FCRL5 exerts binary and compartment-specific influence on innate-like B-cell receptor signaling

Abstract

Innate-like splenic marginal zone (MZ) and peritoneal cavity B1 B lymphocytes share critical responsibilities in humoral responses but have divergent B-cell receptor (BCR) signaling features. A discrete marker of these subsets with tyrosine-based dual regulatory potential termed “Fc receptor-like 5” (FCRL5) was investigated to explore this discrepancy. Although FCRL5 repressed the robust BCR activity that is characteristic of MZ B cells, it had no influence on antigen receptor stimulation that is blunted in peritoneal cavity-derived B1 B cells. The molecular basis for the receptor’s inhibitory function derived from recruitment of the Src homology-2 domain-containing tyrosine phosphatase 1 (SHP-1) to a cytoplasmic immunoreceptor tyrosine-based inhibitory motif. Surprisingly, mutagenesis of this docking site unearthed coactivation properties for FCRL5 that were orchestrated by independent association of the Lyn Src-family kinase with an intracellular immunoreceptor tyrosine-based activation motif-like sequence. FCRL5’s unique binary regulation directly correlated with SHP-1 and Lyn activity, which, like BCR function, differed between MZ and B1 B cells. These findings collectively imply a specialized counterregulatory role for FCRL molecules at the intersection of innate and adaptive immunity.

Innate-like B-lineage cells positioned at strategic microanatomical sites provide the first line of effector defense that bridges host protection until adaptive mechanisms emerge (1). The lymphocytes charged with these responsibilities include splenic marginal zone (MZ) B cells harbored in a location optimal for filtering blood-borne antigens and B1-lineage cells that guard the peritoneal (PEC) and pleural body cavities (2–4). Their capacity for broad neutralization is associated with evolutionarily conserved Ig repertoires, distinct sensitivity to T-cell–independent stimuli, and rapid or spontaneous production of ‘‘natural,’’ polyreactive antibodies (5, 6). These features distinguish MZ and B1 B cells from their more abundant B2-lineage counterparts that recirculate and participate in T-cell–dependent responses.

MZ and B1 B-cell development is strongly influenced by B-cell receptor (BCR) specificity in concert with the composite array of surface and intracellular regulatory proteins that help balance antigenic responses. Mutations in cluster of differentiation 45 (CD45), Bruton's tyrosine kinase (BTK), or phospholipase C gamma 2 (PLCγ2) that dampen BCR signaling favor MZ development and a loss of follicular (FO) B cells (7). However, defects in negative regulatory components, such as Lyn, Src homology-2 (SH2) domain-containing tyrosine phosphatase 1 (SHP-1), or CD22, lead to a loss of MZ B cells, a relative expansion of the B1 compartment, and increased susceptibility to autoimmunity (8). Although many other trophic, migratory, and retention factors instruct their development and positioning, these signals must be integrated in the context of BCR signaling which primarily drives B-cell fate and survival (7, 9–11). Correspondingly, BCR triggering differs markedly between these subpopulations. MZ B cells exhibit more robust whole-cell protein tyrosine phosphorylation, calcium mobilization, and PLCγ2 and spleen tyrosine kinase (Syk) activation than FO B cells but also are more sensitive to apoptosis (12, 13). In contrast, B1 B cells have blunted calcium mobilization, NF-κB activation, and proliferation but also may possess relatively higher rates of apoptosis than PEC B2 cells (14–16). Notably, these properties do not differ according to CD5 expression, because both the B1a (CD5⁺) and B1b (CD5⁻) B-cell subsets respond similarly to BCR ligation (17). Although they differ from anergic cells, it remains unclear whether the unique biology of B1 B cells is secondary to chronic antigenic exposure, suppressive signals present in the coelomic cavity microenvironment, or other causes (4, 14).

An evolutionarily conserved gene family related to the Fc receptors (FCR) for IgG and IgE, termed “FCR-like” (FCRL), is preferentially expressed by B cells and encodes transmembrane proteins with tyrosine-based immunoregulatory motifs (18). Although Ig binding has been detected recently for two human members (19), antibodies do not appear to associate with other FCRL proteins. Intriguingly, two other representatives have been found to interact with MHC-like molecules (20, 21). Because of the growing clinical relevance of FCRLs in infectious diseases, autoimmunity, malignancies, and immunodeficiencies, several groups have investigated FCRL signaling function (22–25). In humans, FCRL1 has two immunoreceptor tyrosine-based activation motif (ITAM)-like sequences and enhances BCR-induced calcium mobilization and cellular proliferation (26). In contrast, FCRL2–5, which feature one or more consensus immunoreceptor tyrosine-based inhibition motifs (ITIM) as well as ITAM-like sequences, all inhibit BCR activation via recruitment of the SH2 domain-containing SHP-1 and/or SHP-2 phosphatases (27–30). We have shown previously that the FCRL5 mouse ortholog discretely marks innate-like B cells and possesses an ITIM as well as an ITAM-like sequence that differs from the canonical motif (D/EX_2–3YXXL/IX_6–8YXXL/I), with a glutamic acid residue rather than an aliphatic residue at the second Y+3 position (31). Although FCRL5 inhibits BCR-mediated calcium mobilization in MZ B cells, the molecular basis for this activity and its function in B1 B cells remains unclear. Furthermore, the conservation of both activating and inhibitory sequences in FCRL5 and other FCRLs suggests they have dual signaling properties, but definitive functional evidence for this bifunctionality is lacking. Because of its distinct distribution and regulatory potential, we investigated the biological role of FCRL5 in these specialized B lymphocytes that have recognized differences in their adaptive signaling capacity.

