Synergistic effect of Tim4 and MFG-E8 null mutations on the development of autoimmunity

Abstract

Phagocytes, including macrophages, recognize phosphatidylserine exposed on apoptotic cells as an “eat me” signal. Milk Fat Globule EGF Factor VIII (MFG-E8) is secreted by one subset of macrophages, whereas Tim4, a type I membrane protein, is expressed by another. These proteins bind tightly to phosphatidylserine on apoptotic cells and enhance their engulfment by macrophages. To study the contribution of these proteins to the engulfment of apoptotic cells, we established a mouse line that was deficient in the genes encoding MFG-E8 and Tim4. The null mutation of Tim4 impaired the ability of resident peritoneal macrophages, but not thioglycollate-elicited macrophages, to engulf apoptotic cells. Mice deficient in either MFG-E8 or Tim4 on the C57BL/6 background developed hardly any autoantibodies, but aged female mice deficient in both MFG-E8 and Tim4 developed autoantibodies in an age-dependent manner. Tumour necrosis factor (TNF) α is known to protect against systemic lupus erythematosus-type autoimmunity, whereas type I IFN accelerates the disease. Indeed, the administration of an anti-TNFα antibody or a reagent that stimulates the IFN-α production [2,6,10,14-tetramethylpentadecane (TMPD; also known as pristane)] enhanced the production of autoantibodies in the MFG-E8- and Tim4-double-deficient mice. These results suggest that the double deficiency of Tim4 and MFG-E8, phosphatidylserine-binding proteins, can trigger autoimmunity and that TNFα and type I IFN regulate reciprocally the development of autoimmune disease.

apoptosis, autoimmune disease, engulfment, macrophages, pristane, TNF, apoptosis

Introduction

Every day, 10⁸ cells in the human body undergo apoptotic cell death (1). At the same time, several billion red blood cells are produced, and an equal number of nuclei (pyrenocytes) is expelled from erythroblasts in this process (2). The dead cells and expelled nuclei must be swiftly removed to prevent them from releasing noxious materials. Failure to remove them appears to activate the immune system and lead to systemic lupus erythematosus (SLE)-type autoimmune diseases (3, 4).

Macrophages, which are professional phagocytes, are responsible for clearing apoptotic cells and pyrenocytes (1). Macrophages recognize phosphatidylserine exposed on the plasma membranes of the dead cells and pyrenocytes as an “eat me” signal and engulf them for degradation in the lysosomes (5, 6). We and others previously identified several molecules that bind phosphatidylserine, resulting in the engulfment of apoptotic cells (6–11). Among them, Milk Fat Globule EGF Factor VIII (MFG-E8) is a soluble protein secreted from a subset of macrophages and immature dendritic cells, including thioglycollate-elicited peritoneal macrophages, GM-CSF-induced bone marrow–derived dendritic cells, and Langerhans cells in the skin (12). MFG-E8 acts as a bridge between dead cells and phagocytes, binding to phosphatidylserine on apoptotic cells and to integrins α_vβ₃ or α_vβ₅ on macrophages or immature dendritic cells (10).

Tim (T-cell immunoglobulin- and mucin-domain-containing)-4 is a type I membrane protein that is expressed by other subsets of macrophages, including resident peritoneal macrophages and splenic MOMA-1⁺-marginal zone macrophages (11, 13). Tim4 binds tightly to phosphatidylserine in vitro and supports the engulfment of apoptotic cells when expressed in mouse NIH3T3 cells.

We previously showed that MFG-E8-deficient mice on a 129/B6 mixed background produce autoantibodies and suffer from autoimmune diseases (14). On the other hand, two groups recently showed that Tim4-deficient mice on the 129 or B6 background developed little or no autoimmunity (13, 15). To clarify the role of MFG-E8 and Tim4, here we established a mouse line deficient in the genes for both MFG-E8 and Tim4 on the B6 background. Mice deficient in either MFG-E8 or Tim4 did not develop autoantibodies. In contrast, the MFG-E8/Tim4-double-deficient mice developed a high level of autoantibodies, and this process was accelerated by the administration of an anti-tumor necrosis factor (TNF) α neutralizing antibody or the hydrocarbon oil 2,6,10,14-tetramethylpentadecane (TMPD; also known as pristane) that stimulates type I IFN production. These results suggested that mice deficient in both MFG-E8 and Tim4, expressed in different sets of the macrophages, may generate sufficient numbers of unengulfed apoptotic cells to activate the autoimmunity, and TNFα and type I IFN reciprocally regulate the process.

