Autoimmune Disease and Impaired Uptake of Apoptotic Cells in MFG-E8-Deficient Mice

See allHide authors and affiliations

Science  21 May 2004:
Vol. 304, Issue 5674, pp. 1147-1150
DOI: 10.1126/science.1094359


Apoptotic cells expose phosphatidylserine and are swiftly engulfed by macrophages. Milk fat globule epidermal growth factor (EGF) factor 8 (MFG-E8) is a protein that binds to apoptotic cells by recognizing phosphatidylserine and that enhances the engulfment of apoptotic cells by macrophages. We report that tingible body macrophages in the germinal centers of the spleen and lymph nodes strongly express MFG-E8. Many apoptotic lymphocytes were found on the MFG-E8–/– tingible body macrophages, but they were not efficiently engulfed. The MFG-E8–/– mice developed splenomegaly, with the formation of numerous germinal centers, and suffered from glomerulonephritis as a result of autoantibody production. These data demonstrate that MFG-E8 has a critical role in removing apoptotic B cells in the germinal centers and that its failure can lead to autoimmune diseases.

Apoptosis is a process for removing harmful or useless cells and is fundamental to the maintenance of mammalian homeostasis (1). Apoptotic cells are rapidly engulfed by phagocytes, a process that should prevent inflammation and the autoimmune response against intracellular antigens that can be released from the dying cells (24). Phosphatidylserine (PS) that is exposed on the surface of apoptotic cells has been identified as a recognition signal for macrophages (5), and several receptors that bind PS have been identified (2, 3). MFG-E8 is secreted from activated macrophages, specifically binds to apoptotic cells by recognizing PS, and enhances the engulfment of apoptotic cells by phagocytes (6).

To examine the in vivo role of MFG-E8, we first examined its expression by Northern blot hybridization (fig. S1). MFG-E8 was strongly expressed in mammary glands (7, 8). Several other tissues such as the spleen, lymph nodes, and brain also expressed MFG-E8 mRNA. Immunohistochemical analysis of spleen sections (9) showed the labeling with antibody to MFG-E8 in cells localized to germinal centers that were stained by peanut agglutinin (PNA) (Fig. 1). Spleen sections from MFG-E8–/– mice (see below) were not stained, which confirmed the specificity of the antibody. Germinal centers contain macrophages called “tingible body” macrophages, which specifically express CD68 but do not express F4/80 (10). Dual staining for MFG-E8 and CD68 or F4/80 showed that the cells expressing MFG-E8 expressed CD68 but did not express F4/80 [Fig. 1, and (11)], which suggests that the tingible body macrophages expressed MFG-E8. The CD68+ tingible body macrophages are present not only in the spleen but also in the lymph nodes (10). Staining with antibody against MFG-E8 indicated that the CD68+ macrophages in the lymph nodes also expressed MFG-E8 (Fig. 1).

Fig. 1.

Expression of MFG-E8 in macrophages in germinal centers. Spleen and lymph node (LN) sections from 10-week-old MFG-E8+/+ (WT) or MFG-E8–/– (KO) mice were stained with antibodies against CD68 (green) and MFG-E8 (red) antibodies, and their staining profiles were merged in the third column. In the fourth column, the adjacent sections were stained with PNA (green) and antibody to MOMA-1 (red), a surface antigen expressed in marginal metallophilic macrophages (14). Scale bar, 100 μm.

