Leukocyte-Associated Ig-like Receptor-1–Deficient Mice Have an Altered Immune Cell Phenotype

Immune responses are tightly controlled by the opposing actions of activating and inhibitory signals. Activation receptors on immune cells recognize stressed cells, including transformed and pathogen-infected, and induce a hyperinflammatory state to combat the danger to the host; inhibitory receptors are expressed to dampen the immune response to prevent unwarranted or excessive inflammation (1). The impairment of inhibitory signals (e.g., the absence of inhibitory immune receptors or the downregulation of ligands for these receptors) can lead to a state of hyperresponsiveness that facilitates the development of autoimmune diseases. For example, CD200^−/− mice develop myeloid cell dysregulation and enhanced susceptibility to autoimmune inflammation, such as experimental autoimmune encephalitis (EAE) and collagen-induced arthritis (2), programmed death 1^−/− mice develop glomerulonephritis and arthritis (3), BTLA^−/− mice develop autoimmune hepatitis-like disease and produce autoantibodies to nuclear Ag (4), and FcγRIIb^−/− mice develop a lupus-like syndrome with fatal glomerulonephritis (5).

The mouse leukocyte-associated Ig-like receptor (LAIR)-1 or CD305 localizes to the leukocyte receptor complex (LRC) at the proximal end of mouse chromosome 7. The human LRC region is syntenic with the mouse, but encodes more genes, including LAIR-2, a secreted protein highly homologous to LAIR-1 (6). LAIR-1 possesses one Ig domain in its extracellular region and two ITIM motifs in the cytoplasmic tail that mediate its inhibitory capacity through interaction with Src homology 2 domain-contain-ing protein tyrosine phosphatase (SHP)-1, SHP-2, and C-terminal Src kinase (7, 8). LAIR-1 is reported to be expressed on the majority of cells of the immune system, including T cells, NK cells, monocytes, dendritic cells (DC), and human B cells, but not on B cells from mice (9–12). In vitro experiments show that ligation of LAIR-1 with mAb or collagen inhibits the cytotoxic activity of NK and CD8 T cells, BCR-induced B cell activation and proliferation, and CD3 signaling and cytokine production by T cells (9–11, 13–16).

Collagens, the most abundant proteins in the body, have been identified as high-affinity functional ligands for LAIR-1 and LAIR-2 (14, 17). LAIR-1 interacts with the glycine-proline-hydroxyproline repeats that are present in all collagens, and a synthetic trimeric peptide of 10 glycine-proline-hydroxyproline repeats alone can inhibit immune cell activation in vitro (14, 18). Twenty-eight different types of collagens have been identified in vertebrates, plus there are >20 other proteins that contain collagenous domains (19). The collagen-rich extracellular matrix is important for maintenance of tissue structures, cell adhesion, and migration during growth, differentiation, morphogenesis, and wound healing (20). Several collagen receptors have been shown to have important biological functions. Examples are the discoidin domain receptor 1 that promotes leukocyte migration and facilitates the differentiation/maturation and cytokine/chemokine production by macrophages and DC (21, 22), GP-VI that plays a central role in the hemostatic plug formation at sites of vascular injury (23, 24), and VLA-1 that plays a role in regulating inflammation during rheumatoid arthri-tis and delayed-type hypersensitivity responses (25). VLA-1 also potentiates CD8 T cell-mediated immune protection against influenza infection (26). Considering the abundance and biological importance of collagens and the broad expression of LAIR-1 on immune cells, it is reasonable to suspect that LAIR-1 plays a role in regulating the responses of immune cells in both normal and pathological situations.

Although the inhibitory potential of LAIR-1 is well known, the actual in vivo function is unknown. To explore the in vivo role of this receptor, we generated LAIR-1–deficient mice. These animals show some phenotypic characteristics distinct from wild-type (wt) mice, but are healthy and show normal longevity, and they develop similar pathology states as wt mice in induced autoimmune diseases.

