6- to 10-week-old wild-type (WT) C57BL/6 mice were obtained from Charles River Laboratories (Sulzfeld, Germany) and were maintained under specific pathogen-free (SPF) or GF conditions. GF mice were kept in adequate isolators (Metall + Plastic, Radolfzell-Stahringen, Germany). The bedding, water, and food were routinely autoclaved, and the sterility of GF mice was checked biweekly. All experimental procedures with GF animals were carried out in a laminar flow hood under sterile conditions. Ffar2^−/−Ffar3^−/−, Slc5a8^−/−, and Tbx21^−/− mice on C57BL/6 background were bred at the animal facility of the Biomedical Research Center, University of Marburg, Germany. Ffar2^−/−Ffar3^−/− mice were generated by Prof. Stefan Offermanns (Max Planck Institute for Heart and Lung Research, Bad Nauheim, Germany) (16). Slc5a8^−/− mice were received from Dr. Thomas Boettger (Max Planck Institute for Heart and Lung Research, Bad Nauheim, Germany). All animal experiments were conducted according to the German animal protection law.

CD4⁺ T cells were purified from spleen and LNs and differentiated into Tregs or Th1 and Th2 cells in vitro as described previously (17). In brief, CD4⁺ T cells were isolated using the kit for negative isolation and were primed with plate-bound anti-CD3 (5 µg/ml, clone 145–2C11) and soluble anti-CD28 (1 µg/ml, clone 37.51) mAbs. In addition, 50 U/ml rh IL-2 (Novartis, Nürnberg, Germany) were provided for Th0, Th1, Th2, and Th17 cell cultures, whereas Tregs received 0.5 µg/ml anti-CD28 and 100 U/ml rhIL-2. Th1 cells were differentiated using 10 ng/ml rmIL-12 (PeproTech, Hamburg, Germany) together with anti-IL-4 (10% culture supernatant of clone 11B11). Th2 cells were generated by using 40 ng/ml rmIL-4 (PeproTech, Hamburg, Germany) together with 10 µg/ml anti-IFN-γ. For Th17 cultures, 0.5 ng/ml rhTGF-β1, 20 ng/ml IL-6 (both PeproTech, Hamburg, Germany), 10 µg/ml anti-IFN-γ, and anti-IL-4 (10% culture supernatant of clone 11B11) were added in the cell cultures. Optimal Treg conditions were obtained by addition of 10 µg/ml anti-IFN-γ, anti-IL-4 (10% culture supernatant of clone 11B11), and 2 ng/ml rhTGF-β1, whereas lower rhTGF-β1 concentrations (0.5 and 1 ng/ml) were used to achieve suboptimal Treg conditions. T cells were stimulated with 0.1–10 mM sodium butyrate, sodium propionate, sodium acetate, or 1–10 nM TSA (all substances from Sigma-Aldrich, München, Germany). Cells were cultured at 37°C for 72 h. Subsequently, the half of the cells was transferred to a new plate with 100 U/ml rhIL-2 plus the corresponding concentration of sodium butyrate, sodium propionate, sodium acetate, or TSA and rested for additional 48–72 h.

Following oral treatment with 100 mM sodium butyrate for 3 weeks, the acute colonic inflammation in GF mice was induced by adding 1.5% (w/v) DSS (MP Biomedicals, Eschwege, Germany) to the drinking water for 5 days (for the experiments with WT mice, 2% DSS was used). Animals were orally treated with 100 mM butyrate throughout the duration of the experiment based on the published protocol (6). In some experiments, mice were orally treated with 100 mM sodium butyrate for 3 weeks without colitis induction. In all SCFA experiments, the drinking water of control mice was pH- and sodium-adjusted. Colitis was scored using a previously published inflammation scoring system (18). Lamina propria mononuclear cells were purified from mice at day 8 after induction of colitis and were analyzed by FACS analysis and RT-PCR.

