Microbial metabolite butyrate facilitates M2 macrophage polarization and function

Metabolites from intestinal microbes modulate the mucosal immune system by regulating the polarization and expansion of T cells. Whether the microbial metabolites influence macrophage polarization, however, is poorly understood. Here, we show that the large bowel microbial fermentation product, butyrate, facilitates M2 macrophage polarization, in vitro and in vivo. The supernatant from butyrate-treated M2 macrophage increased the migration and enhanced the wound closure rate of MLE-12 cells. Butyrate attenuated intestinal inflammation in mice with dextran sulfate sodium (DSS)-induced colitis, with a significant increase in colonic expression of the M2 macrophage-associated protein, Arg1. M2 macrophage treated with butyrate, had increased activation of the H3K9/STAT6 signaling pathway, suggesting a mechanism for butyrate facilitated M2 macrophage polarization. Collectively, our study indicated that commensal microbe-derived butyrate is a novel activator of STAT6-mediated transcription through H3K9 acetylation driving M2 macrophage polarization, and delineated new insights into the immune interplay underlying inflammatory bowel disease.


Results
Butyrate facilitates polarization of M2-BMDMs. The cells were morphologically typical of bone marrow-derived macrophages (BMDMs) (Supplementary Fig. 1a) and 99.9% were F4/80 + as determined by flow cytometry (Supplementary Fig. 1b,c). Flow cytometry and Cell Counting Kit-8 assays showed no changes in BMDMs' viability following butyrate treatment ( Supplementary Fig. 2). To investigate the effect of butyrate on polarization of M2 macrophage, we first examined the hallmarks of M2 macrophage expression in BMDMs directly treated with butyrate. The results showed that treatment of M0 macrophages with butyrate did not stimulate the upregulation of M2 markers (Arg1, Fizz1, Ym1 and MR) ( Supplementary Fig. 3). The data suggested that butyrate did not itself drive polarization of M0 macrophages towards the M2 phenotype. To determine whether butyrate facilitated IL-4-driven M2 polarization, M0 macrophages were activated by IL-4 in the presence of various concentrations (1 to 200 μg/ml) of butyrate. The relative expression of genes encoding Arg1, Fizz1, and Ym1, but not MR, was considerably higher in M2-BMDMs exposed to 50 μg/ml butyrate for 24 h (Fig. 1a). Meanwhile, stimulation with butyrate did not alter the surface expression of costimulatory molecules, CD80 or CD86, on M2-BMDMs ( Supplementary Fig. 4). Butyrate clearly increased the expression of M2 macrophage marker genes in M2-BMDMs and 50 μg/mL was then used to determine its influence the progress of polarization, again triggered by IL-4. Compared to its absence, exposure of M2-BMDMs to butyrate significantly enhanced expression of Arg1, Fizz1, and Ym1 after 12 h treatment with IL-4 and the effect on Arg1 was most striking (Fig. 1b). Western blot analysis of Arg1 protein content at 24 h was consistent with the enhanced gene expression (Fig. 1c). Collectively, these data indicated that butyrate enhanced the IL-4-induced expression of M2-associated markers.
Butyrate enhances migration and wound-healing properties in M2-BMDMs. M2 macrophage promotes the resolution of wound healing by facillating cell migration to damaged tissues 14,15 . To analyze if butyrate influences M2 macrophage function, wound-healing assay was performed using MLE-12 bronchial cells in an in vitro scrape model. As shown in Fig. 2a, supernatants from M2-BMDMs increased wound closure compared with those from the M0-BMDMs (p < 0.05) and this effect on re-epithelialization was improved further when M2-BMDMs had been treated with butyrate. As shown in Fig. 2b, medium conditioned by M2-BMDMs for 24 h promoted the migration of MLE-12 cells in transwell filter assays. The number of MLE-12 cells in the permeating septum after 12 h more than doubled (p < 0.05) in the presence of medium conditioned by M2-BMDMs (a) M0-BMDMs were directly exposed in series concentrations of butyrate for 24 h. Expression of Arg1, Fizz1, Ym1, and MR was measured by quantitative RT-PCR. (b) The relative quantity of Arg1, Fizz1, Ym1, and MR mRNA transcribed by nonpolarized (M0-BMDMs, gray bars) or M2-BMDM after 3, 6, 12, and 24 h with IL-4 in the absence (white bars) or presence (black bars) of 50 μg/mL butyrate. (c) Arg1 protein levels in the absence or presence of 50 μg/mL butyrate in M2-BMDMs after 24 h treatment. Data are mean ± SD for at least three independent experiments. Comparisons between means used t tests (*P < 0.05, **P < 0.01, ***P < 0.001).
