HMGB1 inhibits macrophage activity in efferocytosis through binding to the α_vβ₃-integrin

Abstract

Phagocytosis of apoptotic cells is critical to resolution of inflammation. High mobility group box 1 protein (HMGB1), a mediator of inflammation, has been shown to diminish phagocytosis through binding to phosphatidylserine (PS) exposed on the surface of apoptotic neutrophils. However, it is currently unknown whether HMGB1 also modulates the activity of receptors involved in PS recognition on the surface of phagocytes. In the present studies, we found that preincubation of macrophages with HMGB1 decreased their ability to engulf apoptotic neutrophils or thymocytes. Preincubation of macrophages with HMGB1 prevented the enhancement of efferocytosis resulting from exposure to milk fat globule EGF factor 8 (MFG-E8), an opsonin that bridges PS and α_vβ₃ as well as α_vβ₅-integrins on the surface of phagocytes. The inhibitory effect of HMGB1 on the phagocytic activity of macrophages was prevented by preincubation of HMGB1 with soluble α_vβ₃, but not with soluble α_vβ₅. HMGB1 colocalized with the β₃-integrin on the cell membrane of macrophages and bound to soluble α_vβ₃, but not to soluble α_vβ₅. HMGB1 suppressed the interaction between MFG-E8 and α_vβ₃. HMGB1 also inhibited intracellular signaling events, including ERK phosphorylation and Rac-1 activation, which are activated in macrophages during phagocytosis of apoptotic cells. These results demonstrate that HMGB1 blocks α_vβ₃-dependent recognition and uptake of apoptotic cells.

phagocytosis of apoptotic cells, called efferocytosis, is an essential feature of immune response and critical to resolution of inflammation. Clearance of cells undergoing apoptosis protects surrounding tissue from exposure to proinflammatory intracellular contents released from necrotic cells (19). Ingestion of apoptotic cells also diminishes local inflammation by inducing release of antiinflammatory mediators, including IL-10 and transforming growth factor-β, from phagocytes and through suppressing the production of proinflammatory cytokines by macrophages (15, 26). Impaired clearance of dying cells is associated with autoimmune and inflammatory diseases including acute lung injury, cystic fibrosis, and systemic lupus erythematosus (5, 21). One of the early characteristics of apoptosis is exposure of phosphatidylserine (PS) on the outer leaflet of the cell membrane (32). PS serves as an “eat-me” signal allowing the apoptotic cell to be recognized by phagocytes (22). Opsonins, such as milk fat globule EGF factor 8 (MFG-E8), serve to bridge PS with receptors, including integrins, on the membrane of macrophages and other phagocytic cells, facilitating the ingestion of apoptotic cells (9, 17, 18).

High mobility group box 1 protein (HMGB1) was originally described as a nuclear nonhistone DNA-binding protein (6), but it can also be released into the extracellular milieu. Plasma and tissue levels of HMGB1 are elevated for prolonged periods in chronic and acute inflammatory conditions, including rheumatoid arthritis, acute lung injury, burns, severe infection, and hemorrhage. HMGB1 participates in enhancing inflammatory reactions by potentiating the activity of proinflammatory mediators such as LPS and cytokines, inducing leukocyte accumulation through its direct chemotactic properties, and also by diminishing phagocytosis of apoptotic neutrophils (7, 27, 34, 38, 40).

In previous studies we found that HMGB1 induces inhibitory effects on efferocytosis through binding to PS expressed on apoptotic neutrophils (27). However, those experiments did not examine or identify potential receptors on the macrophage surface that interact with HMGB1 after its binding to PS and that directly participate in diminishing the uptake of apoptotic neutrophils. Bridging molecules, including MFG-E8 and other opsonins, are generally required for PS to interact with ligands on the macrophage surface (9, 17, 18). Our previous studies therefore suggested that the ability of HMGB1 to inhibit phagocytosis might arise either by competing with opsonins for binding to PS or from blocking interactions between opsonins and macrophage receptors. We examined these possibilities in the present experiments and found that HMGB1 can prevent efferocytosis through effectively preventing the binding of the opsonin MFG-E8 to α_Vβ₃-integrins on the macrophage surface.

MATERIALS AND METHODS

Reagents.

Custom cocktail antibodies and negative selection columns for neutrophil isolation were from Stem Cell Technologies. Penicillin-streptomycin and Brewer thioglycollate were from Sigma-Aldrich. Rabbit anti-HMGB1 polyclonal antibodies were from Abcam. Mouse anti human integrin α_vβ₃ or α_vβ₅ antibody was from Millipore. Recombinant mouse MFG-E8 and recombinant human α_vβ₃-integrin were from R&D Systems. Recombinant mouse α_vβ₅ was from Millipore. Purified HMGB1 from pig thymus was a gift from Akitoshi Ishizaka (Department of Medicine, Keio University, School of Medicine, Tokyo, Japan) and had no detectable LPS (1, 29).

Isolation and induction of apoptosis in neutrophils.