LAIR-1^−/− and LAIR-1^fl/fl mice were generated by OZgene as shown in Supplemental Fig. 1. Briefly, exon 4–8 of LAIR-1, which contains two ITIM motifs for binding of SHP, was flanked by loxP sites. A PGK-neo cassette flanked by Flp recombinase target sites was used for selection. Following homologous recombination of the vector in embryonic stem cells, clones bearing the LAIR-1^fl/fl locus were established after deletion of PGK-neo selection cassette by Cre recombination, and clones with the LAIR-1^−/− locus were generated after deletion of the LoxP sites flanking regions (exon 4–8) together with the PGK-neo cassette using Cre recombinase. Identified targeted embryonic stem cell clones were microinjected into the blastocysts of C57BL/6 mice. Chimeric mice were mated with C57BL/6 female mice to produce heterozygous mice and their wt littermates, and LAIR-1^fl/fl mice were bred to CD4 cre mice to get T cell-specific LAIR-1–deficient mice. C57BL/6 and Rag1^−/− mice were obtained from The Jackson Laboratory, and OT-II, CD4 cre transgenic mice were from Taconic Farms. FcγRIIB^−/− mice were a kind gift of Dr. Silvia Bolland (National Institutes of Health, National Institute of Allergy and Infectious Diseases [NIAID]). All experiments were performed with littermate mice: LAIR-1^−/− with LAIR-1^+/− or LAIR-1^+/+ and CD4 cre LAIR-1^fl/fl with LAIR-1^fl/fl. LAIR-1^+/− and LAIR-1^+/+ were shown to express equivalent levels of LAIR-1. All mice were maintained on a C57BL/6 genetic background and housed in a pathogen-free environment in the NIAID animal facility. All experimental protocols were approved by NIAID Animal Care and Use Committee.

Labeled Abs to CD3 (145-2C11), CD4 (RM4-5), CD8 (53-6.7), CD11b (M1/70), CD11c (N418), CD23 (B3B4), CD25 (PC61.5), CD40 (1C10), CD44 (IM7), CD62L (MEL-14), CD69 (H1.2F3), CD80 (16-10A1), CD86 (GL1), B220 (RA3-6B2), c-Kit (ACK2), F4/80 (BM8), Gr-1 (RB6-8C5), I-A^b (AF6-120.1), IFN-γ (XMG1.2), NK1.1 (PK136), DX5 (DX5), Sca-1 (D7), and LAIR-1 (113), along with specific isotype-matched control Abs, were from eBioscience. Labeled anti-CD21 (7G6) and anti-IgM (R6-60.2) were from BD Pharmingen. Labeled anti-plasmacytoid DC (pDC)/lPC (120G8.04) from Imgenex was used for detection of pDC.

Specific cell types were identified by flow cytometry according to the expression of surface molecules as follows: CD4 T (CD4⁺CD8⁻CD3⁺), CD4 naive T (CD4⁺CD8⁻CD44loCD62Lhi), CD4 memory T (CD4⁺CD8⁻CD44hi), CD4 effector memory T (T_EM; CD4⁺CD8⁻CD44hi CD62Llo), CD4 central memory T (T_CM; CD4⁺ CD8⁻CD44hiCD62Lhi), CD8 T (CD8⁺CD4⁻CD3⁺), CD8 naive T (CD8⁺CD4⁻CD44loCD62Lhi), CD8 memory T (CD8⁺CD4⁻CD44hi), CD8 T_EM (CD8⁺CD4⁻CD44hiCD62Llo), CD8 T_CM (CD8⁺CD4⁻CD44hiCD62Lhi), NK (NK1.1⁺CD3⁻ or DX5⁺CD3⁻), NKT (NK1.1⁺ CD3⁺), marginal zone (MZ) B (B220⁺CD21hiCD23lo), transitional 1/B-1 (T1/B1) B (B220⁺CD21⁻CD23⁻), follicular (FO) B (B220⁺CD21loCD23hi), granulocytes (Gr-1⁺), macrophages (F4/80⁺), con-ventional DC (cDC; CD11c⁺B220⁻), pDC (CD11clo120G8⁺), bone mar-row pro-B/pre-B (B220⁺IgM⁻), immature B (B220loIgM⁺), mature B (B220hiIgM⁺), hematopoietic stem (HSC; Lin⁻c-Kit⁺Sca-1⁺), and committed progenitor cells (CPC; Lin⁻c-Kit⁺Sca-1⁻). To determine cytokine production, stimulated cells were incubated in the presence of GolgiStop (BD Biosciences) for the last 4–6 h. Then, cell-surface receptor expression and intracellular cytokine staining were determined with BD Cytofix/Cytoperm Plus kits (BD Biosciences). Data were acquired on an FACSCalibur (BD Biosciences) and analyzed with FlowJo software (Tree Star).