In vitro polarized CD4⁺ T lymphocytes were restimulated for 4 h with PMA (50 ng/mL)/ionomycin (750 ng/mL) in the presence of brefeldin A (10 mg/mL, all three substances from Sigma-Aldrich, München, Germany). For surface staining, fluorochrome-conjugated antibodies were added in appropriate dilutions and incubated for 20 min at 4°C. For intracellular staining of cytokines, CD4⁺ T cells were fixed with 2% formaldehyde solution and stained in 0.3% saponin buffer with the following antibodies: anti-IL-17A (eBio17B7), anti-IL-4 (11B11), and anti-IFN-γ (XMG1.2). For intracellular staining of transcription factors Foxp3 (FJK-16s) and T-bet (eBio4B10), the cells were fixed, permeabilized, and stained using Foxp3 staining kit (eBioscience, Frankfurt, Germany). Flow cytometry analysis was carried out on FACSCalibur (BD Bioscience, Heidelberg, Germany). Data were analyzed with FlowJo analysis software (TreeStar, Ashland, OR, USA).

For the preparation of cell lysates, 5 × 10⁶ cells were harvested from Treg-inducing cell cultures and pelleted at 390 × g for 10 min at 4°C. Cell pellets were lysed with RIPA cell lysis buffer for 20 min on ice. Following the SDS-PAGE, protein samples were transferred to a PVDF membrane, and protein detection was performed in a chemiluminescence image station (MicroChemi, Biostep GmbH, Burkhardtsdorf, Germany). For the detection of histone acetylation in Tregs, anti-acetyl-Histone H3 and H4 Abs (Merck Millipore, Darmstadt, Germany) were used. As a loading control for total cell protein extracts, a monoclonal anti-mouse β-actin Ab (Sigma-Aldrich, Munich, Germany) was used.

CD4⁺ T cells were cultured under Treg-inducing conditions for 72 h. Subsequently, the cells were harvested in the lysis buffer and subjected to HDAC inhibition by adding 5 mM of SCFAs for 10 min at room temperature. Following initial inhibition of HDACs, the peptide substrate Ac-Arg-Gly-Lys-AMC (Bachem, Bubendorf, Switzerland) was added to the reaction tubes for next 30 min and finally the stop solution stopped the reaction mediated by HDAC enzymes. HDAC inhibition activity was determined by measuring the fluorescence intensity of free AMC at the spectrofluorometer FLUOstar Omega (BMG Labtech, Ortenberg, Germany).

Chromatin immunoprecipitation analysis was performed on CD4⁺ T cells which were cultured under optimal Treg-polarizing conditions for 3 days. A total of 1 × 10⁶ Tregs were fixed with 1% formaldehyde for 10 min at room temperature. Subsequently, genomic DNA was extracted from precipitates using specific antibody for pan-acetylated H3 (Merk Millipore, Darmstadt, Germany). Control experiments were carried out with respective isotype antibodies as previously described (19). Samples were probed for the promoter regions of Ifnγ, Il17a, Tbx21, and Foxp3 by quantitative RT-PCR using the following primers: Ifnγ, forward CATACCCTTTCCTTGCTTTTC and reverse TTGTGGGATTCTCTGAAAGCA, Il17a, forward TGGTTCTGTGCTGACCTCAT and reverse GCTCTCCCTGGACTCATGTT, Tbx21 forward AGGTGGCAGGTTGACTCT and reverse CTGCTCCTGGGCTTTCTC, and Foxp3, forward TTCCCATTCAGCTTCA and reverse TGTTTGTGAGTGGAGG.

In a first step, the total RNA is isolated from cell lysates. Subsequently, the RNA was transcribed into cDNA using the RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific) according to the manufacturer’s instructions. Quantitative real-time PCR for Tbx21, Ifnγ, Gata3, and Rorγt was performed using a StepOnePlus device (Applied Biosystems, Darmstadt, Germany) with the following primers: Tbx21 forward CAACAACCCCTTTGCCAAAG, Tbx21 reverse TCCCCCAAGCAGTTGACAGT, Foxp3 forward TTCCTTCCCAGAGTTCTTCCA, Foxp3 reverse CATTGAGTGTCCTCTGCCTCT, Ifnγ forward GCAACAGCAAGGCGAAAAAG, Ifnγ reverse TTCCTGAGGCTGGATTCGG, Gata3 forward AAGGCAGGGAGTGTGTGAAC, Gata3 reverse AGGATGTCCCTGCTCTCCTT, Rorγt forward TCCTGCCACCTTGAGTATAGTCC, and Rorγt reverse GGACTATACTCAAGGTGGCAGGA. Quantitative analysis was performed by normalizing the target gene expression to expression of housekeeping gene Hypoxanthine-guanine phosphoribosyl transferase (Hprt1). The following primers were used for detection of Hprt1: forward, CTGGTGAAAAGGACCTCTCG, reverse, TGAAGTACTCATTATAGTCAAGGGCA.

