If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Laboratory of Immunology and Hematopoiesis, Department of Comparative Pathobiology, College of Veterinary Medicine, Purdue University, West Lafayette, IndianaWeldon School of Biomedical Engineering, Purdue University, West Lafayette, IndianaCenter for Cancer Research, Purdue University, West Lafayette, Indiana
Short-chain fatty acids (SCFAs), the most abundant microbial metabolites in the intestine, activate cells via G-protein−coupled receptors (GPRs), such as GPR41 and GPR43. We studied regulation of the immune response by SCFAs and their receptors in the intestines of mice.
Inflammatory responses were induced in GPR41−/−, GPR43−/−, and C57BL6 (control) mice by administration of ethanol; 2, 4, 6-trinitrobenzene sulfonic-acid (TNBS); or infection with Citrobacter rodentium. We examined the effects of C rodentium infection on control mice fed SCFAs and/or given injections of antibodies that delay the immune response. We also studied the kinetics of cytokine and chemokine production, leukocyte recruitment, intestinal permeability, and T-cell responses. Primary colon epithelial cells were isolated from GPR41−/−, GPR43−/−, and control mice; signaling pathways regulated by SCFAs were identified using immunohistochemical, enzyme-linked immunosorbent assay, and flow cytometry analyses.
GPR41−/− and GPR43−/− mice had reduced inflammatory responses after administration of ethanol or TNBS compared with control mice, and had a slower immune response against C rodentium infection, clearing the bacteria more slowly. SCFAs activated intestinal epithelial cells to produce chemokines and cytokines in culture and mice after administration of ethanol, TNBS, or C rodentium. These processes required GPR41 and GPR43 and were required to recruit leukocytes and activate effector T cells in the intestine. GPR41 and GPR43 activated extracellular signal-regulated kinase 1/2 and p38 mitogen-activated protein kinase signaling pathways in epithelial cells to induce production of chemokines and cytokines during immune responses.
SCFAs activate GPR41 and GPR43 on intestinal epithelial cells, leading to mitogen-activated protein kinase signaling and rapid production of chemokines and cytokines. These pathways mediate protective immunity and tissue inflammation in mice.
The gut immune system normally maintains immune tolerance to harmless antigens but mounts efficient immune responses to infection by pathogens. The factors and mechanisms that regulate immune responses in the intestine remain incompletely understood. The intestine is densely populated with commensal bacteria, which metabolize dietary fibers and other materials in the colon.
Short-chain fatty acids (SCFAs), such as acetate (C2), propionate (C3), and butyrate (C4) are major end products of gut microbial fermentation and are implicated in the regulation of the gut immune system.
The functions of SCFAs and their receptors in regulation of immune responses are poorly understood. We investigated the functions of SCFAs and their receptors in the regulation of immune responses to immunological challenges including breach of the gut barrier function; 2, 4, 6-trinitrobenzene sulfonic-acid (TNBS) treatment; and infection with Citrobacter rodentium. We provide evidence that SCFAs condition gut ECs to mount prompt immune responses to all immunological challenges in a GPR41 and GPR43-dependent manner. SCFAs and their receptors promote acute inflammatory responses in the intestine for protective immunity and tissue inflammation.
Materials and Methods
All experiments with animals in this study were approved by the Purdue Animal Care and Use Committee. C57BL/6 mice were from Harlan Laboratories (Indianapolis, IN). CD45.1 C57BL/6 mice were from The Jackson Laboratory (Bar Harbor, ME). GPR43−/− C57BL/6 mice were from Deltagen (San Mateo, CA), and GPR41−/− C57BL/6 mice were described previously.
Mice were monitored daily for changes in body weight, signs of illness, and stool scores. Stool consistency scores were measured as follows: 0 = normal stool; 1 = soft stool; 2 = diarrhea; 3 = diarrhea and anal bleeding. Mice were sacrificed 1 day after ethanol treatment or 3 days after TNBS rectal administration. The detailed method to score tissue inflammation is described in the Supplementary Material.