Competitive homing assays were performed with isolated LAIR-1^+/+ and LAIR-1^−/− T cells (Miltenyi Biotec) and differentially labeled with 5-(and 6-) carboxyfluorescein, succinimidyl ester (FAM-SE; Molecular Probes) or tetramethylrhodamine isothiocyanate (TRITC; Molecular Probes), as previously described (27). Mice were sacrificed 2.5 and 19 h after i.v. injection of cells, and single-cell suspensions from inguinal and mesenteric lymph nodes, spleen, and blood were isolated and analyzed by flow cytometry. To avoid possible bias related to the dyes toxicity, LAIR-1^+/+ and LAIR-1^−/− cells were labeled with the opposite dyes in different experiments. The homing ratio of LAIR-1^−/− versus LAIR-1^+/+ T cells was calculated as follows: ([LAIR-1^−/− T cells − FAM-SE]/[LAIR-1^+/+ T cells − TRITC] + [LAIR-1^−/− T cells − TRITC]/[LAIR-1^+/+ T cells − FAM-SE])/2.

Experimental colitis was induced by naive CD4 T cells transfer (28). Briefly, FACS-sorted 4 × 10⁵ naive CD4 T cells (CD4⁺CD8⁻CD44lo CD62Lhi) were injected i.p. into Rag1^−/− mice. Mice were weighed weekly and monitored for signs of disease. For EAE induction (29), mice were immunized s.c. with 200 μg MOG 35–55 peptide (Anaspec) emulsified in 200 μl CFA (DIFCO Laboratories) containing killed Mycobacterium (DIFCO Laboratories) plus 200 ng pertussis toxin (List Biological Laboratories) by i.p. injection on days 0 and 2, and EAE disease scoring was determined daily (29). Mononuclear cells from spinal cords were isolated as described previously (30). IFN-γ and IL-17 production by mononuclear cells isolated from spinal cords was determined on day 28 after stimulation with PMA and ionomycin for 4 h.

To elicit a T-dependent immune response, CD4cre LAIR-1^fl/fl and LAIR-1^fl/fl mice were immunized with 100 μg trinitrophenyl (TNP)-OVA in Imject Alum (Pierce) and afterward were rechallenged with 100 μg TNP-OVA on day 28. To elicit a T-independent immune response, LAIR-1^−/− and LAIR-1^+/− mice were immunized with 50 μg TNP-LPS (T-independent type I [TI-I]) or 25 μg TNP-Ficoll (T-independent type II [TI-II]).

Mouse Ig isotype-specific ELISA was carried out by using isotype-specific goat anti-mouse Ig (Southern Biotechnology Associates) to coat 96-well plates. Serially diluted sera from mice were added to the coated plates, and bound Igs were detected by alkaline phosphatase-conjugated detection Abs to specific mouse isotypes (Southern Biotechnology Associates). For T-dependent immunization, sera were collected 2 wk after the primary immunization or 1 wk after rechallenge on day 28. The serum levels of TNP-specific Abs for different Ig subclasses and isotypes were determined using ELISA plates coated with TNP-BSA (Biosearch Technologies). T-independent immunization Ig isotype Abs in serum were determined 7 d after immunization by TNP-specific ELISA. Anti-nuclear Abs (ANA) in the serum of mice at 13 mo of age were measured with an ANA ELISA screen kit (Diamedix).

DC were generated from bone marrow cells according to Inaba et al. (31) with slight modification. In brief, HSC were purified from bone marrow cells obtained from femurs and tibias of LAIR-1^+/+and LAIR-1^−/− mice by c-Kit⁺ isolation kits (Miltenyi Biotec). Cells (5 × 10⁵ cells/well) were incubated in RPMI 1640 complete medium supplemented with 10% FBS, 10 ng/ml recombinant murine (rm)GM-CSF, and 10 ng/ml rmIL-4 (R&D Systems) in 24-well plates. Media containing rmGM-CSF and rmIL-4 was renewed every 2 d. On day 6, cells were harvested by gentle swirling and stimulated with LPS (1 μg/ml), lipoteichoic acid (LTA) (25 μg/ml), or polyinosinic-polycytidylic acid (poly I:C; 25 μg/ml) for 24 h (Sigma-Aldrich).

pDC were isolated from splenocytes by MACS kit (Miltenyi Biotec); purity was at least 90%. A total of 4 × 10⁴ purified pDC were cultured in 100 μl RPMI 1640 supplemented with 10% FBS (Atlanta Biologicals), penicillin (100 U/ml), streptomycin (100 μg/ml), l-glutamine (2 mM), HEPES (10 mM), and sodium pyruvate (1 mM) (Mediatech); 10 μg/ml CpG ODN 1585 (InvivoGen) was added to the cultures for stimulation. Supernatants were collected after 24 h for IFN-α analyses by ELISA according to the manufacturer’s instructions (eBioscience).