Luminal content of colon (~20 mg) was weighted in sterile ceramic bead tubes (NucleoSpin^® Bead Tubes, Macherey-Nagel, Dueren, Germany) and extracted with 1 mL chilled methanol (−20°C; LC-MS CHROMASOLV^®, FLUKA, Sigma Aldrich, St Louis, MO, USA), containing 10 parts per million (ppm) of butyric acid-4,4,4-d3 (98 atom% D, Sigma Aldrich, St Louis, MO, USA) as internal standard. Homogenization and extraction were performed with Precellys^® Evolution Homogenizer (Bertin Corp., Rockville, Maryland, USA; 4,500 rpm, 40×3 s, 2 s pause time). Then, samples were centrifuged for 10 min at 21,000 × g and 4°C. SCFA standards including acetic, propionic, and butyric acids were all purchased from Sigma Aldrich (St Louis, MO, USA) and prepared in methanol to a stock solution concentration of 1,000 ppm. A SCFA working dilution of 100 ppm was prepared together with 100 ppm of butyric acid-4,4,4-d3. Derivatization of SCFAs was performed by using AMP + Mass Spectrometry Kit (Cayman Chemical, Ann Arbor, MI, USA). Eight microliters of SCFA working dilution, 100 ppm of butyric acid-4,4,4-d3, and extracted colon samples were derivatized according to manufacture descriptions of AMP + Mass Spectrometry Kit. Afterward, 50 µL of all samples were diluted with 405 µL of milliQH₂O (milliQH₂O, Milli-Q Integral Water Purification System, Billerica, MA, USA) to a final concentration of 1 ppm. A five-point calibration curve of SCFAs (0.05, 0.1, 0.2, 0.3, 0.4 ppm, each containing 0.1 ppm butyric acid-4,4,4-d3) was prepared in milliQH₂O. Analyses of derivatized SCFAs and samples were performed with UHPLC (Acquity UPLC, Waters, Milford, MA, USA) coupled to MS. A reversed-phase separation was applied using a C8 column (C8: 1.7 µm, 2.1 mm × 150 mm, Acquity™ UPLC BEH™, Waters, Milford, MA, USA). Elution of derivatized SCFAs was ensured by following solvent system of ammonium acetate (5 mM, Sigma Aldrich, St Louis, MO, USA) combined with acetic acid (0.1%, pH 4.2, Biosolve, Valkenswaard, Netherlands) in water (A) (milliQH₂O) and acetonitrile (B) (LC-MS CHROMASOLV^®, FLUKA, Sigma Aldrich, St Louis, MO, USA). The flow rate was set to 0.3 mL/min, injection volume to 5 µL, and column temperature to 30°C. After a pre-runtime of 2 min, a gradient profile was applied by starting at 1% B for 1 min, increasing to 25% B until 10 min, following an increase to 95% B within 7 min. The 95% B was hold for 2 min, returning to initial 1% B within 0.2 min and further holding 2 min, resulting in a total run time of 22 min. The MS analysis (maXis, Bruker, Daltonics, Germany) was performed in positive electrospray ionization mode with the following parameters: nebulizer gas pressure of 2.0 bar, dry gas flow of 10.0 L/min, dry heater of 200°C, capillary voltage of 4,500 V, and end plate offset of −500 V. Theoretical m/z values and corresponding retention times of the derivatized SCFAs were the following: acetic acid ([M]+: 227.118438, 4.1 min), propionic acid ([M]+: 241.134088, 4.9 min), butyric acid ([M]+: 255.149738, 6.2 min), and butyric acid-4,4,4-d3 ([M]+: 258.16802, 6.2 min). Additionally, pseudo multiple reaction monitoring was used for butyric acid and butyric acid-4,4,4-d3 (isolation width of 10 Da). Acetic and propionic acid were also isolated and fragmented (isolation width of 10 Da and collision energy of 40 eV) to monitor typical loss of AMP compound in fragmentation experiments (MS/MS fragment with theoretical mass signal value [M]+: 169.0886). Calibration of mass spectrometer was ensured by injecting ESI-L Low Concentration Tuning Mix 100 mL (Agilent, Santa Clara, CA, USA) before the analysis. Peak areas and concentrations of SCFAs in luminal content of colon were elaborated with QuantAnalysis (Bruker, Daltonics, Bremen, Germany).