Infection with C rodentium
Wild-type (WT), GPR41−/−, and GPR43−/− mice were infected with C rodentium (DBS100, 109 colony-forming unit [CFU]/mouse) via oral gavage as described previously.
For infection of mice after C2 administration, mice were fed with C2 (200 mM, pH 7.5) in drinking water for 4 weeks and infected with a higher dose (1010 CFU/mouse) of C rodentium. When indicated, mice were injected intraperitoneally with antibodies (50 μg/injection for each antibody) to neutralize interleukin (IL)-6 (MP5-20F3 on day 0 and 2) and deplete neutrophils (RB6-8C5 on day −2 and 0). Mice were monitored for weight change and stool consistency. Mice were sacrificed at indicated time points, and the C rodentium load in feces and tissues was determined as described in the Supplementary Material. Indicated tissues were examined for frequencies of neutrophils and T-cell subsets by flow cytometry.
Bone Marrow Reconstitution
Recipient mice were exposed to 960 cGy ionizing radiation and were reconstituted with donor marrow cells (1 × 107/mouse) injected into a tail vein. The mice were kept for 8 weeks for hematopoietic reconstitution and were sensitized and administered with TNBS or orally infected with C rodentium (109 CFU/mouse). CD45.1 mice were used as recipients when GPR41−/− (CD45.2) and GPR43−/− (CD45.2) mice were used as marrow donors. When GPR41−/− and GPR43−/− mice were used as the recipients, CD45.1 mice were used as the marrow donors. The marrow reconstitution efficiency in these animals was >95%.
For immunofluorescence staining, 6-μm–thick frozen and paraffin colon sections were stained with rabbit polyclonal antibodies to ZO-1 (Invitrogen, Carlsbad, CA) and phospho-p44/42 MAPK (Erk1/2) (Cell Signaling Technology, Danvers, MA), respectively. Slides were further stained with fluorescently labeled polyclonal anti-rabbit IgG antibodies (Invitrogen). Tissue sections were stained with anti–Gr-1 (RB6-8C5) and Hoechst 33342 (Invitrogen) when indicated, and the images were acquired with a Leica SP5 II confocal microscope.
Microarray and Real-Time Polymerase Chain Reaction
Colon tissues were frozen in liquid nitrogen and ground using mortar and pestle. The array data were analyzed as described
Mice fasted overnight, received intragastrically with fluorescein isothiocyanate (FITC)–conjugated dextran (0.2 mg/g body weight (g) mean molecular weight 3000–5000; Sigma-Aldrich, St Louis, MO) using a round-tip feeding needle. Mice were sacrificed 3 h later, and the FITC-dextran concentration in the plasma was determined with a fluorescent microplate reader (excitation at 485 nm and emission at 535 nm; BioTek Synergy HT; BioTek, Winooski, VT).
In vitro Culture of Colonic ECs
Lamina propria cells and colonic ECs (viability and purity >95% based on CD45 negativity) were isolated as described previously.
For quantitative real-time polymerase chain reaction analysis of GPR41 and GPR43, CD326+ cells were selected with Miltenyi magnetic beads (Miltenyi Biotec, Bergisch Gladbach, Germany). Colonic ECs (2 × 105 cells/well) were cultured for 6 days, and C2 (10 mM) or C3 (1 mM) was added during this culture period. These cells were further activated with lipopolysaccharide (Sigma-Aldrich; 1 μg/mL) or a commensal bacterial extract (20 μg/mL) for 24 h. Alternatively, colonic ECs, established for 3 days, were treated with kinase inhibitors and SCFAs for 3 days. Please see the Supplementary Material for the experiment with the inhibitors.
Flow cytometry was performed as described previously.
For neutrophils, tissue cells were stained with antibodies to Gr-1 (RB6-8C5) and 7/4 (Ly-6B.2). For phosphorylation of extracellular signal-regulated kinase (ERK) and ATF2 in ECs, cells were fixed and permeabilized with perm III buffer (BD Bioscience, San Diego, CA) and stained with antibodies to phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) or phospho-ATF-2 (Thr71) (Cell Signaling Technology). Donkey anti-rabbit IgG-FITC (BioLegend, San Diego, CA) was used as the secondary antibody.