rather than M0 and doubled again (p < 0.01) when conditioning had occurred in the presence of butyrate. Together, these data demonstrated that butyrate significantly enhanced the functional ability of M2-BMDMs promote cell migration and wound-healing. Consistent with this, the expression of relevant C-C cytokine genes, CCL2, CCL17, and CCL22 was consistently higher in butyrate-treated M2-BMDMs (Fig. 2c).
Butyrate ameliorates dextran sulfate sodium (DSS)-induced colitis in mice. M2 macrophage plays an important role in colonic infection with parasitic nematodes 16 . Furthermore, an increase in intestinal M2 macrophage has been associated with reducing the severity of colitis in mice 17 . We first evaluated the therapeutic effect of butyrate on DSS-induced colitis as a model of inflammatory bowel disease (IBD). Butyrate-treated mice showed much milder symptoms than the untreated colitis control mice, as evidenced by a slower rate of weight loss (Fig. 3a) and a lower disease activity index (DAI; Fig. 3b). Meanwhile, butyrate could promote the growth in mice and the results are consistent with the butyrate can improve the growth performance of animals [18][19][20] . Histological examination of the colons showed less mucosal damage, demonstrating the protective effect of butyrate against colitis in the DSS-induced colitis model (Fig. 3c,d). Following butyrate treatment, reduced levels of proinflammatory cytokines, namely IL-6, TNF-α , and IL-1β , were detected in the serum compared with control mice (Fig. 3e). Moreover, we found that butyrate-treated mice had more Arg1 protein expression in colonic tissues (Fig. 3f). Taken together, these data showed that butyrate-treated mice have reduced clinical disease and histological damage compared with control mice, and this protection during DSS-induced colitis may correlate with increased in Arg1 protein expression in the colon during inflammation.

M2 macrophage confers protection to mice during DSS-induced colitis.
To further elucidate the importance of M2-BMDMs in IBD, we adoptively transferred M0-and M2-BMDMs into colitis-bearing recipient mice. There was no significant difference between the mice treated with M0-BMDMs and the colitis-only control group. In contrast, mice receiving M2-BMDMs showed significantly less weight loss compared with the colitis-only control group, with lowered DAI (Fig. 4a,b) and improved histological condition of the colon (Fig. 3c,d). Moreover, the M2-BMDM recipient mice possessed less IL-6, TNF-α , and IL-1β in the serum than the colitis-only control mice (Fig. 4e). Furthermore, increased Arg1 protein expression was detected in the colonic tissues of M2-BMDM recipient mice than in control tissues (Fig. 4f). These data demonstrated that M2 macrophage can protect against DSS-induced colitis in mice. Migrated cells at the lower surface of the transwell filter were stained with DAPI and counted. (c) Total RNA was isolated from M0-BMDMs, and M2-BMDMs, untreated or treated with butyrate, and analyzed by RT-PCR for the expression of representative M2 macrophage chemokine genes. Data are shown as mean ± SD for three independent experiments, Comparisons between means used t tests (*P < 0.05, **P < 0.01, ***P < 0.001).
Butyrate reduces histone H3K9 deacetylase in M2-BMDMs. Butyrate, as a histone deacetylase (HDAC) inhibitor, is known to regulate gene expression epigenetically 22,23 . Whether butyrate alters the histone acetylation of BMDMs to epigenetically regulate the transcription of genes responsible for M2-BMDM development or function is still unknown. We hypothesized that butyrate enhanced the polarization of M2 macrophage involved the HDACs inhibition. To test this hypothesis, M2-BMDMs were stimulated with butyrate for 6 h and the expression of HDAC genes was analyzed. The expression of HDAC1, HDAC6, HDAC7, and HDAC9 was reduced in M2-BMDMs, while no differences were observed in HDAC2 and HDAC3 gene expression. Butyrate-treated M2-BMDMs had lower HDAC1 gene expression than M2-BMDMs (Fig. 6a). Western blot analysis demonstrated that butyrate treatment of M2-BMDMs enhanced histone 3 lysine 9 acetylation (H3K9) (Fig. 6b). These results suggested that butyrate enhanced M2-BMDMs polarization partly through HDAC1 inhibition and histone H3K9 acetylation.

