MFG-E8 (milk fat globule-epidermal growth factor 8) deficiency is strongly associated with acquisition of immune-mediated disorders due to the loss of tissue homeostasis. However, comparatively little is known regarding its functions in gastrointestinal tract disorders, in which immune homeostasis is a major concern. Herein, we report altered MFG-E8 expression in inflamed colons during the acute phase of murine experimental colitis and found that treatment with recombinant MFG-E8, but not its arginine-glycine-aspartate mutant counterpart, ameliorated colitis by reducing inflammation and improving disease parameters. To reveal the MFG-E8-mediated antiinflammatory mechanism, we employed an in vitro system, which showed the down-regulation of NF-κB in an LPS-dependent manner. Additionally, MFG-E8 altered αvβ3 integrin-mediated focal adhesion kinase phosphorylation by impeding the binding of one of its potent ligands osteopontin, which becomes activated during colitis. Taken together, our results indicated that MFG-E8 has a novel therapeutic potential for treatment of colitis.
Ulcerative colitis (UC)3 and Crohn’s disease (CD) are two major forms of inflammatory bowel disease (IBD), which are characterized by an elevated production of inflammatory mediators to induce inflammation, as well as tissue injury as a result of the migration and infiltration of innate immune cells (1, 2). Activated proinflammatory genes in the intestinal mucosa are in turn under the control of NF-κB, which becomes up-regulated and induces a variety of inflammatory events during colitis (3, 4). Active intestinal inflammation is also accompanied by the up-regulation of certain integrins, including αvβ3 and αvβ5, while blockade of the murine angiogenic endothelial marker αvβ3 effectively decreased both neoangiogenesis and inflammation in a murine model of colitis (5, 6, 7). Considering these inflammatory consequences, several conventional therapeutic approaches for IBD, mainly based on suppression and control of inflammation, have been proposed (8, 9, 10). Also, recently developed novel cytokine antagonist therapies have been found to be effective in certain IBD patients (11, 12, 13). Such molecular targeted inhibition of the inflammatory process may provide better therapeutic options for IBD, and various studies have been conducted to evaluate new innovative approaches.
Milk fat globule-epidermal growth factor 8 (MFG-E8), a glycoprotein secreted from different cell types, participates in phagocytosis of apoptotic cells by forming a link between phosphatidylserine on apoptotic cells and αvβ3 integrin on phagocytes (14, 15, 16, 17). Several lines of evidence show that severe inflammatory and autoimmune consequences in MFG-E8-null mice are due to the infiltration of apoptotic cells, which subsequently causes dysregulated immune functions with abnormal homeostasis (18, 19). An MFG-E8-mediated potential therapeutic benefit is evident in sepsis or GM-CSF-deficient conditions, as it ameliorates the detrimental effects of accumulated apoptotic cells in organ systems (20, 21). In addition to this familiar scavenging function, MFG-E8 was also found to be effective in accelerating mucosal healing during intestinal injury in a phosphatidylserine-dependent manner (22). Although MFG-E8 is involved in several cell surface-mediated regulatory events and modulates immune responses in numerous conditions, the mechanism underlying these features has not been clearly defined. Furthermore, its expression and role in intestinal homeostasis are not fully elucidated.
The aim of the present study was to analyze the expression of MFG-E8 in normal and inflamed intestinal mucosa, and to investigate whether it has a protective role in colitis. We decided to use recombinant MFG-E8 in a model of acute phase of colitis and elucidate its effects by evaluating colitis parameters. Our findings indicate that MFG-E8-mediated novel antiinflammatory effects are generated by NF-κB inhibition through the modulation of αvβ3 integrin signaling.
Materials and Methods
Reagents and Abs
Escherichia coli LPS (0111:B4 strain) and purified flagellin from Salmonella typhimurium (InvivoGen), mouse osteopontin (OPN) and dexamethasone (Sigma-Aldrich), recombinant integrin αvβ3 3 v integrin (Abcam), and anti-phospho-β3 (pY759) integrin (BD Biosciences).
