Bacteroides ovatus Promotes IL-22 Production and Reduces Trinitrobenzene Sulfonic Acid–Driven Colonic Inflammation

Abstract

The intestinal microbiota influences the development and function of the mucosal immune system. However, the exact mechanisms by which commensal microbes modulate immunity is not clear. We previously demonstrated that commensal Bacteroides ovatus ATCC 8384 reduces mucosal inflammation. Herein, we aimed to identify immunomodulatory pathways employed by B. ovatus. In germ-free mice, mono-association with B. ovatus shifted the CD11b⁺/CD11c⁺ and CD103⁺/CD11c⁺ dendritic cell populations. Because indole compounds are known to modulate dendritic cells, B. ovatus cell-free supernatant was screened for tryptophan metabolites by liquid chromatography–tandem mass spectrometry and larger quantities of indole-3-acetic acid were detected. Analysis of cecal and fecal samples from germ-free and B. ovatus mono-associated mice confirmed that B. ovatus could elevate indole-3-acetic acid concentrations in vivo. Indole metabolites have previously been shown to stimulate immune cells to secrete the reparative cytokine IL-22. Addition of B. ovatus cell-free supernatant to immature bone marrow–derived dendritic cells stimulated IL-22 secretion. The ability of IL-22 to drive repair in the intestinal epithelium was confirmed using a physiologically relevant human intestinal enteroid model. Finally, B. ovatus shifted the immune cell populations in trinitrobenzene sulfonic acid–treated mice and up-regulated colonic IL-22 expression, effects that correlated with decreased inflammation. Our data suggest that B. ovatus–produced indole-3-acetic acid promotes IL-22 production by immune cells, yielding beneficial effects on colitis.

Crohn disease and ulcerative colitis are chronic, relapsing and remitting inflammatory bowel diseases (IBDs) that impact over three million Americans.¹ Current treatments include aminosalicylates, systemic steroids, immunomodulators such as 6-mercaptopurine and methotrexate, and biologic medications such as anti–tumor necrosis factor agents. These medications can cause significant adverse effects, including bone marrow suppression and malignancy,² highlighting a critical need for novel and safe adjuvant treatment strategies for IBD.

The complex yet undefined etiology of IBD involves an intricate interplay of genetic predisposition, environmental factors, the host immune system, and the intestinal microbiota. Decreased fecal and mucosal Bacteroides³ are strongly associated with the development of IBD.⁴ Bacteroides species comprise nearly 40% of the average healthy child's gut microbiome, with their abundance increasing with age.⁵ In general, Bacteroides beneficially influence immunoregulatory, physiological, and metabolic homeostasis in the host.⁶ The beneficial effects of Bacteroides species have been demonstrated in several colitis models. Bacteroides thetaiotaomicron ameliorates colitis in both dextran sulfate sodium (DSS) and IL-10 knockout colitis models.⁷ Bacteroides fragilis treatment decreases inflammation in trinitrobenzene sulfonic acid (TNBS) and DSS colitis models.^8,9 Similarly, monocolonization with Bacteroides ovatus in immunodeficient mice decrease mortality associated with DSS-induced colitis.¹⁰

Testing three Bacteroides species (Bacteroides vulgatus, B. thetaiotaomicron, and B. ovatus) in a DSS colitis model¹¹ showed that B. ovatus ATCC 8483 monotherapy was most effective in reducing intestinal inflammation, enhancing epithelial recovery, and prolonging host survival.¹¹ Although commensal Bacteroides species are typically characterized as beneficial microbes, several Bacteroides species possess virulence factors and cause human infection.¹² More importantly, these virulence factors appear to be absent in B. ovatus.¹³ Despite these findings, B. ovatus has not been traditionally considered as a probiotic and its therapeutic potential in the setting of colitis has not been thoroughly investigated. We sought to further characterize the anti-inflammatory properties of B. ovatus in a gnotobiotic mouse model and TNBS model of murine colitis.

