Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Immunol. Author manuscript; available in PMC 2010 November 1.
Published in final edited form as:
PMCID: PMC2787084

Immature Dendritic Cell-Derived Exosomes Rescue Septic Animals via MFGE8


Sepsis, a highly lethal systemic inflammatory syndrome, is associated with increases of proinflammatory cytokines (e.g., TNF-α, HMGB1) and the accumulation of apoptotic cells that have the potential to be detrimental. Depending on the timing and tissue, prevention of apoptosis in sepsis is beneficial; however thwarting the development of secondary necrosis through the active removal of apoptotic cells by phagocytosis may offer a novel anti-sepsis therapy. Immature dendritic cells (IDCs) release exosomes that contain milk fat globule EGF factor 8 (MFGE8), a protein required to opsonize apoptotic cells for phagocytosis. In an experimental sepsis model using cecal ligation and puncture, we found that MFGE8 levels decreased in the spleen and blood, which was associated with impaired apoptotic cell clearance. Administration of IDC-derived exosomes promoted phagocytosis of apoptotic cells and significantly reduced mortality. Treatment with recombinant MFGE8 was equally protective, while MFGE8-deficient mice suffered from increased mortality. IDC exosomes also attenuated the release of proinflammatory cytokines in septic rats. Liberation of HMGB1, a nuclear protein that contributes to inflammation upon release from unengulfed apoptotic cells, was prevented by MFGE8-mediated phagocytosis in vitro. We conclude that IDC-derived exosomes attenuate the acute systemic inflammatory response in sepsis by enhancing apoptotic cell clearance via MFGE8.

Keywords: Phagocytosis, Apoptosis, Inflammation


Phagocytes, including dendritic cells, constitutively secrete exosomes. These are 100-nm vesicles contained and released from so-called multivesicular bodies, intermediates in the process of endocytosis (1). These exosomes contain both exogenic and endogenic proteins that are characteristic for the cells they derive from (2). Immature dendritic cells (IDCs) secrete exosomes, that contain abundant milk fat globule epidermal growth factor-factor VIII (MFGE8), or lactadherin. This protein is commonly found on human milk fat globules (3) and has been recently described to be necessary for the opsonization of apoptotic cells for phagocytosis (4). Hanayama et al. found that while phosphatidylserine and other apoptotic “eat-me signals” can be recognized and bound by phagocytes through other anchoring proteins, MFGE8 is an indispensable factor for the complete engulfment of these dying cells (4).

Dendritic cells play a key role in the interface of innate and adaptive immunity and are strategically placed throughout the body to recognize microbial intruders and promptly react to them (5). Under inflammatory conditions such as an infection, however, these phagocytes become activated and mature into inflammatory cells helping to clear the intruding microbes (5). While this inflammatory response is helpful in minor infections, it becomes overzealous in sepsis, causing more harm than good to the organism. Sepsis is a systemic inflammatory response that is often associated with severe infections. In spite of a growing number of generally effective antibiotics and improved critical care, sepsis still claims a high death toll among affected patients due to cardiovascular shock and multiple organ failure (6). So far, only activated protein C (Xigris®) has been approved as a sepsis-specific drug and it has provided limited success in the treatment of septic patients (7). Under septic conditions, there is a substantial neuroendocrine and immune activation, which leads to an over-stimulation of inflammatory processes (e.g., surge in TNF-α, IL-1β, IL-6 and the nuclear protein HMGB1, acting as a late proinflammatory cytokine) (811), but also to an impairment of vital innate immune functions (e.g., phagocytosis) (1215). One of the problems during sepsis is the strong induction of apoptosis of crucial immune cells, which further impairs the immune function. Upregulation of death receptors and stimulation by cytokines, glucocorticoids, and complement factors (especially factor C5a) lead to an early increase in activation-induced cell death (16, 17). In this inflammatory environment, apoptotic cells are prone to undergoing secondary necrosis if they are not fast removed by phagocytes (17). Without proper clearance, these cell corpses may pose a potential harm to the host, as they release potentially harmful inflammatory and toxic mediators, further impairing the septic condition (1820).

As MFGE8 has been reported to be of crucial importance in the removal of apoptotic cells, we investigated whether this protein may in fact play an important role in sepsis. Here we show that sepsis is associated with suppressed MFGE8, causing impaired clearance of apoptotic cells. We further investigated the beneficial role of IDC-derived exosomes in sepsis and their ability to restore clearance of apoptotic cells, to suppress inflammation, and to improve survival in sepsis.


Experimental sepsis using cecal ligation and puncture (CLP)

Male Sprague-Dawley rats were purchased from Charles River Laboratories (Wilmington, MA). MFGE8−/− mice were a generous gift from Dr. Shigekazu Nagata, Osaka University, Japan. The MFGE8−/− mice were generated by replacing exons 4–6 of MFGE8 gene with a neomycin cassette as described by Hanayama et al (4). The mutant mice were backcrossed to C57BL/6 for at least 9 times. Therefore, C57BL/6 wild-type mice (Taconic, Albany, NY) were used as a control for MFGE8−/− mice. In male Sprague-Dawley rats (275–325 g) or C57BL/6 wild-type and MFGE8−/− mice (20–25g) cecums were ligated and double punctured with an 18-gauge (rats) or 22-gauge (mice) needle as previously described (43, 44). Sham-operated animals underwent the same procedure without the ligation or puncture. The animals were resuscitated with 3 mL/100 g BW normal saline sc. At 5 h and 10 h after surgery, the rats received either 2 × 1 ml PBS (vehicle) or 2 × 1.5 mg/kg exosomes intravenously based on doses used in our own titration studies. Recombinant murine MFGE8 (rmMFGE8) was administered using an osmotic Alzet mini pump (Durect, Cupertino, CA) that was implanted subcutaneously and connected to right jugular vein. The pumps released 8 μl/h of rmMFGE8 over a period of 20 h (20 μg/kg BW) or the same volume PBS. For survival studies, the necrotic cecum was excised in rats and the abdominal cavity was washed with normal saline and animals monitored for 10 days. This procedure produces a consistent mortality approximately LD50. All experiments were performed in accordance with the National Institutes of Health guidelines for the use of experimental animals. This project was approved by the Institutional Animal Care and Use Committee of the Feinstein Institute for Medical Research.

