Fc Receptor-Like A Associates with Intracellular IgG and IgM but Is Dispensable for Antigen-Specific Immune Responses

FcRs bind Igs with a range of specificities and affinities and can either activate or inhibit cellular responses. They assist essential effector functions of Abs by mediating Ab-dependent cell-mediated cytotoxicity, cellular activation, recognition and capture of opsonized pathogens, uptake of Ag for presentation to T cells, and regulation of B cell responses (1). Additionally, some FcRs mediate specialized functions for Ig transport and recycling, including the polymeric Ig receptor and the neonatal FcR (2).

FcR-like (FcRL) proteins were identified as a family of cell surface receptors with homology to FcγRI in one or more domains and are expressed differentially on B cells at various stages of differentiation (3–7). More recently, an additional member of the family (FcRL6) was found primarily on the surfaces of NK cells and cytotoxic T cells (8, 9). FcRL1–6 are type I transmembrane proteins of the Ig superfamily, and all have one or more canonical ITAMs or ITIMs in their cytoplasmic tails. The ability to recruit tyrosine kinases or protein tyrosine phosphatases has been reported for all of these receptors (8–15), with bona fide activating potential found for FcRL1 (12) and potent inhibitory potential found for FcRL4 and FcRL5 (10, 11). This potential to regulate B cell responses appears to be conserved in mouse FcR homolog 3/FcRL5 (16).

The evolutionary divergence of most FcRL family members from humans to rodents has been extensive, with significant structural changes or gene deletions affecting all of the cell surface FcRL proteins (9, 17–21). However, two members of the FcRL family, FcRLA and FcRLB, are highly conserved and share unique features; 1) they have no transmembrane domain and hence are expressed as intracellular receptors; 2) both have a C terminus containing a proline-rich stalk region followed by a leucine-rich coiled-coil motif. Their differences lie in their N-terminal domain structures and expression patterns. Whereas FcRLB has Ig-like domains with homology to all three domains of FcγRI, FcRLA contains Ig-like domains homologous to only the second and third domains of FcγR1, preceded by an N-terminal domain rich in acidic residues as well as potentially unpaired cysteines. In addition, FcRLB seems to be expressed in only a tiny fraction of nondividing B cells, whereas FcRLA is expressed at low levels in mantle zone B cells and at higher levels within germinal centers and splenic marginal zone B cells in humans (13, 22–27).

Currently, the ligands for the ectodomains of the human FcRL proteins are unknown, although anecdotal reports in the literature have implicated Ig binding by FcRL4 and FcRL5 (5, 28). In this article, we demonstrate the clear association of FcRLA with intracellular Ig, illustrate the expression of FcRLA within mouse B cells, and examine the consequence of FcRLA deficiency on humoral immune responses in vivo.

Fcrla^−/− mice were generated in E14.1 (129P2/OlaHsd) embryonic stem cells by targeted disruption and replacement of the third and fourth exons of the Fcrla gene with an MC1-neo^r expression cassette. Chimeras resulting from the injection of targeted embryonic stem cells into C57BL/6 blastocysts were bred to C57BL/6 transgenic mice expressing Cre recombinase under the control of the CMV promoter to excise the MC1-neo^r cassette. Heterozygous Fcrla^−/+ mice were intercrossed, and the resulting homozygous Fcrla^+/+ and Fcrla^−/− progeny were used for analysis of humoral immune responses in vivo. All of the animal procedures in this study have been reviewed and approved by the Washington University Animal Care and Use Committee.

Rabbit antisera specific for human and mouse FcRLA were raised by Pacific Immunology Company (San Diego, CA). Peptides SGHQKPGTTKATAE for human FcRLA and SVYLKPGTTKVADK for mouse FcRLA were conjugated to keyhole limpet hemocyanin (KLH) and immunized according to established protocols. Rabbit antisera (numbers 839 and 840 for human and 1391 and 1392 for mouse) were stored in 50% glycerol at −80°C for preservation and −20°C for working stocks. Purified mAbs specific for human FcRLA have been described previously (24). For immunohistological assessment of lymphoid architecture and flow cytometric analysis of lymphocyte development and FcRLA expression, the following Abs were used. Anti-mouse CD3ε-FITC, CD4-PE, CD8-biotin, IgD-FITC, IgM-PE, CD23-PE, CD21-FITC, B220-biotin, CD11b-FITC, CD5-FITC, and streptavidin-allophycocyanin were obtained from BD Pharmingen (San Diego, CA). Biotinylated peanut agglutinin was obtained from Vector Laboratories (Burlingame, CA) and alkaline phosphatase-conjugated anti-rat IgG was obtained from Southern Biotechnology Associates (Birmingham, AL).