Preparation of recombinant MFG-E8
Mouse wild-type and arginine-glycine-aspartate (RGD) mutant MFG-E8 proteins were prepared as described previously (18). Briefly, coding regions excluding the signal peptide sequence of MFG-E8 (NM_008594) were cloned into the EcoRI and XhoI sites of a 6× His pTriEx-3hygro vector to generate pMFG-E8, and transfected into HEK293 cells using Lipofectamine 2000. At 48 h after transfection, cells were lysed in mild lysis buffer (50 mM NaH2PO4, 300 mM NaCl, 10 mM imidazole, 1 mM (p-amidinophenyl)methanesulphonyl fluoride hydrochloride (pH 8.0)) by sonication on ice and purified using Ni-NTA columns, then the purity was checked by SDS-PAGE and Western blotting assays. For SDS-PAGE, purified proteins were run on a 12% polyacrylamide gel and stained with Coomassie brilliant blue, while for Western blotting the proteins were allowed to react with anti-6× His Ab, and the corresponding signals were detected using ECL (GE Healthcare). Similarly, RGD mutant MFG-E8 was produced by introducing a point mutation to generate D89E containing a p-mutant MFG-E8 expression vector, after which expression was confirmed and purification performed using the mammalian expression systems described above. The functional efficiency of recombinant MFG-E8 protein was evaluated based on its ability to enhance phagocytosis of apoptotic cells using fluorescent microscopy, as described previously (23).
Colitis induction and treatment of recombinant MFG-E8 in mice
Seven-week-old male specific pathogen-free BALB/c mice (Charles River Laboratories) were housed according to our institutional guidelines with the approval of the Ethics Committee of Shimane Medical University. Primarily, a single group of normal mice (n = 3) were euthanized to screen tissue-specific MFG-E8 expression using Northern and Western blotting methods. To produce a DSS colitis model, a group containing five mice was fed 2.5% DSS in drinking water for 9 days, while the control group received only normal drinking water throughout the experiment. Then, purified recombinant wild-type or mutant MFG-E8 was diluted in PBS and 300 μl of the solution (30 μg/kg) was injected i.v. through the tail vein, starting from 2 days before DSS administration, which continued until euthanasia. The parameters for colitis evaluation recorded in the experiments were body weight, colon length, and rectal bleeding, as determined by visual inspection. After stopping DSS treatment, the mice were euthanized and the colon was measured with a ruler on a nonabsorbent surface. For histology, 3-μm-thick formalin-fixed, paraffin-embedded colon tissues were stained with H&E and examined under a light microscope by two investigators in a double-blinded fashion.
Quantitative real-time PCR
Protein extraction, Western blot, and immunoprecipitation
Protein extraction and Western blotting assays were performed as described previously (23). Briefly, after blocking with 10% skim milk (Difco), the membrane was reacted with anti-mouse MFG-E8 Ab at a concentration of 1 μg/ml for 1 h at room temperature and the resulting signals were visualized using ECL reagent. Similarly, Western blotting assays of IκB, p-IκB, αv/β3 integrin, and β-actin were performed using their respective Abs in optimized conditions. For immunoprecipitation, mouse peritoneal macrophages were lysed in Nonidet P-40 buffer containing 150 mM Nacl, 1% Nonidet P-40, 50 mM Tris (pH 8.0), and 1 mM PMSF; after clarification, 200 μg of protein was immunoprecipitated with anti-β3 integrin Ab and protein G-Sepharose (GE Healthcare). Immunoprecipitants were subjected to Western blotting using anti-phospho-β3 and β3 integrin Abs. Similarly, pFAK (phospho-Y397) and FAK assays were performed as described above.
Immunohistochemistry of MFG-E8
Frozen colonic tissue samples were sliced into 3-μm-thick sections and fixed in cold acetone for 20 min. After blocking endogenous peroxidase activity with 0.3% hydrogen peroxide in methanol, sections were incubated for 2 h at room temperature with 1 μg/ml MFG-E8 specific primary Ab and then processed with the corresponding protocols using an immunoperoxidase staining kit (Vectastain).