Quantitative RT-PCR

Quantitative PCR (Q-PCR) was carried out on cDNA samples, reverse transcribed from 2 μg RNA, using the QuantiTect SYBR Green PCR kit (Qiagen, Valencia, CA), reactions were carried out in 24 μl final volume containing 2 pmol of forward and reverse primers, 12 μl QuantiTect Master Mix, and 1 μl cDNA. Amplification was performed according to Qiagen’s recommendations with an ABI Prism 7700 sequence detector (PerkinElmer-Applied Biosystems, Foster City, CA). Primer sequences were as follows: MFGE8 (107 bp, Gene Bank NM_012811) sense 5′ TGA GGA ACA AGG AAC CAG 3′, antisense: 5′ GGA AGG ACA CGC ACA TAG 3′, G3PDH (100 bp, Gene Bank XM_579386), sense: 5′ ATG ACT CTA CCC ACG GCA AG 3′, antisense: 5′ CTG GAA GAT GGT GAT GGG TT 3′. Expression of rat GAPDH mRNA was used to normalize samples and relative expression of mRNA was calculated using the ΔΔCt-method.

Western blotting

Splenic macrophages were collected by digesting spleens with collagenase IV, lysing red blood cells with ammonia-chloride potassium (ACK)-lysing buffer (0.15 M NH4Cl, 10 mM KHCO3, 0.1 mM Na2EDTA in H2O) followed by plastic adherence and thorough washing with PBS. 2×106 Cells were homogenized and dissolved in 1% SDS. Blood plasma was ultrafiltered with Centricon 100 (Millpore) and the elution was concentrated 30× using Centricon YM30 filters. 10 μg of protein or 5 μl of plasma concentrate was fractionated on a 4–12% Bis-Tris gel and transferred to 0.2-μm nitrocellulose membrane. Blots were blocked with 10% bovine serum albumin in Tris-buffered saline-Tween (TBST) and incubated with 1:100 goat anti-MFGE8 IgG (G-17, Santa Cruz Biotechnology, Santa Cruz, CA; specific for a 17 amino acid sequence in the C1 domain shared by human, mouse and rat MFGE8) washed and incubated with horseradish peroxidase-labeled rabbit anti-goat IgG. For HMGB1 detection, plasma samples were directly denatured in Laemmli buffer containing 5% β-mercaptoethanol and cell culture supernatants were concentrated 10× using YM-10 Microcons (Millipore, Bedford, MA) and denatured with 2.5% SDS and 5% β-mercaptoethanol. Protein was transferred to polyvinylidene difluoride membranes (Invitrogen), blocked with 5% BSA in TBST, and blotted with purified, polyclonal rabbit anti-HMGB1 antibody (1:1000) and HRP-conjugated anti-rabbit immunoglobulin G (1:20,000) in 3% BSA in TBST, incubated with ECL (Amersham) and exposed on a radiograph film. Densities of bands were analyzed using the Bio-Rad Imaging system (Hercules, CA).

Generation of bone marrow dendritic cell-derived exosomes

Dendritic cells were generated by culturing rat bone marrow leukocytes for up to 18 days in medium containing IL-4 and GM-CSF as described elsewhere (21). Briefly, bone marrow cells were obtained from healthy rats by flushing freshly isolated femur shafts with Hank’s Balanced Salt Solution (HBSS, CellGro-Mediatech, Herndon, VA), filtering and lysing red blood cells with ACK buffer. Cells were cultured for 6–18 days in Dulbecco’s Modified Eagle’s Medium (DMEM, GIBCO Life Technologies, Carlsbad, CA) containing 10% heat-inactivated exosome-free fetal bovine serum (FBS, obtained by centrifuging at 100,000 g overnight), GM-CSF (1,000 U/ml, Peprotech, Rocky Hill, NJ) and IL-4 (100 U/ml, Peprotech). Generation of BMDCs was verified morphologically by visible dendrites on loosely attached cells and by flow cytometry (>95% of collected cells were CD11b/c+/αE2-integrin+). Conditioned BMDC medium was collected and gradually centrifuged to remove cells and bigger particles and vesicles as described elsewhere (2). Exosomes were retrieved by ultra-centrifuging supernatants at 100,000 g for 3 h followed by a wash with PBS and overnight centrifugation at 100,000 g. Collected pellet was washed and reconstituted in PBS and adjusted to a concentration of 1 mg/ml. To confirm that the secreted MFGE8 is associated with exosomes, the purified exosomes were resolved on a sucrose gradient as described before (45). Concentrated samples (2 mg/500 μl) were mixed with 2.5 ml 85% (w/v) sucrose (in 10 mM Tris/HCl-buffer (pH 7.5) containing 150 mM NaCl and 5 mM EDTA), and placed in centrifuge tubes. The mixtures were layered successively with 4 ml of 60% (w/v), 3 ml of 30% (w/v) and 1 ml of 5% (w/v) sucrose, and centrifuged at 200,000 g for 18 h at 4 °C. Different fractions were collected; their refractive index assessed using a refractometer and samples directly subjected to SDS/PAGE for Western blotting. Please note that exosomes are the extractions from cells and the contents of MFGE8 in exosomes are different in each extraction. According to the Western blot analysis, there is about 9 μg MFGE8 in every mg of immature DC exosomes. Therefore, the amount of MFGE8 in the immature DC exosomes used in this study (i.e., 3.0 mg/kg BW) should be about 27 μg/kg BW. To simplify the calculation and experimental design, we chose 20–30 μg/kg BW rmMFGE8 to be administered to CLP animals.

Apoptosis Assay

Thymuses and spleens were homogenized and cells washed with PBS, and reconstituted in Ca2+-rich annexin V binding buffer (BD Pharmingen) at a concentration of 107 cells/ml. 100 μl of cell suspension was stained with 2.5 μl annexin V-FITC and/or 1 μl propidium iodine (PI) for 15 min and adjusted to a total volume of 500 μl with binding buffer. Cells were then analyzed by flow cytometry using FACS-Calibur.