Abs used for ELISAs to quantify total mouse serum Ig or Ag-specific Ig were as follows. Purified anti-IgM, anti-IgG2b, anti-IgG3, anti-IgA, and anti-IgE were obtained from BD Pharmingen along with biotinylated Abs for detection of IgG3 and IgA. Purified anti-IgG1 and anti-IgG2a were obtained from Southern Biotechnology Associates along with HRP-conjugated Abs specific for mouse Ig-κ, Ig-λ, IgM, IgG1, IgG2a, IgG2c, and IgE as well as biotinylated anti-IgG2b and streptavidin-HRP. Purified anti-IgG2c was obtained from Bethyl Laboratories (Montgomery, TX).

The expression pattern of mouse FcRLA within spleen was determined by immunohistochemistry. Wild-type C57BL/6 and Fcrla^−/− mice were immunized with 50 μg nitrophenyl (NP)-KLH (Biosearch Technologies, Novato, CA), and organs were harvested 10 d later and frozen in OCT compound. Six-micrometer cryosections were stained with the indicated Abs in PBS/0.1% Tween 20/5% goat serum after a 10-min pretreatment with 0.3% hydrogen peroxide to quench endogenous peroxidase activity. Sections were stained with rabbit anti-mouse FcRLA and anti-mouse CD3-biotin. Secondary staining was performed with goat anti-rabbit IgG-alkaline phosphatase and streptavidin-HRP. Blue and red color development was obtained using Vector Blue Alkaline Phosphatase Substrate Kit and Vector Red HRP Substrate Kit (Vector Laboratories). For visualization of mouse splenic architecture in Fcrla^−/− mice, spleens from mice immunized 2 wk previously with 2 × 10⁸ sheep RBCs were harvested and frozen in OCT. Eight-micrometer cryosections were cut and stained with anti-mouse CD21-FITC and anti-mouse CD23-PE (BD Pharmingen).

The presence of soluble FcRLA in cell lysates and supernatants was measured by ELISA. 293T cells were transfected using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) with cDNA encoding FcRLA within the mammalian expression vector pcDNA3.1. MaxiSorp ELISA plates (Nunc, Naperville, IL) were coated with purified monoclonal anti-human FcRLA (24) and blocked with PBS/Tween 20/1% BSA. Supernatants and Triton X-100 lysates were collected from BJAB cells and FcRLA-transfected 293 cells and were incubated in Ab-coated wells for 1 h at 25°C. Bound FcRLA was detected with rabbit anti-human FcRLA (serum 840) followed by goat anti-rabbit IgG-HRP.

For analysis of mouse lymphocyte development, cells were harvested from the indicated organs and stained with anti–CD3-FITC, anti–CD4-PE, anti–CD8-biotin, anti–CD21-FITC, anti–CD23-PE, anti–B220-biotin, anti–CD11b-FITC, anti–CD5-PE, anti–IgD-FITC, anti–IgM-PE, and anti–CD138-PE (BD Pharmingen). Negative gating of dead cells was performed using propidium iodide (Sigma-Aldrich, St. Louis, MO)

Expression of FcRLA was measured by intracellular flow cytometry. Cells were fixed in 2% paraformaldehyde for 10 min, washed, and then permeabilized for 20 min in PBS/0.5% saponin/3% bovine calf serum. FcRLA was stained using a 1:3000 dilution of rabbit antiserum 1392 in saponin buffer. Secondary detection was performed using 2 μg/ml goat anti-rabbit IgG-Alexa Fluor 647 (Invitrogen) in saponin buffer. Cells then were washed sequentially in saponin buffer and FACS buffer (PBS/3% bovine calf serum) and visualized using a BD FACSCalibur flow cytometer (BD Biosciences, San Jose, CA). In the spleen, identification of marginal zone and follicular B cell subsets was performed using anti–CD21-FITC and anti–CD23-PE. Germinal center B cells were identified using anti–Fas-FITC and anti–B220-PE (BD Pharmingen). Cell subsets in the peritoneum were identified by anti–CD11b-FITC and anti–CD5-PE staining.