Here we report that FCRL5 has dual modulatory and compartmental subset-specific effects on antigen receptor signaling. Upon association with the BCR, the ITAM-like and ITIM sequences in FCRL5 are tyrosine phosphorylated and recruit the Lyn Src-family kinase (SFK) and SHP-1 protein tyrosine phosphatase. The nonredundant contributions of these elements to FCRL5’s unique counterregulatory function further revealed that differences in adaptive signaling in MZ and B1 B cells correlate directly with their intrinsic SHP-1 activity.

Results

FCRL5 Attenuates MZ but Not B1 BCR Signaling.

Although MZ and B1 B cells share many similarities, they differ significantly in antigen receptor signaling (13, 14). Consistent with previous work, calcium mobilization among wild-type (WT) C57BL/6 splenic B-cell subsets was most intense in MZ B cells, followed by newly formed (NF) and FO B cells, whereas PEC B2 B cells demonstrated a stronger response than B1 cells (Fig. 1A). Given their dissimilar BCR signaling properties, we investigated whether FCRL5 regulation also might vary in MZ versus B1 B cells. In accord with our earlier findings (31), the strong calcium response observed in gated MZ B cells after BCR engagement alone was diminished significantly by FCRL5 co-crosslinking (Fig. 1B). In contrast, FCRL5 did not alter this impaired activation cascade in B1 B cells. An analysis of global tyrosine phosphorylation by intracellular phospho-specific flow yielded similar results. Although FCRL5 could inhibit this downstream outcome in MZ cells, it had no influence in B1 B cells which showed less robust whole-cell tyrosine phosphorylation (Fig. 1B). Intriguingly, FCRL5 coaggregation reduced BCR-mediated calcium mobilization in MZ B cells to the same extent seen in B1 cells stimulated through the BCR alone. Because this blunted activation response in B1 B cells might be caused by constitutive association of FCRL5 with the BCR, we independently crosslinked FCRL5 before BCR stimulation. However, sequestering FCRL5 did not affect BCR-mediated calcium mobilization in WT B1a B cells (Fig. S1). These data indicate that FCRL5 does not constitutively associate with the BCR and that other mechanisms likely account for observed compartment-specific differences in its regulatory function.

Fig. 1.

FCRL5 differentially regulates innate-like B-cell signaling. (A) Spleen and PEC leukocytes (1 × 10⁶) from WT C57BL/6 mice were labeled with Indo-1/acetoxymethyl (AM), stained for the indicated markers and biotinylated F(ab’)₂ fragments of a rat anti-mouse IgM mAb (SB73a, 10 μg/mL) and were then rested for 30 min. Calcium mobilization in the gated subsets (MZ: CD19^highCD21^highCD23^low; NF: CD19^highCD21^lowCD23^low; FO: CD19^highCD21^lowCD23^high; B1a: CD19^highCD5^high; B1b: CD19^highCD5^lowCD11b^high; and B2: CD19^highCD5^lowCD11b^low) was analyzed before and after crosslinking with 20 μg/mL of streptavidin (SA) by flow cytometry. Arrows indicate the time of SA administration. (B) Spleen and PEC leukocytes from WT mice were stained as in A and were counterstained with biotinylated F(ab’)₂ anti-FCRL5 (3B7) followed by SA-phycoerythrin (SA-PE) to define FCRL5 expression (Left). Indo-1/AM-loaded cells were stained for distinguishing surface markers and with biotinylated F(ab’)₂-digested mAbs including rat anti-mouse IgM (αIgM, SB73a) and anti-FCRL5 (αFCRL5, 3B7) or an isotype control (Ctrl, mouse IgG1κ) (all 10 μg/mL). Ca²⁺ mobilization was monitored in the MZ- or B1 B-cell–gated subsets before and after crosslinking with SA (Center). Whole-cell tyrosine phosphorylation (pTyr) was analyzed in these subpopulations at baseline (NS) and following SA crosslinking for 10 min, fixation, permeabilization, and intracellular staining with anti-pTyr (Right). Results are representative of three independent sets of experiments.

FCRL5 Inhibits BCR-Mediated Tyrosine Phosphorylation.