For the micro-computed tomography (μCT) analysis of trabecular bone, the trabecular volume in the distal femoral metaphysis from 8-wk-old wt and LAIR-1^−/− male mice was measured using a Scanco μCT40 scanner (Scanco Medical AG). A threshold of 200 was used for the evaluation of scans.

Mouse LAIR-1 was reported to be expressed on most immune cells except B cells (9). We found that LAIR-1 is constitutively expressed on all immune cell types examined, including a popu-lation of B cells (Fig. 1A, 1B). The LAIR-1 expression profile of cells isolated from lymph nodes is similar to that of splenocytes (data not shown). We then checked the expression of LAIR-1 on splenic T cell subsets (Fig. 1C) and found that LAIR-1 is constitutively expressed on CD4⁺CD25⁺ T cells and both naive and memory CD4 and CD8 T cells. LAIR-1 is expressed at similar levels on naive and memory CD4 T cells, but is expressed at higher levels on CD8 memory cells than on naive CD8 T cells. Examination of splenic B cell subsets revealed that LAIR-1 is mainly expressed on MZ B, but not on T1/B1 (CD21⁻CD23⁻) and FO B (CD21loCD23hi) cells (Fig. 1D).

LAIR-1^−/− mice were born at the expected Mendelian ratios, developed normally, and appeared healthy in a pathogen-free housing environment. To determine if the absence of LAIR-1 affected lymphoid cell development and homeostasis, we performed flow cytometry analysis of various cell types. Thymi of young (2–4-mo-old) LAIR-1^−/− mice did not show any gross defect in T cell development, having normal levels of CD4⁺/CD8⁺ (double-positive), CD4⁻/CD8⁻ (double-negative), and CD4⁺ or CD8⁺ cells (Fig. 2A). CD69 expression on LAIR-1^−/− double-positive cells is normal, and development of LAIR-1^−/− double-negative thymocytes does not show any sign of abnormality (data not shown). These results indicate that LAIR-1^−/− mice have normal T lymphopoiesis. Likewise, LAIR-1^−/− mice showed normal B cell development in the bone marrow. The pro-B/pre-B, immature, and mature B cell populations in the bone marrow of LAIR-1^−/− mice are comparable to those found in LAIR-1^+/+ mice (Fig. 2B). There are also no defects in the HSC and CPC compartments in LAIR-1^−/− mice (Fig. 2B).

We also checked splenic (Fig. 2C, Table I) and lymph node cell subsets (Fig. 2D) in young mice. In the spleen, the frequency of LAIR-1^−/− CD3 T cells is slightly, but significantly, lower than those of LAIR-1^+/+ mice (∼12% less), which is reflected in both the CD4 (∼10% less) and CD8 T (∼13% less) cells; in contrast, the frequency of CD4⁺CD25⁺ T cells, the majority of which are regulatory T (Treg) cells, is increased ∼23% in LAIR-1^−/− mice. Other more defined T cell populations showed no differences in younger mice.

The percentage of splenic LAIR-1^−/− B cells is increased by ∼10% compared with B cells from LAIR-1^+/+ mice. FO B cells are recirculating cells that home mainly to B cell follicles in secondary lymphoid organs including spleen and lymph nodes and play an important role in T cell-dependent responses in follicles and T cell-independent responses in the bone marrow (32); MZ B cells are noncirculating mature B cells enriched primarily in the MZ of the spleen, which function in both T cell-dependent and -independent responses to blood-borne pathogens (33). Most of the B cell population increase in LAIR-1^−/− mice is due to MZ B cells that are elevated ∼23% (from 8.1% in LAIR-1^+/+ to 10.0% in LAIR-1^−/−), as the frequency of FO and T1/B-1 B cells is similar in LAIR-1^−/− and LAIR-1^+/+ mice (Fig. 2C, Table I).

LAIR-1^−/− mice have an increased frequency of splenic cDC (∼22% more) (Fig. 2C) and bone marrow cDC (∼13% more) (Supplemental Fig. 2A), whereas the frequency of lymph node cDCs is not statistically different compared with LAIR-1^+/+ (Fig. 2D). Human pDC are reported to express high levels of LAIR-1 (34). We confirm that mouse pDC (CD11clo120G8⁺) also express very high levels of LAIR-1 (data not shown). However, when we analyzed the pDC populations in spleen, lymph node, and bone marrow, we found that LAIR-1^−/− mice have similar amounts of pDC as LAIR-1^+/+ mice (Supplemental Fig. 2A).