Cells were pretreated with 10 mM methyl-methanethiosulfonate for 10 min on ice to block free sulfhydryl groups and prevent protein aggregation and disulfide exchange. After treatment, cells were lysed in PBS containing 1% Triton X-100, 10 mM methyl-methanethiosulfonate, and Sigma Protease Inhibitor Cocktail (P8340). Soluble lysates were precleared on either agarose (for CESS cells) or streptavidin (for Daudi and BJAB cells) beads for 1 h at 4°C prior to immunoprecipitations (IPs). For IP from CESS cells, lysates were incubated with protein G-, protein A-, protein L-, or streptavidin-coated agarose beads (Thermo Scientific Pierce, Rockford, IL). As a positive control for FcRLA IP, monoclonal anti-FcRLA (24) on protein G beads was used. For IP of IgM from Daudi and BJAB cell lines, streptavidin-coupled agarose beads were precoated with biotinylated F(ab′)₂ anti-human IgM, anti-κ, or anti-Λ (BD Pharmingen). Western blotting for FcRLA was performed using rabbit anti-FcRLA antiserum 840 and goat anti-rabbit IgG-HRP.

For determination of the effect of FcRLA deficiency on the development of humoral immune responses, mice were immunized by i.p. injection of Ag at the doses and time points indicated in the 8Results section. For consistency and to reduce injection errors, all of the Ags were administered in a total volume of 200 μl in sterile PBS. For assays using sheep RBCs (SRBCs) as an Ag, SRBCs were freshly prepared from whole sheep blood (Colorado Serum Company, Denver, CO). Measurement of total serum Igs was performed by sandwich ELISA. Measurement of Ag-specific Ig was performed using direct ELISA by coating wells with 100 μl of a 5 μg/ml Ag solution. For measurement of anti-SRBC Abs, plates were coated to confluency with glutaraldehyde-fixed SRBCs. Endogenous peroxidase was quenched with 0.3% hydrogen peroxide prior to incubation with diluted mouse sera. The substrate used for all of the ELISAs was 1 mg/ml ortho-phenylenediamine dihydrochloride (Sigma-Aldrich). Reporting of ELISA data occurs in one of three ways: mean absorbance at 490 nm; by titers defined as the mean absorbance plus or minus three standard deviations of the last three (negative) wells in the titration; and relative units derived from direct comparison of the experimental serum with a hyperimmune serum (designated as containing 1,000,000 U) generated by immunizing mice three times with 100 μg NP-OVA in Freund’s adjuvant (1× CFA plus 2× IFA). For experimental results reported in relative units or mean absorbance at 490 nm, experimental sera were first titrated to determine the optimum dilution at which to measure relative Ig responses.

Measurement of IgG1-associated κ-L chains was performed by indirect ELISA on serum from mice immunized twice with SRBCs. Plates were coated with polyclonal goat anti-mouse IgG1. Ab detection was performed using HRP-conjugated anti-mouse IgG1 or anti-mouse κ-chain (Southern Biotechnology Associates). The ability of Abs from Fcrla^−/− mice to activate complement through the classical pathway was analyzed by the following method. Sera from SRBC-immunized mice were titrated and mixed with purified SRBCs in HBSS in 96-well plates. After 1 h, sensitized RBCs were washed once, then resuspended in guinea pig complement (Colorado Serum Company) diluted 1:20 in HBSS, and incubated for 30 min at 37°C. Unlysed RBCs were pelleted by centrifugation, and the supernatant was transferred to 96-well flat-bottom plates for reading of OD at 540 nm. The lytic activity of the mouse sera was defined as the maximum dilution of mouse serum capable of activating complement-mediated lysis. Mann-Whitney tests for differences in medians were performed using Prism software (GraphPad, San Diego, CA) for each of the experiments above.