To define the molecular basis for its differential innate-like B-cell activity, we next dissected the roles of the FCRL5 cytoplasmic tyrosine-based motifs. FcγRIIb/FCRL5 chimeric constructs encoding six different cytoplasmic mutants as well as full-length FcγRIIb and a vector control were stably transduced into the A20-IIA1.6 (IgG2aκ) mouse B-cell line that lacks endogenous FCRL5 and FcγRIIb expression (Fig. 2A) (29). After confirming equivalent levels of surface expression, responses following coligation of the various chimeric proteins with the BCR were compared with antigen receptor ligation alone (Fig. S2 A and B).

Fig. 2.

Coligation of the WT FcγRIIb/FCRL5 chimeric protein with the BCR inhibits whole-cell tyrosine phosphorylation. (A) Schematic illustration of FCRL5 cytoplasmic tyrosine-based motifs and FcγRIIb/FCRL5 chimeric constructs as detailed in SI Materials and Methods. (B) WT or FFF transductants (1 × 10⁷) were stimulated with intact (25 µg/mL) or F(ab’)₂ (16.6 µg/mL) rabbit anti-mouse IgG for the specified times, and whole-cell lysates (WCL) were immunoblotted with the indicated Abs and β-actin as a loading control. Lysates were immunoprecipitated with anti-HA and analyzed for pTyr and HA to verify equal loading. The arrow indicates the position of the chimeric receptor. Results are representative of three independent experiments.

Because the FCRL5 cytoplasmic tail contains both ITAM-like and ITIM sequences, we initially examined whether both motifs could be phosphorylated. After treatment with the sodium pervanadate phosphatase inhibitor, the WT, Y543F, Y556F, Y566F, and Y543F/Y556F (FF) chimeras were all tyrosine phosphorylated, but the Y543F/Y556F/Y566F (FFF) mutant was not (Fig. S3). This last variant confirmed the functional inactivity of a fourth cytoplasmic tyrosine at position 544 and indicates that FCRL5 tyrosine-based signaling is confined to its ITAM-like and ITIM units.

We then investigated the impact of the FCRL5 WT and FFF mutants on BCR-induced whole-cell tyrosine phosphorylation. BCR triggering alone induced rapid tyrosine phosphorylation of multiple intracellular proteins, including the ERK MAP kinase; however, coligation with the WT chimera greatly reduced these effects as a function of time (Fig. 2B). In contrast, global tyrosine phosphorylation and ERK activation were unaffected by FFF crosslinking. Immunoprecipitation experiments indicated that the WT FCRL5 cytoplasmic tail was tyrosine phosphorylated following its association with the BCR but not upon BCR ligation alone. Results in this cell line thus were consistent with FCRL5 regulation in MZ B cells and demonstrated that its cytoplasmic domain can inhibit BCR activation in a tyrosine-dependent manner.

FCRL5 ITIM and ITAM-Like Sequences Counterregulate BCR Signaling.

We next explored the functional contributions of the individual FCRL5 tyrosines in calcium mobilization assays. BCR ligation alone induced a characteristic wave of calcium flux, but crosslinking with the WT chimeric receptor significantly abrogated this effect (Fig. 3A). In fact, the degree of inhibition was even more potent than that of the FcγRIIb-positive control. The function of the ITIM then was examined by disrupting the ITAM-like 543 and 556 tyrosines. Coligation with this FF mutant completely shut down the calcium response. Conversely, mutating the Y566 ITIM residue unearthed coactivation properties for FCRL5 and resulted in enhanced calcium flux compared with antibody receptor triggering alone. Finally, among the ITAM-like tyrosines, the Y543F mutant entirely blocked calcium signaling, whereas the Y556F variant disclosed little difference from that of the WT tail. As expected, the FFF mutant had no influence on BCR-mediated calcium flux. Because our results shown in Fig. 2B indicated that FCRL5 could restrain the MAP kinase cascade, we next deconstructed the behavior of these tyrosines in ERK activation. The Y543F, Y556F, and FF variants, which all retain intact ITIM sequences, inhibited BCR-dependent ERK phosphorylation (Fig. 3B). In contrast, the Y566F mutant possessing an unmodified ITAM-like sequence moderately enhanced ERK activation. Thus FCRL5’s modulation of ERK signaling was similar to its effects on calcium mobilization. These findings collectively indicate that FCRL5 has dual regulatory tyrosine-based signaling properties that are balanced and graded. The ITIM (Y566) exerts strong inhibition individually, whereas the ITAM-like (Y543/Y556) sequence has the inverse effect and can augment downstream BCR-mediated responses. Importantly, within the noncanonical ITAM, Y543 played a critical role, whereas the Y556 residue appeared to be dispensable. To confirm this finding, YFF and FYF chimeric receptor cell lines also were generated (Fig. S4 A and B). Coligation of the YFF mutant with the BCR positively regulated calcium mobilization, whereas the FYF variant had no obvious effect (Fig. S4C). Accordingly, pervanadate treatment demonstrated that the YFF tail could be tyrosine phosphorylated, but the FYF mutant could not (Fig. S4D). The opposing influence of the Y543 ITAM-like and Y566 ITIM residues therefore equips FCRL5 with tyrosine-based counterregulatory function.

Fig. 3.