Transfection, luciferase assay, and RNA interference
For in vitro experiments, mouse macrophage-like cells P388D1 from American Type Culture Collection were grown and plated on 24-well plates (2.5 × 104 cells/well) in RPMI 1640 with 10% FBS in a humidified chamber. After 18–24 h the cells had reached 50% confluence and were transiently transfected with pNF-κB-Luc (200 ng/well) and pRL-TK-Renilla-Luc (20 ng/well) using Lipofectamine 2000 (2.5 μl/well), and then NF-κB activity was measured using a dual-luciferase reporter assay system (Promega). In another experiment, custom FlexiTube small interfering RNAs (siRNAs) for αv/β3 integrin (Qiagen) were transfected into P388D1 cells according to the manufacturer’s protocol, and siRNA effects were checked by real-time PCR and Western blotting assays, as described above.
MFG-E8 binding assay
To investigate the binding of MFG-E8 to αvβ3 integrin, 96-well polyvinyl chloride microtiter plates were coated by pipetting 50 μl of recombinant αvβ3 integrin (20 μg/ml) diluted in coating buffer and incubated overnight at 4°C. Following incubation, the coating solution was removed and the plates were washed twice with PBS. The remaining protein-binding sites in the coated wells were blocked by adding 200 μl of 1% BSA. Recombinant MFG-E8 at different concentrations was then added to the αvβ3 integrin-coated wells and incubated for 2 h at room temperature. After subsequent washing, the wells were treated with HRP-labeled anti-mouse MFG-E8 Ab at a concentration of 1 μg/ml and the resulting absorbance was measured using a plate reader. Similarly, to measure the competitive binding of OPN and MFG-E8 to αvβ3, a fixed amount of exogenous OPN (500 ng/ml) was allowed to bind with αvβ3 integrin in coated wells in the presence of various concentrations of recombinant MFG-E8. After 2 h of incubation at room temperature, the unbound materials were removed by washing with PBS, and HRP-labeled anti-mouse OPN Ab (1 μg/ml) was added to each well. Finally, the resulting signals were measured at a wavelength of 450 nm and analyzed using the CurveExpert 1.3 software package.
Enzyme immune assay (EIA)
Cytokines and MPO contents in colonic tissues and culture supernatants were estimated using EIA according to the manufacturer’s protocol. Briefly, total proteins from distal colonic tissues were extracted using lysis buffer (200 mM NaCl, 5 mM EDTA, 10 mM Tris, 10% glycine, and 1 mM PMSF (pH 7.4)) and subjected to EIA using mouse IL-1β, TNF-α, and MPO EIA kits. In vitro culture supernatants from mouse peritoneal macrophages treated with or without LPS, flagellin, OPN, or recombinant MFG-E8 were also checked for IL-1β and TNF-α contents by EIA, as described above. In another experiment, OPN contents in LPS-treated peritoneal macrophage culture supernatants were assayed using an OPN EIA kit.
All quantitative data are expressed as the mean ± SE. Student’s t test was used for statistical determinations. Values of p of <0.05 were considered statistically significant.
Altered expression of MFG-E8 in acute colitis
We examined MFG-E8 expression in different tissues of normal BALB/c mice and observed the ubiquitous presence of MFG-E8 with its two transcript variants (24) in a tissue-specific manner (Fig. 1⇓). Next, to assess the time course changes of MFG-E8 expression during colitis, mice were given 2.5% DSS in drinking water, after which MFG-E8 expression in colonic tissues was checked at various time points. As shown in Fig. 2⇓, A and B, MFG-E8 was down-regulated in inflamed colons during the DSS induction period (days 1–9), while it gradually became up-regulated during the healing phase (days 10–24), when DSS was no longer added to the drinking water. Moreover, to reveal any link between MFG-E8 down-regulation and DSS-induced colitis characterized by intestinal inflammation, we performed a dose-dependent experiment using low (1.5%) and high (3.5%) concentrations of DSS, and assessed MFG-E8 expression at different time points. Notably, in mild colitis shown by a low grade of inflammation, MFG-E8 expression was not reduced to the same extent as in severe colitis (Fig. 2⇓A). Thus, we concluded that the level of MFG-E8 down-regulation caused by different doses of DSS may be related to inflammation severity. Additionally, to confirm MFG-E8 localization, we performed immunohistochemical analyses using colonic tissue sections, which revealed its presence in lamina propria mononuclear cells that had become infiltrated during colitis (Fig. 2⇓C).