Phagocytosis assay

Spleens were digested for 1 h with 400 U/ml collagenase VI (Worthington, Lakewood, NJ) and homogenized. Red blood cells were lysed with ACK-buffer and macrophages enriched by plastic adherence for 2 h. Cells from Sham and CLP animals were plated at a density of 5×105/well in a 24-well plate and PerCP-anti-CD90 (OX-7, BD Pharmingen, San Diego, CA)-tagged apoptotic thymocytes (induced by 10 μM dexamethasone for 16 h; ≥99% of apoptotic thymocytes [CD90+] both by annexin V/PI and TUNEL) were added at a 4:1 ratio (apoptotic cells/macrophages) for 1.5 h. Non-phagocytized thymocytes were removed by thorough washing with PBS. Macrophages were collected by gentle scraping and stained with FITC-labeled anti-CD11b/c (rats) or APC-labeled anti-CD11b (mice, BD Pharmingen). Analysis was performed by FACS (FACS Calibur, BD Biosciences, San Jose, CA) by gating on CD11b/c+ or CD11b+ cells. Percentage phagocytosis was determined by the ratio of CD90+CD11b(/c)+ to total CD11b(/c)+ cells. Alternatively, macrophages were also cultured on LabTek chamber slides (Nalge Nunc, Rochester, NY), incubated with apoptotic thymocytes for 1.5 h, washed 3× fixed with 4% paraformaldehyde, and stained with TUNEL for fluorescent microscopy using a Nikon Eclipse E600 microscope (Nikon, Melville, NY). To determine apoptotic cell engulfment, a novel phagocytosis assay using pHrodo-labeled apoptotic cells was used. Apoptotic thymocytes were stained with 20 ng/ml pHrodo SE for 30 minand assay performed as described above.

TNF-α assay

Cytokine levels were quantified using enzyme-linked immunosorbent assay (ELISA) kit (Pharmingen, San Diego, CA). The assays were carried out according to the instructions provided by the manufacturer.

Cell culture

Primary rat peritoneal macrophages were cultured at a density of 106/well in a 6-well plate and incubated with apoptotic rmMFGE8-opsonized thymocytes with (10 μg/ml, R&D Systems, Minneapolis, MN). In controls, phagocytosis of apoptotic cells was almost entirely blocked by anti-mouse MFGE8 pAb (10 μg/ml R&D Systems). After 90 minutes, macrophages were washed 3 times with media and followed by stimulation with 100 ng/ml LPS (E. coli 055:B5; Difco Laboratories, Detroit, MI) for 16 h. Supernatants were collected for HMGB1 Western blotting. In a separate experiment, rat peritoneal macrophages were incubated with live, necrotic or apoptotic lymphocytes thymocytes for 90 minutes without LPS-stimulation. Supernatants containing non-engulfed lymphocytes were removed and cultured separately and adherent macrophages were washed and cultured with fresh medium for 16h.


All data are expressed as means ± SEM and compared by analysis of variance (ANOVA, one-way or two-way ANOVA as indicated). Student t-test with a two-tailed distribution and equal variance was used if only two groups were present. Normal distribution of samples was verified using the Kolmogorov-Smirnov test. The survival rate was estimated by the Kaplan-Meier method and compared by the log-rank test. The alpha level for all tests was 0.05.


MFGE8 is suppressed in experimental sepsis

To investigate whether MFGE8 levels are affected in critical disease, we used an experimental sepsis model in rats. 20 h after cecal ligation and puncture (CLP), which produces peritonitis and sepsis in these animals, blood MFGE8 levels decreased by 45% (P = 0.031), indicating the systemic scale of MFGE8 depletion under septic conditions (Fig. 1A). Using quantitative PCR, we found that under normal conditions MFGE8 is produced in various tissues with the highest mRNA expression levels in the spleen, followed by 34–50% of its expression in the lungs, thymus, and skin. Lower mRNA levels of about 13–15% of spleen levels were found in the brain, heart, and liver (data not shown). Due to its high basal MFGE8 expression levels, spleen MFGE8 levels were analyzed after CLP. Indeed, MFGE8 mRNA levels decreased in the septic spleen by 19.9% within the first 5 h (P = 0.294) and by 49% within 20 h of CLP (P = 0.004) (Fig. 1B). This decreased transcription translated into a 48% decrease of MFGE8 protein levels in the spleen 20h after CLP (P = 0.034) (Fig. 1C). MFGE8 Western blotting from isolated CD11b/c+ macrophages from septic spleens revealed that they greatly contributed to the decrease of MFGE8, showing a 51% suppression vs. sham levels (P=0.032, Fig. 1D).

Figure 1
MFGE8 is suppressed in experimental sepsis. (A) Blood was drawn from septic rats 20h after the onset of CLP and MFGE8 levels were compared to sham-operated rats by Western blotting. *P=0.031 vs. Sham, Student t-test, n = 5. (B) Time-dependent decrease ...

IDC-derived exosomes improve apoptotic cell clearance and survival in experimental sepsis