FCRL5 counterregulates BCR signaling. (A) Ca²⁺ mobilization in Fluo-4 AM– and SNARF-1–loaded A20-IIA1.6 transductants (1 × 10⁶) was analyzed before and after BCR stimulation alone (black line) or receptor crosslinking (gray line). Arrows indicate the time of addition of stimulating antibodies. (B) Lysates from transductants (5 × 10⁶) left untreated (−) or stimulated with intact (I) or F(ab’)₂ fragments (F) of rabbit anti-mouse IgG for 10 min were probed for total ERK and pERK. Densitometric quantitation of the pERK/ERK ratio for one representative data set is shown, and pooled results from three independent experiments are indicated below. Error bars specify the mean ± SEM; *P < 0.05; **P < 0.01.

SHP-1 and Lyn Are Recruited to Independent FCRL5 Cytoplasmic Tyrosines.

The cytosolic effector proteins that mediate FCRL5 functions then were examined. To define the elements responsible for its inhibitory activity, chimeric protein immunoprecipitates were probed for SHP-1, SHP-2, and SH2 domain-containing inositol phosphatase (SHIP). Except for the FFF mutant, all the receptors were tyrosine phosphorylated after BCR coengagement (Fig. 4A). Among these candidates, SHP-1 associated with the WT, Y543F, and Y556F single mutants and with the FF mutant but not with the Y566F or FFF receptors. In contrast, neither SHP-2 nor SHIP was coprecipitated with any FCRL5 tail variant. This ITIM-dependent association also was confirmed by SHP-1–specific immunoprecipitation which identified a low level of constitutive FCRL5 binding that was enhanced by BCR coligation (Fig. 4B).

Fig. 4.

FCRL5 ITIM and ITAM-like sequences recruit SHP-1 and Lyn independently. (A) Lysates from A20-IIA1.6 cells (1 × 10⁷) stimulated with intact rabbit anti-mouse IgG for 10 min (+) or left untreated (−) were immunoprecipitated with anti-HA and blotted for pTyr, indicated phosphatases, or HA as a loading control. (B) Lysates from the FF and FFF transductants stimulated as in A were immunoprecipitated with anti–SHP-1 and probed for HA or SHP-1 as a loading control. (C) Immunoprecipitates from the chimeric panel stimulated as in A were blotted for the indicated kinases.

Putative signaling components responsible for FCRL5's activating properties were then assessed. The Syk and PLCγ2 kinases did not associate with the FCRL5 cytoplasmic tail (Fig. 4C). Other kinases, including BTK and PI3K, as well as the adaptor proteins growth factor receptor-bound protein 2 (Grb2) and B-cell linker protein (BLNK), also were undetectable in immunoprecipitates. Because Lyn is the predominant SFK expressed in B cells and is required for the activating function of Igα/Igβ, CD19, and CD180/RP105 as well as the inhibitory properties of CD5, CD22, CD32, and CD72 (8), we explored whether it associated with FCRL5. Lyn indeed coprecipitated with the WT, Y556F, and Y566F chimeras but not with the Y543F, FF, or FFF mutants. We also determined whether Lyn was physiologically active (pY397) or inactive (pY508) (32). Western blotting with a phospho-specific Ab revealed that the SFK associated at the Y543 position was in the active state. Thus, after BCR coligation, the FCRL5 Y566 ITIM residue is phosphorylated and coordinates the receptor’s inhibitory activity through its association with SHP-1. The ITAM-like Y543 tyrosine also is phosphorylated but offsets the receptor’s potent ITIM-mediated repression by recruiting the active form of Lyn.

SHP-1 and Lyn Mediate the Unique Dual Functionality of FCRL5 in Innate-Like B Cells.

To uncouple its regulation in primary B cells, we next validated FCRL5 function in viable motheaten (me^v/me^v) mice, which possess a point mutation that disables SHP-1 phosphatase activity (∼20% of normal), and in Lyn-deficient mice (33, 34). Both models develop severe autoimmunity and have a loss of MZ and FO B cells in the spleen as well as B2 cells in the PEC but develop a relative expansion of B1a and B1b B cells (Fig. S5 A and B) (8, 35). After confirming that FCRL5 is expressed at comparable levels by B1a and B1b B cells from these three strains (Fig. S5B), we examined its regulatory impact on the BCR cascade. In remarkable contrast to WT cells, where it had no clear influence, FCRL5 crosslinking in SHP-1–mutant B1 B cells strongly augmented BCR-mediated calcium mobilization and whole-cell tyrosine phosphorylation (Fig. 5 A and B). Intriguingly, the opposite effect was observed in Lyn^−/− B1 B cells. Here, coligating FCRL5 with the BCR resulted in marked repression of calcium flux and whole-cell tyrosine phosphorylation. Notably, this outcome resembled observations made for WT splenic MZ and A20-IIA1.6 B cells. Responses were similar in the B1a and B1b subsets. Furthermore, FCRL5 engagement alone in these genetically modified B cells failed to regulate calcium flux or tyrosine phosphorylation. These results in mutant mice confirm the inverse contributions of SHP-1 and Lyn to its function and demonstrate that their relative activity has a direct effect on driving FCRL5 biology. These data validate the essential inhibitory role of SHP-1, and an unsuspected nonredundant requirement for Lyn in mediating its activating function also was uncovered. These data collectively indicate that FCRL5 may serve as a cellular sensor for SHP-1 and Lyn function in innate-like B cells.