Student's t test, Mann-Whitney test, and one-way analysis of variance were used to determine significance of the differences between groups. P values ≤.05 were considered significant.
Mice Deficient With GPR41 or GPR43 Are Defective in Mounting the Normal Inflammatory Response After an Ethanol-Induced Breach of the Gut Barrier Function
To determine the functions of SCFAs and their receptors in regulation of acute antibacterial responses, we temporarily breached the gut barrier function through rectal administration of ethanol into WT, GPR41−/−, and GPR43−/− mice. As expected, the ethanol treatment induced a transient decrease in the weight of WT mice at 24 h after the treatment (Figure 1A). However, the 2 knockout (KO) mouse strains were significantly less affected by the treatment. Gross examination revealed that only WT mice, but not the 2 KO strains, had apparent shortening and thickening of the colon (Figure 1B). The ethanol treatment induced infiltration of neutrophils within the colon tissue (Figure 1C). In this regard, the neutrophil frequency in the colonic lamina propria was abnormally low in the SCFA-receptor KO mice compared with WT mice. In line with this, GPR41−/− and GPR43−/− mice, unlike WT mice, did not display severe histological changes and leukocyte infiltration in the mucosa and submucosa (Supplementary Figure 1).
Gut permeability is increased during immune responses.
Unlike WT mice, GPR41−/− and GPR43−/− mice did not lose ZO-1 expression and had relatively mild tissue infiltration with Gr-1+ neutrophils after the ethanol treatment (Figure 1D). Gut barrier function was determined based on FITC-dextran leakage from the colonic lumen into blood circulation. Compared with WT mice, the 2 KO strains had smaller gut permeability changes. As the gut permeability increased, so did the bacterial cells within the tissue. Again, GPR41−/− and GPR43−/− mice had smaller increases in tissue bacteria compared with WT mice (Figure 1E).
To gain more insight into the inflammatory response regulated by the SCFA-receptor signals, we performed an Affymetrix microarray gene expression study. Many genes induced in WT mice were significantly less induced in the 2 SCFA-receptor KO strains (Figure 2). Interestingly, there is a striking similarity in overall gene expression between GPR41−/− and GPR43−/− mice (Figure 2A). GPR41−/− and GPR43−/− mice were defective in up-regulation of certain CCL and CXCL chemokines and inflammatory cytokines (ie, IL-1β, IL-6, and IL-11) (Figure 2B). Also defective was expression of metalloproteinases, including MMP7, MMP9, and MMP10. The defective expression of chemokines, cytokines, and metalloproteinases in SCFA-receptor KO mice was confirmed with quantitative real-time polymerase chain reaction (Figure 2C). These results indicate that SCFA-receptor KO mice are defective in up-regulating key inflammatory mediators in the intestine.
Mice Deficient With GPR41 or GPR43 Are Inefficient in Mounting the Inflammatory Response to TNBS
We further examined the response to TNBS. The TNBS challenge decreased body weight in WT mice by approximately 20% 3 days after the rectal challenge (Figure 3A). The same treatment decreased body weight by only about 10% in SCFA-receptor KO mice. Only WT, but not the KO, mice had soft stool on day 2 (Figure 3B). WT mice had considerable shortening of the colon (∼20% decrease) as expected, but the colon of GPR41−/− and GPR43−/− mice was significantly less affected (∼10% decrease) (Figure 3C). Induction of IL-6, IL-12, and T-helper cell (Th) 1−associated genes (interferon gamma and T-bet) was defective in both GPR41−/− and GPR43−/− mice (Figure 3D and Supplementary Figure 2A). In line with this, the frequency of Th1 cells was decreased in the small intestine, colon, and mesenteric lymph node of the KO mice (Supplementary Figure 2B). In addition, induction of neutrophil infiltration and expression of CXCL1, CXCL2, and CCL2 in the colon were abnormally low in GPR41−/− and GPR43−/− mice (Figure 3E and F). Histologic examination of the distal colon revealed more severe leukocyte infiltration, mucosa hyperplasia, and tissue damage in WT mice than the KO mice (Supplementary Figure 3). These results indicate that the 2 SCFA receptors play positive roles in mounting the acute inflammatory response to TNBS.