Bone marrow generated rat IDCs secreted high amounts of exosomes (1.91±0.28 mg total protein/108 DCs), with a decrease of secretion by over 75% to 0.45±0.07 mg/108 DCs upon DC maturation. By sucrose gradient centrifugation we found that virtually all secreted MFGE8 was associated with exosomes found in the fraction with a density of 1.15 g/ml (refractive index = 1.375) at which exosomes equilibrate (1) (Fig. 2). IDCs secreted exosomes that contained abundant MFGE8, but virtually no co-stimulatory protein B7.2, a marker for mature DCs (Fig. 3A) (21). Mature DCs, marked by the morphological change with the development of visible dendrites, secreted exosomes with high levels of B7.2 but no MFGE8 (Fig. 3A). In septic rats, the systemic decrease in MFGE8 was associated with the accumulation and impaired clearance of apoptototic cells. In untreated animals, apoptotic thymocytes accumulated over time from 4–5% at 0 and 5 h after CLP to up to 14% within 20 h after CLP (P<0.001, data not shown). Phagocytosis of apoptotic cells by CD11b/c+ splenic macrophages was significantly impaired with the average phagocytotic percentage decreasing from 81% to 67% after CLP (P = 0.021, Fig. 2B), while at the same time the total amount of apoptotic cells increased in the spleen from 6.6% to 9.8% (P = 0.003, data not shown) and in the thymus from 6.2 to 12.5% (P<0.001, Fig. 3C). However, IDC-derived exosomes (2×1.5 mg/kg) increased the clearance of apoptotic cells in sepsis and completely restored the CLP-induced suppression of phagocytic capability (iExo, Fig. 3B). Concurrently, CLP-associated accumulation of apoptotic cells was reduced by over 28% (P = 0.027, Fig. 3C). Mature DC-derived exosomes showed no significant influence on the ability of splenic macrophages to clear apoptotic cells (mExo, P = 0.492 vs. Vehicle, Fig. 3B) or to alter the accumulation of apoptotic cells (P = 0.436, Figs. 3C). This experimental sepsis model caused 62.5% lethality within 1–4 days when the necrotic cecum was removed 20 h after CLP (Fig. 3D). Only two out of 16 rats receiving IDC-derived exosomes (2×1.5 mg/kg at 5 and 20 h after CLP) died within 24 h. All other animals survived the following 9-day observation period (P = 0.007 vs. Vehicle Control). Neither treatment with 10% of the effective dose of IDC exosomes, nor treatment with mature DC-derived exosomes affected the outcome in sepsis (Fig. 3D). As demonstrated above, mature DC-derived exosomes did not contain MFGE8 (Fig. 3A), and the beneficial effect of IDC exosomes was likely to be mediated by MFGE8.

Figure 2
Detection of MFGE8 in exosomes isolated from IDCs. (A) The polyclonal goat anti-MFGE8 antibody (clone G-17) is reactive against rat MFGE8 in thioglycolate-elicited peritoneal macrophages (PM) and detects MFGE8 in IDC-derived exosomes (Exo, day 6 of BMDC ...
Figure 3
IDC-derived exosomes improve apoptotic cell clearance and survival in experimental sepsis. (A) Exosomes secreted from immature dendritic cells (DCs) contain MFGE8. Exosomes were collected from the supernatant of primary rat bone marrow dendritic cell ...

MFGE8 reconstitutes the clearance of apoptotic cells in septic rats

The reduced detection of apoptotic cells in septic animals cannot be explained by a direct anti-apoptotic effect of IDC-derived exosomes. In in vitro studies, TNF-α-induced apoptosis of lymphocytes could not be blocked by the pretreatment with exosomes in the absence of macrophages (data not shown). Similar to IDC exosomes, treatment of septic rats with 2×30 μg/kg recombinant murine MFGE8 (rmMFGE8) 5 h and 10 h after CLP, resulted in the reconstitution of apoptotic cell clearance in septic rats (P=0.019, Fig. 4A). Previous publications have shown that MFGE8 is crucial for the engulfment of apoptotic cells, an important step in the removal of apoptotic cells adhering to the surface of macrophages (4). To investigate whether MFGE8 leads to the engulfment of apoptotic cells under septic conditions, we used the approach of staining apoptotic cells with a pH-sensitive dye (phrodo SE) that become detectable only after they are engulfed by a macrophage. Using this method, we found that the rmMFGE8 protein was able to reconstitute engulfment of apoptotic cells in septic rats to levels observed in sham operated animals (P=0.009, Figs. 4B–E).

Figure 4
MFGE8 reconstitutes phagocytosis of apoptotic cells in septic rats. (A) Rats underwent CLP to induce experimental sepsis and were treated with 2×15 μg/kg recombinant murine MFGE8 (rmMFGE8) i.v. at 5h and 10h after CLP. Thymocyte apoptosis ...

IDC-derived exosomes suppress the proinflammatory response in experimental sepsis

Sepsis is associated with a systemic inflammatory response characterized by increases in early (TNF-α) and late proinflammatory cytokines (HMGB1). We investigated whether the treatment with IDC-derived exosomes was able to influence the inflammatory response in experimental sepsis. IDC-derived exosomes suppressed the CLP-induced TNF-α response by 46% (P=0.045, Fig. 5A), while mature DC-derived exosomes did not affect the TNF-α levels in septic rats (Fig. 5A). Interestingly, IDC-derived exosomes failed to suppress TNF-α release from LPS-stimulated macrophages in vitro in the absence of apoptotic cells (data not shown), suggesting an indirect immunosuppressive effect. Similar results could be found in the levels of the late cytokine HMGB1. In vehicle-treated septic animals, blood HMGB1 levels increased by 51% compared to sham-operated animals (P = 0.002). This increase was completely suppressed by the treatment with IDC exosomes (P < 0.001), but not by mature DC exosomes (P = 0.294, Fig. 5B). While phagocytosis of apoptotic cells is known to suppresses TNF-α response of macrophages, its effects on HMGB1 release remains unclear. Apoptotic cells have been previously shown to increase HMGB1 release in a co-culture system with macrophages (22). We therefore investigated the effect of MFGE8-mediated clearance of apoptotic cells on HMGB1-release under inflammatory conditions in vitro. Opsonization of apoptotic thymocytes with rmMFGE8 completely prevented the HMGB1 release from the macrophage/apoptotic cell co-cultured system (P = 0.002, Fig. 5C). Inefficient phagocytosis of apoptotic cells, on the other hand, resulted in the release of increasing amounts of HMGB1 (up to 3.5 times at an apoptotic cell to macrophage ratio of 5:1, P = 0.006, Fig. 5C), depending on the amount of apoptotic cells present. HMGB1 has been shown to be actively released from inflammatory macrophages as well as passively released from necrotic cells (23). Analysis of the release of HMGB1 from separately cultured macrophages and unengulfed apoptotic cells after phagocytosis revealed that macrophages challenged with apoptotic cells released similar levels of HMGB1 as necrotic cell-challenged macrophages, albeit to a significantly lesser degree than unengulfed late apoptotic cells and necrotic cells released themselves. rmMFGE8 treatment suppressed the release of HMGB1 from both macrophages and late apoptotic lymphocytes (Fig. 6). This suggests that MFGE8-mediated clearance of apoptotic cells directly attenuates the release of HMGB1 in sepsis.