Male C57BL/6 mice, 8 wk of age, were purchased from NCI-Frederick. All animal protocols were approved by the Institutional Animal Care and Use Committee at the University of Alabama at Birmingham. Mouse neutrophils were purified from bone marrow cell suspensions essentially as described previously (29). In brief, bone marrow cells were incubated with 20 μl of primary antibody (Ab) cocktail specific to the cell surface markers F4/80, CD4, CD45R, CD5, and TER119 for 15 min at 4°C. Anti-biotin tetrameric Ab complexes (100 μl) were then added into the cells and incubated for 15 min at 4°C followed by incubation with 60 μl of colloidal magnetic dextran iron particles for 15 min at 4°C. The entire cell suspension was then placed into a column surrounded by a magnet. The T cells, B cells, red blood cells, monocytes, and macrophages were captured in the column, allowing the neutrophils to pass through as a result of negative selection. The cells were then pelleted and washed. Neutrophil purity, as determined by HEMA 3-stained cytospin preparations, was >97%. Cell viability, as determined by Trypan blue exclusion, was consistently >98%. Apoptosis was induced by heating 6 × 10⁶ cells/ml of serum-free RPMI media at 43°C for 60 min and followed by incubation at 37°C in 5% CO₂ for 150 min.

Induction of apoptosis in thymocytes.

To induce apoptosis, murine thymocytes were resuspended in serum-free RPMI at a concentration of 6 × 10⁶ cells/ml and exposed to UV (302 nm) light for 5 min, followed by incubation at 37°C in 5% CO₂ for 60 min. More than 90% of thymocytes were apoptotic, as demonstrated by annexin V staining, after being treated in this manner.

Isolation and preparation of mouse peritoneal macrophages.

Peritoneal macrophages were elicited from 8- to 10-wk-old mice by intraperitoneal injection of 1 ml of 3% thioglycollate. Cells were harvested 4 days later by peritoneal lavage. Macrophages (0.3 × 10⁶) were plated on coverslips in 24-well plates in RPMI media containing 5% FBS (Atlanta Biologics). After 1 h at 37°C, nonadherent cells were removed by washing with medium. Macrophages were cultured in RPMI media containing 5% FBS at 37°C and maintained at the same condition by changing the media every 3 days. One hour before the phagocytosis assay, the macrophages were washed several times with fresh serum-free medium.

Isolation and preparation of bone marrow derived macrophages.

Bone marrow cells were obtained from 8- to 10-wk-old mice. Cells were cultured on petri dishes in DMEM-F-12 medium containing 15% FBS and 25 ng/ml granulocyte/macrophage colony-stimulating factor. After 5 days, nonadherent cells were removed by washing and fresh media were added. Cells were allowed to grow for 5 more days, then were trypsinized and plated on coverslips for phagocytosis assays.

In vitro efferocytosis assays.

Phagocytosis of apoptotic neutrophils by macrophages (efferocytosis) was determined by adding 3 × 10⁶ apoptotic neutrophils suspended in 600 μl RPMI medium to each well of a 24-well plate containing macrophage monolayers on coverslips followed by incubation at 37°C for 120 min. To determine the effects of HMGB1, MFG-E8, α_vβ₃, or α_vβ₅, macrophages were preincubated with these proteins at 37°C for 30 min at the indicated concentrations in serum-free RPMI media and the media were then removed before the addition of apoptotic neutrophils for phagocytosis assays. FBS was included at a final concentration of 0.5% during the incubation of macrophages with apoptotic neutrophils. Noningested neutrophils were removed by washing three times with ice-cold PBS. Cells on coverslips were fixed in 100% methanol and then stained with HEMA 3. Phagocytosis was evaluated by counting at least 300 macrophages per slide from duplicate experiments. The phagocytosis index was calculated as the ratio of macrophages containing at least one ingested neutrophil. The fold change in phagocytic index was the ratio of phagocytic index of control groups to experimental groups, where the phagocytic index of the control group was regarded as 1.

Expression and purification of recombinant Flag-tagged HMGB1 from mammalian cells.

Full-length human HMGB1 cDNA was purchased from Open Biosystems and cloned into a mammalian expressing vector, pcDNA3-CMV10–3×FLAG (Sigma). Human embryonic kidney cells (HEK 293) were transfected with the recombinant DNA using Lipofectamine 2000 reagent. After 24 h of transfection, the cells were trypsinized and washed with PBS for protein isolation. The cells were then lysed using mammalian cell lysis buffer [50 mM Tris, pH 8.0, 5 mM EDTA, 100 mM NaCl, 10 mM sodium fluoride, 2 mM dithiothreitol, 1 mM Na₃VO₄, and 0.5% Nonidet P-40 and proteinase inhibitor mixture (Sigma)]. Cellular protein was affinity purified using anti-Flag M2 agarose beads. The beads were washed three times with mammalian cell lysis buffer and twice with buffer B (50 mM Tris, pH 8.0, and 1 mM dithiothreitol) and then eluted with buffer B containing 400 μg/ml 3×Flag peptide (Sigma). 3×Flag peptide was removed by concentrating and purifying the sample with Microcon YM-30 columns (Millipore). Protein concentration was determined by protein assay kit (Bio-Rad).

Coimmunoprecipitation.