MFG-E8 protects mice from DSS-induced colitis
Mice treated with 2.5% (w/v) DSS in their drinking water showed clinical, histological, and immunological signs of colitis (25). Recombinant wild-type and RGD mutant MFG-E8 were separately injected into mice during the DSS period, while other mice were injected with PBS instead of MFG-E8 and received normal drinking water and served as controls (Fig. 3⇓A). As shown in Fig. 3⇓B, in the mice who received DSS with PBS, body weight loss of ∼23% commenced on day 4 and continued to day 9, while mice injected with the wild-type MFG-E8, as opposed to those with mutant MFG-E8, showed significant improvement in body weight loss from day 4, which reached only 14% on day 9. Representative specimens from the dissected colons from the MFG-E8-treated group showed increased length as compared with those from the PBS- or mutant MFG-E8-treated DSS groups (Fig. 3⇓, C and D). Histological examinations also showed that lamina propria infiltration by mononuclear cells as well as crypt epithelial damage were markedly decreased in the MFG-E8-treated colitic mice (Fig. 3⇓E). Furthermore, total histological scores in the distal parts of colon samples from MFG-E8-treated mice after 9 days of DSS administration were significantly lower than in those from the other groups (Table II⇓).
To investigate the effects of MFG-E8 on proinflammatory cytokine production, protein was extracted from colonic tissues, and IL-1β and TNF-α contents were measured. Notably, they were decreased in MFG-E8-treated, but not in the PBS- or mutant MFG-E8-treated, DSS groups (Fig. 3⇑, F and G). Furthermore, MFG-E8 treatment significantly down-regulated the tissue content of MPO as compared with that in PBS- or mutant MFG-E8-treated DSS mice (Fig. 3⇑H). With the present experimental system, we assessed the effects of recombinant MFG-E8 using body weight change as a parameter of colitis, which was apparent at the beginning of day 6 of DSS treatment. Therefore, we also performed histology and immunology experiments using samples from day 6 as an additional time point. As shown in Fig. 3⇑E–H, although the effects of MFG-E8 on day 6 samples were evident, they were more distinct in samples from day 9. Taken together, these results demonstrate that recombinant MFG-E8 has a protective role in colitis.
We also noticed that the colitis was refractory in mice treated with the RGD mutant form of MFG-E8. In a previous study of apoptotic cell phagocytosis by Asano et al., the RGD mutant MFG-E8 (D89E) was shown to be a dominant negative form of MFG-E8 (18). While performing the present experiment to evaluate the functional efficacy of our purified recombinant proteins, we noticed inhibition of phagocytosis of apoptotic cells following addition of purified recombinant RGD mutant MFG-E8 to cocultured cells (macrophage and apoptotic thymocytes) (data not shown), whereas we did not observe any dominant negative effects on DSS-induced colitis or in subsequent in vitro studies. Therefore, it is possible that our RGD mutant protein was nonfunctional rather than having any dominant-negative effects on colitis.
