Butyrate, Valerate, and Niacin Ameliorate Anaphylaxis by Suppressing IgE-Dependent Mast Cell Activation: Roles of GPR109A, PGE₂, and Epigenetic Regulation Free

Dietary fibers that are barely digested in the mammalian digestive system are fermented by the bacteria that constitute the intestinal flora. Short chain fatty acids (SCFAs), including acetate, butyrate, propionate, and valerate, are produced in the colon during the fermentation process as secondary metabolites. The effects of intestinal SCFAs on immune responses have recently become topics of interest. Previous studies indicated that SCFAs contributed to the maintenance of homeostasis by modulating the development and function of immune-related cells (1, 2). SCFAs have been shown to exert beneficial effects on immune-related diseases. Mice kept under germ-free conditions or fed a low-fiber diet showed an exacerbated pathology and symptoms of inflammatory bowel disease, airway hypersensitivity, experimental autoimmune encephalomyelitis, and rheumatoid arthritis due to a deficiency in SCFAs (3–8). SCFAs attenuate these inflammatory diseases by accelerating regulatory T cell (Treg) development (4, 5, 7), increasing IL-10 production by lymphoid cells (4, 8), suppressing the activation of neutrophils and eosinophils (3), and regulating T cell development via the modulation of monocyte functions (6). Intestinal epithelial cells and immune cells take up SCFAs through monocarboxylate transporters and/or sense SCFA signals via G protein–coupled receptors (GPCRs) on the cell surface. SCFAs exhibit an additional function as histone deacetylase inhibitors (HDACis), inducing the expression of genes involved in the anti-inflammatory function of immune cells.

Mast cells (MCs), which develop in bone marrow and terminally differentiate in the mucosa and connective tissues of the whole body, play a key role in IgE-mediated allergic diseases, such as pollinosis and food allergy. The cross-linking of the high-affinity receptor for IgE, FcεRI, by the IgE and Ag (Ag) complex induces rapid degranulation, eicosanoid release, and cytokine production in MCs, resulting in allergic symptoms. Although accumulating evidence recently showed the antiallergic effects of SCFAs targeting MCs (9–11), the molecular mechanisms by which SCFAs modulate MC functions remain unknown. In the current study, we demonstrated that the oral administration of butyrate and valerate ameliorated IgE-mediated anaphylaxis in mice and that SCFAs significantly reduced the degree of IgE-induced degranulation and cytokine production of bone marrow–derived MCs (BMMCs). We also investigated the molecular mechanisms underlying the suppressive effects of SCFAs in vitro and in vivo by examining the roles of GPR109A, HDACi activity, the NRF2 pathway, and PGs in the effects of SCFAs on MCs. Based on the current results, we concluded that butyrate and valerate regulated MCs via GPR109A and by the HDACi activity, with the accelerated synthesis of PGE₂, resulting in the amelioration of anaphylaxis.

BMMCs were generated from the BM cells of C57BL/6 mice (Japan SLC, Hamamatsu, Japan) by cultivation in the presence of 5 ng/ml of mouse IL-3 (BioLegend) as previously reported (12–14). Peritoneal MCs were isolated from the peritoneal exudate cells of mice according to a method described in previous studies (15, 16). All animal experiments were performed in accordance with the approved guidelines of the Institutional Review Board of Tokyo University of Science, Tokyo, Japan. The Animal Care and Use Committees of Tokyo University of Science approved this study (K22005, K21004, K20005, K19006, K18006, K17009, K17012, K16007, and K16010).

SCFAs (acetate, butyrate, sodium butyrate, isobutyrate, propionate, valerate, ethylvalerate, and isovalerate), a monocarboxylate transporter inhibitor (α-cyano-4-hydroxycinnamic acid) (catalog no. 476870), and acetylsalicylic acid (ASA) (catalog no. A5376) were purchased from Sigma-Aldrich (St. Louis, MO). Pertussis toxin (PTX), which is used as a G_i/o-type G protein inactivator, was purchased from Calbiochem (catalog no. 516561, San Diego, CA), and indomethacin was obtained from Fujifilm Wako Pure Chemical (catalog no. 095-02472, Japan). PGD₂ (catalog no. 12010), PGE₂ (catalog no. 14010), and an inhibitor of COX-1, SC-560 (CAY-70340), were supplied by Cayman Chemical (Ann Arbor, MI). Nicotinic acid/niacin (catalog no. 72309, Sigma-Aldrich) and nicotinamide/niacinamide (catalog no. N0078, Tokyo Chemical Industry Co., Tokyo, Japan) were obtained. ONO-AE5-599, an antagonist of PGE₂ receptor 3 (EP3), and ONO-AE3-208, an antagonist of EP4, were kindly provided by Ono Pharmaceutical Co. (Osaka, Japan).