GPR41 and GPR43 Signals Are Required for Undelayed Induction of the Immune Response to C rodentium
We used a C rodentium infection model to determine the role of SCFA receptors in mounting immune responses to bacterial pathogens. GPR41−/− and GPR43−/− mice were more susceptible than WT mice based on weight change and stool consistency scores (Figure 4A and B). Clearance of C rodentium from the intestine was delayed in the KO mice (Figure 4C). There was increased translocation of the pathogen into the liver and spleen of the KO mice. The KO mice had low responses in blood IL-6 (Supplementary Figure 4), neutrophil recruitment (Figure 4D), and frequencies of Th1 and Th17 cells in the colon (Figure 4E). We examined the expression kinetics of inflammatory cytokines (ie, IL-6, IL-17A, and IL-12), and interferon gamma and chemokines (CXCL1 and CXCL2) and found that induction of these cytokines and chemokines was delayed by 1−2 weeks (Supplementary Figure 5A and B). We next examined the gut permeability change regulated by SCFA receptors. There was no difference in gut permeability or expression of tight junction protein genes (ZO-1 and Occludin) between WT and any of the 2 SCFA receptor KO mice in the absence of any immunological challenges (Supplementary Figure 6). However, on infection with C rodentium, GPR41, or GPR43 KO mice had a delayed kinetics in gut permeability change (Figure 4F). Although the tissue bacteria level was lower at early time points (before 14 days) in the KO mice compared with WT mice, it was considerably higher at later time points, reflecting increased pathogen invasion and/or bacterial translocation into the tissue.
Immune Response Promoted by SCFAs Is Required for Timely Clearance of C Rodentium
We next examined the effect of an increased C2 concentration on the antibacterial immune response. For this, we fed mice with C2 and examined the kinetics of the antibacterial response to C rodentium. C2-fed mice suffered less than WT mice from the infection based on weight change and stool consistency (Figure 5A) and the intestinal C rodentium load (Figure 5B). C2 administration accelerated the expression of IL-6, CXCL1, and CXCL2 (Figure 5C). Also increased upon the C2 administration was recruitment of neutrophils (Figure 5D) and Th17 cells (Figure 5E) in the large bowel (cecum). C2 administration somewhat accelerated the gut permeability change (Figure 5F). C2 administration accelerated the overall immune response during the infection with C rodentium.
To assess the importance of the immune response promoted by SCFAs, we included animal groups where key immune components (IL-6 and neutrophils) were neutralized or depleted just before or after the infection with C rodentium. As expected, the antibody-injected mice were more susceptible than control mice to C rodentium (Figure 5A and B), demonstrating the important roles of IL-6 and neutrophils in induction of antibacterial immunity. The antibody treatment delayed the induction of IL-6 and neutrophil-attracting chemokines (Figure 5C). In line with this, the increase of neutrophils (Figure 5D) and Th17 cells (Figure 5E) in the large bowel was significantly delayed with the antibody treatment. The boosting effect of C2 administration was mostly abolished by the antibody treatment. Overall, the immune responses promoted by SCFAs are critical for timely clearance of pathogens.
We also examined the impact of the antibody-induced delay in immune responses on gut permeability change. This early injection of antibodies somewhat suppressed the increase in gut permeability at an early time point (day 1−12) (Figure 5F). At later time points (day 15 and later), however, a greater change in gut permeability occurred in the antibody-injected animals (Figure 5F). The hyper-gut permeability at later time points is in line with the higher C rodentium load and the subsequently increased cytokine/chemokine response to the bacteria (Figure 5B and C). Overall, these results suggest that the gut permeability change is induced by the immune response promoted by the SCFA signal.
SCFA receptor deficiency can affect the commensal bacteria profile in the gut to indirectly affect the immune response. To determine the possibility, we examined the levels of 7 bacterial groups in the gut of GPR41−/− and GPR43−/− mice. We found no significant change in the gut commensal bacteria profile (Supplementary Figure 7).