Methods

Mice

C57BL/6 mice were from Japan SLC (Shizuoka, Japan). CAD^-/-mice were described (16). MFG-E8^-/- mice generated from a 129 ES cell line (14) were backcrossed to the C57BL/6 mice eight times. Tim4^-/- MFG-E8^-/- mice were generated by crossing Tim4^+/- MFG-E8^-/- mice. All mice were housed in a specific pathogen-free facility at Kyoto University Graduate School of Medicine. All animal experiments were carried out in accordance with protocols approved by the Kyoto University Animal Care and Use Committee.

Monoclonal antibodies and reagents

To produce an mAb against Tim4, Armenian hamsters were immunized with Tim4-Fc, in which the extracellular region of mouse Tim4 was joined to human IgG Fc region (11). Lymphocytes from the immunized hamsters were fused with a mouse myeloma NSO^bcl2 (17). One hybridoma (Mat4-9-5) that was suitable for immunohistochemical staining, FACS analysis, immunoprecipitation, western blotting, and neutralization of the engulfment was grown in serum-free GIT medium (Nihon Pharmaceutical, Tokyo, Japan), and the secreted mAb was purified using Protein A-Sepharose. The hamster Kat5-18 mAb was described (11). A rat anti-TNFα mAb (clone: MP6-XT22 (18)) was produced by culturing the hybridoma in GIT medium and purified by fractionation with (NH₄)₂SO₃ precipitation. A leucine zipper-containing human Fas ligand (FasL) was prepared in 293T cells and partially purified as described (19).

Targeted disruption of the Tim4 gene

The Tim4-targeted mice were generated as custom order by Unitech (Kashiwa, Chiba, Japan). In brief, the Tim4 chromosomal gene was isolated from Bacterial Artificial Chromosome (BAC) clones of C57BL/6 mice, and its exon 1 and most of exon 2 were replaced by a coding sequence for Venus and neo genes (Fig. 1). To increase the efficiency of homologous recombination, a coding sequence for the diphtheria toxin A fragment (DT-A) was inserted downstream of exon 5 in the targeting vector. The targeting vector was transfected into Embryonic Stem (ES) cells (clone: Bruce 4) derived from C57BL/6 mice (20). The ES clones carrying the Tim4-deficient allele were introduced into host embryos, and chimeric mice were produced. The chimeric mice with a high ES cell contribution were crossed with C57BL/6 mice to obtain heterozygous Tim4-deficient mice.

Fig. 1.

Establishment of Tim4-deficient mice. (A) Targeting of the Tim4 gene. Schematic representation of the wild-type and mutant loci of the Tim4 gene and the targeting vector. Exons are represented by closed boxes. The targeting vector carries the neomycin resistance gene (neo) and the gene encoding the Diphtheria toxin A fragment (DT-A). (B) Expression of Tim4 in Mac-1⁺ peritoneal cells. Peritoneal cells from the wild-type and Tim4-deficient mice were stained with biotin-labeled anti-Tim4, followed by APC-Cy7-streptavidin and APC-CD11b, and analyzed on a FACSAria.

The genotype of the Tim4 and MFG-E8 genes was determined by PCR using TaKaRa LA Taq (Takara Bio, Kyoto, Japan) or Taq DNA polymerase from Ampliqon (Skovlunde, Denmark). For the wild-type and mutant alleles of the Tim4 gene, a sense primer specific for the wild-type (5′-ATTGAGGGAGGTCATTCAGGA-3′) or mutant allele (5′-GACCCTGAAGCTGATCTGCA-3′) was used with a common antisense primer (5′-AGCGCACATTCTTCTTGACA-3′). The wild-type and mutant alleles of the MFG-E8 gene were detected by a similar method using the wild-type (5′-CTTTGGAGGATGTACACAGA-3′) or mutant (5′-CGTGGGATCATTGTTTTTCT-3′) specific sense primer and the common antisense primer (5′-CTTTATGTACTGTGCCTCCA-3′).