MFG-E8 down-regulates proinflammatory cytokines through αvβ3 integrin-mediated NF-κB inhibition
To elucidate the detailed mechanism of the MFG-E8-dependent antiinflammatory effect on experimental colitis, we used bacterial LPS, which is a potent inducer of inflammatory signals in cultured murine peritoneal macrophages treated with or without MFG-E8 in vitro. As expected, LPS markedly induced IL-1β and TNF-α in culture supernatants, while cells pretreated with recombinant MFG-E8 significantly down-regulated the effects of LPS on cytokine production, whereas mutant MFG-E8 had no such effects on LPS-stimulated macrophages (Fig. 4⇓, A and B). We then examined the effects of recombinant MFG-E8 on NF-κB status in LPS-treated P388D1 cells. Following LPS treatment, NF-κB activity became considerably elevated, while cells exogenously pretreated with recombinant MFG-E8 showed significant down-regulation of the LPS-induced NF-κB activity, whereas mutant MFG-E8 had no significant effects on this event (Fig. 4⇓C). Other than LPS, several pathogen-associated molecular patterns are also known to activate the innate immune responses in immune-reactive cells via TLRs, and therefore it is of interest to check whether recombinant MFG-E8 inhibits the flagellin (as an additional stimulant)-induced effects on mononuclear cells. Interestingly, we noticed an MFG-E8-dependent significant down-regulation of proinflamatory cytokine production, as well as NF-κB activity in flagellin-treated macrophages, although the RGD mutant protein had no such effects in these events (Fig. 4⇓D–F). According to these findings, macrophage cells are less responsive to flagellin compared with LPS stimulation, and we therefore conducted our subsequent in vitro studies using LPS as one of the stimulants for the initiation of inflammation in macrophages. We also observed that the optimum degradation and phosphorylation of IκB occurred at 30 min after LPS treatment, and that treatment with the wild-type MFG-E8, but not RGD mutant MFG-E8, considerably reduced the effects of LPS on NF-κB at that time point (Fig. 4⇓, G and H).
Next, to confirm the involvement of αvβ3 integrin in MFG-E8 functions, we utilized αvβ3 integrin siRNA in P388D1 cells and assessed NF-κB activity in LPS-treated conditions, after validating the siRNA efficiency of corresponding gene knockdown at both the mRNA and protein levels (Fig. 5⇓, A and B). In P388D1 cells transfected with the control siRNA, MFG-E8 inhibited LPS-induced NF-κB activation, which was similar to the results observed in cells without siRNA transfection. On the other hand, in the presence of both αv and β3 integrin siRNAs, MFG-E8 did not show a significant inhibitory effect on NF-κB activation in LPS-treated cells (Fig. 5⇓C). These results suggest that the antiinflammatory role of MFG-E8 via inhibition of NF-κB is dependent on αvβ3 integrin signaling in P388D1 cells. Based on these observations, we speculated that MFG-E8 may competitively bind to αvβ3 integrin with certain potential ligands existing in the culture medium and inhibit αvβ3 integrin-mediated NF-κB activation in macrophages.
MFG-E8 interrupts OPN effects for NF-κB activation during inflammation
Previous studies have shown that several extracellular matrix proteins (ECMs), including OPN, vitronectin (VN), and fibronectin (FN), are potential ligands for αvβ3 integrin and regulate a variety of physiological and pathological functions in various organs (26, 27). To explore potential ligands related to the association of αvβ3 integrin with the antiinflammatory role of MFG-E8, we initially examined the expression of colonic ECMs during experimental colitis. As shown in Fig. 6⇓A, a robust induction of OPN was observed in the early phase of DSS-induced colitis, while other ligands (VN and FN) were increased during the regeneration phase of colitis after the end of DSS administration. Consistent with these in vivo results, we also noticed an increased production of OPN by peritoneal macrophages in an LPS-dependent manner (Fig. 6⇓, B and C). Recently, several studies have shown that OPN is one of the potential ligands used by αvβ3 integrin to participate in innate immune responses during colitis or other autoimmune diseases via the activation of NF-κB (5, 28, 29, 30, 31). In consideration of those findings, we focused on OPN function to evaluate the antiinflammatory role of MFG-E8 in macrophages and designed several kinds of in vitro experiments. In this regard, we first confirmed MFG-E8 binding to αvβ3 integrin and investigated its relative binding affinity as compared with OPN based on a dose-response curve of varying MFG-E8 amounts with a fixed amount of OPN. As shown in Fig. 7⇓A, MFG-E8 exhibited a strong binding affinity to αvβ3 integrin with an apparent Kd value of ∼3.05 nM. Moreover, we also found that OPN binding to αvβ3 become decreased with increased concentrations of recombinant MFG-E8. In contrast, when the relative Kd value was ∼4.45 nM, it inhibited OPN by competitive binding to αvβ3 integrin (Fig. 7⇓B).