Discussion
In this study, we have identified that a metabolite derived from commensal microbes, butyrate, effectively facilitates M2 macrophage polarization. Although butyrate did not directly affect the polarization of M0-BMDMs to M2-BMDMs, it could promote a greater degree of polarization in IL-4-induced M2 macrophage. Additionally, butyrate enhanced the critical functions of M2 macrophage by improving re-epithelialization and the migration of MLE-12 cells. Furthermore, we demonstrated that butyrate attenuates intestinal inflammation in mice DSS-induced colitis model partly via M2 polarized macrophage.
The DSS-induced colitis model is suitable to study the contribution of innate immune cells to colitis, particularly to epithelial repair mechanisms 24,25 . After DSS-induced gut injury, macrophages are recruited to the colon and play a central role in disease pathogenesis 26,27 . Previous studies showed that a defect in M2 macrophage polarization led to increased severity in DSS-induced colitis, whereas adoptive transfer of in vitro-derived M2 macrophage could attenuate colonic inflammation in mice 17,28,29 . Recent evidence showed that another SCFA, acetate, in the drinking water markedly attenuated disease indices, with increased colon length, decreased DAI, and reduced TNF-α levels in germ-free mice treated with DSS 30 . Here, we employed the DSS-induced colitis  model and demonstrated that M2 macrophage were indeed responsible for the therapeutic effects of butyrate on IBD. These findings supported the role of M2 macrophage polarization in the development of inflammatory diseases, and they showed that commensal microbe-derived butyrate, as a new effector molecule, could ameliorate disease severity in DSS-induced colitis by enhancing the polarization of M2 macrophage. It has been suggested that Arg1, expressed by M2 macrophage, is important for tissue repair in diverse models and tissues 14,15,[31][32][33] . Our results show increased Arg1 protein in butyrate-treated mice, associated with less mucosal damage and lower levels of proinflammatory cytokines.
It is well known that the classical pathway for IL-4-induced M2 macrophage polarization involves the STAT6 signaling pathway. The polarization of M2 macrophage is mediated by IL-4-dependent activation of STAT6, which leads to the expression of Fizz1, Ym-1, and Arg1 in macrophages 34 . These results provided a molecular mechanism for the butyrate-mediated protection against DSS-induced colitis, possibly by enhancing STAT6 signaling in intestinal tissues. Another possible mechanism of butyrate-enhanced M2 polarization may be by inhibiting the activity of HDACs. Previous studies have demonstrated that butyrate inhibited the activity of HDACs in many different cell types 22 . H3K9 acetylation is considered a hallmark of transcriptional activation. Another SCFA propionate treatment of Ffar2 +/+ mice reduced Treg HDAC6 and HDAC9 expression and enhanced H3K9 acetylation 6 . Treatment of BMDMs with butyrate led to a downregulation of LPS-induced proinflammatory mediators by inhibition of H3K9 acetylation 12 . To determine whether butyrate behaved as an HDAC inhibitor involved in the polarization of M2-BMDM, we checked the levels of HDACs in M2-BMDMs without butyrate. Compared with M0-BMDMs, the expression of HDAC1, HDAC6, and HDAC9 were decreased. Western blot analysis also demonstrated that H3K9 acetylation was enhanced in M2-BMDMs. Furthermore, HDAC1 expression was further reduced and the H3K9 acetylation was increased in butyrate-treated M2-BMDMs. The level of histone acetylation is tightly regulated by the opposing activities of HAT and HDACs 35 . C646 is capable of attenuating HDACs, such as in butyrate, trichostatin, and cAMP-induced H3K9 acetylation 36,37 . To further determine whether acetylation of H3K9 mediated the polarization of M2 macrophage, we treated M2-BMDMs with C646. We showed that the expression of Arg1 and H3K9 acetylation was decreased in C646-treated M2-BMDMs. Moreover, C646 efficiently inhibited the effect of butyrate on facilitating M2-BMDMs polarization. These results suggest that butyrate may facilitate M2 macrophage partly through HDAC1 inhibition.