Administration of pristane and anti-TNFα mAb

Female mice at 7–10 weeks of age were injected i.p. with 0.5 ml Pristane (Sigma-Aldrich,, St. Louis, MO) as described (21), and the anti-TNFα mAb or control rat IgG (200 μg/head) was administered twice a week by i.p. injection as described (22).

Isolation of tissue macrophages

Resident and thioglycollate-elicited peritoneal macrophages were isolated from female C57BL/6 mice at 6 weeks of age as described (10, 11). Alveolar macrophages were prepared from the lungs by bronchoalveolar lavage with PBS containing 0.5 mM EDTA, cultured on plastic dish for 2 h, and washed with PBS to remove nonadherent cells. Kupffer cells were prepared as described (23). In brief, the liver was perfused in situ at a flow rate of 5 ml/min for 5 min with prewarmed (37°C) Hank’s Balanced Salt Solution (HBSS, Invitrogen, Carlsbad, CA) containing 0.5 mM EGTA, then with HBSS containing 1.8 mM CaCl₂, 2.0 mM MgCl₂, 1.0 mg/ml collagenase D, and 0.1 mg/ml DNase I for 5 min. The liver was removed and minced at 4°C using scissors in 10 ml HBSS. After filtrating through stainless steel mesh (diameter: 106 μm), the cell suspension was centrifuged at 50 × g for 1 min at 4°C to remove aggregates. The cells in the supernatant were collected by centrifugation at 280 × g for 10 min and suspended in cold HBSS containing 10% FCS. The cells were then layered on 17% metrizamide (Sigma-Aldrich) and centrifuged at 1800 × g for 30 min at 4°C. The cells at the interphase was collected, suspended in HBSS containing 10% FCS, and precipitated by centrifugation at 780 × g for 10 min. Cells were suspended in serum-free RPMI, cultured at 37°C for 15 min in noncoating dish (Iwaki, Chiba, Japan), and washed to remove nonadherent cells.

Flow cytometry

Cells were incubated on ice for 30 min with 1 μg/ml biotinylated Kat5-18 in 100 μl staining solution (PBS containing 2% FCS), followed by incubation with 2 μg/ml Allophycocyanin (APC)-Cyanine 7 (Cy7)-conjugated streptavidin (BD Bioscience, Franklin Lakes, NJ) and 2 μg/ml APC-conjugated anti-CD11b (BioLegend, San Diego, CA). The cells were then stained with 0.5 μM SYTOX Blue (Invitrogen) to exclude dead cells and analyzed by flow cytometry using a FACSAria (BD Bioscience).

Phagocytosis assay

The engulfment of apoptotic cells was assayed by TUNEL staining with CAD^-/- thymocytes as prey as described (10). In brief, 1 × 10⁶ peritoneal cells were incubated in 12-well plates at 37°C for 2 h and washed with PBS. Thymocytes from CAD^-/- mice were induced to undergo apoptosis by treatment with 4 units per ml FasL for 2 h, and 5 × 10⁶ apoptotic cells were cultured at 37°C for 30 min with the peritoneal cells. After staining with APC–anti-CD11b, the cells were fixed with 1% paraformaldehyde, subjected to TUNEL staining with FITC-labeled dUTP (Roche Diagnostics, Indianapolis, IN), and analyzed by flow cytometry on a FACSAria (BD Biosciences). In some cases, 2 × 10⁵ peritoneal cells were cultured in glass dishes (Iwaki) that had been coated with 1.0% gelatin. Mouse thymocytes were labeled with CellTracker^TM Orange (Invitrogen) and treated with FasL as described above; apoptotic cells (1 × 10⁶) were co-incubated with resident peritoneal macrophages for 30 min. After staining with FITC–anti-CD11b, the cells were observed using fluorescence microscopy (BioRevo BZ-9000, Keyence, Osaka, Japan).