To explore the possible role of OPN in production of proinflammatory cytokines by macrophages, we blocked the function of secreted OPN using a neutralizing Ab with LPS-treated cells. Treatment with the neutralizing Ab for OPN significantly decreased the production of IL-1β and TNF-α (Fig. 7⇑, C and D), indicating that OPN secreted by macrophages induces cytokine production via αvβ3 integrin in an autocrine-dependent manner. These results also support our speculation that MFG-E8 may inhibit αvβ3 integrin signaling mediated by other ligands (OPN) existing in culture medium treated with LPS. To precisely elucidate the antiinflammatory role of MFG-E8, we used recombinant OPN in various in vitro experiments (Fig. 8⇓). Exogenous stimulation with recombinant OPN induced the production of IL-1β and TNF-α in LPS-treated macrophages, which was clearly related to NF-κB activation and IκB phosphorylation. As expected, treatment with MFG-E8 significantly decreased OPN-induced cytokine production and NF-κB activation in the cells in a dose-dependent manner (Fig. 8⇓A), whereas no effect was observed in cells treated with mutant MFG-E8. Taken together, these results clarify the potential role of OPN to induce inflammatory cascades via αvβ3 integrin-mediated NF-κB activation, whereas those are interrupted in the presence of MFG-E8.
MFG-E8 modulates FAK via αvβ3 integrin
Although our in vitro results indicate that OPN induces proinflammatory cytokine production via αvβ3 integrin signaling, its effect was clearly observed only in the LPS-treated conditions. Tyrosine phosphorylation of the β3 integrin subunit is an essential event for αvβ3 integrin-mediated functions (32, 33), as it facilitates the binding of its potential ligands and subsequently activates FAK to mediate several downstream signaling pathways (34, 35, 36). Thus, we examined whether time course changes of LPS stimulation alter the status of β3 integrin and FAK status in peritoneal macrophages. Following 15–60 min of stimulation with LPS, increased phosphorylation of the β3 integrin subunit as well as FAK were clearly observed in the cells (Fig. 9⇓A), suggesting that LPS-induced activation of the integrin further promotes the binding of its potential ligands to induce integrin-dependent cellular responses. To further investigate whether OPN binding to activated αvβ3 integrin generates the phosphorylation of FAK, cells were treated with LPS then exposed to recombinant OPN with or without MFG-E8 proteins, after which the phosphorylation of FAK was examined in total FAK pools. As shown in Fig. 9⇓B, OPN stimulation induced abundant phosphorylation of FAK (lanes 3 and 7), while that was markedly inhibited by wild-type MFG-E8 (lanes 4 and 8), but not by mutant MFG-E8 (lanes 5 and 9). These findings indicate the possible role of MFG-E8 to interfere with αvβ3 integrin-mediated OPN functions by modulating FAK phosphorylation. Interestingly, phosphorylation of FAK was also observed in LPS-treated cells without OPN stimulation (lanes 2 and 6).
MFG-E8 was originally identified in the process of phagocytic clearance of apoptotic cells (17), while the present novel findings indicate its antiinflammatory function via the modulation of NF-κB activity during acute colitis. Our results showed a decreased production of MFG-E8 during the initiation of colitis, whereas it was up-regulated during the healing phase to restore homeostasis mechanisms. In the present study, we investigated whether down-regulation of MFG-E8 is dependent on the severity of inflammation. Our results were consistent with those of other studies of atherosclerotic mice characterized by severe inflammation (37), and they indicate a reduced level of suppressive activity by regulatory T cells and enhanced atherosclerotic plaque formation. We also noted that MFG-E8 effectively ameliorates the development of experimental colitis by attenuating inflammation and disease status. Using an in vitro model, we further confirmed that the antiinflammatory effects of MFG-E8 are mediated via αvβ3 integrin by modulating the family of protein tyrosine kinases, which finally inhibit NF-κB activation.