To compare the mean between two groups, statistical analysis was performed using an unpaired Student’s t-test with Prism 5 software (GraphPad, La Jolla, CA, USA). P-Values of P < 0.05 were considered statistically significant. For the comparisons of the multiple groups, the one-way analysis of variance was used. The following P-values were used: ***P < 0.001, **P = 0.001–0.01, *P = 0.01–0.05. Where appropriate, the mean ± SEM is represented in graphs.

Treatment of CD4⁺ T cells with butyrate has been shown to promote the differentiation of mucosal Tregs (3, 4). Accordingly, we observed that WT mice treated orally with 100 mM sodium butyrate for 3 weeks were capable of locally expanding Tregs in colonic lamina propria but not in mesenteric lymph nodes (mLN) or spleen (Figures 1A,B). A similar increase in the frequency of colonic Tregs was observed when mice received 200 mM of sodium butyrate in the drinking water. In contrast, no enhanced Treg responses were observed after oral treatment of WT animals with 50 mM sodium butyrate (Figure S1 in Supplementary Material). Furthermore, neither frequencies nor cell numbers of colonic Th1 and Th17 cells were affected by the butyrate treatment, suggesting that under steady-state conditions only colonic Tregs and not effector T cells are regulated by SCFAs (Figure 1C and data not shown). Finally, oral administration of butyrate resulted in increased colonic mRNA levels of Il10 and Foxp3 but not of pro-inflammatory genes such as Tbx21 (Figure 1D).

In in vitro experiments, we confirmed that low concentrations of butyrate (0.1–0.25 mM) enhance expression of Foxp3 and differentiation of inducible (i) Tregs in the presence of TGF-β1 (Figure 1E). Interestingly, suboptimal Treg-inducing TGF-β1 concentrations were essential for butyrate-mediated impact on Treg differentiation. The treatment of CD4⁺ T cells with butyrate in the absence of TGF-β1 did not lead to conversion into Foxp3⁺ Tregs, whereas optimal Treg-inducing TGF-β1 concentrations masked the effects of butyrate (Figure 1F). To exclude potential contamination by low numbers of already pre-existing Tregs in the CD4⁺ T cell culture, we sorted and cultured highly purified naïve GFP^-CD62L⁺CD4⁺ T cells from DEREG mice in the presence of butyrate and TGF-β1. After 5 days of culture, we observed an increase in GFP⁺CD4⁺ T lymphocytes in butyrate-treated cells as compared to the control cell cultures indicating a direct effect of butyrate on de novo generation of Tregs from naïve T cells (Figure S2 in Supplementary Material). Taken together, butyrate cooperates with TGF-β1 to enhance the expression of Foxp3 and differentiation into Tregs. This effect of butyrate was proposed to be based on its potent HDAC-inhibitory activity and ability to epigenetically regulate Foxp3 gene expression. In addition, butyrate was also able to stabilize Foxp3 protein by promoting its acetylation (3, 20).