The purification of total RNA and reverse transcription to synthesize cDNA were performed using a ReliaPrep RNA cell miniprep system (Promega) and ReverTra Ace qPCR RT Master Mix (TOYOBO, Osaka, Japan), respectively. The mRNA levels were measured by quantitative PCR using a Step-One real-time PCR system (Applied Biosystems) with Thunderbird probe qPCR Mix (TOYOBO) or Thunderbird SYBR qPCR Mix (TOYOBO). The TaqMan primers Gata1 (catalog no. Mm00484678_m1), Gata2 (catalog no. Mm00492300_m1), Spi1 (catalog no. Mm01270606_m1), Fcer1a (catalog no. Mm00438867_m1), Ms4a2 (catalog no. Mm00442780_m1), Fcer1g (catalog no. Mm00438869_m1), and Gapdh (catalog no. 4352339E) were purchased from Applied Biosystems. The following oligonucleotides were synthesized for PCR primers: Gpr41 forward, 5′-GTGACCATGGGGACAAGCTTC-3′, and reverse, 5′-CCCTGGCTGTAGGTTGCATT-3′; Gpr43 forward, 5′-GGCTTCTACAGCAGCATCTA-3′, and reverse, 5′-AAGCACACCAGGAAATTAAG-3′; Gpr109a forward, 5′-ATGGCGAGGCATATCTGTGTAGCA-3′, and reverse, 5′-TCCTGCCTGAGCAGAACAAGATGA-3′; Slc5a8 forward, 5′-CATTCGTCTCTGTGGCACAATC-3′, and reverse, 5′-GGGCATAAATCACAATTCCAGTGT-3′; Slc5a12 forward, 5′-ACAACAACAGTAGCCCCACAGA-3′, and reverse, 5′-GGTAGGAGAGTGAGTACCATGTGTCA-3′; and Slc16a1 forward, 5′-ACAACAACAGTAGCCCCACAGA-3′, and reverse, 5′-GGTAGGAGAGTGAGTACCATGTGTCA-3′.

Total of 200 ng of anti-TNP mouse IgE (clone IgE-3, BD Bioscience, San Jose, CA) was added to 1 ml of culture medium containing BMMCs (5 × 10⁵). After a 2-h incubation, the cells were washed with Tyrode’s buffer and resuspended in Tyrode’s buffer containing TNP-BSA (final concentration, 3 ng/ml) (LSL, Tokyo, Japan). Thirty minutes after the TNP-BSA stimulation, the culture supernatant was harvested, and β-hexosaminidase activity in the supernatant was assessed as previously reported (17).

Cell surface FcεRI and c-kit stained by PE-labeled anti-mouse FcεRIα Ab (MAR-1, eBioscience) and FITC-labeled anti-mouse CD117 Ab (2B2, BioLegend) were detected by a MACS Quant (Miltenyi Biotech). Cycloheximide (catalog no. C1988, Sigma-Aldrich) and brefeldin A (catalog no. 420601, BioLegend) were used as inhibitors of protein synthesis and protein transporters, respectively. A fixation buffer (catalog no. 420801, BioLegend) and intracellular staining perm wash buffer (10×) (catalog no. 421002, BioLegend) were used for the intracellular staining of FcεRIα. To detect Foxp3⁺ cells in CD4⁺ T cells, cells treated with a Foxp3/transcription factor staining buffer kit (catalog no. TNB-0607, TOMBO Bioscience) were stained with FITC-labeled anti-mouse CD4 Ab (GK1.5, BioLegend) and allophycocyanin-labeled anti-mouse Foxp3 Ab (3G3, TOMBO Bioscience).

The data on the mRNA levels of GPCRs in human MCs were obtained from cap analysis of gene expression (CAGE) human PRJDB1099 (FANTOM5) with the following URL: https://figshare.com/articles/dataset/RefEx_expression_CAGE_all_human_FANTOM5_tsv_zip/4028613.

Small interfering RNA (siRNA) for mouse Gpr109a (MSS234551), and control siRNA (Stealth RNAi Negative Universal Control, catalog no. 12935) were purchased from Invitrogen (Carlsbad, CA). Ten microliters of 20 μM siRNA was introduced into BMMCs (5 × 10⁶) by a Neon Transfection System (Invitrogen) set at Program 5 using a Neon 100 μL kit.