Recent reports have revealed the essential role of MFG-E8 to maintain normal tissue homeostasis by clearing apoptotic cells from the body, and preserving the balance between pro- and antiinflammatory cytokines during inflammation (18, 19). Because the pathogenesis of IBD is mainly due to a disorder of intestinal immune homeostasis, a therapeutic role of MFG-E8 without notable side effects is an important finding. Based on our speculation, we treated mice with recombinant MFG-E8 and observed its benefits to control intestinal inflammation, which led us to consider that the antiinflammatory role of recombinant MFG-E8 might be generated via enhanced clearance of apoptotic cells during acute colitis. However, previous studies that used an MFG-E8 knockout murine model reported no increase in accumulated apoptotic cells in small intestinal crypts (22), while apoptotic B cells were found infiltrated only in the germinal centers in spleens of MFG-E8-null mice (19). Those results suggest that the in vivo molecular mechanisms used by MFG-E8 to remove apoptotic cells vary depending on tissue type. Thus, we speculated that MFG-E8-mediated antiinflammatory effects are facilitated not only by removal of apoptotic cells, but there may be some other mechanisms by which recombinant MFG-E8 performs an antiinflammatory role during colitis.
Expression of proinflammatory cytokines is under the control of the potent transcription factor NF-κB and is increased in intestinal lamina propria of patients with IBD (38, 39). Recently, we reported that decoy oligodeoxynucleotides targeting NF-κB attenuate intestinal inflammation in murine experimental colitis, indicating that blockade of NF-κB-mediated signaling is a potent therapeutic strategy for IBD (25). In the present study, lower levels of proinflammatory cytokines in inflamed colonic mucosa were observed in MFG-E8-treated mice, and thus we speculated that MFG-E8 may inhibit intestinal inflammation via modulation of NF-κB-related pathways. To reveal a plausible antiinflammatory mechanism of MFG-E8, we established an in vitro model by utilizing LPS, one of the potent inducers of the inflammatory cascade in immune-reactive cells, and observed that treatment with recombinant MFG-E8 significantly down-regulated LPS-induced proinflammatory cytokines by modulating NF-κB. Once the involvement of NF-κB in this event was confirmed, we investigated how MFG-E8 inhibits NF-κB activation in LPS-dependent conditions. Several reports have shown that RGD mutant MFG-E8 protein decreases phagocytosis of apoptotic cells by macrophages due to the inability of binding to αvβ3 integrin (18, 19). In the present study, RGD mutant MFG-E8 protein did not show any inhibitory effects toward NF-κB activation in LPS-treated cells, suggesting that the antiinflammatory role of MFG-E8 may depend on αvβ3 integrin-mediated intracellular signaling. To explore this, we investigated the relevance of the αvβ3 integrin using an RNA interference technique and noticed that blocking of αvβ3 integrin reversed the potential effects of MFG-E8 for inhibiting NF-κB in LPS-treated in vitro conditions. These findings clarify the two major points of our study: first, the involvement of αvβ3 integrin to generate MFG-E8 effects; and second, the possibility of the presence of certain ligands that competitively share the same αvβ3 integrin with MFG-E8.
OPN is an extracellular matrix phosphoprotein and contains the RGD domain, which is predominantly expressed in macrophages, activated T cells, osteoblasts, and epithelial cells (40) and induces the production of NF-κB-mediated inflammatory cytokines after binding to αvβ3 integrin (29). In the present study, we observed increased OPN expression in DSS-induced inflamed colonic tissues, and our in vitro experiments with the neutralizing Abs for OPN and recombinant OPN clearly showed that OPN is a potent mediator to induce proinflammatory cytokines in peritoneal macrophages via NF-κB activation. Recently, Zhong et al. reported that OPN-null mice demonstrated significantly inhibited disease activity of DSS-induced colitis as compared with wild-type mice, as shown by reduced levels of rectal bleeding, weight loss, and histological intestinal injury (41), which supports our speculation that OPN plays an important role in the development of intestinal inflammation. Additionally, the present findings clarified that wild-type MFG-E8 protein but not the RGD mutant significantly inhibited OPN-induced production of proinflammatory cytokines in both autocrine (Fig. 7⇑, C and D) and paracrine (Fig. 8⇑) models using cultured macrophages. Taken together, these results suggest the possibility that MFG-E8 interferes with OPN binding to αvβ3 integrin for downstream signaling toward NF-κB activation.