Fig. 5.

FCRL5 has unique SHP-1– and Lyn-dependent dual functionality in B1 B cells. (A and B) PEC cells isolated from me^v/me^v and Lyn^−/− mice were stained and stimulated as in Fig. 1, and Ca²⁺ mobilization and intracellular pTyr were analyzed in the gated subsets. (C) PEC cells from the indicated mice were preloaded with Indo-1/AM and blocked with anti-CD16/32 (2.4G2) before staining with discriminating surface markers, biotinylated F(ab’)₂ rat anti-mouse IgM (SB73a), and intact CD5, CD22, CD32, CD72 or isotype-matched control mAbs (all at 10 μg/mL). Ca²⁺ mobilization in gated B1a cells was monitored before and after the addition of SA (20 μg/mL). Arrows indicate the time of SA administration. The data are representative of at least two independent experiments.

To determine whether dual functionality was a feature unique to FCRL5, other well-characterized regulatory proteins including CD5, CD22, CD32, and CD72 were analyzed also. Cell-surface staining demonstrated their similar expression levels on PEC B1a B cells derived from WT, me^v/me^v, and Lyn^−/− mice (Fig. S6). IgM expression was slightly lower on me^v/me^v B1a B cells, and IgD was uniformly dim. In contrast to FCRL5, all four of the receptors tested could inhibit BCR-mediated calcium mobilization in WT B1a B cells, albeit to different extents (Fig. 5C). In SHP-1–mutant B1a B cells, CD22 and CD32 retained inhibitory function, likely through their association with the SHIP inositol phosphatase (36, 37). However, the dampening capacity of CD5 and CD72 was abolished, confirming that their inhibitory properties are SHP-1 dependent (38, 39). Furthermore, unlike FCRL5, none of these molecules acquired enhancing function in SHP-1–mutant B1a B cells, including CD5, CD22, and CD72, each of which has been shown to play both positive and negative roles (40–42). Nevertheless, these receptors could all block BCR calcium signaling in Lyn^−/− B1a B cells. These data indicate that ITIM tyrosine phosphorylation, which is required for SHP-1 and SHIP docking, and the consequent inhibitory function of each of these proteins, including FCRL5, could be compensated by SFKs other than Lyn that are coexpressed in these cells. However, the FCRL5 ITAM-like Y543 residue does not appear susceptible to promiscuous SFK activity in this fashion and instead is primarily Lyn dependent. Taken together, these results indicate that FCRL5 possesses a unique counterregulatory function in B cells that is directly mediated by SHP-1 and Lyn.

Divergent FCRL5 Function in Innate-Like B Cells Reflects Differences in SHP-1 Activity.

To confirm further the association of SHP-1 and Lyn with FCRL5, we examined its biochemical properties in primary cells. Because innate-like B cells represent only a minority of the total B lymphocytes in normal adult mice, we used a human soluble B-cell activating factor (BAFF) transgenic (Tg) mouse model as a source of larger quantities of MZ B cells (Fig. S7A). The resulting human BAFF (hBAFF) Tg mice possessed a phenotype similar to previous established models and developed markedly enlarged spleens compared with WT littermate controls (Fig. S7B) (43). As a fundamental B-cell survival factor, mice overexpressing soluble hBAFF dramatically expanded their total B-cell compartment and MZ B-cell frequency (Fig. S7C). To corroborate these flow cytometric studies, immunohistology was performed on the spleens of adult WT and hBAFF Tg mice. Fig. S7D demonstrates that spleens from hBAFF Tg mice have a substantially expanded MZ outside the MOMA-1⁺ ring that demarcates metallophilic macrophages positioned in the marginal sinus. Splenic MZ B cells from these hBAFF Tg mice then were sorted by FACS and used for biochemical validation (Fig. S7E). As expected, BCR coligation in purified MZ B cells induced tyrosine phosphorylation of FCRL5 and the coincident association of SHP-1 and Lyn (Fig. 6A). These BCR-dependent interactions with the FCRL5 cytoplasmic tail also were confirmed in PEC B1 B cells derived from T-cell leukemia/lymphoma 1 (TCL1) Tg mice that have a marked expansion of these cells (Fig. S8) (44).

Fig. 6.