Materials and Methods

Bacterial Culture

Bacteroides ovatus ATCC 8483 (ATCC, Manassas, VA) was cultured in an anaerobic workstation (AS-580; Anaerobe Systems, Morgan Hill, CA) with a mixture of 5% CO₂, 5% H₂, and 90% N₂. Bacteroides ovatus was grown in Brain-Heart-Infusion medium (Difco; Fisher Scientific, Hampton, NH) supplemented with 2% yeast extract and 0.2% cysteine (BHIS) from single colonies at 37°C overnight anaerobically. Bacteroides ovatus was subcultured into a chemically defined minimal media, termed CDMM,^14,15 at an OD (OD_600nm) = 0.1. Bacteroides ovatus CDMM cultures were grown anaerobically for 18 hours at 37°C. After incubation, cultures were centrifuged at 5000 × g for 5 minutes, and the resulting supernatant was sterile filtered through a 0.2-μm pore polyvinylidene fluoride membrane (Millipore, Burlington, MA). This supernatant is termed conditioned media. For animal experiments, B. ovatus was grown overnight anaerobically in BHIS, centrifuged at 5000 × g for 5 minutes, and washed 2× with sterile anaerobic phosphate-buffered saline (PBS) to remove residual BHIS. Then, the cultures were adjusted to 2 × 10⁸ cells/mL in sterile anaerobic PBS and used for oral gavage. Bacterial viability was confirmed for each gavage session by serially plating B. ovatus on BHIS agar to calculate colony-forming units.

Mouse Models

Germ-free and specific pathogen-free (SPF) animal experimental procedures were approved by the Institutional Animal Care and Use Committee at Baylor College of Medicine, Houston, TX. All gnotobiotic mouse experiments were performed in sterile isolators in the Baylor College of Medicine germ-free facility. For experiments, C57BL/6 germ-free mice (aged 8 to 12 weeks) were gavaged with sterile PBS (germ-free controls) or 2 × 10⁸ colony-forming units/mL B. ovatus ATCC 8483 grown in Brain-Heart-Infusion medium and resuspended in sterile PBS (B. ovatus mono-associated). Mice received one oral gavage treatment at the start of the experiment and were maintained in gnotobiotic isolators for 3 weeks. Colonization was confirmed by culturing fecal samples from mice before treatment and 3 weeks after treatment on BHIS agar plates (Hardy Diagnostics, Santa Maria, CA) anaerobically at 37°C for 48 hours. To confirm the absence of other bacteria, BHIS plates were incubated aerobically at 37°C for 48 hours. Experiments were repeated four separate times with n = 5 males/5 females per experiment (total 20 males/20 females). Two independent experiments were used for flow cytometry analysis, and two independent experiments were used for histology and quantitative real-time PCR (qPCR) analysis.

For TNBS experiments, BALB/c mice (aged 8 to 12 weeks) were purchased from Taconic (Germantown, NY). Mice were pretreated by oral gavage with sterile PBS or B. ovatus at 2 × 10⁹ colony-forming units/mL for 1 week. Then, mice were anesthetized by isoflurane inhalation, and 5% (w/v) TNBS in ethanol was rectally instilled by catheter into the midcolon. Following TNBS administration, mice received daily oral gavage of PBS or B. ovatus until euthanasia (1 or 3 days after TNBS). Animals were sacrificed on day 1 after TNBS treatment for immune cell analysis by flow cytometry. Alternatively, animals were sacrificed on day 3 after TNBS treatment for histologic, serum, and colon gene expression analysis. Hematoxylin and eosin (H&E) stains were imaged on the Nikon Eclipse 90i (Nikon, Tokyo, Japan) microscope using a 20× Plan Apo (numerical aperture, 0.75) differential interference contrast objective with a DS-Fi1-U2 camera (Nikon). Histologic scores of H&E-stained slides were assessed by a board-certified anatomic pathologist (D.A.S.) on an Olympus BX41 microscope with a DP71 Olympus camera (Olympus, Tokyo, Japan).