Figure 5
IDC-derived exosomes suppress the proinflammatory response in sepsis. (A) TNF-α levels in septic rats 20 h after CLP. Rats underwent either sham operation or CLP with either PBS (Vehicle), IDC-derived exosome (iExo) or mature DC-derived exosome ...
Figure 6
Post-phagocytic release of HMGB1 from macrophages and late apoptotic cells is suppressed by rmMFGE8 treatment. Rat peritoneal macrophages were incubated with live, necrotic (necr.), or apoptotic lymphocytes (induced by 10 μM dexomethasone for ...

MFGE8-deficiency is associated with poor apoptotic cell clearance in sepsis

To address whether impaired phagocytosis due to MFGE8-deficiency affects apoptotic cell accumulation in sepsis, we compared MFGE8 knockout (MFGE8−/−) mice and their C57BL6/J wild-type controls 20h after CLP. As expected, spleen macrophages from MFGE8−/− mice showed a dramatically decreased ability to phagocytose apoptotic cells under normal conditions (22% of WT Sham, P<0.001, Fig. 7A). To an even stronger degree than in rats, septic WT mice showed a 74% suppression of phagocytosis of apoptotic cells 20h after CLP, while CLP had no further impact on phagocytosis in MFGE8−/− mice (Fig. 7A). Mice deficient in MFGE8 also accumulated higher amounts of apoptotic cells (19%) compared to WT mice at the same time point (12%, P < 0.001, Fig. 7B). This indicates that the clearance of apoptotic cells in sepsis is positively regulated by MFGE8.

Figure 7
MFGE8-deficiency is associated with poor clearance of apoptotic cells in sepsis. (A) Phagocytosis of apoptotic cells is severely impaired in MFGE8-deficient mice. C57BL6/J wild-type (WT) or MFGE8 knockout (KO) mice underwent CLP and 20 h later, splenic ...

MFGE8 protects from sepsis-associated mortality

Finally, we were interested in whether MFGE8 influences survival in experimental sepsis. Indeed, MFGE8-deficient mice were more susceptible to sepsis-mediated mortality. While 50% of wild-type mice died within 10 days of CLP, 82% of MFGE8−/− mice died in the same period (P = 0.045, Fig. 8A). On the other hand, a continuous infusion of rmMFGE8 over 20 h protected 83% of septic rats from dying from CLP-induced sepsis compared to 50% of rats that died without treatment (P = 0.042, Fig. 8B). Thus, administration of rmMFGE8 provided similar results as the treatment with IDC-derived exosomes (Fig. 3D).

Figure 8
MFGE8 protects from sepsis-associated mortality. (A) MFGE8 knockout (KO) mice display dramatically reduced survival in sepsis. MFGE8 KO mice and C57BL6/J wild-type (WT) mice underwent CLP and were observed for the following 10 days. Kaplan-Meyer LogRank ...


IDCs constitutively secrete exosomes that contain MFGE8. In mature DCs, exosome and MFGE8 production and release are reduced. In our studies, we have presented that MFGE8 is systemically downregulated in sepsis, which leads to a wide-spread impairment of apoptotic cell clearance. The associated proinflammatory response in sepsis can be prevented by the administration of exogenous exosomes from IDCs. These exosomes improve apoptotic cell clearance, prevent the excessive release of proinflammatory cytokines and protect septic animals from dying. While exosomes from mature DCs do not contain MFGE8 and fail to be protective, the protein MFGE8 itself has been shown to be an indispensable factor for the prevention of accumulating apoptotic cells and mortality in sepsis.

Sepsis is marked by a systemic inflammatory response, mediated by innate immune cells. An increase in proinflammatory cytokines is normally beneficial to fight microbes in minor infections (24). In sepsis however, this cytokine response is extensive and prolonged (25), leading to multiple organ damage and septic shock (6, 25). Systemic increases of the cytokines TNF-α, IL-1β, IL-6, and HMGB1 in sepsis have been previously associated with a high mortality rate. Being equally the source and the target of these mediators, antigen-presenting cells become activated and mature, thereby shutting down the endocytotic machinery in favor of an immunostimulatory response (26). During this maturation process, the secretion of exosomes and the production of MFGE8 are reduced (27). A number of mediators are responsible for the modulation of MFGE8 production in sepsis. Endotoxin can suppress MFGE8 production in vitro (28), and GM-CSF + IL-4-mediated maturation of DC in vitro in the absence of bacteria or endotoxin indicates that cytokines may equally play a role. In this regard, GM-CSF has been shown to induce MFGE8 expression in vivo (29). GM-CSF is required for the expression of MFGE8 in antigen-presenting cells, and that MFGE8-mediated uptake of apoptotic cells is a key determinant of GM-CSF-triggered tolerance and immunity (29). This is interesting because GM-CSF has been investigated as a promising treatment option for septic patients due to its immune-modulating function (3033). Whether other cytokines such as TNF-α or IFN-γ, also influence MFGE8 expression is unclear and needs further investigation. We have shown in this study that in sepsis CD11b/c expressing macrophages and DCs contribute to the reduction of MFGE8 production in the spleen which is mirrored by a decrease of this protein in the circulation.

As we have shown further, a deficiency in MFGE8 is detrimental in sepsis. Mice lacking MFGE8 accumulate 2–3 times as many apoptotic cells above basal level and have a 60% higher mortality rate than wild-type mice. Similarly, the administration of rmMFGE8 to septic rats protected the majority from sepsis-mediated lethality. The protective effect of MFGE8 is evident in this acute inflammatory model of sepsis and provides further evidence that it is a crucial part of the protective effect of exosomes derived from IDCs in septic animals.