Mouse Gpr109a cDNA was amplified by PCR using mouse BM macrophage cDNA as a template, and the following oligonucleotides as primers: 5′-ggcggatccatgagcaagtcagaccattttctag-3′ (the inserted BamHI sequence is shown in italics, and the initiation codon is underlined) and 5′-gggctcgagttaacgagatgtggaagccag-3′ (the inserted XhoI site and termination codon are shown in italics and underlined, respectively). The BamHI/XhoI-digested PCR fragment encoding Gpr109a cDNA was inserted into the BamHI/XhoI region in the multicloning site of pMXs-puro (18) to obtain pMXs-puro-mGPR109A. The preparation of retroviral vectors using pMX plasmids and packaging cells, the transfection of BMMCs by retroviral vectors, and the selection of transfectants were performed as previously described (19, 20).

The concentrations of IL-6, IL-13, and TNF-α in culture media were measured by using ELISA MAX deluxe set mouse IL-6 (BioLegend), mouse IL-13 DuoSet ELISA (R&D Systems, Minneapolis, MN), and ELISA MAX deluxe set mouse TNF-α (BioLegend), respectively. A PGE₂ ELISA kit–monoclonal (Cayman Chemical, catalog no. 514010) was used to assess the concentration of PGE₂.

C57BL/6 mice were orally administered 435 mg/kg/day butyrate or 510 mg/kg/day valerate for 4 to 6 d using saline and ethanol-polyethylene glycol (PEG) as a vehicle. On the last day of the administration, the mice received an i.v. injection of 3 μg of TNP-specific mouse IgE. Twenty-four hours after the IgE injection, the mice were i.v. transfused with 20 μg of TNP-BSA. The body temperature of each mouse was measured every 10 min for 1 h.

After 4 d of the administration of butyrate, 0.02 μg of TNP-specific mouse IgE or saline as a control was injected into the footpads of the mice to establish passive cutaneous anaphylaxis (PCA). Twenty-four hours after the IgE injection, the mice were i.v. transfused with 20 μg of TNP-BSA. Footpad thickness was measured 0.5 h after the TNP-BSA injection.

The mice were orally administered ASA (1.6 mg/ml in drinking water), or indomethacin (40 μg/ml in drinking water), Oral administration of ONO-AE5-599 (10 mg/kg), ONO-AE3-208 (10 mg/kg), or vehicle (0.5% methyl cellulose as vehicle of AE5-599, and 2.5 mM NaOH as vehicle of AE3-208, 0.2 ml) twice per a day was performed by using a disposable feeding needle with a 200-μL scale (catalog no. 7202, Fuchigami Co., Kyoto, Japan). Niacin/nicotinic acid (100 mg/kg), niacinamide (100 mg/kg), or vehicle (saline, 0.2 ml) was administered via an i.p. injection once per day.

A two-tailed Student t test was used to compare two samples, and a one-way ANOVA followed by Tukey’s multiple comparison test, Dunnett’s multiple comparison test, and Sidak’s multiple comparison test were used to compare more than three samples. p values < 0.05 were considered to be significant.

To evaluate the efficacy of SCFAs to prevent MC-mediated allergic responses in vivo, we used the passive systemic anaphylaxis (PSA) model and found that the decrease in body temperature caused by the IgE-mediated activation of MCs was significantly abated by the administration of butyrate (Fig. 1A) and valerate (Fig. 1B). To confirm whether the intake of SCFAs under this experimental condition affected the development of Treg, which is known to be enhanced by SCFAs, we examined the frequency of Treg cells in the mesenteric lymph node (MLN) and spleen. In the mice that received 100 μmol butyrate or valerate once per day for 4 to 6 d, an increase in Treg cells was not observed in the MLN or spleen (Fig. 1C, 1D). Footpad swelling following a s.c. injection of IgE, and i.v. injection of Ag was also reduced in mice that received orally administered butyrate (Fig. 1E), suggesting that the uptake of SCFAs via the digestive tract suppressed not only systemic anaphylaxis but also peripheral cutaneous anaphylaxis. These results indicate that butyrate and valerate exerted suppressive effects on the MC-mediated allergic responses in vivo independent of Treg development.