Real-time PCR

Total RNA was prepared using ISOGEN (NIPPON GENE, Tokyo, Japan), followed by RNeasy Micro Kit (QIAGEN, Hilden, Germany), and was reverse-transcribed using Superscript III (Invitrogen). Aliquots of the products were amplified in a reaction mixture containing LightCycler^TM 480 SYBR Green I Master (Roche Diagnostics). The primers used for real-time PCR were as follows. Tim4, 5′-GGCTCCTTCTCACAAGAAACCACA and 5′-TCAGCTGTGAAGTGGATGGGAGA; MFG-E8, 5′-GATCTTTCCAACAACCTAGCCTCC and 5′-ACCGCTTTCATCCTGGATGAACTC; GAPDH, 5′-AACGACCCCTTCATTGAC and 5′-TCCACGACATACTCAGCAC.

Solid-phase ELISA

The serum level of anti-dsDNA antibodies and antinuclear antibodies (ANA) was determined as described (24). In brief, for the anti-dsDNA, the linearized plasmid DNA (5 μg/ml) was immobilized on NucleoLink plates (Nalge Nunc, Rochester, NY) by treating it at 50°C for 5 h with 10 mM 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide in 10 mM 1-methylimidazole. The plates were washed with 5× Saline-sodium citrate (SSC: 150 mM sodium chloride, and 15 mM trisodium citrate, pH 7.0) containing 0.25% SDS, and nonspecific binding sites were blocked with 1% BSA in PBS. The serum was diluted to the appropriate concentration, added to the wells, and incubated at room temperature for 1 h. After washing with PBS containing 0.1% Tween 20, the concentration of autoantibodies bound to each well was quantified using ELISA with Alkaline phosphatase (AP)-conjugated rabbit anti-mouse Ig (Dako, Copenhagen, Denmark). Phosphatase activity was detected using the BluePhos Microwell Phosphatase Substrate System (KPL, Gaithersburg, MA) and quantified by measuring the absorbance at 620 nm. The ANA level was determined using the Mesacup Enzyme immuno assay (EIA) system from MBL (Nagoya, Japan) for human ANA, except that the HRP-conjugated goat anti-human Ig was replaced by AP-conjugated rabbit anti-mouse Ig and the plate was incubated with the mouse serum for 2 h.

Results

Requirement of Tim4 by peritoneal macrophages for the engulfment of apoptotic cells

In addition to Tim4 and MFG-E8, several molecules, such as BAI1, stabillin-2, Tim1, and Tim3, are reported to recognize phosphatidylserine and mediate the engulfment of apoptotic cells (8, 9, 11, 25). We previously showed that MFG-E8 and Tim4 are rather complementarily expressed in different sets of macrophages. That is, Tim4 is expressed in resident peritoneal macrophages, whereas MFG-E8 is expressed in thioglycollate-elicited peritoneal macrophages (11). This was further confirmed with other tissue macrophages. Alveolar macrophages expressed MFG-E8 but not Tim4, whereas Kupffer cells expressed Tim4 but not MFG-E8 (Supplementary Figure 1 is available at International Immunology Online).

We previously established MFG-E8-null mice and showed that MFG-E8 is required for the efficient engulfment of apoptotic cells in vitro by thioglycollate-elicited macrophages and in vivo by the tingible-body macrophages in the spleen (14). To examine the role of Tim4 in the engulfment of apoptotic cells by a distinct set of macrophages, we established a Tim4-deficient mouse line (Fig. 1A). As shown in Fig. 1(B), about 40–50% of the cells in the mouse peritoneum were strongly stained with CD11b, and most of them expressed Tim4. As expected, the peritoneal cells from Tim4-null mice did not stain with anti-Tim4 mAb. When resident peritoneal macrophages were co-incubated with apoptotic thymocytes, they efficiently engulfed them (Fig. 2). The Tim4-null mutation almost completely abolished this ability. In agreement with MFG-E8 not being expressed in resident peritoneal macrophages (11), the null mutation of the MFG-E8 gene did not have any effect on the ability of these cells to engulf apoptotic cells. These results indicated that Tim4 plays an indispensable role in the engulfment of apoptotic cells in resident peritoneal macrophages.