FCRL5 function in innate-like B cells correlates directly with SHP-1 activity. (A) MZ B cells (5 × 10⁶) sorted from hBAFF Tg mice were washed in serum-free RPMI medium 1640 and rested for 2 h. Cells were then lysed directly or stained with biotinylated F(ab’)₂ mAb fragments specific for IgM (SB73a) and FCRL5 (3B7) or an isotype control and were crosslinked with SA for 10 min at 37 °C. WCL or anti-FCRL5 (9D10) immunoprecipitates were immunoblotted for the indicated proteins. Blots were stripped and reprobed with anti-FCRL5 (5-3B2) as a loading control. (B) Whole-cell protein tyrosine phosphorylation in the specified B-cell subsets, discriminated by surface staining WT splenocytes with anti-CD19, CD21, and CD23 or PEC leukocytes with anti-CD19, CD5, and CD11b, was determined at homeostasis, after stimulation with a biotinylated F(ab’)₂ rat anti-mouse IgM mAb (SB73a) crosslinked with SA or after pervanadate treatment for 10 min. Cells were stained intracellularly with anti-pTyr or an isotype control mAb before analysis by flow cytometry. (C) Quantitative comparisons of constitutive pTyr among the indicated B-cell subsets. (D) Relative whole-cell protein pTyr in response to anti-IgM stimulation. (E) Comparison of basal SHP-1 abundance determined by intracellular staining. (F) Whole-cell protein pTyr in response to pervanadate treatment. (G) Assessment of total as well as active (Y397) or inactive (Y508) states of Lyn analyzed by intracellular staining with phospho-specific antibodies. Data in C–G are expressed as mean ± SEM for three independent experiments, and isotype control staining represents combined analysis from the six subsets. *P < 0.05; **P < 0.01; ***P < 0.001.

To determine the nature of FCRL5’s disparate function in MZ and B1 B cells, we compared the signaling properties of these subsets. Previous biochemical studies had found that MZ B cells express comparatively more SHP-1 than FO B cells but express similar levels of total Lyn (12). Furthermore, splenic MZ and PEC B1 cells show higher constitutive whole-cell tyrosine phosphorylation than the conventional B2 cells present in their respective microenvironments (12, 45). To examine compartment-specific differences in these variables among these cell types directly, we used a flow cytometry-based analysis for quantitative comparisons at baseline and in response to different stimuli (Fig. 6B). Consistent with previous work, both innate-like B-cell subpopulations had significantly higher constitutive and BCR-induced tyrosine phosphorylation relative to their B2-lineage counterparts; however, activation was greatest in MZ B cells (Fig. 6 C and D). Correspondingly, basal SHP-1 expression was significantly elevated in MZ and B1 B cells compared with conventional B2 cells but was twofold greater in MZ than in B1a or B1b B cells (Fig. 6E). Relative SHP-1 abundance also correlated strongly with global tyrosine phosphorylation in response to pervanadate treatment (Fig. 6F). Although total Lyn expression was similar among these subsets, its active and inactive forms were slightly higher in PEC B1 than in B2 cells (Fig. 6G). Along with our findings in mutant mice defining the requisite contributions of Lyn and SHP-1 to FCRL5's function, these data indicate that divergent FCRL5 biology in MZ and B1 B cells correlates directly with SHP-1 activity.

FCRL5 Binary Immunoregulation Discloses the Contributions of SHP-1 and Lyn to MZ and B1 BCR-Induced Apoptosis.

Because BCR stimulation markedly induces apoptosis in MZ B cells but has more modest effects on B1 B-cell viability (13, 16, 46), we next investigated the differential influence of Lyn and SHP-1 on antigen receptor-mediated survival. Following culture alone, the survival of WT MZ B cells was much shorter than that of B1 B cells purified from WT, me^v/me^v, or Lyn^−/− mice (Fig. S9). Although SHP-1–mutant B1 B cells were more sensitive to spontaneous apoptosis than WT B1 B cells, Lyn^−/− B-cell viability was higher overall (Fig. 7). Thus, in the absence of stimulation, Lyn deficiency favored B1 B-cell survival, and SHP-1 deficiency promoted apoptosis. We then examined the impact of FCRL5 on BCR-mediated survival. Anti-IgM crosslinking markedly increased MZ B-cell apoptosis; however, consistent with the relatively elevated SHP-1 activity in this subset, FCRL5 coengagement rescued a significant number of these cells. BCR ligation also enhanced apoptosis in WT PEC B1 B cells. But, in contrast to its failure to modulate calcium or tyrosine phosphorylation, here FCRL5 augmented apoptosis. This effect was in line with the comparatively lower SHP-1 levels in these cells and was magnified in me^v/me^v B1 B cells. Conversely, in the absence of Lyn, FCRL5 demonstrated regulation akin to that of MZ B cells and suppressed apoptosis. These findings collectively indicate that compartmental differences in SHP-1 and Lyn activity play a key role in influencing the differential adaptive function of innate-like MZ and B1 B cells.

Fig. 7.

FCRL5 reveals opposing roles for SHP-1 and Lyn in MZ- and B1 BCR-mediated survival. Splenic MZ or PEC B1 B cells (1 × 10⁵) purified from the indicated mice were stained with biotinylated F(ab’)₂ mAbs specific for IgM (SB73a) and FCRL5 (3B7) or an isotype control, crosslinked with SA, and plated for culture in triplicate. Apoptosis was determined for MZ B cells at 12 h and for B1 B cells at 24 h by annexin V and propidium iodide staining. Total apoptotic cells include both the early (annexin V⁺PI⁻) and late fractions (annexin V⁺PI⁺). Data are shown as mean ± SEM for three independent experiments. **P < 0.01; ***P < 0.001.