Generation of Mouse Bone Marrow–Derived Dendritic Cells

To isolate bone marrow cells, germ-free mice or SPF conventional C57BL/6 mice (aged 12 to 16 weeks) were euthanized according to an Institutional Animal Care and Use Committee–approved protocol and processed, as previously described.¹⁶ Briefly, bone marrow cells were collected from femurs and tibias by flushing with RPMI 1640 cell culture media. The resulting cell solution was centrifuged at 500 × g for 5 minutes, and the cell pellet was treated with red blood cell lysis buffer for 1 minute. Following an additional wash, 10-cm Petri dishes were seeded with 10⁵ mL⁻¹ bone marrow cells in RPMI 1640 medium supplemented with 10% (v/v) heat-inactivated fetal bovine serum, 100 ng/mL IL-4 (Peprotech, Rocky Hill, NJ; number 214-14), and 100 ng/mL murine granulocyte-macrophage colony-stimulating factor (Peprotech number 315 to 03). Mouse bone marrow cells were incubated for 6 days at 37°C, 5% CO₂ with a medium change on day 3. On day 6, immature dendritic cells were trypsinized, assessed for viability and cell number on a Countess (Invitrogen, Carlsbad, CA), and seeded in new quadrant Petri dishes at 2 × 10⁵ cells/mL. After allowing the dendritic cells to adhere for >3 hours, dendritic cells were treated with RPMI 1640 (medium), 100 ng/mL lipopolysaccharide (derived from Escherichia coli O111:B4; InvivoGen, San Diego, CA; number tlrl-3pelps) supplemented with RPMI 1640 medium, 25% uninoculated CDMM in RPMI 1640 medium, or 25% B. ovatus CDMM conditioned medium in RPMI 1640 medium overnight (16 hours). After the incubation, supernatants were collected for cytokine examination by Luminex MagPix analysis (Luminex Corp., Austin, TX).

Mouse Cytokine Studies by Multiplex Liquid Bead Arrays or Single-Plex Enzyme-Linked Immunosorbent Assays

Serum samples were collected from cardiac puncture after centrifugation of blood in K₂-EDTA–coated tubes, according to manufacturer's protocol (BD Microtainer tubes number 365974; BD Biosciences, Franklin Lakes, NJ). Plasma and mouse bone marrow–derived dendritic cell supernatants were assayed using a Cytokine Magnetic bead panel (Millipore; catalog number MCYTOMAG) with a MagPix instrument (Luminex Corp.). All panels were performed by the Functional Genomics and Microbiome Core of the Texas Medical Center Digestive Diseases Center. Data were obtained with Luminex xPONENT for MAGPIX version 4.2 Build 1324 and analyzed with MILLIPLEX Analyst version 5.1.0.0 standard Build 10/27/2012. Dendritic cell supernatants were examined for IL-22 using the mouse IL-22 DuoSet enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN; number DY582-05), according to the manufacturer's protocols.

Immune Cell Staining for Flow Cytometry Analysis

Single-cell suspensions from the colon were obtained following multiple steps, including clearance of fecal matter, removing epithelial cells, enzymatic digestion, flow through from 70-μm cell strainers (BD Biosciences, San Jose, CA), buffy layer (containing immune cells) isolation by Ficoll-Paque Premium (GE Healthcare, Chicago, IL), and finally washing and resuspending cells in MACS buffer (1× PBS, 0.5% bovine serum albumin, and 2 mmol/L EDTA) for staining. Immune cell isolation was based on the bio-protocol¹⁷ with modifications. Epithelial cell dissociation solution consists of 5 mmol/L EDTA and 10 mmol/L HEPES in RPMI 1640 medium (Sigma-Aldrich, St. Louis, MO). Enzymatic digestion solution is RPMI 1640 complete medium supplemented with collagenase V (0.1 mg/mL) from Clostridium histolyticum, DNase I (30 U/mL), and dispase (50 U/mL; Sigma-Aldrich). Specific antibodies for surface and intracellular staining included CD4 (GK1.5) conjugated with fluorescein isothiocyanate, CD8 (53 to 6.7) conjugated with Pacific Blue, intracellular transcription factor forkhead box protein P3 (Foxp3) (FJK-16a) conjugated with allophycocyanin for T cells, and CD11c (N418) conjugated with phycoerythrin cyanine 7, CD11b (M1/70) conjugated with peridinin-chlorophyll protein cyanine 5.5, and CD103 (2E7) conjugated with phycoerythrin. All conjugated antibodies were purchased from BioLegend (San Diego, CA). Intracellular staining was performed using a fixation and permeabilization kit, according to the manufacturer's protocol (eBioscience, San Diego, CA). Prepared samples were analyzed by flow cytometry using a BD LSRFortessa Cell Analyzer (BD Biosciences) for data collection and FlowJo software version 10 for data analysis (FlowJo LLC, BD, Ashland, OR).

