereby modulating the functional activity of lymphocyte surface antigens or receptors by FcµR (see Fig. 4).

To explore this possibility, we compared Ca2+ mobilization upon ligation of lymphocyte surface antigen alone with co-ligation of lymphocyte surface antigen plus FcµR (Honjo et al. 2015). In the first experiment, an equal mixture of WT (i.e., FcµR-negative/GFP-negative) and FcµR-positive/GFP-positive Jurkat cells was preloaded with Ca2+ dye and then simultaneously stimulated by IgM anti-CD2 mAb. As shown in Fig. 3a, the rise in intracellular Ca2+ concentrations ([Ca2+]i) occurred significantly faster in FcµR+ cells than in FcµR− WT cells when ligated with the anti-CD2 mAb. In contrast, the ionomycin-induced [Ca2+]i increase occurred at the same time in both cell types. These findings suggest that the co-ligation of CD2 and FcµR induces a more rapid increase in [Ca2+]i, presumably release from the intracellular store, than does the ligation of CD2 alone. In the second experiment, Ca2+ mobilization by freshly prepared blood B cells was assessed following stimulation with a mitogenic IgM anti-κ mAb (END-5C1 clone) in the presence of IgG2bκ anti-FcµR mAb with blocking (HM7 clone) or non-blocking (HM3 clone) activity for IgM-ligand binding (Kubagawa et al. 2014b). IgM anti-κ mAb-induced Ca2+ mobilization was the same in the absence or presence of FcµR non-blocking mAb. In contrast, the FcµR-blocking mAb diminished the IgM anti-κ mAb-induced Ca2+ mobilization of blood B cells (Fig. 3b), suggesting that FcµR provides a stimulatory signals upon B cell receptor (BCR) cross-linkage with IgM mAbs. Collectively, FcµR expressed on B, T, and NK cells may thus have a potential to modulate the function of target antigens or receptors when they are recognized by natural or immune IgM antibodies, on the same cell surface (see the model shown in Fig. 4).

Fig. 3
figure 3

Ca2+ mobilization by IgM mAbs against CD2 or Igκ. a An equal mixture of WT (FcµR−/GFP−; blue lines) and FcµR+/GFP+ (red lines) Jurkat cells preloaded with the Ca2+ dye Indo-1/AM was stimulated with IgM anti-CD2 (373 clone) mAb at 10 µg/ml (left) or by 1 µM ionomycin (right) at the time points indicated by arrows. The intracellular Ca2+ concentration ([Ca2+] i ) was assessed by the 405/485 nm fluorescence ratio in each viable cell population using an LSR II flow cytometer. Note that the [Ca2+]i rise occurs faster in co-ligation of CD2 and FcµR than ligation of CD2 alone and that the ionomycin-induced [Ca2+]i rise occurs at the same time in both cell types. b The experimental design is depicted in a cartoon fashion (left). Fluo-4-loaded blood B cells were treated with IgM anti-human Igκ mAb (END-5C1; 10 µg/ml) in the absence (green line) or presence of IgG2b, FcµR-blocking (HM7 clone; red line) mAb, or FcµR-non-blocking (HM3 clone; blue line) mAb (100 µg/ml) at the time point indicated by an arrow. The [Ca2+] i levels were assessed by fluorescence intensity during a 5-min period. Note that FcµR provides a stimulatory signal upon BCR cross-linkage with IgM mAbs

Full size image
Fig. 4
figure 4

Working hypothesis for FcµR function in humans. FcµR expressed on the plasma membrane of lymphocytes (B, T, and NK cells) consists of two subunits. The α chain (blue badminton racket shape) has ligand-binding activity via its Ig-like domain (head part) and signal-transducing ability via its conserved Tyr (three small yellow circles) and Ser (not shown) residues in the cytoplasmic tail. The other subunit termed the adaptor is non-covalently associated with the α chain and contains an ITAM (green square) signaling motif. (The molecular identity of this adaptor is presently unknown.) FcµR can either negatively or positively modulate the functional activity of lymphocyte surface protein/receptors (brown sausage shape) recognized by natural or immune IgM (purple broom shape) through the cis engagement