HMGB1 or Flag-tagged HMGB1 (HMGB1-Flag; 50 ng) was incubated with 100 ng of α_vβ₃ or α_vβ₅ protein in 0.5 ml of buffer containing 50 mM Tris (pH 7.4), 150 mM NaCl, 10% (vol/vol) glycerol, and 1 mM DTT for 1 h at 4°C, then 1 μg anti-α_vβ₃ or anti-α_vβ₅ antibody was added to each sample and incubated overnight at 4°C. The samples were incubated for additional 2 h with 30 μl protein A-Sepharose-4B beads. The immunocomplexes were resolved by 15% SDS-PAGE and analyzed by Western blotting using anti-HMGB1 or anti-Flag antibodies.

Coimmunoprecipitation assay to detect competition of HMGB1 with MFG-E8.

HMGB1-Flag (50 ng) or MFG-E8 (100 ng) was incubated with 100 ng of α_vβ₃ protein for 30 min and then 100 ng of MFG-E8 or 50 ng HMGB1 were added to the samples and incubated for additional 30 min. α_vβ₃-bound HMGB1-Flag was then immunoprecipitated by anti-α_vβ₃ or anti-α_vβ₅ antibodies and detected by Western blotting using anti-Flag antibodies. In each assay, replacement of HMGB1-Flag or MFG-E8 with 100 ng BSA was used as a negative control.

Confocal microscopy.

Peritoneal macrophages grown on coverslips were preincubated with 1 μg/ml HMGB1-Flag for 1 h. Cells were then fixed with 4% formaldehyde in PBS for 5 min. After being washed with PBS and blocked with 5% BSA in PBS, the cells were incubated with anti-Flag and anti-integrin-β₃ antibodies at 4°C overnight. After being washed three times with PBS, the cells were incubated with Alexa Fluor 488- or Alexa Fluor 594-conjugated secondary antibodies for 1 h at room temperature. The coverslips were than mounted on slides, and the localization of HMGB1 and integrin-β₃ was examined by a confocal microscope (Nikon).

Determination of Rac-1 activation.

GTP-bound Rac-1 (active Rac-1) was pulled down by p21 binding domain-conjugated agarose beads (Millipore). Rac-1 precipitated by the beads was resolved by 12% SDS-PAGE and detected by anti-Rac-1 antibody.

ELISA assay.

The concentrations of TNF-α and IL-6 in the supernatants of macrophages that were treated with BSA or HMGB1 were determined by ELISA assays according to the manufacturer's instructions (R&D Systems).

Statistical analysis.

Data are presented as means ± SD for each experimental group. One-way ANOVA followed by analysis with the Tukey-Kramer test was performed for comparisons among multiple groups, and Student's t-test was used for comparisons between two groups. A value of P < 0.05 was considered significant.

RESULTS

Preincubation of macrophages with HMGB1 inhibits their ability to engulf apoptotic neutrophils.

In our previous experiments, we found that exposure of apoptotic neutrophils to HMGB1 inhibits efferocytosis and that HMGB1 binds to PS exposed on the surface of apoptotic neutrophils (27). However, those experiments did not examine potential interactions of HMGB1 with macrophage receptors, such as integrins, that occupy central roles in modulating efferocytosis (13). A recent study demonstrated that HMGB1 is involved in integrin, such as Mac-1, mediated migration of hematopoietic cells (28). As various integrins, particularly α_vβ₃ and α_vβ₅, have been shown to participate in efferocytosis (13), we hypothesized that HMGB1 might inhibit efferocytosis through interactions with macrophage-based integrins. To test this hypothesis, peritoneal macrophages were preincubated with HMGB1 before exposure to apoptotic neutrophils.

As shown in Fig. 1A, HMGB1 dose dependently inhibited the ability of macrophages to phagocytose apoptotic neutrophils. Figure 1B shows a representative field where macrophages that were preincubated with BSA or HMGB1 were then cultured with apoptotic neutrophils. In our previous study (27), we found that preincubation of apoptotic neutrophils with at least 100 ng/ml HMGB1 was required to diminish their uptake by macrophages. The decreased phagocytosis in the present experiments of apoptotic neutrophils by macrophages exposed to HMGB1 is therefore unlikely to be caused by binding of residual HMGB1 to apoptotic neutrophils since the HMGB1 containing medium used for preincubation was removed from the cell cultures and replaced with apoptotic neutrophils in medium without HMGB1 before initiation of the phagocytosis assays. To rule out a possibility that the observed inhibitory effects of HMGB1 were caused by contaminants in thymus-derived HMGB1, we repeated the experiments with recombinant HMGB1-Flag purified from HEK 293 cells and found that preincubation of macrophages with HMGB1-Flag also significantly decreased their ability to phagocytose apoptotic neutrophils (Fig. 1C). These data suggest that HMGB1 regulates the phagocytic activity of macrophages independent of its binding to PS or other receptors on the surface of apoptotic cells. Neither the thymus-derived HMGB1 nor the recombinant HMGB1-Flag purified from HEK 293 cells activated macrophages to produce proinflammatory cytokines, such as TNF-α and IL-6, as determined by ELISA (data not shown). Of note, the concentrations of HMGB1 used in these experiments, i.e., 0.1–1 μg/ml, and that resulted in significant decreases in phagocytosis of neutrophils by peritoneal macrophages are similar to those found in the circulation during acute inflammatory conditions, such as severe infection (1, 3, 35, 42).