Mouse Colonic Gene Expression and Fecal Bacterial Studies

RNA was extracted from mouse colon in TRIZOL, according to manufacturer details (ThermoFisher number 15596018; Thermo Fisher, Waltham, MA). cDNA was generated from 1 μg RNA using the SensiFAST cDNA synthesis kit (Bioline USA Inc., Taunton, MA). qPCR was performed on a QuantStudio 3 qPCR machine (Applied Biosystems, Foster City, CA) using FastSYBR Green (ThermoFisher) and 10 nmol/L primers designed using PrimerDesign (ThermoFisher). The ΔddCT method was used to calculate the relative fold change with the housekeeping gene 18S. For analysis of B. ovatus colonization, genomic DNA was extracted from fecal samples using the Zymo Quick-DNA Fecal/Soil Microbe Kit (catalog number D6010; Zymo Research, Irvine, CA), and genomic DNA was examined by qPCR using the following primers: B. ovatus, 5′-AAGTCGAGGGGCAGCATTTT-3′ (forward) and 5′-CACAACTGACTTAACAATCC-3′ (reverse), as previously described.¹⁸

Scratch Assays

Human intestinal enteroids (HIEs) were obtained from the Texas Medical Center Digestive Diseases Center Gastrointestinal Experimental Model Systems Core and grown, as previously described.19, 20, 21 Jejunal HIE culture J3 (adult, female, deidentified) was used for these studies. HIEs were transduced with LifeActRuby (Addgene plasmid number 51009; Addgene, Watertown, MA), as described previously.21, 22, 23 jHIE-LifeActRuby enteroids were grown in three dimensions in phenol red-free, growth factor–reduced Matrigel (Corning, Corning, NY) on 24-well plates (Corning) with complete media with growth factors with high Wnt3a.19, 20, 21^,24 Monolayers (two dimensional) were prepared from three-dimensional HIE cultures, as previously described.^21,24,25 Briefly, three-dimensional HIEs were disrupted and seeded in Matrigel-coated 10-well CELLview chamber slides (Greiner Bio-One number 543979; Greiner Bio-One, Monroe, NC) and incubated in complete media with growth factors with 10 μmol/L Y-27632 Rock inhibitor at 37°C with 5% CO₂. After 24 hours, HIE medium was replaced with differentiation medium and monolayers were differentiated for 4 to 5 days. For live-cell imaging assays, HIE medium was replaced with optically clear Dulbecco’s modified Eagle’s medium (DMEM) Fluorobrite supplemented with 1× Glutamax, 1× sodium pyruvate, 15 mmol/L HEPES, and 1× nonessential amino acid (ThermoFisher). Then, HIEs were treated with DMEM Fluorobrite control or 10 ng/mL recombinant human IL-22 (Peprotech number 200-22) in DMEM Fluorobrite. While in the appropriate treatments, monolayers were scratched using a 10-μL pipette and then placed in an Okolabs stage-top incubation chamber at 37°C with CO₂ and humidity control. The stage was then placed on a Nikon TiE inverted widefield epifluorescence microscope (Nikon) with a motorized X, Y, and Z stage for multiposition live imaging. Videos and images were obtained with widefield epifluorescence using a 20× Plan Apo (numerical aperture, 0.75) differential interference contrast objective, using a SPECTRA X LED light source (Lumencor, Beaverton, OR) and an ORCA-Flash 4.0 sCMOS camera (Hamamatsu, Shizuoka, Japan). Images at various time points were generated using Nikon Elements software version 4.5.

In a similar approach, mouse colonic CMT93 cells were grown as monolayers in 10-well CELLview chamber slides (Greiner Bio-One number 543979). Monolayers were divided into five groups (two wells per group): i) control (no scratch), ii) scratch with DMEM, iii) scratch with 25% B. ovatus CDMM conditioned media, iv) scratch with supernatant from SPF mouse bone marrow–derived dendritic cells, or v) scratch with supernatant from SPF mouse bone marrow–derived dendritic cells treated with 25% B. ovatus CDMM conditioned media. CMT93 monolayers were scratched using a 10-μL pipette and incubated at 37°C, 5% CO₂. Monolayers were imaged every 6 hours for 24 hours on a Nikon TiE inverted widefield epifluorescence microscope (Nikon). Images were obtained with widefield epifluorescence using a 20× Plan Apo (numerical aperture, 0.75) differential interference contrast objective, using a SPECTRA X LED light source. For wound-healing analysis of HIEs and CMT93 cells, FIJI software (NIH, Bethesda, MD; https://imagej.net/fiji) was used to define the wound edges (as denoted by actin), and the wound diameter was recorded in three regions per well, with two to three wells per slide, and repeated three independent times.