BMMCs pretreated with various SCFAs were stimulated with IgE and Ag to evaluate the effects of SCFAs on MC activation. The activity of β-hexosaminidase released from MCs was assessed as an index of the immediate response level (Fig. 2A). As shown in Fig. 2B, the frequency of DAPI-stained cells remained unchanged in the presence of 10 mM SCFAs, suggesting that SCFAs did not induce apparent toxicity. Under this experimental condition, when BMMCs were incubated with 1 mM SCFAs, five of the six SCFAs examined, namely propionate, butyrate, valerate, isobutyrate, and isovalerate, significantly suppressed the degranulation of IgE-stimulated BMMCs (Fig. 2A). The degranulation of BMMCs was completely inhibited by the treatment with 10 mM of propionate, butyrate, valerate, and isovalerate. At a concentration of 10 mM, acetate significantly suppressed degranulation. We also examined sodium butyrate to exclude the effects of pH changes caused by the acids and confirmed that sodium butyrate significantly reduced IgE-stimulated degranulation as well as butyrate (Fig. 2C).

The effects of SCFAs on cytokine production by MCs were then examined. As demonstrated in Fig. 2D, pretreatments with all of the SCFAs tested significantly reduced the release of IL-6 and IL-13 from IgE-stimulated MCs. The IgE-induced production of TNF-α was significantly inhibited by the treatment with valerate, isobutyrate, and isovalerate, whereas the butyrate treatment promoted the production of TNF-α.

The IgE-induced activation of MCs was initiated by the binding of IgE to FcεRI. The expression level of FcεRI is associated with the degree of degranulation and is a risk factor for allergic diseases (21–24). To evaluate the effects of SCFAs on FcεRI expression, we performed a flow cytometric analysis and found that FcεRI levels on the cell surface were decreased by the pretreatment with butyrate and valerate (Fig. 3A). FcεRI comprises three subunits, namely, α, β, and γ, and the transcription factors PU.1, GATA1, and GATA2 regulate the cell type–specific expression of α and β (12, 17, 25–27). To clarify whether the downregulation of surface FcεRI occurred in a transcription-dependent manner, we measured the mRNA levels of FcεRI subunits (Fig. 3B) and related transcription factors (Fig. 3C) by quantitative PCR. The obtained results revealed that butyrate and valerate did not reduce but rather tend to increase the mRNA levels of Fcer1a, Ms4a2, FceR1g, Spi1, Gata1, and Gata2. Therefore, valerate and butyrate may have reduced FcεRI levels on the cell surface without inhibiting the transcription of FcεRI subunit genes. When BMMCs were treated with cycloheximide, the cell surface level of FcεRI was markedly reduced, and butyrate did not enhance the cycloheximide-induced downregulation of surface FcεRI (Fig. 3D). In addition, butyrate reduced not only the surface level but also the intracellular level of the FcεRIα protein (Fig. 3E). Butyrate-induced decreases in intracellular FcεRIα protein levels were still observed in the presence of brefeldin A under the condition that the transport of FcεRIα toward the cell surface was inhibited by brefeldin A (Fig. 3E). These results indicate that butyrate and valerate suppressed the IgE-mediated degranulation and cytokine release of MCs and also that the decrease observed in the cell surface level of FcεRI, which may have occurred with an event of protein synthesis, was partly involved in the suppressive effects of SCFAs on the IgE-mediated activation of MCs.