Roles of SCFA Receptors Expressed by Bone Marrow vs Non−Marrow-Derived Cells in Promoting the Antibacterial Immune Response
To determine the roles of SCFA receptors expressed by marrow vs non−marrow-derived cells, we prepared bone marrow chimera mice with WT recipients and GPR41−/− or GPR43−/− marrow donors or vice versa. At 8 weeks post-transplantation, these mice were infected with C rodentium, and clinical scores and immune responses after the infection were examined (Figure 6). WT mice reconstituted with WT marrow (WT→WT) or KO marrow (41KO→WT or 43KO→WT) did not lose weight significantly after infection. In contrast, the KO recipient mice, reconstituted with WT marrow (WT→41KO or WT→43KO), lost weight significantly (Figure 6A). Stool consistency score and C rodentium load were significantly higher in WT→41KO mice and WT→43KO mice compared with WT→WT mice (Figure 6B and C). In line with the defective clearance of C rodentium, frequencies of Th17 cells and neutrophils were lower in the colon of WT→41KO mice and WT→43KO mice compared with other mice (Figure 6D and E). These data indicate that the SCFA receptors expressed by non−marrow-derived cells play bigger roles than those of marrow-derived cells in promoting the antibacterial responses.
SCFAs Condition ECs for Optimal Production of Chemokines and Inflammatory Cytokines in a MEK-ERK−Dependent Manner
The bone marrow chimera study in Figure 6 reveals that non−marrow-derived cells play important roles in mediating the SCFA-dependent inflammatory response. Enteroendocrine cells and colonic enterocytes express GPR43 and GPR41.
We confirmed that both GPR41 and GPR43 messenger RNAs were highly expressed by CD326+ (epithelia cell adhesion molecule) ECs in the intestine (Figure 6F). We determined the possibility that SCFAs condition ECs for more efficient production of chemokines and inflammatory cytokines in response to bacterial products. We established primary colonic ECs in vitro and activated them for 24 h with lipopolysaccharide or a gut commensal bacterial extract. Interestingly, small but detectable induction of IL-6 expression in response to C2 or C3 alone was observed (Figure 7A and Supplementary Figure 9). In addition, expression of IL-6, CXCL1, and CXCL10 in response to the bacterial products was increased in the presence of C2 or C3. The induction was abolished when colonic ECs were pretreated with pertussis toxin, which is an inhibitor of G0/i-coupled receptors, such as GPR41 and GPR43 (Supplementary Figure 9). The ECs, isolated from GPR41−/− or GPR43−/− mice, were largely unresponsive to C2 or C3 in expression of the cytokines (Figure 7A) and chemokines beyond basal levels (Figure 7B).
We studied the intracellular pathways important for SCFA-dependent expression of the cytokines. SCFAs activate the MEK-ERK pathway in neutrophils.
We confirmed that C2 and C3 induced phosphorylation of ERK in primary colonic ECs (Figure 7C). The SCFA-dependent phosphorylation of ERK did not occur in GPR41−/− and GPR43−/− ECs (Figure 7C), suggesting that ERK activation is induced by the SCFA signals through GPR41 and GPR43. This ERK activation was suppressed by the following inhibitors of ERK: PD98059 (a MEK1 inhibitor) and U0126 (a MEK1/2 inhibitor) (Figure 7D). These inhibitors also suppressed the expression of IL-6, CXCL1, and CXCL2 induced in response to C2 or C3 (Supplementary Figure 10). Also, p38 MAPK is implicated in the SCFA receptor signaling in a breast cancer cell line.
SB203508 (a p38 inhibitor) was suppressive on induction of IL-6 and chemokines (Figure 7D and Supplementary Figure 10). However, a c-Jun N-terminal kinases/c-Jun N-terminal kinase inhibitor, SP600125, was ineffective. We confirmed, using an immunohistochemistry technique, that activation of ERK was greatly enhanced with C2 administration through drinking water in the colon of WT, but not GPR41 or GPR43 KO, mice infected with C rodentium (Figure 7E). Activating transcription factor 2 (ATF2) is activated downstream of ERK and MAPK pathways and involved in expression of inflammatory cytokines and chemokines.