Full size image

3.4 Key Residues in the Transmembrane and Cytoplasmic Tail for FcµR Function

In the case of pairs of activating and inhibitory receptors with highly homologous ectodomains, such as FcRs, killer cell Ig-Like receptors (KIRs), and PIRs, a general receptor structural rule has become appreciated. Namely, when the ligand-binding α chain has a short cytoplasmic tail with no signal-transmitting potential, then it contains a charged residue in the transmembrane segment that facilitates non-covalent association with another transmembrane protein containing immunoreceptor tyrosine-based activation motifs (ITAMs). This association allows for the transmission of activating signals to cells, as seen with FcγRI, FcγRIII, FcαR, FcεRI, KIR2DS or KIR3DS, and PIR-A (Blank et al. 1989; Clevers et al. 1988; Ernst et al. 1993; Kubagawa et al. 1999; Morton et al. 1995; Ravetch 1994). By contrast, when the ligand-binding α chain has a long cytoplasmic tail, then it contains a conventional hydrophobic transmembrane segment and ITIMs in the cytoplasmic tail (Vély and Vivier 1997). Upon phosphorylation of ITIMs after receptor ligation, the ITIM recruits either the polyphosphate inositol 5-phosphatase or the tyrosine phosphatases 1 and 2 to attenuate signaling, as seen with FcγRIIb, KIR2DL or KIR3DL, and PIR-B.

Unlike this general consensus, FcµR is unusual in having both features (Fig. 5): a charged His residue (H253) in the transmembrane segment and a long cytoplasmic tail containing three conserved Tyr (Y315, Y366, Y385) and five conserved Ser (S283, S299, S359, S368, S382 ) residues, when comparing seven different species (Kubagawa et al. 2009). The C-terminal Tyr matches the recently described Ig-tail tyrosine (ITT) motif (DYxN) (Engels et al. 2009; Engels and Wienands 2011), but the other two Tyr residues do not correspond to an ITAM (D/Ex2Yx2L/Ix6-8Yx2V/I), ITIM (I/VxYx2L/V), or switch motif (TxYx2V/I). Collectively, these characteristics suggest a dual signaling ability of FcµR: one from a potential as yet unidentified adaptor protein non-covalently associating with the FcµR via the H253 residue and the other from its own Tyr and/or Ser residues in the cytoplasmic tail (see Figs. 4 and 5).

Fig. 5
figure 5

Summary of FcµR mutational analysis. The human FcµR cDNA encodes a type I transmembrane protein that consists of a single V-set Ig-like domain (blue oval shape), an additional extracellular region (stalk) with no known domain structure, a transmembrane (TM) segment (between two thick lines) containing a charged His residue (purple circle), and a relatively long cytoplasmic tail containing three conserved Tyr residues (yellow circles). Point mutations are indicated, and the extent of the deletion of the cytoplasmic tail is shown by the green bracket. Hatch marks indicate exon boundaries in the FCMR gene. In representative flow cytometric profiles (upper left), cells stably expressing FcµR WT (blue) and ΔCy (red) were stained with biotin-labeled anti-FcµR (open) or isotype-matched control (shaded) mAb for cell surface expression of FcµR (left panel) and with biotin-IgM (open) or PBS (shaded) for ligand binding (right panel). Because profiles with control mAb or PBS were the same between FcµR WT and ΔCy cells, only one shaded profile is shown. Note the enhanced IgM binding by FcµR ΔCy cells as compared to FcµR WT cells, despite their equivalent levels of surface FcµR. In representative epi-fluorescence microscopic images (upper right), the FcµR WT (left) and H253F cells (right) were incubated with Alexa Fluor 555-IgM (without NaN3) on ice, washed, and cytocentrifuged. Fluorescence images were combined with phase contrast cell images (scale bars = 10 µm). Altered phenotypes observed in mutant FcµR cells or potential function of the indicated residues are shown in the yellow-filled boxes. The unique sequences around three conserved Tyr residues are also shown with underlines indicating conserved amino acid residues