To elucidate the molecular mechanisms by which butyrate and valerate modify the function of MCs, we examined cell surface molecules to identify a candidate transporter and/or receptor for SCFA. A quantitative PCR analysis showed that BMMCs expressed detectable amounts of mRNAs for the solute carrier group of membrane transport proteins (Slc5a8, Slc5a12, and Slc16a1) and GPCRs (Gpr41, Gpr43, and Gpr109a). Moreover, the mRNA expression levels of Slc16a1 and Gpr109a were higher than those of other mRNAs (Fig. 4A, 4B). To confirm the involvement of the transporter and receptor in SCFA signaling, we pretreated BMMCs with a reagent that inhibits the transporter or receptor for 1 h prior to the addition of SCFAs. As shown in Fig. 4C, the presence of 2-cyano-4-hydroxyphenyl acrylic acid (monocarboxylate transporter inhibitor) did not affect the SCFA-mediated suppression of degranulation. In contrast, PTX, an inhibitor of G_i/o proteins, counteracted the suppressive effects of butyrate and valerate (Fig. 4D), suggesting that G_i/o-type GPCR functions are required for the butyrate- and valerate-mediated suppression of MC activation. An analysis of FANTOM5 data revealed that mRNAs for GPR43 and GPR109A were expressed in human MCs (Fig. 4E). Gpr109a mRNA was expressed in peritoneal mouse MCs, whereas that of Gpr41 and Gpr43 was not (Fig. 4F). To clarify the involvement of GPR109A, which was a G_i-GPCR that was expressed at higher levels than other G_i-GPCRs (GPR41 and GPR43) in mouse and human MCs (Fig. 4B, 4E, and 4F), we performed a knockdown experiment using siRNA. When butyrate and valerate significantly suppressed the degranulation of control siRNA-introduced BMMCs, Gpr109a siRNA transfectants in which Gpr109a mRNA was effectively knocked down (Supplemental Fig. 1A) exhibited markedly enhanced degranulation (Fig. 4G). This result supports the hypothesis that GPR109A is a receptor for butyrate and valerate on MCs. The increase in degranulation by the knockdown of Gpr109 was more restrictive for butyrate-treated MCs than for valerate-treated MCs, suggesting that the effects of butyrate were partly dependent on GPR109A. The knockdown of Gpr109a upregulated the degranulation of nontreated control BMMCs. This result suggests that GPR109A signaling constitutively suppressed MC activation. Therefore, to clarify the role of GPR109A in the IgE-dependent activation of MCs, we conducted a GPR109A overexpression experiment. When GPR109A was constitutively overexpressed in BMMCs using a retroviral vector (Supplemental Fig. 1B), the suppressive effects of SCFAs were enhanced, and the degree of degranulation was markedly reduced even in the absence of SCFAs (Fig. 4H). Based on these results, we concluded that GPR109A was involved in the suppression of the IgE-dependent degranulation of MCs and that valerate and butyrate functioned as ligands for GPR109A.

SCFAs exhibit HDACi activity, which enhances the anti-inflammatory functions of immune cells by inducing the expression of genes including Foxp3 and IL-10 (5, 8). The order of antiallergic effects in Fig. 2A (butyrate > valerate ≫ acetate) is consistent with the order of HDACi activity among SCFAs (8). To evaluate the effects of HDACi activity on the IgE-mediated activation of MCs, we analyzed MCs that were treated with trichostatin A (TSA), an inhibitor of class I and II HDACs. The pretreatment of MCs with TSA (5–20 nM) for 24 h before the IgE-mediated stimulation significantly and dose-dependently suppressed degranulation (Fig. 5A) without affecting the frequency of DAPI-stained dead cells (Fig. 5B). The release of IL-6, IL-13, and TNF-α from IgE-stimulated MCs was significantly inhibited by the pretreatment with 20 nM TSA (Fig. 5C).

The TSA treatment reduced the surface expression levels of FcεRI (Fig. 5D), even though the mRNA levels of FcεRIα, β, and γ were not decreased in TSA-treated MCs (Fig. 5E), which is consistent with the results obtained from butyrate-treated MCs (Fig. 2). We also found that the suppressive effects of butyrate on FcεRI expression levels in MCs were not inhibited by the PTX treatment (Fig. 5F). These results suggest that SCFA-induced reductions in the cell surface expression level of FcεRI were mediated by the HDAC inhibition rather than by the Gi-GPCR stimulation.

Butyrate has been shown to activate the transcriptional regulator NRF2, which alleviates oxidative stress, in various cells (28–30). Furthermore, the activation of NRF2 is frequently observed in MCs treated with anti-allergic compounds (31–33). Hmox1, a target gene of NRF2, exhibited the greatest increase in its mRNA level following the treatment of human MCs with butyrate (9); however, the role of NRF2 in human MCs was not analyzed. Based on these findings, we investigated whether NRF2 was activated in SCFA-treated MCs and played a role in the suppressive effects of SCFAs on MCs. As shown in Fig. 6A and 6B, the mRNA levels of Hmox1 and Nfe2l2 (gene ID of NRF2) were elevated in butyrate-treated BMMCs, and these increases were sustained for at least 48 h under our experimental conditions. In addition, we found that Hmox1 mRNA levels, were increased by the IgE-mediated stimulation and were further upregulated in the presence of butyrate (Fig. 6C) and the TSA treatment increased Hmox1 mRNA levels in MCs (Fig. 6D). To clarify whether the SCFA-induced activation of NRF2 in MCs was involved in the suppression of IgE-mediated activation, we conducted an experiment using an NRF2 inhibitor and found that the inhibition of the NRF2 pathway did not reduce the suppressive effects of butyrate on the degranulation and cytokine release (Fig. 6E). These results indicate that the activation of the NRF2–HO-1 pathway was not involved in the suppression of IgE-mediated activation in SCFA-treated MCs; however, SCFAs and TSA activated this pathway in MCs.