In the present in vivo and in vitro experiments, the effects of recombinant MFG-E8 were evident only in DSS colitis or LPS-treated conditions, which indicates that an inflammatory environment is required for effective MFG-E8 functioning. Although the essential pathway of TLR-4-mediated signaling for NF-κB activation has been widely recognized (42), there may also be other surface molecules that play a complementary role in TLR-4-induced events. Integrins are a class of receptors that have been linked to LPS signaling, and a recent study revealed that TLR-4 signaling mediates FAK phosphorylation, with subsequent phosphorylation of integrin leading to increasing binding of its ligands (43). After ligand binding, the activation of integrin also results in recruitment of phosphorylated FAK (pFAK), leading to NF-κB activation (35, 36). Consistent with the above reports, we found that in vitro treatment of macrophages with LPS resulted in increased phosphorylation of FAK and β3 integrin to generate “inside-out signaling” (TLR-4 signaling → pFAK → αvβ3 integrin activation), and such treatment facilitated enhanced binding of exogenous OPN to generate downstream “outside-in signaling” (OPN binding to activated αvβ3 integrin → pFAK → NF-κB activation) by augmenting FAK phosphorylation. By targeting this pathway, we found that MFG-E8 markedly reduced LPS-induced NF-κB activation by blocking OPN binding and modulated αvβ3 integrin-dependent and FAK-mediated downstream signaling (Fig. 9⇑C).
Recent studies have reported that high levels of both plasma and tissue OPN were observed in patients with IBD (44, 40, 29), suggesting that OPN, which regulates proinflammatory and T cell-mediated immune responses, may be part of a new therapeutic strategy for the disease. Although our present findings showed a beneficial antiinflammatory role of recombinant MFG-E8 in acute murine model of colitis, there are several points that require further clarification before clinical use. Since IBD refers to a chronic, relapsing form of intestinal disorder, the role of MFG-E8 in chronic models of colitis must be addressed further. Additionally, although we assessed the effects of MFG-E8 with noncolitic healthy mice, it was given for only a short period, and the long-term effects of MFG-E8 in regard to physiological, immunological, and clinical aspects should be evaluated in the future. The present findings also suggest the possibility of a downstream pathway initiated by activation of the family of protein tyrosine kinases, which implicates the need of future studies to determine the nature of the downstream pathway that links the LPS-induced activation of αvβ3 integrin and outside-in-mediated signaling for NF-κB during inflammation. Furthermore, LPS-mediated integrin activation leads to induction of other signaling pathways, including the MAPK family. Herein, we evaluated NF-κB as one of the LPS-dependent integrin-mediated signaling events and also focused on other pathways.
In summary, we investigated the antiinflammatory effects of MFG-E8 against experimental colitis in mice and report for the first time that MFG-E8 attenuated intestinal inflammation, indicating the potential of targeting NF-κB-mediated proinflammatory cytokines in the development of new therapies for IBD.
We thank Rika Tohma and Keiko Masuzaki for their technical support.
The authors have no financial conflicts of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Science, Sports, and Culture of Japan.
↵2 Address correspondence and reprint requests to Dr. Shunji Ishihara, Department of Internal Medicine II, Shimane University, Faculty of Medicine, 89-1, Enya-cho, Izumo, Shimane, Japan. E-mail address:
↵3 Abbreviations used in this paper: UC, ulcerative colitis; CD, Crohn’s disease; IBD, inflammatory bowel disease; MFG-E8, milk fat globule-epidermal growth factor 8; DSS, dextran sodium sulfate; OPN, osteopontin; MPO, myeloperoxidase; FAK, focal adhesion kinase; RGD, arginine-glycine-aspartate; siRNA, small interfering RNA; EIA, enzyme immune assay; ECM, extracellular matrix protein; VN, vitronectin; FN, fibronectin.
- Received November 11, 2008.
- Accepted February 27, 2009.
- Copyright © 2009 by The American Association of Immunologists, Inc.