Full size image

To explore whether the aforementioned amino acid residues in FcµR are responsible for the receptor function, we made human FcµR cDNA constructs with point mutations (H253F, Y315F, Y366F, or Y385F) or a deletion of most of the cytoplasmic tail (A281–A390; ΔCy) and expressed them in Jurkat T cells. The results (Honjo et al. 2015) are summarized as follows (see Fig. 5). (i) Although non-mutated and mutated FcµR-bearing cells expressed comparable levels of cell surface FcµR as judged by receptor-specific mAbs, IgM-binding activity was significantly increased in the ΔCy mutant. [The ΔCy mutant lacks most of the cytoplasmic tail but includes an eight post-transmembrane basic amino acid-rich region (K273–K280)]. Our subsequent data suggested that this enhancement was likely due to the formation of oligomeric FcµR due to its presumably mobile nature within the plasma membrane, rather than to the inside-out regulation of FcµR ligand binding by its cytoplasmic tail as seen in adhesion molecules (Kinashi 2005). (ii) The His253 residue was found to be important in the anchoring of FcµR in the plasma membrane. When examining the fate of IgM bound to FcµR by immunofluorescence microscopy, enhanced cap formation was clearly observed with the H253F mutant even at 4 °C as compared to the cells expressing non-mutant or other mutant FcµRs (except the ΔCy mutant), which exhibited a more broadly localized staining pattern. Notably, unlike other multi-chain FcRs, the FcµR H253F mutant was expressed on the surface of Jurkat cells without a potentially associated membrane protein. (iii) Consistent with the findings of Vire et al. (2011), the two C-terminal conserved Tyr residues were involved in receptor-mediated endocytosis. (iv) The FcµR-mediated protection from IgM anti-Fas mAb-induced apoptosis was significantly diminished in the Y315F and ΔCy mutants, as the frequency of apoptotic cells in these mutants was indistinguishable from those in FcµR− control cells. This is of interest, given the unique sequence around the Y315 residue: 308 P∙R∙S/T∙Q·N·N·I/V∙Y∙S/T∙A∙C∙P∙R∙R∙A· R 323 (bold type indicates conserved amino acids). This does not match any known Tyr-based signaling motif. The mechanism of Toso-/FcµR-mediated protection from apoptosis was suggested to result from potentiation of the cellular FLIP [FADD-like IL-1ß-converting enzyme (FLICE)-like inhibitory protein], a master anti-apoptotic regulator (Hitoshi et al. 1998), or alternatively by prevention of internalization of Fas, an important step for apoptosis signaling (Vignaux et al. 1995; Yamauchi et al. 1996), owing to simultaneous cross-linkage of both Fas and FcµR with IgM Fas mAb (Murakami et al. 2012).

In addition to the mutational analysis, our previous findings indicated that ligation of FcµR with preformed IgM immune complexes induced the phosphorylation of both Tyr and Ser residues of the receptor. Intriguingly, the phosphorylated FcµR migrated on sodium dodecyl sulfate-polyacrylamide gel electrophoresis faster than the unphosphorylated form, unlike most proteins, which usually run slower when phosphorylated (Kubagawa et al. 2009). It remains unclear whether the phosphorylation causes a global structural change of FcµR leading to increased mobility as seen, e.g., in CD45 on activated myeloid cells (Buzzi et al. 1992) or if proteolytic cleavage occurs in the cytoplasmic tail of FcµR after receptor ligation as observed in FcγRIIa on platelets (Gardiner et al. 2008). Ligation of FcµR on NK cells with IgM immune complexes was shown to lead to the phosphorylation of PLCγ2 and ERK1/2 (Murakami et al. 2012).