The frequencies of granulocytes and macrophages are not perturbed by the absence of LAIR-1, and the expression of the activation markers CD80 and CD86 on macrophages is comparable between LAIR-1^−/− and LAIR-1^+/+ mice (Fig. 2C). As shown in Fig. 2D, the phenotypes of lymph node cells in LAIR-1^−/− mice also tend to have less T cells and more Treg cells, even though not statistically significant. The altered peripheral populations in young LAIR-1^−/− mice do not lead to pathological conditions (X. Tang and F. Borrego, unpublished observations).

We also analyzed splenocytes from >10 pairs of littermates that were older than 12 mo. We found that in comparison with LAIR-1^+/+, as in younger mice, LAIR-1^−/− mice tend to have less CD3⁺ T cells with significantly less CD8 T cells. More strikingly, whereas the analyses of CD69 expression by peripheral T cells revealed no differences between LAIR-1^−/− and LAIR-1^+/− mice in their activation state when the mice are young (2–4 mo old) (Fig. 3A, Table I), older LAIR-1^−/− mice have markedly increased CD69 expression on both splenic CD4 and CD8 T cell populations (Fig. 3A, Table II). The frequency of CD69⁺CD4 and CD69⁺CD8 cells was elevated ∼20 and 40%, respectively, in LAIR-1^−/− compared with LAIR-1^+/− and ^+/+ mice (Fig. 3A, Table II). Compared to wt mice, young LAIR-1^−/− mice did not show any difference in the frequencies of peri-pheral CD4 T_EM, CD4 T_CM, CD8 T_EM, and CD8 T_CM cell subsets (Fig. 3B, Table I). In contrast, older LAIR-1^−/− mice have increased population within the memory CD4 T cell pool, due to an ∼20% increase in CD4 T_EM population, as the CD4 T_CM population is less in older mice (Fig. 3B, Table II). A similar accumulation of CD8 memory T cells was evident in older LAIR-1^−/− with increased fre-quencies of both CD8 T_EM and CD8 T_CM (Fig. 3B, Table II). Re-flecting this, the LAIR-1^−/− naive CD4 and CD8 T cell populations were smaller in older LAIR-1^−/− than those in the LAIR-1^+/− and ^+/+ mice (Fig. 3B, Table II). As with younger mice, splenic B cells tend to be elevated in older LAIR-1^−/− mice, which is accounted for by the presence of more MZ B cells but not FO and T1/B-1 B cells (Table II). The frequencies of LAIR-1^−/− granulocytes and macrophages and their activated state do not show differences in aged mice compared with LAIR-1^+/+ populations. Despite the presence of higher percentages of activated and effector T cell subsets, older LAIR-1^−/− animals appear to be healthy and normal.

Two-year-old LAIR-1^−/− mice housed in a pathogen-free environment are visibly normal, and 18-mo-old LAIR-1^−/− mice necropsy reports did not show significant differences from wt mice (data not shown). Complete blood count analyses of young (2–5-mo-old) versus older (9-mo-old) mice revealed that eosinophils were significantly elevated (2-fold) in younger LAIR-1^−/− mice, but not different in older mice. Older LAIR-1^−/− mice had slightly elevated platelet levels. There were no other differences in the blood cell numbers between LAIR-1^−/− and wt (Supplemental Tables I, II). We also compared LAIR-1^−/− to wt mice for serum Ig levels and found that the levels of IgA, IgG2a, IgG2b, and IgM are not different, whereas the level of IgG1 is significantly less (23%) in LAIR-1^−/− mice. The lower amounts of IgG3 in LAIR-1^−/− mice are not significant (Fig. 4A).