Metabolite Analysis of Mouse Intestinal Contents and B. ovatus Cultures

Chemicals, Reagents, and Durable Supplies

Optima LC/MS-grade water, acetonitrile (ACN), and formic acid were obtained from Thermo Fisher Scientific (Waltham, MA). Ion chromatography–grade heptafluorobutyric acid was from Millipore-Sigma (Billerica, MA). Authentic analytical reference standards for tryptophan, 5-hydroxytryptophan, serotonin, N-acetylserotonin, and melatonin were all from Thermo Fisher Scientific; and indole-3-acetic acid, 5-hydroxyindoleacetic acid, and tryptamine were obtained from Millipore-Sigma. Authentic deuterated internal standard reference materials for d5- tryptophan, d4-serotonin, d4-melatonin, and d5–5-hydroxyindoleacetic acid were all purchased from CDN isotopes (Pointe-Claire, QC, Canada). Chromatographic separations were performed using a Luna C18 (2) (150-mm length × 1-mm internal diameter; 3 μm; 100-Å pore size) analytical column equipped with a C18 SecurityGuard (4-mm length × 2-mm internal diameter) cartridge from Phenomenex (Torrance, CA).

LC-MS/MS Equipment

The liquid chromatography–tandem mass spectrometry (LC-MS/MS) system was composed of a Shimadzu Nexera X2 MP high-performance liquid chromatography system (Kyoto, Japan) coupled to a Sciex 6500 QTrap hybrid triple-quadrupole/linear ion trap MS system from Danaher (Washington, DC). Operational control of the LC-MS/MS was performed with Analyst version 1.6.2 (SciEx, Framingham, MA), and quantitative analysis was performed using MultiQuant version 3.0.1 (SciEx).

Critical Solution Preparations and Sample Preparation

Stock solutions for each authentic analytical and internal standard reference compound were prepared at concentrations of 10.0 mg/mL each in water. A working internal standard (WIS) A solution was prepared at concentrations of 500 ng/mL each for d5- tryptophan, d4-serotonin, and d4-melatonin, and 1500 ng/mL for d5–5-hydroxyindoleacetic acid in water; this solution is used in the preparation of microbiome samples. A WIS-B solution was prepared by diluting 9 mL of WIS-A solution with 1 mL of water to accommodate the 10% dilution of the internal standard compounds during sample preparation; this solution was used in the preparation of the combined intermediate solution and calibration standards. A combined intermediate for the nondeuterated analytes was prepared at a concentration of 100 μg/mL each in WIS-B solution. Calibration standards were prepared by serial dilutions at concentrations of 1000, 250, 62.5, 15.6, 3.90, and 0.977 ng/mL using the WIS-B solution as the diluent.

Microbiome samples analyzed in this study included blank control and B. ovatus–conditioned CDMM media, and the cecal contents and feces collected from germ-free and B. ovatus–treated mice. For media sample preparation, a 10-μL volume of filter-sterilized media was diluted in a 90-μL volume of the WIS-A solution directly in autosampler vials. Cecal metabolites were extracted by the addition of a known mass of wet cecal material to a 500-μL volume of ice-cold methanol, followed by homogenization on an MP Fastprep Classic Instrument (MP Biomedical, Santa Ana, CA) of vortex mixing to disperse the solids. Fecal metabolites were extracted by the addition of a known mass of wet fecal material to a 10 μL/mg volume equivalent of an ice-cold ACN/water (1:1 v/v) solution, followed by 5 minutes of vortex mixing and brief sonication to disperse the solid fecal pellet. After solid dispersion, the fecal and cecal samples were centrifuged at 17,000 × g for 5 minutes, and a 10-μL volume of the extracted supernatants was diluted in a 90-μL volume of the WIS-A solution directly in autosampler vials. A 5-μL volume of each sample was injected onto the LC-MS/MS system.