Fig. 2

. Phagocytosis of apoptotic cells by resident peritoneal macrophages. (A) Resident peritoneal macrophages from wild-type, Tim4^-/-, MFG-E8^-/-, and Tim4^-/-MFG-E8^-/- mice were incubated at 37°C for 30 min with apoptotic CAD^-/- thymocytes, and the phagocytosis was assayed by TUNEL staining, followed by FACS analysis. The percentage of TUNEL-positive cells in the Mac-1⁺ population is plotted. (B) Resident peritoneal macrophages from wild-type, Tim4^-/-, MFG-E8^-/-, and Tim4^-/-MFG-E8^-/- mice were incubated at 37°C for 30 min with Celltracker^TM Orange-labeled-apoptotic thymocytes, washed with PBS, stained with FITC-anti-CD11b, and observed using fluorescence microscopy.

Synergistic effect of MFG-E8 and Tim4 on the development of autoimmunity

We previously established MFG-E8-null mice in a 129/B6 mixed background and showed that they, in particular the females, age-dependently developed an SLE-type autoimmunity (14). On the other hand, when mice bearing this mutation were back-crossed to C57BL/6 mice more than eight times, this phenotype and the level of autoantibodies in the serum were significantly reduced (Fig. 3). Recently, two groups established Tim4-null mice in the 129 and B6 mouse strains, respectively. The Tim4-null mice on the 129 background do not produce autoantibodies (13), but those on the B6 background produce a low but significant level of them (15).

Fig. 3

. Production of autoantibodies in MFG-E8^-/-Tim4^-/- mice. The concentration of the anti-dsDNA antibody (A and B) and ANA (C and D) was determined for the male and female mice of the wild-type, Tim4^-/-, MFG-E8^-/-, and Tim4^-/-MFG-E8^-/- genotypes at 10 (A and C) and 30 (B and D) weeks of age. The average value was determined for each group (more than five mice per group) and is shown as bar. The P-values between the indicated pairs are shown.

As shown in Fig. 3, the Tim4-null mice that we established on the B6 background did not produce a significant level of autoantibodies (anti-dsDNA or ANA). We previously reported that the injection of recombinant MFG-E8 mutant (D89E) into B6 mice, which masks the phosphatidylserine on apoptotic cells, causes strong autoimmunity (26). As discussed above, Tim4 and MFG-E8 are expressed in distinct sets of macrophages and probably function nonredundantly. We, therefore, examined whether mice bearing double knockout mutations of the MFG-E8 and Tim4 genes would exhibit any effect on the development of autoimmunity. As shown in Fig. 3, no autoantibodies (anti-ds DNA or ANA) were found in the serum of the MFG-E8/Tim4-double-deficient mice at 10 weeks of age. On the other hand, by 30 weeks of age, the female MFG-E8/Tim4-double-deficient mice developed a significant level of autoantibodies. These results suggested that the dose of unengulfed apoptotic cells generated by the lack of either the Tim4 or the MFG-E8 gene alone, but not by their double deficiency, might be insufficient to trigger the development of autoimmunity.

Reciprocal regulation of the autoimmunity by TNFα and type I IFN

Previous studies by a number of groups showed that an SLE-type autoimmunity is reciprocally regulated by TNFα and type I IFN in mouse and human (27–31). That is, blocking the circulating TNFα or increasing the IFNα level in the serum accelerates the development of the disease. To examine whether TNFα has any effects on the regulation of the SLE-type disease in the Tim4^-/-, MFG-E8^-/-, and Tim4^-/-MFG-E8^-/- mice, a TNFα-neutralizing mAb was administered twice a week to these mice. As shown in Fig. 4, a significant level of anti-dsDNA antibodies could be detected even in the serum of Tim4 or MFG-E8-single-null mice after 6 weeks and an enhanced synergistic effect on the development of the autoimmunity was observed in the MFG-E8/Tim4-double-deficient mice, whereas control rat IgG had no effect on autoimmunity in Tim4^-/- mice (data not shown). These results confirmed that the constitutive exposure to TNFα prevented the development of autoimmunity in these mice.