4 FcµR Deficiency in Mice

Fcmr KO mice have been independently generated by three different laboratories [K. H. Lee (Hannover Medical School, Hannover, Germany); H. Ohno (RIKEN, Yokohama, Japan); and T. W. Mak (Princes Margaret Cancer Center/University Health Network, Toronto, Canada)] and have been characterized by five different groups of investigators (Brenner et al. 2014; Choi et al. 2013; Honjo et al. 2014; Honjo et al. 2012a; Lang et al. 2013a; Nguyen et al. 2011; Ouchida et al. 2012). Clear differences in the reported phenotypes exist among these mice that are well summarized in a recent review article (Wang et al. 2016). While the basis for these differences requires further investigation, they could in part be attributed to: (i) different strategies for gene targeting (i.e., deletion of exon 4-7, exon 2-4 versus exon 2-8 and/or the absence versus presence of the Neo gene in the mouse genome); (ii) embryonic stem (ES) cells of C57BL/6 versus 129 origin as well as the extent of the 129 mouse-origin DNA around the Fcmr gene remaining after backcrossing onto C57BL/6. [In this regard, the region closely flanking the targeted gene, called the passenger genome, remains of donor origin and typically contains mutations called passenger mutations. Annotating these passenger mutations to the reported genetically modified congenic mice generated using 129-ES cells revealed that nearly all these mice possess multiple passenger mutations potentially influencing the phenotypic outcome (Vanden Berghe et al. 2015)]; (iii) investigators’ ideas regarding the cellular distribution of FcµR/Toso in B cells versus myeloid and T cells and its function as an IgM Fc-binding protein versus an anti-apoptotic protein; and/or (iv) other factors (e.g., age of the mice, environments including intestinal microbiota or reagents used). Nevertheless, the abnormal phenotypes commonly observed in Fcmr KO mice are as follows: (i) alterations in B-lineage cell subpopulations; (ii) dysregulation of humoral immune responses; and (iii) predisposition to autoantibody production (Choi et al. 2013; Honjo et al. 2012a; Ouchida et al. 2012). Notably, many of the abnormalities seen in Fcmr KO mice mirror those observed in mice deficient in IgM secretion, suggesting that FcμR is a critical sensor of secreted IgM. In Fcmr KO mice on autoimmune backgrounds, FcμR was shown to play important regulatory roles in: (i) the autoantibody production; (ii) the differentiation of MZ B cells into plasma cells; and (iii) the formation of Mott cells, aberrant plasma cells with intracytoplasmic Ig inclusion bodies (Honjo et al. 2014).

Apart from these mutant mice, the laboratory of N. Baumgarth has recently generated a mouse strain in which the second stalk region exon (exon 4) of Fcmr is flanked by loxP sites allowing the B cell-specific deletion of Fcmr by crossing with Cre-Cd19 transgenic mice (Nguyen et al. 2017). Several interesting findings were obtained by comparative analysis between mice with such a B cell-specific FcµR deletion and control mice. FcμR directly interacts with membrane-bound IgM BCR in the trans-Golgi network of BM immature B cells, thereby regulating the surface expression of BCR and eventually resulting in limiting tonic BCR signaling. B cell-specific FcμR deficiency results in dysregulated spontaneous activation and differentiation of B-1 and B-2 cells and development of a lympho-proliferative disorder. This suggests that FcμR constrains BCR expression to regulate the fundamental homeostasis and biology of B cells. [It is worth noting, however, that deletion of exon 4 could still allow production of soluble form of FcμR by the mutant B cells via a reading frame shift in exon 5 (TM) (see hatch marks in Fig. 5). Such a soluble FcμR could have unexpected consequences on B cell function. In fact, we have identified another splice variant in CLL patients that results from the direct splicing of exon 4 to exon 6 (1st cytoplasmic), skipping exon 5. This splice event results in a reading frame shift in exon 6 and generates a novel 70-amino acid hydrophilic, carboxyl-terminal tail and the resultant soluble FcμR protein is clearly elevated in many patient’s sera as determined by enzyme-linked immunosorbent assays (Li et al. 2011)]. Contrary to the above FcµR-mediated suppression of BCR signaling, FcµR has also been shown to enhance survival of mature B cells upon BCR cross-linkage ex vivo with F(ab’)2γ anti-µ antibodies by activation of the non-canonical NF-κB pathway, but not upon CD40 ligation or LPS stimulation (Ouchida et al. 2015). Thus, FcµR may have a potential to transmit both positive and negative signals to cells.