The lack of a transmembrane domain but presence of a leader peptide suggests that FcRLA may be secreted. Although we failed previously to find evidence for FcRLA secretion (24), intracellular FACS analyses indicated that the Abs used had potentially overlapping epitopes (data not shown). Therefore, any soluble FcRLA may have been undetectable in this system. Using our newly developed rabbit antisera specific for the C terminus of the protein along with the previously reported monoclonal reagents specific for the Ig-like domains of FcRLA, we tested for FcRLA secretion by B cells and transfected 293T cells using an indirect ELISA. Very little FcRLA was detectable in supernatants from BJAB cells or FcRLA-transfected 293 cells. In contrast, large amounts of FcRLA were readily detectable in lysates from both BJAB cells and FcRLA-transfected 293 cells (Fig. 1A). These data confirm that FcRLA is not secreted but is retained as an intracellular receptor.

Because FcRLA is most highly homologous to FcγRI, it seems likely that it also can bind Igs. To determine whether FcRLA is an FcR, we previously generated a chimeric fusion protein of FcRLA with CD4 to target the protein to the cell surface. However, the transfected cells did not bind heat-aggregated IgG (24). Subsequent attempts using the expression vector pDisplay also failed to detect any IgG binding. However, it should be noted that neither full-length FcRLA nor a shorter form expressing only the acidic N-terminal domain plus the two Ig-like domains was expressed successfully on the cell surface, even in the vector pDisplay. This suggests that rFcRLA either contains sequence motifs that prevent surface expression or is unstable as either a cell surface protein or an overexpressed protein in 293T cells. Given this, it seems possible that forced cell surface expression of FcRLA could preclude functional binding activity.

For these reasons, it became necessary to determine if FcRLA bound to Ig when expressed as an intracellular protein at endogenous levels. To directly test if FcRLA binds to IgG within the cell, Igs were immunoprecipitated from CESS (a human B cell line expressing IgG-λ) cell lysates using agarose beads conjugated to either protein G or protein A. Immunoblotting for either human IgG or FcRLA revealed that FcRLA coprecipitated with IgG (Fig. 1B). To determine whether FcRLA associates with other Ig isotypes, we performed further IPs from IgM-κ–expressing BJAB and Daudi cells. FcRLA coprecipitated with IgM when using biotinylated F(ab′)₂ fragments specific for either IgM or κ-L chain but not λ-L chain, indicating a specific interaction with endogenous IgM within these cells (Fig. 1C). These data establish the association of FcRLA with intracellular Igs.

To understand how FcRLA binding to intracellular Igs might influence the development of humoral immune responses, Fcrla^−/− mice were generated by targeted disruption of the third and fourth exons of the Fcrla gene in E14.1 embryonic stem cells (Fig. 2A). Fcrla^−/− and wild-type control mice were obtained from F2–F4 intercrosses of Fcrla^−/+ mice.

Prior to analyzing the response of Fcrla^−/− mice to Ag challenge, we first examined basic lymphoid development and differentiation. The major T cell subsets were present in normal frequencies and numbers in the thymus of FcRLA knockout mice (Fig. 2B). Within the bone marrow, development of B cells, as visualized with the markers B220 and IgM, did not appear to be inhibited in any way. All of the major B lineage populations, including B220⁺IgM⁻ pro-B cells, B220⁺IgM^high pre-B cells/immature cells, and B220^highIgM^low mature B cells were present. No abnormal accumulation of cells was apparent at any developmental stage in Fcrla^−/− mice, and the minor differences observed were not consistent from one animal to the next (Fig. 2B). Among splenocytes, the relative proportion of B and T lymphocytes was equivalent in Fcrla^−/− and wild-type mice. Staining B cells with the markers B220, IgM, and IgD indicated intact development of B cells from immature through transitional stages and finally to mature cells (Fig. 2B). The presence of equivalent numbers of CD21^highCD23^low marginal zone B cells also was observed in naive mice (data not shown).

Histological analysis of lymphoid tissues was performed to verify the structural integrity of central lymphoid organs, with specific regards to B cell differentiation. After immunization with SRBCs, no major disruptions in lymphoid organ structure or cellular distribution were discovered in the Fcrla^−/− mice. Organization and segregation of white pulp in the spleen, the organization of follicular dendritic cells, and germinal center morphology were visualized using Abs to IgD, CD3, CD21, CD23, CD35, or biotinylated peanut agglutinin (Fig. 2C and data not shown). No gross abnormalities were observed in Fcrla^−/− mice.