There are reports that deficiencies in inhibitory receptors can cause development of autoimmune diseases that are characterized by the presence of ANA in the serum (35, 36). We found that aged (13 mo) LAIR-1^−/− mice have similar amounts of ANA compared with wt mice (Fig. 4B). Inhibitory receptors can cooperate to reg-ulate autoimmune diseases. For instance, LAG-3 can synergis-tically act with programmed death 1 to prevent autoimmune diseases (37). FcγRIIB is another ITIM-containing receptor, and FcγRIIB^−/− mice on C57BL/6 background develop autoantibodies and autoimmune glomerulonephritis (5). Moreover, duplication of the TLR7 gene accelerates the development of this autoimmune disease (38). We wondered if the absence of LAIR-1 would also exacerbate the disease process in FcγRIIB^−/− mice; however, LAIR-1^−/−FcγRIIB^−/− mice did not show any differences in the levels of ANA, urine protein, or the activated lymphocyte phenotype (data not shown), and the pathology of double knockout mice was similar to the FcγRIIB^−/− mice, as were the survival rates (Fig. 4C, 4D). Taken together, we found no evidence that LAIR-1^−/− mice develop spontaneous autoimmune diseases, and LAIR-1 deficiency does not accelerate the development of autoimmunity in FcγRIIB^−/− mice.

Because LAIR-1^−/− T cells from older mice have elevated populations of activated T cells (Fig. 3, Table II), we pursued the possibility that LAIR-1 deficiency might make mice more susceptible to the induction of autoimmune diseases. To study colitis susceptibility, we employed a model in which naive CD4⁺ cells are transferred into lymphopenic hosts (Rag1^−/− mice) and colitis development is monitored by measuring body weight changes (39). We found that transfer of LAIR-1^−/− or LAIR-1^+/− naive T cells resulted in equally severe body weight loss through the ninth week after transfer (Fig. 5A), which indicates that LAIR-1^−/− naive CD4 T cells induce similar colitis as LAIR-1^+/− naive CD4 T cells.

We also investigated the role of LAIR-1 in an EAE model because a critical step in the pathogenesis of EAE is the extravasation of leukocytes from the bloodstream into the CNS parenchyma, which involves autoaggressive T cell adhesion to and migration through the endothelial monolayer of the postcapillary venules, where they encounter the endothelial cell basement membrane that includes collagen type IV (29). Thus, it is reasonable to suspect that T cells and other leukocytes will interact with collagens in their trafficking toward the CNS and that this interaction might somehow affect their activation status as a consequence of the engagement of LAIR-1. We found that similar EAE pathogenesis progression was induced in LAIR-1^−/− and LAIR-1^+/− mice (Fig. 5B), and IFN-γ and IL-17 production by mononuclear cells, isolated from the spinal cords on day 21 after disease induction, was also similar (data not shown). We also did not observe any difference in the induction of serum (K/BxN)-induced arthritis between LAIR-1^−/− and LAIR-1^+/+ mice (data not shown).

Several surface collagen receptors, like integrins and discoidin domain receptor tyrosine kinases, are reported to be involved in cell adhesion and migration (19). Therefore, because we observed some small alterations in the peripheral LAIR-1^−/− T cell subsets in the spleen (Tables I, II), we did competitive homing experiments to compare the migration of LAIR-1^−/− and LAIR-1^+/+ T cells to peripheral lymphoid organs. The results show that LAIR-1^−/− T cells are not altered in their entrance to peripheral lymphoid tissues (measurement at 2.5 h) or their recirculation (measurement at 19 h) through the peripheral lymphoid tissues, like lymph nodes, spleen, and peripheral blood (Fig. 5C).

We compared the in vitro functional status of LAIR-1^−/− and LAIR-1^+/− NK cells by degranulation assays. Briefly, splenocytes from such mice were stimulated for 5 h with the anti-NK1.1 or mouse IgG2a in the presence of soluble anti-CD107a Abs. Cells were then stained with anti-DX5 and anti-CD3 mAb and for intracellular IFN-γ. The frequency of DX5⁺CD3⁻ cells that were CD107a⁺IFN-γ⁺ from LAIR-1^−/− mice was comparable to that from LAIR-1^+/− controls (Supplemental Fig. 2B).

The ability of LAIR-1^−/− T cells to provide B cell help was determined by comparing the responses of CD4 cre LAIR-1^fl/fl and LAIR-1^fl/fl mice to immunization with TNP-OVA in Imject Alum which induces a T-dependent response. CD4 cre LAIR-1^fl/fl mice showed a significantly diminished TNP-specific IgG2a and IgG2b response (Fig. 6A), indicating that LAIR-1^−/− T cells are less efficient in helping B cells to switch to produce these Ag-specific IgG isotypes.