We generated MFG-E8–/– mice by gene targeting. The murine MFG-E8 gene is encoded by 10 exons within 16 kb of genomic DNA on chromosome 7. A targeting vector was constructed by replacing exons 4 to 6 with the neo-resistance gene (fig. S2). The vector was introduced into mouse embryonic stem (ES) cells, and ES clones carrying the mutation were identified by polymerase chain reaction (PCR). Mice derived from two ES clones had identical phenotypes, and those from one representative clone were characterized in detail. Thioglycollate activates macrophages (12), and we previously showed that peritoneal macrophages obtained after an intraperitoneal injection of thioglycollate secrete abundant MFG-E8 (6). Accordingly, immunoprecipitation and Western blotting with antibody against MFG-E8 revealed a 74-kD protein in the culture supernatant of thioglycollate-elicited peritoneal macrophages from wild-type mice (Fig. 2A). This protein was not detected in the supernatant of MFG-E8–/– macrophages, which confirmed that the MFG-E8–/– allele is a null allele for MFG-E8. To examine the ability of thioglycollate-elicited macrophages to engulf apoptotic cells, we used as prey apoptotic thymocytes from CAD (caspase-activated DNase)–deficient mice. CAD–/– thymocytes do not undergo DNA degradation when exposed to apoptotic stimuli (13). However, when the CAD–/– thymocytes were induced to apoptosis and incubated with MFG-E8+/+ macrophages, the thymocytes became TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling)–positive, and the phagocytosis index (the number of engulfed apoptotic cells per macrophage) was 1.20 (Fig. 2, B and C), which indicated that the MFG-E8+/+ macrophages efficiently engulfed the apoptotic cells and digested their DNA. However, macrophages from MFG-E8–/– mice engulfed a few apoptotic cells, and the phagocytosis index was 0.31. Addition of recombinant MFG-E8 to the cultures restored the ability of MFG-E8–/– macrophages to engulf apoptotic cells in a dose-dependent manner (Fig. 2C). A similar result, showing a requirement for MFG-E8, was obtained with the apoptotic CAD+/+ wild-type thymocytes as prey (fig. S3, A and C). Nevertheless, MFG-E8 was not required for engulfment of microspheres (fig. S3, B and C), which indicated that the phagocytic activity of the macrophages was not impaired by the deficiency of MFG-E8.

Fig. 2.

Targeted disruption of the MFG-E8 gene. (A) No expression of MFG-E8 protein in MFG-E8–deficient macrophages. Thioglycollate-elicited peritoneal macrophages from 12-week-old MFG-E8+/+ or MFG-E8–/– mice were cultured for 48 hours. Proteins from cell lysates and culture supernatants were immunoprecipitated with the 2422 antibody to MFG-E8 and Western blotted with the 18A2 antibody to MFG-E8. (B) Engulfment of apoptotic cells by macrophages. Thioglycollate-elicited peritoneal macrophages from MFG-E8+/+ and MFG-E8–/– mice were cultured with apoptotic thymocytes from CAD-null mice in the absence or presence of 0.1 μg/ml MFG-E8, stained for TUNEL (brown), and observed by light microscopy. Magnification, ×400. (C) Phagocytosis index. Engulfment of apoptotic cells by MFG-E8+/+ (+/+) or MFG-E8–/– (–/–) macrophages in the absence (–) or presence (0.03, 0.1, and 0.3 μg/ml) of MFG-E8 was carried out as described above. The experiments were done three times, and the average value (±SD) of the phagocytosis index (the number of engulfed apoptotic cells per macrophage) is shown. (D) Splenomegaly in MFG-E8–null mice. Photographed spleens of 40-week-old MFG-E8+/+ and MFG-E8–/– mice. Scale bar, 1 cm. (E) Age-dependent splenomegaly. The spleens of 10- or 40-week-old MFG-E8+/+ (white bar) and MFG-E8–/– mice (black bar) were weighed, and the average values (±SD) from 10 mice are shown. The probability of statistical differences was determined by Student's t test. P < 0.01. (F) Histology of the spleen. Spleen sections from 40-week-old MFG-E8+/+ and MFG-E8–/– mice were stained with hematoxylin and eosin. Scale bar, 100 μm. (G and H) Immunohistochemical analysis of the spleen. The spleen sections of 40-week-old MFG-E8+/+ and MFG-E8–/– mice were stained with PNA (green) and antibody against MOMA-1 (red) (G), or with antibodies against IgG (green) and MOMA-1 (red) (H).