To determine expression of FcRLA in mice, rabbit antisera were raised against a peptide representing the last 15 aa of mouse FcRLA. Cryosections from spleens and lymph nodes of control and Fcrla^−/− mice were stained with anti-FcRLA antisera and counterstained for CD3 to show T cell zones. We found that FcRLA is localized within the entire B cell compartment, including mantle and marginal zones. In a departure from previously reported expression patterns of FcRLA in humans (24, 25), mouse FcRLA appeared to be somewhat downregulated within germinal centers (Fig. 3A). No FcRLA expression was evident by immunohistochemistry in nonlymphoid tissues from which FcRLA transcript had been detected previously by Northern blotting, including kidney and liver (data not shown).

In addition to histological analysis, intracellular FcRLA expression was assessed using the anti-FcRLA antiserum for flow cytometry. Among splenocytes, all of the B220⁺ cells were found to express FcRLA, including follicular, marginal zone, and germinal center B cells (Fig. 3B). Within the peritoneum, all of the populations of B cells, CD11a^lowCD5⁺ B1a, CD11a^lowCD5⁻ B1b, and CD11a⁻CD5⁻ B2, expressed FcRLA at comparable levels. CD5^high T cells and CD11b^high macrophages did not express FcRLA (Fig. 3C). In the bone marrow, all of the B lineage cells expressed FcRLA, beginning at the immature B cell stage (data not shown). In summary, all of the circulating peripheral B cell subsets express FcRLA throughout their differentiation, and non-B cells do not express FcRLA within any lymphoid organ.

Having found FcRLA expression within all of the B cells and knowing that Fcrla^−/− mice did not have any developmental defects, we set out to analyze B cell responses to both T-dependent and T-independent Ag challenge. To analyze the B cell response to particulate Ags, mice were immunized i.p. with 2 × 10⁸ SRBCs, and sera were obtained at days 14 and 21 to monitor the development of the immune response. Five weeks after initial immunization, mice were given a secondary challenge of 2 × 10⁸ SRBCs, and serum was collected 7 d later. Minimal or no difference between wild-type and Fcrla^−/− mice was noted after quantifying anti-SRBC Abs that developed during the primary response. However, after secondary challenge with 2 × 10⁸ SRBCs, the Fcrla^−/− mice had significantly (p < 0.01) higher levels of anti-SRBC IgG1 (Fig. 4A). In addition, total IgG1 was elevated consistently in these mice after secondary challenge (data not shown). These data may indicate that FcRLA is necessary for regulating responses to particulate Ags.

To measure Ab responses to a different T-dependent Ag, mice were immunized i.p. with 50 μg NP₁₂-KLH. Sera were collected to measure the primary response at 14 and 21 d postimmunization. Mice were boosted again after 5 wk. In this case, responses to different doses of Ag were assayed. Mice were boosted with either 5 or 50 μg NP₁₂-KLH. In these experiments, although Fcrla^−/− mice seemed to lag slightly behind wild-type mice in development of IgG1 responses, mice had equivalent levels of KLH-specific IgG1 by 21 d postimmunization (Fig. 4B). After secondary Ag challenge, no difference in IgG1 responses was seen, regardless of whether a high or low Ag dose was given (Fig. 4C). No significant differences were found for other isotypes, including IgM responses at any time point (data not shown).

Next, the ability of Fcrla^−/− mice to form adequate long-term memory responses was tested. Mice were administered 50 μg NP₁₂-KLH i.p. and allowed to rest for 5.5 mo without additional intervention. At that time, sera were collected, and KLH-specific IgM and IgG1 were measured. Although KLH-specific IgG was weakly detectable in these mice, no consistent difference in the two cohorts of mice was seen between wild-type and Fcrla^−/− mice (data not shown). Six months after the primary immunization, mice were given a secondary immunization of 50 μg NP-KLH, and after 7 d, serum was collected and Ab responses were measured. Wild-type mice were found to have significantly (p < 0.05) higher titers of NP-KLH–specific IgM, although this level of significance was not maintained when measuring the high-affinity NP-specific Abs (Fig. 4D), and it is possible that this difference was a result of the mixed genetic background of the mice and not indicative of a need for FcRLA in the development of memory IgM responses. Similarly, no significant difference in KLH- or NP-specific IgG produced during the memory response was found between wild-type and Fcrla^−/− mice (Fig. 4E).