GPR109A is classically known as a receptor for niacin/nicotinic acid/vitamin B₃. Moreover, the intake of excess amounts of niacin induces a niacin flash due to the enhanced release of PGs. Based on these findings, we hypothesized that the treatment of MCs with butyrate and valerate may induce the production of PGs, which affects the activation degree of MCs. To evaluate the roles of PGs in MC activation, we analyzed the degranulation degree of SCFA-treated MCs in the presence of the nonsteroidal anti-inflammatory drugs (NSAIDs), ASA, and indomethacin. The suppression of degranulation in valerate-treated MCs was counteracted by the addition of ASA, whereas the effects of ASA and indomethacin on butyrate-induced suppression were moderate (Fig. 7A, 7B).

PGD₂ and PGE₂ are preferentially produced by activated MCs. Therefore, we evaluated the effects of PGD₂ and PGE₂ on the degranulation levels. Although pretreatment with 1–100 nM PGD₂ for 24 h did not affect the degranulation of MCs (Fig. 7C), that with PGE₂ exerted inhibitory effects (Fig. 7D). Similar suppressive effects of PGE₂ were observed following its addition 2 h prior to the stimulation (Fig. 7F), whereas PGD₂ did not inhibit degranulation at either time point (Fig. 7E). The measurement of PGE₂ concentrations in culture media revealed that the IgE stimulation strongly promoted the release of PGE₂, which was already enhanced by butyrate, valerate, and propionate (Fig. 7G). Furthermore, SCFA-treated MCs released slightly more PGE₂ than nontreated MCs, even in the steady state (Fig. 7G).

We recently reported that treatment with ASA or the knockdown of cyclooxygenases (COXs), particularly COX-1, promoted the IgE-induced activation of BMMCs (34). To reveal the role of COX-1 in the suppression of degranulation in SCFA-treated MCs, we evaluated the effects of the COX-1–specific inhibitor SC-560 on MC degranulation. When BMMCs were pretreated with SC-560 prior to the addition of SCFAs, SC-560 attenuated the suppressive effects of SCFAs in a dose-dependent manner (Fig. 7H). Collectively, these results demonstrated that the SCFAs increased the de novo synthesis and release of PGE₂, which suppressed the activation of MCs.

To investigate whether prostanoid production is involved in the effects of SCFAs in vivo, we performed anaphylaxis analyses of mice that were administered an NSAID. When PSA was induced in mice administered butyrate and/or indomethacin, the effects of the administration of butyrate were significantly inhibited by supplementation with indomethacin (Fig. 8A). In mice administered ethylvalerate (to avoid the corrosivity of valerate) and/or ASA, the decrease of body temperature following the Ag injection was reduced by the administration of ethylvalerate, and the suppression of PSA by ethylvalerate was counteracted by additional supplementation with ASA (Fig. 8B).

The results showing that NSAIDs suppressed the SCFA-mediated amelioration of PSA suggest the involvement of PGs in the effects of SCFAs on passive anaphylaxis. To further clarify the role of PGE₂, we conducted a PGE₂ receptor antagonism experiment in vivo (Fig. 8C) and found that the administration of the EP3 antagonist inhibited the attenuating effects of butyrate on PSA (Fig. 8D) and PCA (Fig. 8E), whereas the EP4 antagonist did not affect the butyrate-mediated suppression of PCA (Fig. 8F).

To elucidate the roles of the GPR109A stimulation on IgE-mediated anaphylaxis, we evaluated the effects of niacin on PCA. Niacin is general term for two kinds of vitamin B₃: nicotinic acid and nicotinamide. Although both compounds play similarly important roles in nutrition as vitamin B₃, nicotinic acid and nicotinamide are distinguishable on the basis of GPR109A stimulation activity; namely, nicotinic acid stimulates GPR109A, but nicotinamide does not. As shown in Fig. 8G, nicotinic acid significantly reduced footpad swelling induced by PCA, whereas nicotinamide did not. The suppression of PCA by the nicotinic acid treatment was inhibited by the additional administration of NSAIDs (Fig. 8H) or the EP3 antagonist (Fig. 8I). These results demonstrated that the stimulation of GPR109A was involved in the suppression of anaphylaxis by modulating the production of prostanoids, particularly PGE₂.