Fig. 4

. Enhanced production of the autoantibodies. (A) Effect of anti-TNFα mAb on the autoantibody production. Wild-type, Tim4^-/-, MFG-E8^-/-, and Tim4^-/-MFG-E8^-/- female mice (five mice each) at 7–11 weeks of age were injected i.p. twice a week with the anti-TNFα mAb (200 μg per head). Before the injection (0 weeks) and at 2, 4, and 6 weeks after the injection, the anti-dsDNA antibody level of the serum was determined. The average values are indicated by horizontal bars, and the P-values for the differences between the indicated pairs are shown. (B) Effect of pristane on the autoantibody production. Wild-type, Tim4^-/-, MFG-E8^-/-, and Tim4^-/-MFG-E8^-/- female mice (5–7 mice each) at 7–10 weeks of age were injected i.p. with pristane (0.5 ml per head). Before and at 7 weeks after the injection, the serum was collected, and the level of the anti-dsDNA antibody was determined. Bars indicate the mean values.

When pristane is introduced into mouse peritoneal cavity, it stimulates the production of type I IFN in a TLR-dependent manner, causing chronic inflammation (32). To examine the effect of type I IFN on the development of the autoimmune disease in Tim4^-/-, MFG-E8^-/-, and Tim4^-/-MFG-E8^-/- mice, pristane was injected i.p into these mice at 7–10 weeks of age. As shown in Fig. 4(B), a high level of anti-dsDNA could be detected at week 7 after the pristane injection in the Tim4^-/-MFG-E8^-/--double-deficient mice. These data suggested that the pristane-induced inflammation, probably type I IFN system, accelerated the production of the autoantibody in Tim4^-/-MFG-E8^-/- mice.

Discussion

A failure of the efficient engulfment of apoptotic cells has been believed to cause autoimmune disease because noxious materials are released from unengulfed dead cells that undergo secondary necrosis (1, 33). Many molecules have been proposed to regulate the engulfment of apoptotic cells, and mice with a deficiency of MER (a tyrosine kinase receptor) developed a severe autoimmune response, including splenomegaly, proteinuria, and elevated autoantibodies (34). Here, we showed that although a single deficiency of the MFG-E8 or Tim4 gene did not activate autoimmunity in the B6 mouse strain, the double deficiency of the Tim4 and MFG-E8 genes did. This finding is different from our previous one showing that MFG-E8-null mice develop autoimmunity (14). We now attribute this observation to the 129/B6 mixed background of the MFG-E8-null mice that we analyzed in the earlier study because the mouse strain of 129 and B6 chimeric background often spontaneously develops autoimmunity (35). The experiments in this report were, therefore, performed with MFG-E8-null mice that were backcrossed at least eight times to B6 mice. Our results also do not agree with Rodriguez-Manzanet et al. (15), which showed a significant development of autoantibodies in Tim4-null mice on the B6 background. Because SLE-type autoimmunity is greatly influenced not only by genetics but also by environmental factors (36), it is possible that although our mice and those studied by Rodriguez-Manzanet et al. (15) were kept under specific pathogen-free conditions, some small difference in the environment might result in a different outcome in the development of the autoimmunity.

In any case, we showed here that the double mutation of the MFG-E8 and Tim4 genes activated autoimmunity with high frequency. MFG-E8 and Tim4 are expressed in different sets of macrophages and are required for the engulfment of apoptotic cells differentially among macrophages. Tim4 was reported to stimulate proliferation of T cells (37). However, no apparent abnormality (splenomegaly, lymphadenopathy, or lymphopenia) was observed in the Tim4-null mutation. We also did not observe unengulfed apoptotic cells in Tim4-, MFG-E8-, and MFG-E8/Tim4-double-deficient mice. Whereas our preliminary analysis of the dexamethasone-treated mice suggested that the MFG-E8/Tim4-double-deficient mice may have the reduced ability to clear apoptotic thymocytes (K.S. and S.N., unpublished results). Thus, although a possibility that the abnormal T-cell development caused by the Tim4 deficiency enhances the autoimmunity cannot be ruled out, we postulate that the double mutations of the Tim4 and MFG-E8 genes may generate a higher dose of unengulfed apoptotic cells that can be immunogenic. This appears to agree with the hypothesis that autoimmunity develops only when the determining “factors” for SLE reach a certain threshold (38).