Although T-dependent responses are characterized by the ability to develop high-affinity Abs, sterilizing immunity, and long-term memory, many primary infections are kept under control or eliminated by the rapid generation of T-independent Ab responses. To analyze T-independent type II responses in the absence of FcRLA, we immunized Fcrla^−/− mice and wild-type controls with 50 μg NP-Ficoll. Sera were collected on days 5 and 7 after immunization, and NP-Ficoll–specific IgM, IgG1, and IgG3 were measured. Although Fcrla^−/− mice seemed to develop serum Ab more rapidly than wild-type controls, little difference was seen in the presence of any of these isotypes by 7 d postimmunization, indicating that T-independent responses to Ag were also intact (Fig. 5).

Because FcRLA was found to bind intracellular Igs (Fig. 1) and a previous yeast two-hybrid screen had implicated a potential interaction between FcRLA and Ig H and L chains (T. Wilson and M. Colonna, unpublished observations), we hypothesized that FcRLA could act as a molecular chaperone for Ab assembly. However, because the concentration of total and Ag-specific Abs was rarely different between Fcrla^−/− mice and wild-type controls, it seemed necessary to assess the functional quality of the Ab generated in these mice. To confirm that Fcrla^−/− B cells were secreting fully intact Ab, we measured amount of IgG-associated κ-L chains in relation to the total amount of IgG H chain in the serum using indirect ELISA. The relative frequency of κ-chains detected was proportional to the observed amount of total IgG1 H chain in the two population samples, indicating that intact Ab was being generated in these mice (Fig. 6A).

In addition to Ig assembly, the ability of the secreted Abs to activate the complement cascade was measured. Serum from SRBC-immunized mice was titrated and used to opsonize SRBCs prior to incubation with guinea pig complement. The lytic activity of the Abs was defined as the largest dilution of Ab capable of activating complement-mediated lysis of RBCs. The lytic activity of Abs produced in control and Fcrla^−/− mice (Fig. 6B) was found to be consistent with the relative concentration of SRBC-specific Abs measured earlier (Fig. 4A). In conclusion, Fcrla^−/− mice produce intact and functionally active Ab in response to Ag challenge.

In this study, we have examined the expression patterns and function of FcRLA in B cell responses. We have shown for the first time, at the protein level, that FcRLA can be found in all of the peripheral B cells in mice and is completely absent in cells that do not express the marker B220. The strong baseline expression in follicular B cells and diminished expression of FcRLA in mouse germinal center B cells differs somewhat from the pattern observed in humans, where FcRLA expression is the highest in germinal centers and splenic marginal zone but only weakly expressed by mantle zone B cells. This difference between mouse and human expression may implicate a difference in functional requirements of B cells between the two species. However, it is possible that even a low level of FcRLA expression is sufficient for function and that the differential expression among B cell subsets and between species is simply coincident with promoter activity.

Through generation of additional polyclonal reagents specific for human and mouse FcRLA, we have been able to confirm that FcRLA is an intracellular receptor. Full-length receptor is not secreted from either human B cell lines or FcRLA-transfected fibroblasts. Previous studies have failed to demonstrate binding of surface-targeted FcRLA to extracellular Igs, and further efforts on our part have confirmed the lack of association. However, IP of Igs from human B cell lines resulted in the coprecipitation of FcRLA with both IgG and IgM, indicating that this protein binds Igs before they are expressed on the surface or secreted. This implicates FcRLA as a potential chaperone for the generation of intact Igs by B cells. It is interesting to note that FcRLA was recruited by both Ig μ- and Ig γ-chain–containing Igs, indicating that it may be important for the production of Abs, regardless of B cell isotype differentiation.