We examined if C2 and C3 would activate ATF2 in a GPR41/GPR43-dependent manner in ECs. ATF2 activation was induced by C2 and C3, but this activation did not occur in GPR41 or GPR43 KO cells (Figure 7F).
The mechanism of action for probiotics, prebiotics, and SCFAs to promote immunity in the gut is incompletely understood.
We studied the roles of SCFA receptors in regulating immune responses in the intestine. The results identified positive roles of SCFAs and their receptors in preparing ECs to promptly mount immune responses during immunological challenges.
Our study demonstrated that SCFA signals are required for mounting acute inflammatory responses in the colon after ethanol treatment. Expression of key mediators of inflammatory responses, including chemokines and cytokines, was defective in the absence of GPR41 and GPR43. There have been mixed reports regarding the roles of SCFAs in regulation of chronic gut inflammation. Some reported that parenteral administration of butyrate increased gut expression of IL-1β and IL-6.
Our results indicate that both delayed and exaggerated immune responses can occur if SCFA signals are deficient. The exaggerated immune responses can occur at later time points and can be caused by delayed immune responses and subsequently increased bacterial load in the intestinal tissue. Different interpretations are likely, depending on the experimental models and time points.
The TNBS model is a good method to study acute immune responses to a defined antigen (ie, hapten) in the intestine. GPR41−/− and GPR43−/− mice experienced abnormally low inflammatory responses in the colon, as evidenced by low induction of inflammatory chemokines and cytokines and leukocyte infiltration. In addition, mice deficient with GPR41 or GPR43 could not mount a normal Th1 response to TNBS. The TNBS study corroborates the conclusion from the ethanol treatment study that SCFA signals are required for optimal acute inflammatory responses in the gut.
Utilizing an infection model with C rodentium that examines immune responses occurring during a 4-week period, we found that mice deficient with GPR41 or GPR43 failed to induce the acute inflammatory response to clear the pathogen at early time points. This would lead to delayed clearance and increased invasion of the pathogen into tissues. On the other hand, in vivo C2 administration facilitated clearance of the pathogen. Induction of inflammatory cytokines and chemokines in the GPR41 or GPR43 KO mice was significantly delayed. Consistently, neutrophil infiltration and the Th17 response to bacteria were also delayed. These results suggest that the early inflammatory response promoted by SCFA signals plays a beneficial role in clearance of pathogens.
SCFA receptors are differentially expressed by ECs, mast cells, phagocytes, and other cell types in the intestine.
CD326+ ECs express both GPR41 and GPR43 at high levels. Neutrophils express GPR43, but not GPR41, and T and B cells do not express the receptors. Along with the results from the bone marrow reconstitution study, this indicates that ECs are the most likely cells to mediate the observed SCFA effects. This interpretation is further supported by the results that SCFAs activate colonic ECs in a GPR41/43-dependent manner and condition the ECs to effectively respond to bacterial products. The results indicate that timely expression of chemokines and inflammatory cytokines by colonic ECs requires activation of both GPR41 and GPR43 by SCFAs. Although not significantly expressing GPR41, the GPR43-expressing phagocytes can contribute to the overall immune responses regulated by SCFAs. Additional cell types can also mediate the SCFA effects, and this is a subject of importance for future studies. As an intracellular mechanism, we found that the MEK-ERK and p38 MAPK pathways and the ATF2 transcription factor are activated by SCFA signals downstream of GPR41 or GPR43 in gut ECs. Activation of these pathways is required for normal expression of inflammatory cytokines and chemokines by ECs in response to SCFAs.