We next investigated how butyrate modulates STAT6 phosphorylation, which is required for IL-4 induced M2-BMDMs polarization. We hypothesized that H3K9 may play an important role in regulating the level of phosphorylated STAT6 and controlling M2 macrophage polarization. Our results showed that C646 could effectively reduce the level of phosphorylated STAT6 in both M2-BMDMs and butyrate-treated-M2-BMDMs. These data demonstrated that butyrate enhanced the STAT6 phosphorylation partly via inhibiting HDAC1 gene expression and increasing H3K9 acetylation. The STAT6 signaling pathway requires various co-activators or co-repressors to promote and regulate transcription of endogenous genes. Consistent with this concept, The polypyrimidine tract binding protein (PTB)-associated splicing factor (PSF) recruits HDAC1 to the STAT6 complex, which reduces H3 acetylation at the promoter regions of the Igε gene and inhibits STAT6 signing in lymphomas 38 . IL-4 induced expression of the reticulocyte-type 15-lipoxygenase-1 (15-LOX-1) gene requires the enhancement of H3 acetylation and the activation of the STAT6 signaling pathway in A549 lung epithelial cells 39 . Our findings also support the concept that histone acetylation correlates with their post-translational modifications and coordinately regulate STAT6-mediated gene transcription. Butyrate-mediated activation of H3K9 acetylation and the STAT6 signaling pathway, and its influence on the polarization of M2 macrophage are still unclear. SCFAs bind the G-protein-coupled receptor 43 (GPR43; also known as FFAR2), which is expressed on immune cells, mediating resolution of inflammatory responses 6,30,40 . GPR43 is expressed on the colonic epithelium and butyrate provides a major source of energy for colonocytes 41 . Gpr43 −/− mice showed aggravated or unresolving inflammation in DSS-induced colitis. In a T-cell-dependent model of colitis, Gpr43 −/− mice also had more severe disease with increased colon histological scores 30 . SCFAs mediate the size and function of the colonic Treg pool and protect against colitis in a Gpr43-dependent manner in mice 5,6 . Therefore, using Gpr43 −/− mice might enable the dissection of a precise mechanism of action for butyrate on H3K9 acetylation/STAT6 signaling in the polarization of M2-BMDMs. Moreover, the inflammatory process is highly complicated, and considerably more information is required for an improved understanding about macrophages in vivo. Because macrophages are able to perform mixed M1/M2 phenotypes in pathological conditions, further studies may focus not only on M2 polarization but also on the changes in M1/M2 polarization treated with butyrate during the development of inflammatory bowel disease. In addition, the polarization of either M2 macrophage or Treg could be facilitated or induced by butyrate. Future studies may delineate a dynamic view of this inflammatory process, take M2 macrophage and Treg in systemic and local milieu into account, and define the kinetics, plasticity, reversibility, and memory of their responses in order to full elucidate the mechanism that butyrate ameliorates inflammatory bowel disease in mice.
In sum, our studies indicated that commensal microbe-derived butyrate enhanced the IL-4-mediated STAT6 transcription, achieved through H3K9 acetylation to enhance M2-BMDMs polarization, and delineated new insights into the immune interplay underlying inflammatory bowel disease. Co., Ltd (SLAC; Shanghai, China). All mice were maintained in the Experimental Animal Center of Zhejiang University. Experiments were conducted in accordance with the Chinese guidelines for animal welfare, approved by the Animal Welfare Committee of Animal Science College, Zhejiang University. M2 macrophage culture. Bone marrow-derived macrophages (BMDMs) were generated by isolating bone marrow from femurs and culturing in high glucose Dulbecco's modified Eagle's medium (DMEM, Hyclone, Logan, UT, USA) containing 10% (vol/vol) fetal bovine serum (Gibco) with macrophage colony-stimulating factor (M-CSF; 15 ng/mL; Peprotech) at 37 °C with 5% CO 2 .Half of the medium was replaced with fresh medium containing M-CSF every 3 days. After 6 days of culture, IL-4 (15 ng/mL) was added to generate full alternatively activated M2-BMDMs for 24 h. Untreated BMDMs were retained as nonpolarized macrophages (M0-BMDMs). The BMDMs were subsequently cultured for 1.5-24 h in the presence of sodium butyrate (Sigma-Aldrich) or histone acetyltransferase (HAT) inhibitor C646 (Selleck).