Surprisingly, when performing a titration experiment with butyrate-treated T cells cultured under Treg-inducing conditions, we found that the frequencies of Foxp3 peaked at 0.25 mM butyrate concentration, while higher butyrate doses suppressed the differentiation of Foxp3⁺ Tregs (Figures 2A,B). Recently, it has been shown that SCFAs not only promote expansion of Tregs but impact also on Th1 and Th17 cells (21). Therefore, we examined whether SCFAs might trigger the production of pro-inflammatory cytokines under Treg-polarizing conditions. Notably, butyrate and to a lesser extent propionate were able to selectively induce expression of IFN-γ but not of IL-17A or IL-4 in a dose-dependent manner (Figures 2C,D; Figure S3A,B in Supplementary Material). Interestingly, this effect was independent of FFA2 and FFA3 receptors and of Slc5a8 transporter as the production of IFN-γ was not impaired in Ffar2^−/−Ffar3^−/− and Slc5a8^−/− T cells during Treg differentiation upon butyrate treatment (Figure 2E). The percentage of Foxp3⁺ Tregs was similar in the spleen of WT, Ffar2^−/−Ffar3^−/−, and Slc5a8^−/− mice and was also comparable in in vitro generated Treg cultures (Figure S4A–C in Supplementary Material). Both, Ffar2^−/−Ffar3^−/− and Slc5a8^−/− CD4⁺ T cells cultured under Treg-inducing conditions and treated with butyrate behaved similar to butyrate-treated WT Tregs with respect to the Foxp3 expression (Figure S4D in Supplementary Material). Furthermore, quantitative RT-PCR data demonstrated that butyrate promoted induction of Th1-associated genes Ifnγ and Tbx21 in T cells cultured under Treg-inducing conditions while the expression of Th17- and Th2-promoting transcription factors Rorγt and Gata3 was not induced (Figure 2F). To test whether the increase in IFN-γ production correlates with the HDAC inhibitory effects of SCFAs, we performed an HDAC activity assay on iTregs in the presence of butyrate, propionate, and acetate. While butyrate and propionate exerted a potent HDAC inhibitory effect, acetate almost completely lacked this inhibitory activity (Figure 2G).

As previously described, CD4⁺ T cells cultured under Treg-polarizing conditions and treated with 1 mM butyrate showed significantly increased acetylation of histones H3 and H4 as compared to untreated Tregs (Figure 3A). Under unpolarized conditions, histones are hypoacetylated at the Ifnγ locus in CD4⁺ T cells. During Th1 differentiation, proximal promoter sites of the Ifnγ gene become hyperacetylated (22, 23). To analyze if the inhibition of HDAC activity by butyrate induces histone acetylation directly at the Ifnγ and Tbx21 locus during Treg differentiation, we performed ChIP assay using an anti-acetyl-H3 antibody. Our results revealed that butyrate was capable of promoting H3 acetylation at the Ifng and Tbx21 but not at Il17a locus in T cells cultured under Treg-inducing conditions (Figure 3B). Of note, at a concentration of 0.25 mM, butyrate was already able to increase the pan-H3 acetylation during Treg differentiation (Figure S5A in Supplementary Material). In the absence of butyrate, at the Foxp3 promoter region, H3 acetylation was observed only in TGF-β1-generated iTregs but not in naïve CD4⁺ T cells. This TGF-β1-mediated effect was enhanced after treatment of iTregs with 0.25 or 1 mM butyrate. Interestingly, at a concentration of 0.25 mM, butyrate was capable of upregulating H3 acetylation at the Foxp3 but not at the Ifnγ and Tbx21 promoter (Figure S5B,C in Supplementary Material).

T-bet and IFN-γ have been shown to directly oppose the induction of Foxp3 expression and negatively regulate differentiation of Tregs (24, 25). FACS analysis confirmed induction of T-bet protein and partial reduction of Foxp3 expression in butyrate-treated cells during Treg differentiation (Figure 3C). By comparing the ability of WT and Tbx21^−/− CD4⁺ T cells cultured under Treg-inducing conditions to induce IFN-γ after butyrate treatment, we found that T-bet was required for production of endogenous IFN-γ (Figure 3D). Furthermore, we observed almost no reduction of Foxp3 frequencies in the absence of T-bet in T cells cultured under Treg-inducing conditions and treated with 1 mM butyrate. These data point to the role of this transcription factor in suppressing Foxp3 expression during the butyrate treatment (Figure 3E). To further prove the requirement of HDAC inhibition for IFN-γ induction, CD4⁺ T cells cultured under Treg-inducing conditions were treated with increasing concentrations of pan-HDAC inhibitor TSA. Indeed, the treatment with TSA caused the induction of IFN-γ production and reduction of Foxp3 expression during Treg differentiation (Figures 3F–I). In conclusion, our findings indicate that SCFAs might exhibit multiple effects during Treg differentiation including the induction of pro-inflammatory molecules.