Homozygous MFG-E8–/– and heterozygous MFG-E8+/– animals were represented according to normal Mendelian inheritance. MFG-E8–/– mice were fertile with normal fecundity (11). However, MFG-E8–/– mice developed splenomegaly in an age-dependent manner (Fig. 2D). At the age of 40 weeks, the spleen of MFG-E8–/– mice was about 3 times heavier than that of wild-type mice (Fig. 2E). Histochemical analysis of the spleen indicated that the white pulps were greatly enlarged in the MFG-E8–/– mice and carried numerous germinal centers that were stained with PNA (Fig. 2, F and G). Metallophilic macrophages, defined by the marker MOMA-1, are located at the border of the marginal and follicular zones of the white pulp (14). The staining of spleen sections with antibody against MOMA-1 confirmed that the follicular zones were enlarged in the spleen of MFG-E8–/– mice (Fig. 2G). An increased number of immunoglobulin G (IgG)–producing cells were found not only in the follicular zones but also in the marginal zones of MFG-E8–/– mice (Fig. 2H). There were 2 to 3 times more lymphocytes in the MFG-E8–/– spleen than in the wild-type spleen, but the ratio of B cells to T cells in the spleen was similar between the MFG-E8+/+ and MFG-E8–/– mice (fig. S4).

Tingible body macrophages engulf apoptotic B cells generated in the germinal center (15). In the spleen of 40-week-old MFG-E8+/+ mice, few TUNEL-stained cells (0.6 cells per macrophage) were associated with macrophages expressing CD68 in the germinal center (Fig. 3A). In MFG-E8–/– mice, the CD68-expressing cells were larger than those of wild-type mice and were associated with many TUNEL-stained cells (3.6 cells per macrophage). The enlargement of the CD68-expressing macrophages and their enhanced association with TUNEL-stained cells in the MFG-E8–/– spleen became apparent (10.3 versus 2.1 cells per macrophage) when B cells were activated in mice immunized with keyhole limpet hemocyanin (KLH) (Fig. 3A). CD68 was predominantly located intracellularly within late endosomes or lysosomes (16). Accordingly, when a macrophage carrying TUNEL-stained cells in the spleen of KLH-treated MFG-E8–/– mice was analyzed by confocal microscopy (Fig. 3B), CD68 was found intracellularly. The TUNEL-stained material, that is, the nuclei of apoptotic cells, was not localized with CD68, but seemed to be present extracellularly on the surface of macrophages. To confirm the extracellular localization of the apoptotic material associated with the MFG-E8–/– macrophages, the spleen sections were analyzed by electron transmission microscopy. Tingible body macrophages in wild-type mice had condensed nuclei from apoptotic cells inside them, and some of the apoptotic materials were degraded (Fig. 3C). In the MFG-E8–/– spleen, apoptotic cells were also localized to the tingible body macrophages, but they had intact plasma membranes. These results indicated that these apoptotic cells were located outside the MFG-E8–/– macrophages or that the macrophages just wrapped many apoptotic cells without engulfment, which may explain the apparent enlargement of the germinal center macrophages in MFG-E8–/– mice. Lymphocytes prepared from the spleens of MFG-E8–/– mice underwent apoptotic cell death as efficiently as those prepared from the wild-type mice, and PS was exposed on their surface (fig. S5). From these results, we conclude that the ability of tingible body macrophages to engulf apoptotic cells appeared to be impaired by the lack of MFG-E8.

Fig. 3.

Impaired engulfment of apoptotic cells by MFG-E8–deficient tingible body macrophages in the germinal center. (A) Immunohistochemical analysis of the spleen. Spleen sections were prepared from 40-week-old MFG-E8+/+ or MFG-E8–/– mice, or KLH-immunized 10-week-old mice, and stained with antibody against CD68 (red) and TUNEL (green); staining profiles were merged in the third column. Scale bar, 100 μm. Enlarged images of the dotted region are shown in the fourth column. Scale bar, 25 μm. Number of TUNEL-positive cells associated with tingible body macrophages was counted for 50 to 100 macrophages, and is presented as the number per macrophages at the upper corner of panels in the third column. (B) A confocal image of a tingible body macrophage. A CD68-containing (red) tingible body macrophage associated with TUNEL-positive cells (green) located in the spleen of KLH-immunized 10-week-old MFG-E8–/– mice was observed by confocal microscopy (LSM510, Zeiss), and a three-dimensional image was produced using a three-dimensional projection program. (C) An analysis of tingible body macrophages by EM. Spleen sections from KLH-immunized 10-week-old MFG-E8+/+ or MFG-E8–/– mice were analyzed by EM. Tingible body macrophages (TBMs) and remnants of apoptotic cells (arrows) are labeled. Scale bar, 2 μm. The right panel shows the tingible body macrophages with unengulfed apoptotic cells in MFG-E8–/– mice at higher magnification. Scale bar, 2 μm.