Discussion

In these studies a distinguishing marker of innate-like B cells was found to exert compartment-specific regulatory function. Although FCRL5 could inhibit the striking BCR activation typical of splenic MZ B cells, it had no obvious influence on this cascade which is muted in PEC B1 B cells. These findings led to broader questions concerning the molecular basis for FCRL5 function and its significance in these specialized subsets. Although it failed to initiate downstream responses by itself, FCRL5 as a coreceptor had a unique dualistic impact on BCR signaling that derived from its tyrosine-based recruitment of Lyn and SHP-1. The incongruent biology evident for FCRL5 as well as the BCR in MZ versus B1 B cells correlated directly with the distinct activity of SHP-1 in these subsets. Collectively these findings demonstrate that FCRL5 is a discrete counterregulatory biomarker of innate-like B cells directly coupled to the Lyn–SHP-1 biochemical circuit.

Through their direct and indirect associations with the BCR complex, Lyn and SHP-1 have profound effects on B-cell selection and function (47–49). Deficiency of either of these proteins leads to a loss of MZ B cells, expansion of the B1 lineage, and a breakdown in peripheral tolerance (33, 35, 50, 51). We found that basal calcium influx, whole-cell tyrosine phosphorylation, and SHP-1 levels were significantly higher in MZ B cells than in B1 B cells but that Lyn activity did not vary substantially. These differences indicate that splenic MZ B cells are more globally activated at homeostasis and thus are primed for stronger signal transduction once triggered. Given its governing effects on BCR signaling strength (52), the effector responsibilities of these cells, and their developmental loss in its absence, SHP-1 likely is up-regulated to raise the triggering threshold and offset the preactivation induced by tonic BCR and other environmental signals that these lymphocytes experience in the splenic MZ niche (7, 53). Dampening their preamplified resting state by means of elevated phosphatase activity not only would suspend BCR activation but also would heighten their overall potential for explosive responsiveness once this restriction is lifted. The relationship between SHP-1 expression, preactivation status, and BCR responsivity also was apparent in the other splenic subpopulations, and recent work by the DeFranco group has demonstrated that increased signaling sensitivity among transitional B2 cells declines in a maturation-dependent fashion as the Lyn–SHP-1 pathway becomes active (54). As a distinctive bifunctional substrate in this circuit, FCRL5 inhibitory function in MZ B cells reflects more elevated SHP-1 than Lyn activity. A pivotal role for SHP-1 was also substantiated by findings that neither total Lyn abundance nor its physiologic state differed dramatically among splenic subsets. Thus compensatory SHP-1 up-regulation appears critical for buffering MZ B-cell excitation and fate. Although what modulates its expression remains unclear, factors originating from the MZ milieu likely contribute given the extreme sensitivity of MZ B cells to spontaneous apoptosis ex vivo and the marked elevation of SHP-1 in this subset.

B1 B cells were less constitutively activated than splenic MZ B cells; however, their global tyrosine phosphorylation and SHP-1 and Lyn activity were slightly higher than in PEC B2 cells. These features, along with their attenuated BCR responsiveness, again highlight their unconventional biology. The inverse functionality of FCRL5 in me^v/me^v and Lyn^−/− B1 B cells but relative indolence in WT cells indicates the SHP-1-Lyn circuit is more balanced in this subset at homeostasis. Under these conditions, Lyn binding to the FCRL5 ITAM-like Y543 residue could be stoichiometrically compensated by SHP-1 recruitment to the FCRL5 Y566 ITIM. Importantly, lower SHP-1 activity would yield relatively higher Lyn kinase function. Elevated Lyn activity was apparent in the augmentation of FCRL5-induced apoptosis in me^v/me^v as compared with WT lymphocytes. Diminished SHP-1 activity also might be related to elevated basal levels of phosphorylated ERK (55, 56). Despite its enigmatic source of stimulation in B1 B cells, ERK activation in response to agonistic T-cell receptor ligation can modify SFK activity in a feedback loop that blocks SHP-1 function (57). Perhaps additional signals native to the coelomic cavities trigger Toll-like and/or cytokine receptors and commandeer B1 B-cell function. Sustained BCR triggering in combination with these stimuli may down-modulate adaptive signaling as well as SHP-1 expression, leading to their exhaustion, as seen for T cells in chronic inflammatory states (58).