MCs play important roles in IgE-dependent allergic responses. The cross-linking of IgE-binding FcεRI on MCs via allergens induces rapid responses, including degranulation and the release of eicosanoids, and late-phase inflammation accompanied by the production of cytokines. Current therapeutic medicines for allergies, such as histamine receptor antagonists, leukotriene receptor antagonists, and humanized anti-human IgE, target steps in the IgE–MC axis, suggesting that the inhibition of MC activation is useful for the prevention of allergic diseases.

In the current study, we identified GPR109A-mediated signaling as a key event for the suppression of MC-mediated allergic responses. A recent study demonstrated that GPR109A was involved in colonic homeostasis based on a Gpr109a gene deficiency exacerbating dextran sodium sulfate (DSS)–induced colonic inflammation (35). Furthermore, the aforementioned study indicated that the activation of GPR109A by niacin or butyrate suppressed colitis and colon cancer, and colonic macrophages, dendritic cells (DCs), and the epithelium were considered to be the main target cells expressing GPR109A. GPR109A signaling amplified the production of retinoic acid and anti-inflammatory cytokines from macrophages and DCs and subsequently accelerated the development of Tregs. The involvement of butyrate in CD103⁺ DC-mediated Treg differentiation was supported by another study using a food allergy mouse model with high-fiber feeding (36). High-fiber feeding enhanced retinaldehyde dehydrogenase activity in CD103⁺ DCs in a GPR109A-dependent manner. As reported in previous studies, DCs and macrophages were identified as GPR109A-expressing myeloid cells. In contrast, the role of GPR109A in MCs was largely unknown. In the current study, we found that GPR109A mRNA levels were higher than those of GPR43 (and GPR41) in BMMCs, mouse peritoneal MCs, and human MCs (Fig. 4B, 4E, 4F). The knockdown of Gpr109a completely and partially canceled the suppressive effects of valerate and butyrate, respectively, and exacerbated the degranulation of control MCs. The overexpression of GPR109A reduced the degranulation degree and enhanced the sensitivity of MCs to SCFAs. These results demonstrated the following: 1) GPR109A signaling inhibits MC activation, 2) valerate is one of the ligands for GPR109A, and 3) endogenous ligands for GPR109A are present in the MC culture medium. Furthermore, the administration of butyrate, valerate, or niacin prior to passive anaphylaxis significantly ameliorated IgE-dependent anaphylaxis (Figs. 1, 8). The results of in vitro and in vivo experiments revealed the potential of the butyrate/valerate/niacin–GPR109A axis as a therapeutic target for MC-mediated allergy. In a previous study, the provision of butyrate in drinking water ad libitum for 3 wk reduced the anaphylaxis score of food allergy in mice, and this was accompanied by a decrease in serum IgE levels and increase in CD103⁺ DCs and the Treg population in MLNs (36). Although the effects of butyrate on MC functions were not evaluated (36), the oral intake of butyrate or valerate for 1 wk suppressed PSA and PCA in the current study (Figs. 1, 8); therefore, butyrate may have contributed to the decrease observed in the anaphylaxis score with the suppression of MC functions in the previous study. In the current study, Treg levels were not elevated in mice orally administered butyrate or valerate for 1 wk (Fig. 1). Therefore, we concluded that butyrate and valerate ameliorated anaphylaxis by directly modulating MC functions even when Treg development was not affected by SCFAs.

The protective effects of niacin against colitis were also reported in a study using a DSS- or 2,4,6-trinitrobenzene sulfonic acid–induced mouse model (37). The administration of niacin upregulated PGD₂ levels in the colon and subsequently ameliorated DSS/2,4,6-trinitrobenzene sulfonic acid–induced colitis in mice via the DP1 receptor (37). MCs are a major source of PGD₂. A previous study using mice with food allergy revealed an antiallergic role for PGD₂ derived from MCs (38). However, under our experimental conditions, PGE₂, but not PGD₂, suppressed MC activation in an autocrine manner (Fig. 7). Another study using ptges^−/− mice demonstrated that a PGE₂ deficiency caused aspirin-exacerbated respiratory diseases (AERD), in which overproduced cysteinyl leukotrienes and activated MCs were involved in airway inflammatory disorders (39). PGE₂ has been shown to attenuate the cysteinyl leukotriene–mediated IL-33–dependent activation of MCs (39, 40), whereas MC-derived PGD₂ is one of the hallmarks of the severity of AERD (41, 42). Furthermore, PGE₂ exhibited immunosuppressive effects in the colonic mucosa (43). SCFAs, which enhanced PGE₂ production by MCs (Fig. 7), may contribute to the prevention and/or treatment of AERD and inflammatory bowel disease. Although further detailed investigations are required to clarify the roles of MCs in the SCFA-mediated alleviation of colitis, the physiological levels of SCFAs reported, namely 0.1–10 mM in serum, 200 mM in the colon (in the Discussion of Ref. 3), and 1–10 μmol/g in cecal contents and in the colonic tissue (5), are expected to be sufficient to regulate the function of MCs according to our in vitro experiments.