Fig. 1.

Preincubation of macrophages with high-mobility group box 1 protein (HMGB1) inhibits their ability to engulf apoptotic neutrophils, but not other targets. A, C, and D: peritoneal (A and C) or bone marrow-derived (D) macrophages were washed five times with serum-free medium and exposed for 30 min to 0.1, 0.5, and 1 μg/ml HMGB1 purified from porcine thymus (A and D), or to 1 μg/ml Flag-tagged HMGB1 purified from human embryonic kidney (HEK)-293 cells transiently transfected with Flag-tagged HMGB1 (HMGB1-Flag)-expressing constructs (C). The medium was then removed and 3 × 10⁶ apoptotic neutrophils were added in a total volume of 600 μl for 90 min. Phagocytosis assays were performed as described in materials and methods. As a control, macrophages were pretreated with 1 μg/ml bovine serum albumin (BSA). *P < 0.05, compared with controls. Independent experiments were performed at least three times with similar results. B: the experiments were performed as in A with 1 μg/ml BSA or HMGB1. A representative field demonstrating efferocytosis is shown. Arrows point to engulfed neutrophils. E: experiments were performed as in B except that apoptotic neutrophils were replaced by apoptotic thymocytes. F: peritoneal macrophages were washed five times with serum-free medium and exposed for 30 min to Chromeo642-labeled HMGB1 or BSA (1 μg/ml). The cells were then washed five times with serum-free medium and flow cytometry was performed. G: peritoneal macrophages were washed five times with serum-free medium and then exposed for 30 min to 1 μg/ml BSA, 1 μg/ml HMGB1, or 1 μg/ml Chromeo642-labeled HMGB1. The medium was removed, and 3 × 10⁶ apoptotic neutrophils were added in a total volume of 600 μl for 90 min for phagocytosis assays. **P < 0.01, compared with controls. The experiments were performed twice with similar results.

To determine whether the inhibitory effect of HMGB1 on the phagocytic activity of macrophages is limited to peritoneal macrophages or involves other macrophage populations as well, we examined the ability of bone marrow-derived macrophages to ingest apoptotic neutrophils. Consistent with the results from peritoneal macrophages, preincubation of bone marrow-derived macrophages with HMGB1 decreased their ability to phagocytose apoptotic neutrophils (Fig. 1D). To determine whether HMGB1 decreases the ability of macrophages to phagocytose apoptotic cells other than neutrophils, we performed a similar experiment with apoptotic thymocytes. As shown in Fig. 1D, preincubation of HMGB1 decreased the ability of peritoneal macrophages to phagocytose apoptotic thymocytes. These data suggest that the inhibitory effect of HMGB1 on the activity of macrophages to phagocytose apoptotic cells is a general phenomenon.

To determine whether HMGB1 binds to the macrophage surface, we incubated Chromeo642-labeled HMGB1 with macrophages and found increased association of HMGB1 with the surface of macrophages, compared with Chromeo642-labeled BSA, which served as a negative control (Fig. 1E). Of note, similar to unlabeled HMGB1, Chromeo642-labeled HMGB1 decreased the ability of macrophages to phagocytose apoptotic neutrophils (Fig. 1G).

Preincubation of macrophages with HMGB1 abolishes the ability of MFG-E8 to enhance efferocytosis.

To delineate the mechanism by which HMGB1 regulates the phagocytic activity of macrophages, initial experiments were performed to examine if HMGB1 modulates efferocytosis mediated by α_vβ₃- and α_vβ₅-integrins. Previous studies demonstrated that α_vβ₃ and α_vβ₅ do not bind directly to PS on the surface of apoptotic cells, but rather interact with PS via the opsonin MFG-E8 which acts as a bridging molecule between PS and α_vβ₃ or α_vβ₅ (20, 25, 33, 45).

Incubation of macrophages with MFG-E8 enhanced their ability to phagocytose apoptotic neutrophils (Fig. 2A). However, preincubation of macrophages with HMGB1 decreased (compare bar 3 and bar 4), although did not abolish (compare bar 2 and bar 4) the enhanced phagocytosis mediated by MFG-E8 (Fig. 2A). These data suggest that HMGB1 and MFG-E8 may competitively bind to a common receptor, such as α_vβ₃ or α_vβ₅, and that preincubation of macrophages with HMGB1 blocks subsequent binding of MFG-E8 to the receptor. If this hypothesis is correct, incubation of macrophages with MFG-E8 before exposure to HMGB1 would render HMGB1 ineffective in affecting efferocytosis. To examine this question, macrophages were preincubated with MFG-E8 alone or incubated with MFG-E8 followed by exposure to HMGB1. As expected, preincubation with MFG-E8 enhanced the phagocytic activity of the macrophages (Fig. 2B). However, in contrast to the inhibitory effects of preincubation with HMGB1 before MFG-E8, exposure of macrophages to HMGB1 after preincubation with MFG-E8 did not affect the ability of MFG-E8 to enhance the phagocytic activity of macrophages (Fig. 2B).