Repeated systemic injection of apoptotic cells induces autoantibody production (17). The MFG-E8–/– mice spontaneously produced autoantibodies in an age-dependent manner. The concentrations of antibodies to double-stranded DNA (dsDNA) or nuclear proteins (ANA, antinuclear antibody) in the serum of MFG-E8–/– mice were not different from those of wild-type mice at the age of 10 weeks (Fig. 4A). However, at the age of 40 weeks, MFG-E8–/– mice had high concentrations of dsDNA antibody and ANA in their serum. In particular, about half of MFG-E8–/– female mice carried more than 1000 times as much ANA in their serum as the wild-type mice. When female mice at the age of 10 weeks were immunized twice by KLH, MFG-E8–/– mice developed ANA in 20 days, whereas no ANA was found in the serum of MFG-E8+/+ mice under the same conditions (Fig. 4A). This finding indicates that the inefficient engulfment of apoptotic B cells can lead to the development of autoimmunity. A major pathogenic consequence of circulating autoantibodies is the deposition of immune complex within the kidney, leading to glomerulonephritis (18). The examination of kidney cryosections by immunofluorescence revealed that all glomeruli in six different 40-week-old MFG-E8–/– female mice had a massive deposition of IgG (Fig. 4B), which was accompanied by hypercellularity of the glomeruli in 84% cases (Fig. 4C). Accordingly, most of the MFG-E8–/– female mice (8 out of 10 mice) at the age of 40 weeks suffered from high concentration of protein in the urine (Fig. 4D).

Fig. 4.

Development of glomerulonephritis in MFG-E8–/– mice. (A) Production of autoantibodies. The titers of the dsDNA-specific antibody (left) and ANA (middle) in the serum of male (M) and female (F) mice (eight mice for each group) at the age of 10 or 40 weeks (w) are plotted with the average (bar). Right, 10-week-old MFG-E8+/+ and MFG-E8–/– female mice (eight mice for each group) were immunized twice with KLH, and the ANA titer was determined at day 20. P < 0.001, *P < 0.01 and **P < 0.05. (B) Accumulation of immune complex in glomeruli. Cryosections of kidneys from 40-week-old MFG-E8+/+ and MFG-E8–/– female mice were stained with Cy3-conjugated antibody to mouse IgG (Jackson) (red). Scale bar, 100 μm. (C) Kidney sections from 40-week-old MFG-E8+/+ and MFG-E8–/– female mice were fixed with paraformaldehyde and stained with periodic acid-Schiff (PAS). (D) High concentrations of protein in the urine. Urinary albumin levels per milligram of urinary creatinine were determined in 40-week-old MFG-E8+/+ and MFG-E8–/– female mice, and their average values ± SD from 10 mice are shown. P < 0.01.

The removal of apoptotic cells is usually very efficient, and few apoptotic cells can be found even in tissues with a very high rate of apoptosis (19). Although several molecules in macrophages have been proposed as receptors for apoptotic cells or bridging molecules between apoptotic cells and macrophages (2), the physiological roles of these molecules have been elusive. In the germinal centers, B cells with a low affinity for the antigen undergo apoptosis (20). Tingible body macrophages are responsible for the clearance of these apoptotic cells, and our results using MFG-E8–/– mice indicate that MFG-E8 has an essential role in this process. The MFG-E8–deficient tingible body macrophages were, however, associated with apoptotic cells, which suggests that the MFG-E8–/– macrophages could still recognize apoptotic cells. The molecules that mediate this tethering step (21) remain to be identified. We recently found that MFG-E8 is expressed in immature dendritic cells such as Langerhans cells (22). But it was not expressed in the thymic macrophages and was not apparently involved in the clearance of apoptotic thymocytes (fig. S6), which suggests that macrophages in different tissues may use different molecules for the engulfment of apoptotic cells (2325). Finally, human patients with systemic lupus erythematosus often have a defect in the engulfment of apoptotic cells by the tingible body macrophages in the germinal centers (26). The autoimmune properties observed in MFG-E8–/– mice may provide a good model system for the study of human systemic lupus erythematosus.

Supporting Online Material

Materials and Methods

Figs. S1 to S6


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article