The possession of both functional ITAM-like and ITIM sequences distinguishes FCRL5 from other coreceptors. These experiments showed that CD5, CD22, CD32, and CD72 could each inhibit BCR-mediated calcium mobilization in WT and Lyn-deficient B1 B cells but, unlike FCRL5, failed to acquire activation properties in the absence of SHP-1. Relative to FCRL5’s balanced function in WT cells, me^v/me^v B1 B cells validated the bimodal qualities of FCRL5 and also indirectly exposed Lyn’s preferential affinity for the ITAM-like Y543 residue. The lack of SFK redundancy at this location is suggestive of sequence-specific characteristics that distinctly accommodate Lyn there. Although the molecular features conferring SFK tyrosine phosphorylation at the Y566 ITIM appear more indiscriminate, phosphatase docking was restricted to SHP-1. Thus, although Lyn and SHP-1 can concomitantly associate with FCRL5, their relative cytosolic activity balances this receptor’s overall function. As a scaffold that directs their tandem positioning, it is possible that Lyn occupancy at Y543 could promote processive phosphorylation at the Y566 ITIM (59). Lyn-dependent transphosphorylation, which has been shown for CD19, would maintain FCRL5 binary function at balance unless quantitative differences or other intracellular factors compromised Lyn and/or SHP-1 recruitment to these respective sites (60). Conversely, SHP-1 conceivably could interfere intramolecularly with Lyn’s activity. Lyn also imbues FCRL5 with properties analogous to other activation receptors such as Igα, Igβ, CD19, and CD180/RP105 (8). When altered or untethered from SHP-1 countersuppression, this Lyn-driven cascade potentially could fuel FCRL5-mediated B-cell pathogenesis.

Very recent studies by the Colonna group identified that human FCRL4 and FCRL5 bind heat-aggregated IgA and IgG (19), but the endogenous ligands for other B-cell–expressed FCRLs, including mouse FCRL5, have yet to be elucidated. Notably, two additional members interact with MHC-like proteins. In humans FCRL6, expressed by cytotoxic CD8⁺ T and natural killer cells, interacts with MHC class II; whereas in mice an immunoevasin termed “orthopoxvirus MHC class I-like protein” was identified as an FCRL5 ligand (20, 21). The consequences of these interactions are under investigation; however, their associations have intriguing implications for these receptors in mechanisms of tolerance induction. Furthermore, the signaling characteristics defined here for FCRL5 have relevance for other FCRL proteins. Thus far these molecules have been found to inhibit BCR signaling generally, but their conserved cytoplasmic motifs along with evidence from mutagenesis studies has revealed subtle hints of their possible bifunctionality (27, 28). Thus, unearthing the FCRL5–Lyn relationship raises the likelihood that other FCRLs possess similar activation features.

One underlying question relates to the integration of FCRL5 signaling with its discrete expression by MZ and B1 B cells. Given their front-line access to potential pathogens, a hallmark aspect of these lymphocytes is their preferential responsiveness to innate agonists and T-cell–independent antigenic stimulation. Accordingly, recent work demonstrated fascinating regulatory potential for FCRL4 in human B cells. Despite its possession of a tyrosine-based switch motif and two ITIMs that inhibit adaptive BCR function via SHP-1 and SHP-2 binding (29), Sohn et al. have shown that FCRL4 also can enhance innate Toll-like receptor 9-dependent signaling (61). Its variable influence on these pathways was transmitted in the absence of direct FCRL4 engagement, indicating that its expression alone may modulate B-cell responsivity differentially to adaptive or innate stimuli. The identification of these features in another FCRL, along with the ability of FCRL5 to recruit the active form of Lyn shown here, strongly indicates that these molecules are important modifiers of B-cell adaptive and innate signaling. Moreover, in addition to its constitutive expression by MZ and B1 B cells, FCRL5’s role as an innate regulator also is endorsed by its detection on subsets of B2 B cells that vary among mice and its sensitivity to induction by innate stimuli such as LPS (31). Importantly, Lyn also is required for CD180/RP-105 TLR activation in response to LPS binding (62). Thus, future studies of mouse deficiency models should better clarify FCRL5’s roles in these two facets of B-cell biology.

Finally, our clearer understanding of FCRL5 function provides insight into the possible pathogenic importance of these molecules in disorders with which they already have been widely associated. For example, FCRL2 and FCRL3 are expressed by a putative MZ effector memory population that circulates in human blood as well as by chronic lymphocytic leukemia B cells that are believed to originate from this subset (22, 63). In addition to this lymphoproliferative disorder, FCRL3 has been linked to multiple autoimmune diseases including rheumatoid arthritis (25), and FCRL4 is associated with dysfunctional humoral responses found in HIV- and malaria-infected patients (24, 64). These observations underscore the probability that FCRL proteins participate in perturbed immunity. The current findings therefore provide insight into the fundamental regulatory contributions of the FCRL family in normal and pathogenic B-cell function.

Materials and Methods

Details of the mice strains used and their maintenance are described in SI Materials and Methods. Similarly, the construction of chimeric receptors, the retroviral transduction methodology, and the cell lines and antibodies used are outlined there. Also included in SI Materials and Methods are the protocols for immunoblotting, coimmunoprecipitation, calcium mobilization, intracellular-staining, phospho-flow analysis, histologic staining, confocal microscopic imaging, apoptosis assays, and statistical analysis.