The reduced acute immune response in the SCFA-receptor KO mice is mainly due to defective SCFA signaling in ECs. Induction of chemokines and cytokines in ECs by SCFAs in vitro in a GPR41/43-dependent manner supports this interpretation. Our results indicate that the abnormally low cytokine and neutrophil response is the primary reason for the ineffective clearance of pathogens in the GPR41/43 KO mice. Our results ruled out the possibilities that SCFAs increase gut permeability in a steady condition, and that the low immune response would be due simply to reduced gut permeability and bacterial translocation. Gut permeability was even increased in GPR41/43 KO mice at later time points after infection due to the delayed clearance of pathogens and increased antibacterial response. Gut permeability change is an integral part of the overall immune response, which is positively regulated by SCFAs and their receptors through immune responses. In sum, our study revealed novel functions of SCFAs and their receptors in promoting acute inflammatory responses (Supplementary Figure 11), which is beneficial for mounting the immune response to pathogens, but can mediate tissue inflammation.
The authors thank B. Ulrich and F. Chu (Purdue University) for their excellent assistance and Dr. B. Vallance (University of British Columbia) for providing a C rodentium strain.
In vitro Culture of Colonic ECs
Colonic ECs (2 × 105 cells/well) established in 48-well plates coated with Matrigel basement membrane matrix (BD Bioscience, San Diego, CA). C2 (10 mM) or C3 (1 mM) were activated with lipopolysaccharide (Sigma-Aldrich; 1 μg/mL) or commensal bacterial extract (20 μg/mL) for 24 h. The commensal extract was prepared from the cecal content of 8-week-old C57BL/6 mice with bead beating of the fecal suspension layer. Alternatively, colonic ECs, established for 3 days, were treated with kinase inhibitors (PD98059, U0126, SB203580, and SP600125; Enzo Life Sciences, Inc., Farmingdale, NY) for 1 h before treatment with SCFAs for 3 days. Phosphorylation of ERK in SCFA-treated ECs was examined by flow cytometry as described here. Concentrations of indicated cytokines in serum and conditioned medium were measured in triplicate with a Luminex assay (Luminex Corporation, Austin, TX) or an enzyme-linked immunosorbent assay method. SCFA-treated ECs were harvested and examined by quantitative real-time polymerase chain reaction for expression of chemokines and cytokines.
Scoring of Intestinal Inflammation
Histological changes in the distal colon after rectal administration with 50% ethanol with or without TNBS (Sigma-Aldrich). Colon tissues in paraffin were cut into 6-μm−thick sections and stained with H&E. Sections were semi-quantitatively graded on a 0−4 scale based on pathological features, including the extent of leukocyte infiltration, crypt elongation/abscess, bowel wall thickening, mucosa hyperplasia, and ulceration: 0 = no evidence of inflammation; 1 = low level of leukocyte infiltration in <10% high power fields (hpf), no structural changes observed; 2 = moderate leukocyte infiltration in 10%−25% hpf, crypt elongation, and bowel wall thickening; 3 = high level of leukocyte infiltration in 25%−50% hpf, bowel wall thickening and mucosa hyperplasia; and 4 = severe leukocyte infiltration in >50% hpf, crypt elongation, bowel wall thickening, mucosa hyperplasia, and ulceration.
Bacterial 16S Ribosomal RNA Gene Analysis
Colon tissues, washed to completely remove the luminal content and feces, were frozen in liquid nitrogen and ground using mortar and pestle. Total DNA was isolated from colon tissues and fecal pellets using QIAamp DNA Stool Mini Kit (Qiagen, Valencia, CA). For total eubacteria in tissues, a pair of primers was used to amplify the 16S ribosomal RNA (rRNA) gene sequence conserved in all commensal bacteria. The polymerase chain reaction was performed for 40 cycles (95°C for 30 s and 63°C for 45 s) using Maxima SYBR Green/ROX qPCR Master Mix (Thermo Scientific, Logan, UT). Tissue-associated eubacteria levels based on the Ct values of the amplified conserved 16S rRNA gene sequence were normalized with the signal of amplified mouse genomic Actb (β-actin) gene in each sample. The abundance of specific bacterial groups was measured with group-specific 16S rRNA gene primers (synthesized by Integrated DNA Technologies) (Supplementary Table 1). The signals of group-specific 16S rRNA gene were normalized with that of the total eubacteria 16S rRNA gene in each sample. The primers used are described in Supplementary Table 1.