Exosomes have been previously shown to transfer cell-mediated immunity from one cell to another (26). Mature antigen-pulsed DCs secrete exosomes that act as cross-presenting carriers of MHC-antigen complexes and hence confer functional immunity to the recipient cell (26). The absence of antigen-presenting and co-stimulatory molecules on exosomes from IDCs suggests that these exosomes hold a minor role in directly modulating immune responses. However, IDC-derived exosomes are not at all functionally inert. As we have shown, the administration of exosomes from immature, but not from mature DCs conferred protection in sepsis, highlighting the important role of MFGE8 in these vesicles. Unfortunately, technical reasons precluded conclusive experiments using exosomes from wild type and MFGE8 knock out exosomes.

MFGE8 contains two important regions to function as an opsonin for apoptotic cells; two epidermal growth factor-like domains contain an RGD-motif necessary for the binding of αvβ3- or αvβ5-integrins, and two coagulation factor V/VIII like domains that bind to phosphatidylserine (PS) exposed on the surface of apoptotic cells (27). Binding of MFGE8 to PS on apoptotic cells opsonizes them for a complete engulfment by macrophages via αvβ3- or αvβ5-integrins. MFGE8 has been shown to be important for the removal of apoptotic lymphocytes in the spleen and the prevention of a systemic lupus erythematosus-like disease in mice (4). In our septic model, we found a similar phenomenon in an acute inflammatory environment.

The beneficial effect of IDC-derived exosomes is mediated by the promotion of apoptotic cell clearance. Apoptosis is often found in sepsis, with lymphoid CD4 T and B cells and DCs being most commonly affected (3437). Particularly apoptosis of DCs may contribute to the depletion of MFGE8 in sepsis as these are one source of this protein. Overall, the occurrence of apoptosis has been associated with poor outcome in sepsis. Targeted inhibition of the pro-apoptotic Fas-signaling or overexpression of anti-apoptotic proteins, such as BH4, Bcl-2, or Bcl-XL, have been shown to prevent apoptosis and protect from associated lethality in sepsis (35, 38). Historically, apoptosis has been seen as an orderly process of cell suicide that, unlike necrosis, does not elicit inflammation (39). Recently it has become clear, however, that apoptotic cells eventually undergo secondary necrosis and stimulate an inflammatory response if they are not removed by phagocytosis (19, 20). By using the pHrodo labeling system of apoptotic cells we have clearly demonstrated that rmMFGE8 promotes the engulfment of apoptotic cells also under septic conditions, which is at least in part responsible for the decrease in apoptotic cell number. Hence, the sepsis-associated decrease of MFGE8 contributes to the accumulation of apoptotic cells, resulting in a surge in proinflammatory cytokines, such as TNF-α and HMGB1 which by itself promotes the progression and deterioration in sepsis (11, 40).

We have previously shown that the pretreatment with bone marrow-derived DC exosomes was beneficial in septic animals and possibly dependent on the presence of MFGE8 (41). The current study shows that the clearance of apoptotic cells in septic animals is impaired in the absence of MFGE8 and that this is detrimental in the acute inflammatory disease model. Furthermore, we now show that even treatment with IDC exosomes at a later time point is beneficial (i.e., when apoptotic cells start to accumulate in sepsis, long after the initiation of inflammatory responses). The present report also demonstrates how MFGE8-mediated internalization of apoptotic cells prevents the release of proinflammatory mediators, which leads to a suppression of the septic systemic inflammatory response.

The beneficial effect of IDC-derived exosomes can be particularly attributed to the immune suppressive effect of phagocytosis of apoptotic cells (39). Exosomes from IDCs were neither able to suppress TNF-α release from endotoxin-stimulated macrophages, nor to prevent TNF-α-induced apoptosis of lymphocytes in vitro (our unpublished observations). Thus the reduction in apoptosis in vivo is most likely mediated via enhanced clearance of apoptotic cells.

In our experimental sepsis model, early increases in TNF-α and IL-6 (1–4 h after CLP) are followed by a significant release of HMGB1 in the late phase of sepsis (16–24 h after CLP) (11). These cytokines play a central role in the morbidity and mortality in experimental sepsis as well as in septic patients (42). Studies using inhibitors of these cytokines demonstrated increased survival of septic mice treated with TNF-α or HMGB1 blocking antibodies (8, 22). We have shown here that IDC-derived exosomes suppressed both TNF-α and HMGB1 in experimental sepsis, which was associated with a dramatically improved survival.

Exosomes secreted from IDCs are not merely a byproduct of excessive endocytosis but they have, as we have shown, the ability to suppress a once-established systemic proinflammatory response. The opsonizing protein MFGE8 plays a key role in this pro-phagocytic and secondary immunosuppressive effect. This distinguishes them from exosomes secreted from mature, antigen-pulsed DCs that confer cellular immunity by cross priming. This novel finding should open a new option of sepsis therapy in which the clearance of apoptotic cells is targeted. The restoration of this basic immunological function and the induction of the reparative phase may ultimately contribute to the attenuation of the life-threatening systemic inflammatory response in sepsis. It is likely that a combination of apoptosis prevention and the promotion of apoptotic cell clearance can be used in conjunction to the current treatment regime to the benefit of critically ill patients in the future.


This study was supported by NIH grants R01 GM057468, R01 GM053008, and R01 AG028352 (P. Wang).

Max Brenner, Herb Borrero, and Thomas McCloskey were helpful with inputs and in performing FACS analysis for the phagocytosis assay and detection of apoptotic cells. Enesa Paric and James Mason assisted in the isolation and gradient centrifugation of exosomes. Margot Puerta, Wei Li, and Tianpen Cui were cooperative in the methodology of HMGB1 detection. Maowen Hu and Yingjie Cui were instrumental in establishing the real time PCR for MFGE8 and detecting MFGE8 in the blood, respectively. Kavin Shah was instrumental in the rat sepsis model. Shigekazu Nagata of Osaka University, Japan, provided us with MFGE8−/− mice for this study, for which we are thankful.