Fig. 2.

Preincubation of macrophages with HMGB1 abolishes the enhancement of efferocytosis induced by milk fat globule EGF factor 8 (MFG-E8). A: preincubation of macrophages with HMGB1 abrogated MFG-E8-enhanced phagocytic activity of macrophages. Macrophages were exposed for 30 min to 1 μg/ml HMGB1 or 1 μg/ml BSA. The medium was then removed and replaced with fresh medium containing 1 μg/ml MFG-E8 or 1 μg/ml BSA for 30 min. Finally, the medium was removed and 3 × 10⁶ apoptotic neutrophils were added in a total volume of 600 μl for 90 min. Phagocytosis assays were performed. **P < 0.01, ***P < 0.001, compared with the control group that was pretreated with 1 μg/ml BSA. B: preincubation of macrophages with MFG-E8 abolishes the inhibitory effect of HMGB1 on the phagocytic activity of macrophages. Macrophages were exposed for 30 min to 1 μg/ml MFG-E8 or 1 μg/ml BSA. The medium was then removed and replaced with fresh medium containing 1 μg/ml HMGB1 or 1 μg/ml BSA for 30 min. Finally, the medium was removed and 3 × 10⁶ apoptotic neutrophils were added in a total volume of 600 μl for 90 min. Phagocytosis assays were performed. **P < 0.01, ***P < 0.001, compared with the control group that was only pretreated with 1 μg/ml BSA. Experiments were performed at least three independent times with similar results.

The inhibitory effects of HMGB1 on efferocytosis are abolished by preincubation with soluble α_vβ₃, but not with soluble α_vβ₅.

Interactions between MFG-E8 and the integrins α_vβ₃ and α_vβ₅ on the macrophage surface are responsible for the ability of MFG-E8 to enhance phagocytosis of apoptotic neutrophils (2, 4, 17). Given the ability of HMGB1 to block the enhanced efferocytosis mediated by MFG-E8, we hypothesized that HMGB1 might interact with α_vβ₃ and/or α_vβ₅ on the macrophage as does MFG-E8. To explore this hypothesis, initial experiments were performed to examine whether incubation with soluble α_vβ₃ or α_vβ₅, which include only the extracellular domain of α_vβ₃ or α_vβ₅, can modulate the effects of HMGB1 on efferocytosis.

Concentrations of α_vβ₃ of 0.5 μg/ml or less or of α_vβ₅ of 1 μg/ml or less had no detectable effects on efferocytosis (Fig. 3, A and C). However, incubation of 0.5 μg/ml α_vβ₃ or 1 μg/ml α_vβ₅ with MFG-E8 abolished the enhanced efferocytosis mediated by MFG-E8 (Fig. 3, B and D). These data suggest that soluble α_vβ₃ or α_vβ₅ binds to MFG-E8 and acts to prevent binding of MFG-E8 to α_vβ₃ or α_vβ₅ on the macrophage surface. We next determined whether incubation with soluble α_vβ₃ or α_vβ₅ had similar abilities to prevent the inhibitory effects of HMGB1 on efferocytosis. As shown in Fig. 4A, incubation of 0.5 μg/ml α_vβ₃ with HMGB1 abolished the inhibitory effects of HMGB1 on the phagocytic activity of macrophages. These findings suggest that binding between soluble α_vβ₃ and HMGB1 prevents HMGB1 binding to α_vβ₃ on the membrane of macrophages. However, the same concentration of α_vβ₅, i.e., 1 μg/ml, that blocked the potentiating activity of MFG-E8 on efferocytosis had no effect on that of HMGB1 (Fig. 4B). These data suggest that α_vβ₃ is specifically involved in the ability of HMGB1 to modulate macrophage uptake of apoptotic neutrophils.

Fig. 3.

Soluble integrin-α_vβ₃ or -α_vβ₅ abrogate MFG-E8-enhanced efferocytosis. A: macrophages were incubated with 1 μg/ml human serum albumin (HSA) or α_vβ₃ for 30 min. The medium was then removed, and 3×10⁶ apoptotic neutrophils were added to the macrophages. B: MFG-E8 was incubated with HSA or α_vβ₃ for 30 min in serum-free medium at room temperature, and the mixture of proteins was then added to macrophages and incubated for 30 min. The medium was then removed, and apoptotic neutrophils were added to the macrophages. C and D: experiments were performed essentially the same as in A and B except that α_vβ₃ was replaced by α_vβ_5. *P < 0.05, ***P < 0.001, compared with the group incubated with murine serum albumin (MSA) alone; †P < 0.05, compared with the control group that was pretreated with HSA alone. Experiments were performed at least three independent times with similar results.

Fig. 4.