A recent study reported that PGE₂ levels inversely correlated with the severity of anaphylaxis in humans and mice (44). In that study, an EP2 agonist and EP4 agonist suppressed the IgE-induced activation of BMMCs and IgE-mediated PSA. In contrast, the EP3 antagonist, but not the EP4 antagonist, inhibited the protective effects of butyrate and niacin against anaphylaxis under our experimental conditions (Fig. 8). Although we cannot explain this discrepancy, EP3 plays protective roles in MC-mediated allergic responses because Ptger3^−/− mice had more severe airway inflammation accompanied by the hyperactivation of MCs (45).

The characteristics of butyrate and valerate differed in the current study. For example, TNF-α production by MCs was decreased by valerate and increased by butyrate (Fig. 2). Valerate primarily suppressed MC activation via GPR109A, whereas butyrate affected multiple pathways in MCs, including those associated with other GPCRs, in addition to GPR109A (Fig. 4). Two other GPCRs, GPR41 and GPR43, which are receptors for SCFAs (2–4, 6), may be involved in MC functions. However, we were unable to clarify this aspect because siRNAs for GPR41 and GPR43 were not functional in MCs, in which mRNAs for GPR41 and GPR43 were rarely detected. Furthermore, a previous study demonstrated that the deficiencies in GPR41 and GPR43 in mouse MCs did not affect the butyrate-mediated suppression of degranulation (9). The observation that acetate, the specificity of which is restricted to GPR43 (EC₅₀ for acetate is ∼250–500 μM) and to a lesser extent to GPR41 (propionate > butyrate ≫ acetate) but not to GPR109A (46), did not suppress the IgE-mediated activation of MCs in a previous study (9) and was the least effective in the current study (Fig. 2) may support the irrelevance of GPR41 and GPR43 in the activation of MCs.

In vitro experiments using TSA revealed that HDACi, which is one of the activities of butyrate and valerate, inhibited the IgE-induced degranulation and cytokine release of MCs (Fig. 5). The suppressive effects of HDACi on the IgE-dependent activation of MCs were reported in previous studies in which the downregulation of tyrosine kinases (9) and the suppression of transcription factor recruitment (47) were observed. In the current study, the cell surface expression level of FcεRI was decreased by both SCFAs and TSA (Figs. 3, 5). Although the downregulation of BTK, SYK, and LAT in butyrate-treated MCs was transcriptional due to a decrease in the acetylation of these promoters (9), the mRNA levels of the three subunits of FcεRI slightly increased in SCFA-treated MCs (Fig. 3) and TSA-treated MCs (Fig. 5). Further studies are warranted to clarify the mechanisms by which SCFAs reduce the surface expression of FcεRI on MCs via HDACi activity.

The composition of the daily diet has been shown to affect the microbiota population and SCFA concentrations, which are closely associated with human and animal health (48, 49). In contrast to acetate, butyrate, and propionate, which are major SCFAs in the colon, limited information is currently available on valerate, isobutyrate, and isovalerate, the colonic concentrations of which are low. Clostridia and Bacteroides have the potential to produce valerate (48, 50, 51). Isobutyrate and isovalerate, which suppressed MC activation (Fig. 2), are produced as secondary metabolites from amino acids in Bacillus (52). Both of these branched SCFAs are present in the Japanese traditional food “nattō,” which is made of soybeans through a fermentation process by Bacillus subtilis natto. Although branched SCFAs in nattō have been recognized merely as the source of a unique flavor, they may exert antiallergic effects and/or contribute to colonic homeostasis. Changes in dietary habits in Japan in the past few decades, from a typical Japanese diet including nattō and high-fiber dishes to Western diets may be associated with the incidence of allergic diseases. We intend to investigate the relationship between the mucosal environment and immunoregulation by focusing on valerate, butyrate, and branched SCFAs, which may lead to the proposal of the health benefits of certain foods in the near future.

The authors have no financial conflicts of interest.

We are grateful to the members of the Laboratory of Molecular Biology and Immunology, Department of Biological Science and Technology, Tokyo University of Science, for constructive discussions and technical support. We greatly appreciate the consideration from Yayoi Yasuda, Masako Yasuda, and Kimihiko Yasuda.