1. Thery C, Boussac M, Veron P, Ricciardi-Castagnoli P, Raposo G, Garin J, Amigorena S. Proteomic analysis of dendritic cell-derived exosomes: a secreted subcellular compartment distinct from apoptotic vesicles. J Immunol. 2001;166:7309–7318. [PubMed]
2. Thery C, Regnault A, Garin J, Wolfers J, Zitvogel L, Ricciardi-Castagnoli P, Raposo G, Amigorena S. Molecular characterization of dendritic cell-derived exosomes. Selective accumulation of the heat shock protein hsc73. J Cell Biol. 1999;147:599–610. [PMC free article] [PubMed]
3. Newburg DS, Peterson JA, Ruiz-Palacios GM, Matson DO, Morrow AL, Shults J, Guerrero ML, Chaturvedi P, Newburg SO, Scallan CD, Taylor MR, Ceriani RL, Pickering LK. Role of human-milk lactadherin in protection against symptomatic rotavirus infection. Lancet. 1998;351:1160–1164. [PubMed]
4. Hanayama R, Tanaka M, Miyasaka K, Aozasa K, Koike M, Uchiyama Y, Nagata S. Autoimmune disease and impaired uptake of apoptotic cells in MFG-E8-deficient mice. Science. 2004;304:1147–1150. [PubMed]
5. Degli-Esposti MA, Smyth MJ. Close encounters of different kinds: dendritic cells and NK cells take centre stage. Nat Rev Immunol. 2005;5:112–124. [PubMed]
6. Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med. 2003;348:1546–1554. [PubMed]
7. Marti-Carvajal A, Salanti G, Cardona AF. Human recombinant activated protein C for severe sepsis. Cochrane Database Syst Rev. 2008:CD004388. [PubMed]
8. Tracey KJ, Fong Y, Hesse DG, Manogue KR, Lee AT, Kuo GC, Lowry SF, Cerami A. Anti-cachectin/TNF monoclonal antibodies prevent septic shock during lethal bacteremia. Nature. 1987;330:662–664. [PubMed]
9. Li P, Allen H, Banerjee S, Franklin S, Herzog L, Johnston C, McDowell J, Paskind M, Rodman L, Salfeld J. Mice deficient in IL-1 beta-converting enzyme are defective in production of mature IL-1 beta and resistant to endotoxic shock. Cell. 1995;80:401–411. [PubMed]
10. Meyer TA, Wang J, Tiao GM, Ogle CK, Fischer JE, Hasselgren PO. Sepsis and endotoxemia stimulate intestinal interleukin-6 production. Surgery. 1995;118:336–342. [PubMed]
11. Wang H, Bloom O, Zhang M, Vishnubhakat JM, Ombrellino M, Che J, Frazier A, Yang H, Ivanova S, Borovikova L, Manogue KR, Faist E, Abraham E, Andersson J, Andersson U, Molina PE, Abumrad NN, Sama A, Tracey KJ. HMG-1 as a late mediator of endotoxin lethality in mice. Science. 1999;285:248–251. [PubMed]
12. Alves-Filho JC, de FA, Russo M, Cunha FQ. Toll-like receptor 4 signaling leads to neutrophil migration impairment in polymicrobial sepsis. Crit Care Med. 2006;34:461–470. [PubMed]
13. Huber-Lang MS, Younkin EM, Sarma JV, McGuire SR, Lu KT, Guo RF, Padgaonkar VA, Curnutte JT, Erickson R, Ward PA. Complement-induced impairment of innate immunity during sepsis. J Immunol. 2002;169:3223–3231. [PubMed]
14. Kaufmann I, Hoelzl A, Schliephake F, Hummel T, Chouker A, Peter K, Thiel M. Polymorphonuclear leukocyte dysfunction syndrome in patients with increasing sepsis severity. Shock. 2006;26:254–261. [PubMed]
15. Hallwirth U, Pomberger G, Zaknun D, Szepfalusi Z, Horcher E, Pollak A, Roth E, Spittler A. Monocyte phagocytosis as a reliable parameter for predicting early-onset sepsis in very low birthweight infants. Early Hum Dev. 2002;67:1–9. [PubMed]
16. Gershov D, Kim S, Brot N, Elkon KB. C-Reactive protein binds to apoptotic cells, protects the cells from assembly of the terminal complement components, and sustains an antiinflammatory innate immune response: implications for systemic autoimmunity. J Exp Med. 2000;192:1353–1364. [PMC free article] [PubMed]
17. Fink SL, Cookson BT. Apoptosis, pyroptosis, and necrosis: mechanistic description of dead and dying eukaryotic cells. Infect Immun. 2005;73:1907–1916. [PMC free article] [PubMed]
18. Hanayama R, Tanaka M, Miwa K, Shinohara A, Iwamatsu A, Nagata S. Identification of a factor that links apoptotic cells to phagocytes. Nature. 2002;417:182–187. [PubMed]
19. Bell CW, Jiang W, Reich CF, III, Pisetsky DS. The extracellular release of HMGB1 during apoptotic cell death. Am J Physiol Cell Physiol. 2006;291:C1318–C1325. [PubMed]
20. Scaffidi P, Misteli T, Bianchi ME. Release of chromatin protein HMGB1 by necrotic cells triggers inflammation. Nature. 2002;418:191–195. [PubMed]
21. Muthana M, Fairburn B, Mirza S, Slack LK, Pockley AG. Systematic evaluation of the conditions required for the generation of immature rat bone marrow-derived dendritic cells and their phenotypic and functional characterization. J Immunol Methods. 2004;294:165–179. [PubMed]
22. Qin S, Wang H, Yuan R, Li H, Ochani M, Ochani K, Rosas-Ballina M, Czura CJ, Huston JM, Miller E, Lin X, Sherry B, Kumar A, Larosa G, Newman W, Tracey KJ, Yang H. Role of HMGB1 in apoptosis-mediated sepsis lethality. J Exp Med. 2006;203:1637–1642. [PMC free article] [PubMed]
23. Ulloa L, Messmer D. High-mobility group box 1 (HMGB1) protein: friend and foe. Cytokine Growth Factor Rev. 2006;17:189–201. [PubMed]