α_vβ_3, but not α_vβ_5, abolishes the inhibitory effect of HMGB1 on efferocytosis. A: 1 μg/ml HMGB-1 was incubated with increasing doses of soluble α_vβ₃ or 1 μg/ml HSA for 30 min. The mixture of proteins was then added to macrophages for 30 min and then removed. Apoptotic neutrophils (3 × 10⁶) were added to the macrophages for assessment of the phagocytic index. B: the experiment was the same as in A except α_vβ₃ was replaced by α_vβ_5. ***P < 0.001, compared with the group that only preincubated with BSA.

HMGB1 binds to α_vβ₃ but not to α_vβ₅.

Given that soluble α_vβ₃ abolished the inhibitory effects of HMGB1 on the phagocytic activity of macrophages, we next explored whether soluble α_vβ₃ binds to HMGB1 and acts to prevent binding of HMGB1 to α_vβ₃ on the membrane of macrophages. To do this, we incubated soluble α_vβ₃ with purified HMGB1 from thymus or purified HMGB1-Flag and then immunoprecipitated α_vβ₃ with anti-α_vβ₃ antibodies. We found that HMGB1 does interact with α_vβ₃ (Fig. 5A). To determine whether HMGB1 colocalizes with integrin-α_vβ₃ on the macrophage surface, we performed confocal microscopy and found that HMGB1 did indeed colocalize with the β₃-integrin (Fig. 5B). In addition, we found that, in contrast to α_vβ_3, there were no apparent interactions between α_vβ₅ and HMGB1 (Fig. 5C). These data suggest that soluble α_vβ₃, but not α_vβ₅, specifically binds to HMGB1 thereby preventing binding of HMGB1 to α_vβ₃ on the membrane of macrophages; this finding is consistent with our results that incubation of HMGB1 with soluble α_vβ₃, but not with soluble α_vβ₅, is able to abolish the inhibitory effects of HMGB1 on phagocytic activity of macrophages (Fig. 4).

Fig. 5.

HMGB1 binds to α_vβ_3, but not to α_vβ₅. A: HMGB1 or HMGB1-Flag (500 ng) was incubated with 1 μg of recombinant soluble α_vβ₃ for 1 h. Anti-α_vβ₃ antibody (2 μg/ml) was then added to the mixture of the two proteins and incubated overnight. The immunocomplexes were pulled down by protein A-Sepharose-4B beads and resolved by 15% SDS-PAGE. HMGB1 in the complexes was analyzed by Western blotting using anti-HMGB1 antibodies. IP, immunoprecipitation. B: peritoneal macrophages grown on coverslips were preincubated with 1 μg/ml HMGB1-Flag for 1 h. The cells were then fixed with 4% formaldehyde, blocked with 5% BSA in PBS, and incubated with anti-Flag and anti-integrin-β₃ antibodies at 4°C overnight, followed by incubation with Alexa Fluor 488- or Alexa Fluor 594-conjugated secondary antibodies for 1 h at room temperature. Confocal microscopy analysis was then performed. C: HMGB1 or HMGB1-Flag (500 ng) was incubated with 1 μg of recombinant soluble α_vβ₅ for 1 h. Anti-α_vβ₅ antibody (2 μg/ml) was then added to the mixture of the two proteins and incubated overnight. The immunocomplexes were pulled down by protein A-Sepharose-4B beads and resolved by 15% SDS-PAGE. HMGB1 in the complexes was determined by Western blotting using anti-HMGB1 antibodies. Representative experiments are shown. Second experiments with independent samples produced similar results.

HMGB1 competes with MFG-E8 for binding to α_vβ₃.

In experiments, such as those shown in Fig. 2, we found that HMGB1 abrogates enhanced efferocytosis mediated by MFG-E8. To assess whether HMGB1 and MFG-E8 directly compete for binding to α_vβ₃, which appears to be the macrophage receptor for both proteins, HMGB1 was incubated with soluble α_vβ₃ followed by addition of MFG-E8 to the cultures. We found that the binding between HMGB1 and α_vβ₃ was not affected by the later addition of MFG-E8 (compare lanes 2 and 4 in Fig. 6). However, preincubation of α_vβ₃ with MFG-E8 prevented binding of α_vβ₃ to HMGB1 when added subsequently to the cultures (compare lanes 3 and 5 in Fig. 6). These data are consistent with our findings that the inhibitory effects of HMGB1 on the ability of macrophages to ingest apoptotic neutrophils was not reversed by subsequent exposure of the macrophages to MFG-E8 and that the enhancing effects of MFG-E8 on macrophage phagocytic activity was not reversed by the subsequent addition of HMGB1 to the macrophage cultures (Fig. 2).

Fig. 6.

HMGB1 and MFG-E8 compete for binding to α_vβ₃. Soluble α_vβ₃ (1 μg/ml) was preincubated with 50 ng/ml HMGB1-Flag (lanes 2 and 4) for 30 min. MFG-E8 (100 ng/ml; lane 2) or BSA (lane 4) was then added into the mixture of two proteins and incubated for another 30 min. Antibodies to α_vβ₃ were then added to each sample and incubated overnight. The immunocomplexes were precipitated with protein A-Sepharose-4B beads and resolved by 15% SDS-PAGE. The coprecipitated HMGB1-Flag was detected by anti-Flag antibody. Alternatively, 1 μg of soluble α_vβ₃ was preincubated with 100 ng MFG-E8 (lane 3) or 100 ng BSA (lane 5) for 30 min. HMGB1 (100 ng) was then added to the mixture of the two proteins and incubated for another 30 min. Anti-α_vβ₃ antibodies were then added to each sample and incubated overnight. The immunocomplexes were precipitated with protein A-Sepharose-4B beads and resolved by 15% SDS-PAGE. The coprecipitated HMGB1-Flag was detected by anti-Flag antibody. Results from a representative experiment are shown. A second independent experiment provided similar findings.