MZ B cells are LAIR-1⁺, and they have been functionally linked to TI-II immune responses to multivalent Ags due to an established requirement for the spleen in these responses (40, 41). We tested if LAIR-1 plays a role in TI-I or TI-II immune responses. The ELISA results showed that upon both TI-I and TI-II immunizations, TNP-specific Ig Abs in serum were similar between LAIR-1^−/− and control mice. The level of TNP-IgG3 tended to be higher in LAIR-1^−/− mice with the TI-I immunization; however, it was not statistically significant (Fig. 6B). Taken together, the lack of LAIR-1 did not affect the T-independent B cell responses.

Human LAIR-1 was reported to inhibit the differentiation of CD14⁺ peripheral blood precursors to DC (42). We addressed if LAIR-1 deficiency affects the differentiation of bone marrow precursors into DC. We found that the in vitro generation of bone marrow-derived DCs (BMDC) was not affected by LAIR-1 deficiency (Fig. 7A) and that upon stimulation with LPS, LTA, or poly I:C, the upregulation of CD11b, CD40, CD80, CD86, and MHC class II is comparable between LAIR-1^+/+ and LAIR-1^−/− BMDC (Fig. 7B). LAIR-1 alone or in combination with NKp44 was also reported to inhibit IFN-α production by human pDC (34). Therefore, we investigated if LAIR-1 absence has an effect on the production of IFN-α by pDC. Our results show that LAIR-1^−/− and LAIR-1^+/+ pDC produce similar amounts of IFN-α when stimulated with CpG (Fig. 7C).

Osteoclasts are giant multinucleated cells that resorb the bone matrix. Osteoclasts are dependent on ITAM signals for differentiation (43). Interestingly, the osteoclast-associated receptor (OSCAR), which associates with the FcRγ, has recently been shown to be a collagen receptor and contribute to osteoclastogenesis in vivo (44). Like OSCAR, LAIR-1 is also expressed on osteoclast precursors and encoded in the LRC. LAIR-1 could therefore negatively regulate OSCAR signaling during osteoclastogenesis with a possible effect on bone mass (45). We therefore analyzed the bones of wt and LAIR-1^−/− mice by μCT to see if there were any differences in bone mass. No significant differences in bone mass could be observed between wt and LAIR-1^−/− mice (Fig. 8).

Immune cell responses are tightly regulated by the balance of sig-nals from activating and inhibitory receptors that they express. All else being equal, inhibitory signals tend to predominate over activat-ing signals serving to prevent hyperresponsiveness/autoimmunity against self-Ags that could damage the host. A large number of ITIM-bearing inhibitory molecules with diverse tissue distribution and ligand recognition have been shown to negatively regulate cell activation. ITIM-containing molecules are involved in the control of a large spectrum of immune functions (46). LAIR-1 is an ITIM inhibitory-bearing receptor expressed by the majority of immune cells, for which cross-linking in vitro by Ab or collagens delivers an inhibitory signal that can downregulate activating signals (9, 10, 47); however, the function of LAIR-1 in vivo remains largely a mystery. To begin to address this, we generated and characterized LAIR-1^−/− and LAIR-1^fl/fl mice.

We found that the deficiency of LAIR-1 did not affect the Mendelian ratio of the offspring nor their longevity or visible normality in a pathogen-free housing environment. Old (18 mo) LAIR-1^−/− mice necropsy reports did not show significant pathologic differences from LAIR-1^+/+ mice. We performed complete blood count tests on young (2–5-mo-old) and older (9-mo-old) mice (Supplemental Tables I, II) as a means of evaluating the physiological peripheral immune profile (48). Younger LAIR-1^−/− mice have 2-fold elevated eosinophils compared with LAIR-1^+/+ mice. Increased eosinophil counts (eosinophilia) can signify infections or allergic or autoimmune diseases (49). In LAIR-1^−/− mice, the elevated eosinophil counts subside with age with no evidence of other pathologic conditions. LAIR-1 was reported to negatively regulate the in vitro maturation of primary megakaryocytic progenitors (50) that, upon maturation, produce platelets. However, we observed only a slightly increased platelet count in 9-mo-old LAIR-1^−/− mice (Supplemental Tables I, II).

The absence of LAIR-1 expression has no evident pathological consequences. However, a sign that LAIR-1 may have an inhibitory role in vivo is reflected by the fact that with age, LAIR-1^−/− mice exhibit a higher frequency of activated and effector/memory T cells (Fig. 3). The importance of the expression of LAIR-1 on T cells is also shown by the low efficiency that LAIR-1^−/− T cells have in helping B cells to effectively switch to Ag-specific IgG2a and -2b (Fig. 6). Altogether, these data indicate that LAIR-1 has a small role, albeit noticeable, in controlling T cell functions, although its absence does not result in an observable pathologic effect.