24. Moretta A. Natural killer cells and dendritic cells: rendezvous in abused tissues. Nat Rev Immunol. 2002;2:957–964. [PubMed]
25. Riedemann NC, Ward PA. Anti-inflammatory strategies for the treatment of sepsis. Expert Opin Biol Ther. 2003;3:339–350. [PubMed]
26. Utsugi-Kobukai S, Fujimaki H, Hotta C, Nakazawa M, Minami M. MHC class I-mediated exogenous antigen presentation by exosomes secreted from immature and mature bone marrow derived dendritic cells. Immunol Lett. 2003;89:125–131. [PubMed]
27. Veron P, Segura E, Sugano G, Amigorena S, Thery C. Accumulation of MFG-E8/lactadherin on exosomes from immature dendritic cells. Blood Cells Mol Dis. 2005;35:81–88. [PubMed]
28. Asano K, Miwa M, Miwa K, Hanayama R, Nagase H, Nagata S, Tanaka M. Masking of phosphatidylserine inhibits apoptotic cell engulfment and induces autoantibody production in mice. J Exp Med. 2004;200:459–467. [PMC free article] [PubMed]
29. Jinushi M, Nakazaki Y, Dougan M, Carrasco DR, Mihm M, Dranoff G. MFG-E8-mediated uptake of apoptotic cells by APCs links the pro- and antiinflammatory activities of GM-CSF. J Clin Invest. 2007;117:1902–1913. [PMC free article] [PubMed]
30. Orozco H, Arch J, Medina-Franco H, Pantoja JP, Gonzalez QH, Vilatoba M, Hinojosa C, Vargas-Vorackova F, Sifuentes-Osornio J. Molgramostim (GM-CSF) associated with antibiotic treatment in nontraumatic abdominal sepsis: a randomized, double-blind, placebo-controlled clinical trial. Arch Surg. 2006;141:150–153. [PubMed]
31. Presneill JJ, Harris T, Stewart AG, Cade JF, Wilson JW. A randomized phase II trial of granulocyte-macrophage colony-stimulating factor therapy in severe sepsis with respiratory dysfunction. Am J Respir Crit Care Med. 2002;166:138–143. [PubMed]
32. Rosenbloom AJ, Linden PK, Dorrance A, Penkosky N, Cohen-Melamed MH, Pinsky MR. Effect of granulocyte-monocyte colony-stimulating factor therapy on leukocyte function and clearance of serious infection in nonneutropenic patients. Chest. 2005;127:2139–2150. [PubMed]
33. Ahmad M, Fleit HB, Golightly MG, La Gamma EF. In vivo effect of recombinant human granulocyte colony-stimulating factor on phagocytic function and oxidative burst activity in septic neutropenic neonates. Biol Neonate. 2004;86:48–54. [PubMed]
34. Ayala A, Chung CS, Xu YX, Evans TA, Redmond KM, Chaudry IH. Increased inducible apoptosis in CD4+ T lymphocytes during polymicrobial sepsis is mediated by Fas ligand and not endotoxin. Immunology. 1999;97:45–55. [PubMed]
35. Ayala A, Lomas JL, Grutkoski PS, Chung S. Fas-ligand mediated apoptosis in severe sepsis and shock. Scand J Infect Dis. 2003;35:593–600. [PubMed]
36. Ayala A, Xin XY, Ayala CA, Sonefeld DE, Karr SM, Evans TA, Chaudry IH. Increased mucosal B-lymphocyte apoptosis during polymicrobial sepsis is a Fas ligand but not an endotoxin-mediated process. Blood. 1998;91:1362–1372. [PubMed]
37. Tinsley KW, Grayson MH, Swanson PE, Drewry AM, Chang KC, Karl IE, Hotchkiss RS. Sepsis induces apoptosis and profound depletion of splenic interdigitating and follicular dendritic cells. J Immunol. 2003;171:909–914. [PubMed]
38. Hotchkiss RS, Chang KC, Grayson MH, Tinsley KW, Dunne BS, Davis CG, Osborne DF, Karl IE. Adoptive transfer of apoptotic splenocytes worsens survival, whereas adoptive transfer of necrotic splenocytes improves survival in sepsis. Proc Natl Acad Sci U S A. 2003;100:6724–6729. [PubMed]
39. Fadok VA, Bratton DL, Konowal A, Freed PW, Westcott JY, Henson PM. Macrophages that have ingested apoptotic cells in vitro inhibit proinflammatory cytokine production through autocrine/paracrine mechanisms involving TGF-beta, PGE2, and PAF. J Clin Invest. 1998;101:890–898. [PMC free article] [PubMed]
40. Yang H, Ochani M, Li J, Qiang X, Tanovic M, Harris HE, Susarla SM, Ulloa L, Wang H, DiRaimo R, Czura CJ, Wang H, Roth J, Warren HS, Fink MP, Fenton MJ, Andersson U, Tracey KJ. Reversing established sepsis with antagonists of endogenous high-mobility group box 1. Proc Natl Acad Sci U S A. 2004;101:296–301. [PubMed]
41. Miksa M, Wu R, Dong W, Das P, Yang D, Wang P. Dendritic cell-derived exosomes containing milk fat globule epidermal growth factor-factor VIII attenuate proinflammatory responses in sepsis. Shock. 2006;25:586–593. [PubMed]
42. Lotze MT, Tracey KJ. High-mobility group box 1 protein (HMGB1): nuclear weapon in the immune arsenal. Nat Rev Immunol. 2005;5:331–342. [PubMed]
43. Wu R, Dong W, Zhou M, Zhang F, Marini CP, Ravikumar TS, Wang P. Ghrelin attenuates sepsis-induced acute lung injury and mortality in rats. Am J Respir Crit Care Med. 2007;176:805–813. [PMC free article] [PubMed]
44. Yang S, Chung CS, Ayala A, Chaudry IH, Wang P. Differential alterations in cardiovascular responses during the progression of polymicrobial sepsis in the mouse. Shock. 2002;17:55–60. [PubMed]
45. Raposo G, Nijman HW, Stoorvogel W, Liejendekker R, Harding CV, Melief CJ, Geuze HJ. B lymphocytes secrete antigen-presenting vesicles. J Exp Med. 1996;183:1161–1172. [PMC free article] [PubMed]