HMGB1 inhibits downstream signaling events induced in macrophages by efferocytosis.

Exposure of macrophages to apoptotic cells enhances ERK phosphorylation and activates Rac-1 (20, 25, 33, 45), events that are dependent, in part, on interactions between integrins on the macrophage surface with apoptotic cells. To determine whether exposure of macrophages to HMGB1 diminishes integrin-associated downstream signaling events that are activated during efferocytosis, we examined ERK phosphorylation and Rac-1 activation in macrophages incubated with either HMGB1 or human serum albumin followed by exposure to apoptotic neutrophils.

As shown in Fig. 7, A and B, addition of apoptotic neutrophils to macrophages induced phosphorylation of ERK. However, preincubation of macrophages with HMGB1 attenuated such efferocytosis-associated ERK activation. Similarly, preincubation of macrophages with HMGB1-attenuated efferocytosis induced Rac-1 activation (Fig. 7C). Because activation of ERK and Rac-1 are involved in facilitating alterations in macrophage morphology required for ingestion of apoptotic cells (20, 25, 33, 45), these results suggest that the mechanism by which HMGB1 prevents macrophages from engulfing apoptotic cells is through preventing α_vβ₃-induced activation of intracellular pathways, including those involving ERK and Rac-1, required for optimal phagocytic function during efferocytosis.

Fig. 7.

HMGB1 inhibits efferocytosis induced ERK phosphorylation and Rac-1 activation. A: macrophages were washed five times with serum-free medium and exposed for 30 min to 1 μg/ml HMGB1 or BSA (control). The medium was then replaced by fresh medium without or with addition of apoptotic neutrophils. Thirty minutes later, cells were collected and Western blot (WB) analysis was performed to determine the levels of phosphorylated (p-ERK) and total ERK. Two additional experiments provided similar results. B: densitometry was performed using Western blots from independent experiments performed as described in A. The relative ratio of p-ERK to total ERK in the BSA-exposed control group after apoptotic neutrophils was regarded as 1. *P < 0.05, compared with the BSA-exposed control group. C: macrophages were washed five times with serum-free medium and exposed for 30 min to 1 μg/ml HMGB1 or BSA (control). The medium was then replaced by fresh medium with or without apoptotic neutrophils for 15 min. Cells were then collected and Rac-1 activation was determined. The lanes that are separated by dashed lines are from the same blot. A second independent experiment provided similar findings. D: densitometry was performed using Western blots from independent experiments performed as described in C. The relative ratio of activated Rac-1 to input Rac-1 in the BSA-exposed control group after addition of apoptotic cells was regarded as 1. **P < 0.01, compared with the BSA-exposed group (control).

DISCUSSION

Impairment of phagocytic activity of macrophages during efferocytosis and resultant diminished clearance of apoptotic cells can potentiate inflammation through allowing apoptotic cells to progress to necrosis with associated release of proinflammatory intracellular contents into the extracellular milieu and interstitium. Alterations in efferocytosis have been demonstrated in acute and chronic disease processes, including atherosclerosis, cystic fibrosis, and chronic obstructive pulmonary disease (8, 23, 39, 43, 44). Elevations in local and systemic concentrations HMGB1 are present during inflammatory conditions, and increased extracellular levels of HMGB1 appear to contribute to the development and progression of tissue injury and organ dysfunction in diseases such as rheumatoid arthritis, cystic fibrosis, sepsis, and acute lung injury (1, 3, 35, 42). While HMGB1 was initially described to have intrinsic proinflammatory properties, other studies, however, demonstrated that HMGB1 itself is unable to activate macrophages or other cellular populations but can potentiate inflammation through binding to cytokines, LPS, DNA, and other proinflammatory mediators (6, 31, 38, 40). HMGB1 also appears to contribute to inflammatory processes by functioning as a chemotactin and by inhibiting efferocytosis (27, 28).

In previous studies, we demonstrated that HMGB1 diminishes phagocytosis of apoptotic neutrophils through binding to PS on the neutrophil surface (31). Those experiments did not identify potential receptors on the macrophage that interact with HMGB1 and that are responsible for the ability of HMGB1 to diminish uptake of apoptotic neutrophils. In this work we show that HMGB1 binds to macrophage-based α_vβ₃-integrin and that interactions between HMGB1 and α_vβ₃ are responsible for the inhibitory effects of HMGB1 on efferocytosis.