It is not clear why FcRLA is capable of binding to Igs within the cell but exhibits no Ab binding ability when expressed on the cell surface. It is possible that FcRLA is unstable when expressed on the surface, due to either pH conditions or the lack of a stabilizing accessory molecule, and that it quickly loses biological activity in these conditions. However, it is intriguing to consider that FcRLA may bind preferentially an immature form of Ig and may assist in the final assembly of Abs prior to extracellular expression. For example, FcRLA may show a preference for unpaired or incorrectly paired H chains. In support of this idea, IP experiments exhibited more efficient coprecipitation of FcRLA through the Ig H chains than through L chains, but because we cannot rule out the possibility that this phenomenon is of a technical nature because of the polyclonal Abs used, further studies will be required to determine the biochemical nature of the interaction.

Given that FcRLA is expressed in all of the mouse B cells and binds intracellular Igs, we expected to find a profound defect in the ability of Fcrla^−/− mice to generate Ag-specific immune responses. Instead, we found that mice lacking FcRLA and their wild-type controls were equally capable of generating secreted Abs. When differences were found, such as in secondary responses to SRBCs, it was overproduction of IgG that was most evident in Fcrla^−/− mice. This may indicate an unanticipated regulatory ability of FcRLA. For instance, if FcRLA associates with a signaling partner, then it may limit the production of IgG in activated B cells. However, it is also possible that genes closely linked with Fcrla on chromosome 1 are responsible for this effect. For instance, the failure of mice carrying the 129 allele of Fcgr2b to upregulate this inhibitory FcR, with a resulting increase in Ab production, has been documented (29–31). Therefore, it is possible that the increase in IgG1 in Fcrla^−/− mice after secondary immunization with SRBCs could be attributed to lower expression of FcγRIIb in the germinal centers of these mice.

The overall absence of a dramatic phenotype in mice lacking FcRLA has two principle explanations. The first is that the stimuli used were insufficient to illustrate the importance of this protein in Ab production and organism survival. For instance, it is conceivable that given an infection by a specific pathogen, the functional consequence of FcRLA deficiency may become obvious. Alternatively, FcRLA may mitigate or enhance pathogenesis of certain autoimmune disorders, and targeting of the gene on a different genetic background would be required for a major phenotype to be produced.

A second possibility is that another protein is compensating for the lack of FcRLA in vivo. The most obvious candidate for such a compensatory effect is FcRLB. These two proteins share many structural features and are expressed as intracellular proteins in B cells. However, their expression levels and patterns seem to differ greatly. FcRLA is quite abundant within B cell populations in mice and humans, whereas FcRLB transcripts and protein are quite rare (13, 26). In addition, although FcRLB protein could be detected in human tonsillar B cells, its expression was mutually exclusive with that of FcRLA in this tissue (13). Furthermore, FcRLA and FcRLB have divergent N-terminal domains, which may reflect a nonoverlapping function.

Although the primary association of FcRLA with Igs is likely to be mediated by the two Ig C2 domains homologous with FcγRI, the key to its function may reside within the N-terminal domain, which is comprised of 20% acidic amino acids and a pair of conserved and potentially unpaired cysteines and is unrelated to other known protein domains. Our most favored hypothesis remains that FcRLA is a chaperone for the assembly of Abs by B cells. If this is the case, then one of many other endoplasmic reticulum chaperones, such as Bip or calnexin family members (32), may be partially redundant with FcRLA. In addition, the conserved cysteines and acidic residues within the FcRLA N-terminal domain may give the protein enzymatic activity, either directly or through coordination of metal ions. Because a pentameric IgM molecule will require formation of 96 disulfide bonds to be assembled correctly, additional sulfhydryl group-modifying activity might be required by B cells at certain stages of differentiation and Ab production. The pair of conserved cysteines within the N-terminal domain of FcRLA may function as a thiol isomerase during the assembly of Ab chains in a manner analogous to that used by protein disulfide isomerases (33). Alternatively, a previous study has isolated a protein from B cells with sulfhydryl oxidase activity that catalyzes the polymerization of IgM. Although the precise identity of the protein was not determined, it was found to have a low isoelectric point and be expressed specifically by B cells (34). These properties are consistent with FcRLA, and hence future studies will focus on examining the biochemical activity of FcRLA within B cells.

We thank John Atkinson, Hector Molina, and Emil Unanue for experimental advice, Mike White for embryonic stem cell microinjection, and Carey Strader for technical assistance.

Disclosures The authors have no financial conflicts of interest.