To extend our analysis from the Treg-polarizing conditions to other effector T cells, we first examined possible effects of SCFA-induced IFN-γ expression under unpolarized conditions (Th0 cells). Butyrate was capable of selectively upregulating IFN-γ without inducing the expression of IL-17A (Figure 4A). The treatment of Th0 cells with butyrate increased the expression of IFN-γ in a dose-dependent manner peaking at the concentration of 1 mM. Propionate was less effective than butyrate, while acetate was not able to induce the expression of IFN-γ in unpolarized T cells (Figures 4A,B). Furthermore, we were wondering whether pan-HDAC inhibitor TSA is capable of inducing IFN-γ expression in Th0 cells. Similar to butyrate, a strong induction of IFN-γ but not of Foxp3 or IL-17A was observed after treatment of cells with TSA (Figures 4C,D). Of note, this effect of HDAC inhibitors butyrate and TSA was mediated without contribution of additional factors, while the butyrate-triggered expansion of Tregs was TGF-β1 dependent (Figure 1F).

Moreover, in the presence of butyrate, the IFN-γ expression was increased even in Th1 cells that already display high levels of this cytokine (Figure S6 in Supplementary Material). Remarkably, we also observed a strong, butyrate-triggered induction of IFN-γ and T-bet and reduced IL-4 and IL-17A expression even under Th2- and Th17-promoting conditions, respectively (Figures 5A–D; Figure S7A in Supplementary Material). This effect seems to be mediated via T-bet as Tbx21^−/− Th17 and Th2 cells did not switch toward IFN-γ production (Figures 5E,F). In accordance with these results, the inhibitory effect of butyrate on the expression of Gata3 and Rorγt, genes encoding for transcription factors controlling Th2 and Th17 development, was detected in WT but not in T-bet-deficient cells (Figure S7B in Supplementary Material).

A recent study has suggested that butyrate treatment is ineffective in ameliorating dextran sodium sulfate (DSS)-induced colitis in conventional mice pretreated with antibiotics (14). Although specific in vivo effects of SCFAs on Tregs have been proposed, the induction of Th1-associated genes by butyrate in an inflammatory milieu is also conceivable. To address this possibility, GF mice were orally treated with 100 mM butyrate [based on the published protocol (6)], during the course of DSS-induced colitis. As expected, in naïve GF mice, no SCFAs were found, owing to the absence of microbiota, whereas conventional (SPF) mice exhibited high luminal levels of acetate, propionate, and butyrate in the colon (Figure 6A). Our data demonstrate that butyrate treatment does not attenuate acute intestinal inflammation in GF animals. Although the severity of colitis was similar in both groups, GF mice treated with butyrate and DSS had even slightly increased weigh loss as compared to only DSS-administered mice (Figures 6B,C). While the treatment of animals with butyrate did not impact on the Foxp3 levels during the course of inflammation, the expression of pro-inflammatory molecules Tbx21 and Ifnγ was significantly increased after butyrate administration as compared to only DSS-treated group (Figure 6D). In addition, the FACS analysis confirmed a higher frequency of T-bet⁺ cells within colonic CD4⁺ T lymphocytes after butyrate treatment of GF mice (Figures 6E,F). A similar trend was observed in WT mice treated with butyrate and DSS (Figure S8A,B in Supplementary Material). In summary, our present data challenge the current paradigm that butyrate-mediated effects on the host immune system are only beneficial. We conclude that, in the context of inflammatory bowel disease (IBD), the elevated levels of butyrate in lamina propria might be more harmful than previously thought.