While PS is exposed on apoptotic cells and functions as a well-characterized “eat me” signal during efferocytosis, its interaction with receptors on the macrophage appear to occur in part through binding to opsonins, collectins, and other bridging molecules (14). MFG-E8 is a well-characterized opsonin that interacts with both PS on the surface of apoptotic cells and integrins, including α_vβ₃ and α_vβ₅, on macrophages and other phagocytes (2, 16). Binding between MFG-E8 and α_vβ₃ has been shown to reduce the severity of intestinal inflammation in experimental models of colitis, and administration of MFG-E8 diminishes the severity of lung injury and improves survival in the setting of acute lung injury that develops after intestinal ischemia and reperfusion (4, 9). In the present experiments, we found that HMGB1 and MFG-E8 competed for binding to α_vβ₃, and also that incubation of macrophages with HMGB1 diminished the ability of MFG-E8 to enhance uptake of apoptotic neutrophils. In contrast, while MFG-E8 is known to bind to α_vβ₅, there was no evidence of interaction between HMGB1 and α_vβ₅, demonstrating the specificity of HMGB1 interactions with integrins.

Although both MFG-E8 and HMGB1 are able to bind to PS and α_vβ₃, they function in opposite ways during efferocytosis, in that MFG-E8 serves as a bridging molecule and has enhancing properties, whereas HMGB1 plays an inhibitory role. Such differing actions of MFG-E8 and HMGB1 on efferocytosis may occur directly through affecting the association of apoptotic neutrophils and macrophages or may be caused by modification of efferocytosis associated downstream signaling events, including activation of ERK and Rac-1 when MFG-E8, but not HMGB1, binds to PS and α_vβ₃.

HMGB1 has been shown to associate with many proteins by recognizing multiple, nonhomologous short amino acid sequences (12). Recent studies have also demonstrated that HMGB1 can bind to cytokines, DNA, and LPS and that such interactions with HMGB1 enhance the ability of these proinflammatory mediators to activate macrophages and other cell populations (6, 31, 38, 40). Despite the ability of HMGB1 to associate with multiple proteins, this is clearly not an indiscriminate process, because the absence of binding between HMGB1 and specific proteins, such as the α_vβ₅-integrin in the present studies, has also been demonstrated (12). Although direct interaction between HMGB1 and integrins has not previously been reported, several studies have found that integrins, including Mac-1, are involved in modulating proinflammatory activities of HMGB1, including neutrophil recruitment and activation of NF-κB (28). HMGB1 does not contain an arginine-glycine-aspartate (RGD) sequence able to bind to integrins, suggesting that the interaction between HMGB1 and α_vβ₃ found in the present experiments must occur through alternate mechanisms. Further work must be carried out to examine the domains in HMGB1 that are involved in its binding to α_vβ₃.

The ability of HMGB1 to bind to α_vβ₃ may affect inflammatory processes beyond those involved in efferocytosis. For example, α_vβ₃ is involved in chemotaxis as well as potentiation of Toll-like receptor 4-induced neutrophil activation through its interaction with fibrinolytic proteins, such as urokinase (24). α_vβ₃ also binds to vitronectin, a protein that circulates in high concentrations and that is directly involved in chemotaxis and enhancing inflammatory responses, including LPS-induced acute lung injury (11, 41). Vitronectin itself acts as a “don't eat me” signal in efferocytosis and plays a critical role in mediating the activity of plasminogen activator inhibitor 1 in diminishing efferocytosis of apoptotic neutrophils (30, 36). Interaction of HMGB1 with α_vβ₃ may therefore potentiate inflammation through several mechanisms including enhancement of cellular responses to microbial products, increasing neutrophil accumulation in inflammatory foci, and diminishing the uptake of apoptotic neutrophils.

Opsonins, collectins, and other proteins enhance the uptake and clearance of apoptotic cells through bridging PS with receptors on the surface of macrophages and other phagocytes (14). The present experiments show that HMGB1 competes with the opsonin MFG-E8 for binding to α_vβ₃, but not to α_vβ₅. Other proteins, such as thrombospondin, can facilitate efferocytosis through bridging between PS and α_vβ₃ either directly, or indirectly via association with vitronectin (37). It is presently unknown whether HMGB1 competitively interferes with the function of opsonins other than MFG-E8 on efferocytosis. However, given the ability of HMGB1 to bind directly to α_vβ₃ and to compete with MFG-E8 for binding to α_vβ₃, it is likely that HMGB1 can also prevent association with α_vβ₃ of other opsonins involved with efferocytosis. If this is indeed the case, then the mechanism of action for the inhibitory role of HMGB1 in phagocytosis would extend beyond interference with the binding of MFG-E8 to α_vβ₃ on macrophages to blocking association of multiple opsonins and other bridging proteins with both PS and α_vβ₃. Therapeutic approaches that are aimed at inhibiting binding of HMGB1 to either PS or α_vβ₃ may therefore be useful in enhancing the uptake and clearance of neutrophils and improving outcome from acute inflammatory processes, such as acute lung injury, as well as more chronic conditions, such as chronic obstructive pulmonary disease or cystic fibrosis, in which the accumulation of activated neutrophils in the lungs plays a major role in mediating tissue damage and organ dysfunction.

GRANTS

This work was supported by National Institutes of Health Grants GM-087748 and HL-076206 (to E. Abraham).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.