Another difference in Fcmr KO mice between T. W. Mak and H. Ohno is that splenic B cells from our (the latter) Fcmr KO mice produce significantly less IL-10, but comparable amounts of IL-6, ex vivo upon stimulation with Salmonella bacteria or with ligands for Toll-like receptor 4 (TLR4), TLR7, or TLR9 (Fig. 6). Since B cell-derived IL-10 has been implicated as an important negative regulator of MOG-induced EAE (Fillatreau 2015), we thought that our Fcmr KO mice would be more susceptible than WT to MOG-induced EAE. However, the results reported by Brenner et al. (2014) suggest that this is not the case. Their Fcmr/Toso KO mice are resistant to MOG-induced EAE, because their Fcmr/Toso KO DCs are immature and tolerogenic and weak stimulators of inflammatory T cell responses. Fcmr/Toso KO Th17-polarizing cells secrete significantly less IL-17 and IL-10 than WT control mice (Gaublomme et al. 2015). Intriguingly, passive administration of a recombinant, soluble human FcµR–IgG fusion protein ameliorates MOG-induced EAE in WT mice (Brenner et al. 2014). The basis for this discrepancy (susceptible versus resistant) remains unclear at the moment, because we have never examined the susceptibility of our Fcmr KO mice to MOG-induced EAE and Brenner et al. have never assessed IL-10 production by B cells in their Fcmr/Toso KO mice. A side-by-side analysis of these two different strains of Fcmr/Toso KO mice would facilitate the resolution of these conflicting results, and it is highly likely that this discrepancy results from different strategies for gene targeting (deletion of exon 2-4 without Neo for ours versus deletion of exon 2-8 with remaining of Neo in the mouse genome for Brenner et al.).

Fig. 6
figure 6

Diminished IL-10 production by FcµR-deficient B cells. Splenic B (left and right) and bone marrow myeloid (middle) cells (4 × 105 cells) from sex and age-matched Fcmr-deficient (black columns) or littermate control (white columns) mice were cultured for 2 days in the absence (none) or presence of heat-killed, non-opsonized (−) or serum-opsonized (+) BW335, an LT2 strain of Salmonella enterica serovar Typhimurium (4 × 105 cfu), or the indicated TLR ligands (TLRL): LPS (10 µg/ml) for TLR4, Gardiquimod (1 µg/ml) for TLR7, and ODN1826 (2 µM) for TLR9. The concentration (pg/ml) of IL-10 (left and middle) and IL-6 (right) in the culture supernatants was assessed by ELISA in triplicate. The symbols ** and *** and “ns” indicate p < 0.01, p < 0.001, and “not significant,” respectively, as assessed by Student’s t test. Note (i) diminished production of IL-10, but not IL-6, by FcµR-deficient B cells and (ii) comparable production of IL-10 by marrow myeloid cells from both groups of mice. Identical results were obtained seven times for IL-10 and twice for IL-6, and representative experiments are shown

Full size image

5 Epilogue

FcRs for switched Igs are expressed by many different cell types, including myeloid cells, and are considered to be central mediators coupling innate and adaptive immune responses. Rewardingly, much of the knowledge gained from studies of these FcRs has been translated into clinical fields. On the other hand, the long elusive IgM FcR was finally identified eight years ago by functional cloning. However, since the cloned FcµR cDNA was identical to the cDNA-encoding TOSO or Fas apoptosis inhibitory molecule 3 (FAIM3), which was also previously identified by functional cloning as a potent inhibitor for Fas-mediated apoptosis, there have been spirited debates regarding the real function of this receptor, Fcµ-binding versus anti-apoptotic activity. Notably, there is now a general consensus that TOSO/FAIM3 is an authentic Fcµ-binding protein and not a Fas inhibitory protein per se (Kubagawa et al. 2015). Several interesting findings about the FcµR have recently been reported: FcµR binds more efficiently to the Fc portion of IgM antibody when it is attached to a membrane component via its Fab region on the same cell surface (i.e., cis interaction) than to the Fc portion of IgM in solution (trans interaction). FcµR directly interacts with membrane-bound IgM BCR in the trans-Golgi network of BM immature B cells, thereby regulating the surface expression of IgM BCR and eventually resulting in limiting tonic BCR signaling. FcµR can also regulate the differentiation of MZ B and B1 cells. By contrast, FcµR may enhance survival of mature B cells upon BCR cross-linkage via activation of the non-canonical NF-κB pathway. Immunotherapy targeting the FcµR is now designed for CLL cells. Many conflicting results still exist, but we hope that this short article may help to resolve these existing puzzles and will open new avenues of investigation.