Although our data showed that LAIR-1 expression on T cells may have a role in helping B cells in switching to certain Ag-specific IgG isotypes, we do not know if the expression of LAIR-1 on B cells may also have a role. Our data showed that LAIR-1 expression is restricted to MZ B cells (Fig. 1D). MZ B cells are sessile and reside in the outer white pulp of the spleen between the marginal sinus and the red pulp. They have been functionally linked to TI-II immune responses to multivalent Ags due to an established requirement for the spleen in these responses (40, 41). We showed the deficiency of LAIR-1 had no effect in the generation of Ag-specific Igs upon either TI-I or TI-II immunizations, except that TNP-IgG3 tended to be higher in LAIR-1^−/− mice with the TI-I immunization (Fig. 6B). This indicates that the ef-fect of LAIR-1 deficiency on MZ B cells has a marginal role in T-independent B cell responses.

It was very surprising to us not to find a role for LAIR-1 in the autoimmune disease models that we tested, specifically in EAE (Fig. 5B). It is well known that during extravasation to inflamed tissues, autoaggressive T cells migrate through the endothelial monolayer and encounter the endothelial cell basement membrane (29). Collagen type IV is an important component of the basement membranes, and it has been shown before that LAIR-1 is able to interact with collagen type IV that is in matrigel, a preparation of basement membranes (18). Encephalitogenic T cells entering the CNS apparently interact with components of the basement membranes, as laminins have been shown to regulate the extravasation of T cells into the brain (51, 52). Why the interaction of LAIR-1 with collagens present in the basement membranes does not occur or has no impact remains to be explained. In this same vein, it was surprising to find that LAIR-1 has no obvious impact on bone mass stimulated by OSCAR-generated ITAM signaling in osteoclasts or their precursors, as both receptors are expressed by these cells and both recognize collagens (44). One possible explanation is that there is redundancy in inhibitory receptor pathways that can negatively regulate OSCAR (e.g., PECAM-1, PIR-B, or SIRPα) (53–55).

The deficiency of the ITIM-containing receptor LAIR-1 results in higher frequencies of activated and effector/memory T cells in aged mice, yet this apparently does not lead to a hyperactivated immune response. In apparent contradiction, LAIR-1 deficiency results in lower levels of serum IgG1 and IgG3 in LAIR-1^−/− mice and a lower efficiency of LAIR-1^−/− T cells in providing help to B cells for switching to Ag-specific IgG2a and IgG2b following immunization. There are several potential explanations for these observations: 1) LAIR-1 is not a pivotal regulator of the immune system, and/or other inhibitory receptors compensate for its in-hibitory function; 2) LAIR-1 needs to cooperate with other receptors to effect its inhibitory function; 3) LAIR-1 functions only if mice are faced with overwhelming challenges like infections and septic shock (our preliminary data do not show any differences between LAIR-1^+/+ and LAIR-1^−/− mice after lymphocytic choriomeningitis virus infection and LPS injection [data not shown]); 4) higher numbers of Treg cells observed in LAIR-1^−/− mice may suppress activation of the immune system; and 5) mouse LAIR-1 may not function in vivo as an inhibitory receptor. Unlike human LAIR-1, mouse LAIR-1 recruits SHP-2 but not SHP-1 when tyrosine phosphorylated (6), and it has been shown that SHP-2 may play a positive regulatory role in T cells (56). Additional studies are required to investigate these or other possibilities. In summary, although murine LAIR-1 plays a role in the natural development of the immune system, its absence does not affect survival in a controlled pathogen-free environment or even the severity of induced ailments in which exposure to collagen might be expected to be involved. Similar findings were made for CD94-deficient mice in that unexpectedly, these mice developed normally and responded normally in a variety of disease models (57); however, recently, these mice have been shown to lack resistance to mousepox caused by the Orthopoxvirus ectromelia virus infection (58). Whether a critical circumstantial role for LAIR-1 will ultimately unveil itself remains to be determined.

We thank Dr. Silvia Bolland for providing FcγRIIB^−/− mice and constant helpful advice, Lily I. Cheng, Robert Valas, Mirna Pena, Tatyana Tarasenko, and Prapaporn Pisitkun for technical assistance, and Aleksandra Gil-Krzewski, Konrad Krzewski, Yousuke Murakami, Giovanna Peruzzi, and Jennifer Weck for critical review of the manuscript.

The authors have no financial conflicts of interest.