Resident peritoneal macrophages express Tim4 and require Tim4 to engulf apoptotic cells. MER is also expressed in mouse resident peritoneal macrophages and seems to be indispensable for this ability (39), suggesting that Tim4 and MER co-operate in the engulfment of apoptotic cells. On the other hand, tingible body macrophages in the spleen express Tim4, MFG-E8, and MER, although MFG-E8 and MER, but not Tim4, seem to be required for the engulfment of apoptotic cells by tingible body macrophages (40). However, it is not clear how MFG-E8, Tim4 and MER in different macrophages trigger the engulfment of apoptotic cells. To fully understand the molecular mechanism for the recognition of apoptotic cells, it will be necessary to reconstitute the engulfment system with these and other phosphatidylserine-binding proteins.

TNFα is a pleiotropic cytokine with many physiological and pathological functions (41). It is responsible for the inflammation caused by endotoxin-induced septic shock and rheumatoid arthritis. On the other hand, the chronic exposure of mice to a low level of TNFα suppresses immune reactions, for example, by inhibiting cytokine production from T cells and activating regulatory T cells (42). The administration of an anti-TNFα mAb to human patients with rheumatoid arthritis sometimes triggers an SLE-type autoimmune disease (43). Similarly, the administration of the anti-TNFα mAb accelerated the development of autoantibodies in MFG-E8-, Tim4-, and MFG-E8/Tim4-double-deficient mice, supporting the previous report that TNFα inhibits the development of SLE (44). Tim4^-/- macrophages were reported to produce TNFα constitutively (13). Although we could not detect the difference in the serum TNFα level between the wild-type and Tim4^-/- mice (data not shown), it is possible that the local chronic exposure of the Tim4^-/- immune system to a low concentration of TNFα may explain the difference between the lack of autoantibody production in Tim4^-/- mice and the strong autoantibody production in adult mice treated with an anti-Tim4 mAb (11). The 3′ noncoding region of TNFα, which has a regulatory role in its gene expression, shows a strong polymorphism among mouse strains, which may explain the different susceptibilities of various MFG-E8-null mouse strains to autoantibody production.

IFNα, a type I IFN, has recently been recognized as a critical mediator of human SLE (30, 31). Administration of pristane into mice induces the IFN gene expression leading to the development of SLE-type autoimmune disease in a Toll-like receptor (TLR)7-dependent manner (21). TLR7 is a receptor for nucleic acids. Because the mutations (lpr and gld) in Fas and FasL, which are involved in the activation-induced cell death of lymphocytes (45), reduces the pristane-induced SLE phenotype (46), it is believed that the nucleic acids released from dead cells trigger the disease (32). The accelerated development of the pristane-induced autoimmunity in Tim4^-/-MFG-E8^-/- mice may support the idea that the lack of Tim4 and MFG-E8 generates more unengulfed dead cells, which undergo secondary necrosis to release nucleic acids. About 90% of human SLE patients are female (47), and an involvement of female hormone is considered to play a role. Like human patients, the SLE phenotype in Tim4^-/-MFG-E8^-/- mice is prevalent in female, indicating that the Tim4/MFG-E8-null mice developed for this report may be useful for studying SLE-type autoimmune diseases.

Supplementary data

Supplementary data are available at International Immunology Online.

Funding

This work was supported in part by a grant-in-aid for Specially Promoted Research from the Japan Society for the Promotion of Science (JSPS), Japan. M. M. was a Research Fellow of JSPS, and a Post-Doctoral Fellow for Kyoto University Global COE program (Center for Frontier Medicine).