Cell viability assays. Cells (5, 000 cells/well) were seeded into a 96-well plate and treated with variable concentrations of butyrate for 24 h. Subsequently, the cells were labeled using a Cell Counting Kit-8 (CKK8) (Dojindo, Tabaru, Japan) for 2 h. Absorbance at 450 nm was measured with a plate reader and cell viability was expressed as the percentage of the absorbance of treated cells versus untreated cells.
Western blotting. Cell lysates containing extracted proteins were separated on 10% SDS-polyacrylamide gels and transferred to nitrocellulose membranes. Membranes were blocked at room temperature for 1 h in 5% fat-free milk, incubated overnight at 4 °C with a primary antibody, followed by incubation with the appropriate HRP-conjugated reporter antibody for 1 h at room temperature. The enhanced chemiluminescence Fluorchem E system (Cell Biosciences, Santa Clara, CA) was used for developing. Wound-healing assay. The wound-healing assay was performed as previously described in vitro 42 .
Transformed mouse lung epithelial cells (MLE-12) were grown to confluency in a 24-well plate. Confluent cell monolayers were mechanically wounded by scraping a sterile conventional pipette tip across the monolayer. Wounded MLE-12 monolayers were washed three times with PBS (pH 7.4) to remove loose cells and debris, and then were incubated for 12 h at 37 °C with 5% CO 2 in the presence of supernatants collected from M2-BMDMs untreated or stimulated with butyrate for 24 h. A magnification of 10× was used to allow a major surface area to be viewed. Photographs were taken and analyzed using Image J software to measure the re-epithelialization area.
Migration assays. A migration assay of MLE-12 cells was performed using matrix-coated polycarbonate filters (8-μm pore size, 24 well, Transwell; Becton Dickinson). MLE-12 cells were resuspended in 500 μL DMEM without fetal bovine serum and seeded to the upper chamber. Supernatant (500 μL) from M2-BMDMs cultured in the presence or absence of butyrate was added to the lower chamber. After 24 h, cells that had migrated into the lower wells were collected, stained with DAPI (Roche Diagnostics), and counted under a microscope.
Cytokine assay. Cytokine levels (IL-1β , IL-6, and TNF-α ) in sera of mice were measured by cytometric bead array immunoassay (CBA; BD Biosciences) according to the manufacturer's instructions.
Histology and immunohistochemistry. Colon tissue samples were fixed in PBS containing 10% neutral-buffered formalin. Paraffin-embedded sections (5 μm) were stained with hematoxylin and eosin. The degree of histological damage and inflammation was graded by a blinded veterinary pathologist. The following variables were scored: (i) amount of inflammation (0, none;1, mild; 2, moderate; 3, severe; 4, accumulation of inflammatory cells in the gut lumen), (ii) distribution of lesions (0, none; 1, focal; 2, multifocal; 3, nearly diffuse; 4, diffuse), and (iii) depth of inflammation and layers involved (0, none; 1, mucosa only; 2, mucosa and submucosa; 3, limited transmural involvement; 4, transmural). The overall histological score was the sum of the three variables (maximum score 12). Consecutive cryostat colon sections were used for immunohistochemistry to detect deposited Arg1 protein. The mean density was shown as an average integrated optical density/area using the Image-Pro Plus software (version 6) (Media Cybernetics, USA).

Statistical analysis.
All data are expressed as means ± SD. Comparisons between two groups was performed using the Student's t test, whereas differences between > 2 groups were performed using one-way ANOVA, (with post hoc comparisons using Dunnett tests) with GraphPad Prism version 4.0 statistical software. Differences were considered significant if P < 0.05.