LC-MS/MS Method

Ion-pairing reverse-phase chromatography was used for this separation, and includes a mobile phase A solution composed of water/ACN/formic acid/heptafluorobutyric acid (99.3:0.5:0.1:0.1 v/v/v/v), a mobile phase B solvent consisting of 100% ACN, and a needle wash composed of ACN/water (1:1 v/v). The chromatographic method included column heating at 50°C, autosampler tray chilling at 10°C, a mobile phase flow rate of 0.100 mL/minute, and a gradient elution program specified as follows: 0 to 9 minutes, 5% to 60% mobile phase B; 9 to 9.1 minutes, 60% to 5% mobile phase B; and 9.1 to 12 minutes, 5% mobile phase B. The 6500 QTrap divert valve (SciEx) was enabled and specified to divert the chromatographic eluate, according to the following program: initial position, divert to waste (position B); 0 to 1 minute, divert to waste (position B); and 1 to 12 minutes, divert to MS source (position A). The overall cycle time for a single injection was approximately 12.4 minutes, and the operational back pressure for the chromatographic system was approximately 2000 psi at initial conditions.

The TurboIonSpray electrospray ionization probe was installed in the Turbo V ion source, and was operated with the following source conditions: ionization mode polarity, positive; curtain gas, 20; TurboIonSpray voltage (internal standard), 5000 V; source temperature, 200°C; ion source gas 1 (nebulization gas), 25 psi; and ion source gas 2 (heater gas), 25 psi. MS/MS operational parameters include the following: collisionally activated dissociation gas pressure, HIGH; Q1 and Q3 quadrupole resolutions were set to unit/unit; and the molecule-specific instrument parameters are included in Table 1.

Table 1.

Molecule-Specific MS/MS Parameters for the Targeted Metabolomics Method

Metabolite	Q1, m/z	Q3, m/z∗	Declustering potential, V	Entrance potential, V	Collison energy, eV	Collision-cell exit potential, V
Tryptophan	205.1	188.1/146.1	35	7	15/25	15
D5-tryptophan (IS)	210.1	192.1/150.1	35	7	15/25	15
Serotonin	177.1	160.1/115.1	20	8	17/38	10
D4-serotonin (IS)	181.1	164.1/119.1	20	8	17/38	10
Melatonin	233.1	174.1/159.1	45	6	22/39	15
D4-melatonin (IS)	237.1	178.1/163.1	45	6	22/39	15
5-Hydroxy-indoleacetic acid	192.1	146.1/118.1	70	8	23/40	17
D5–5-hydroxy-indoleacetic acid (IS)	197.1	150.1/122.1	70	8	23/40	17
5-Hydroxy-tryptophan	221.1	204.1/162.1	40	7	16/26	23
N-acetylserotonin	219.1	160.1/115.1	40	8	23/49	10
Tryptamine	161.1	144.1/115.1	30	9	18/49	20
Indoleacetic acid	176.1	130.1/77.0	70	4	25/56	11

IS, internal standard; MS/MS, tandem mass spectrometry; m/z, mass/charge.

^∗The ratio corresponds to the m/z of the quantifying fragment ion, over the m/z of the qualifying fragment ion.

Statistical Analysis

Comparisons between groups were made with t-test, one-way analysis of variance, or repeated-measures analysis of variance, using the Holm-Sidak post-hoc test. Graphs and statistics were performed using GraphPad (GraphPad Software, Inc., La Jolla, CA). P < 0.05 was considered statistically significant.

Results

B. ovatus ATCC 8483 Modulates CD11c⁺ Immune Cell Populations and Intestinal IL-22 Production in a Gnotobiotic Mouse Model

Bacteroides species influence the immune system and dampen intestinal inflammation.^8,12,26,27 B. ovatus ATCC 8483 treatment reduces colitis severity in a DSS model of colitis.¹¹ However, the mechanism(s) behind how B. ovatus reduced inflammation remain(s) unclear. To address this, germ-free C57BL/6 mice were mono-associated with B. ovatus ATCC 8483 by a single oral gavage and compared with germ-free controls. Three weeks after colonization, the ileum and colon of germ-free mice and B. ovatus mono-associated mice were examined by H&E staining (Supplemental Figure 1). On the basis of H&E staining, the ileum and colon appear histologically similar in germ-free and B. ovatus mono-associated mice, with no observed differences in villi height or crypt depth (Supplemental Figure S1). Similarly, no changes were noted in mouse weight between groups (germ-free, 30.9 ± 2.5 g; B. ovatus, 31.1 ± 2.1 g).

Next, the effects of B. ovatus mono-association on immune cells were examined. Flow cytometry was used to analyze immune cell composition of the colonic mucosa (Figure 1). Bacteroides ovatus ATCC 8483 colonized mice harbored increased populations of CD11c⁺ immune cells compared with germ-free controls (Figure 1, A and B). CD11c is a membrane protein that is highly expressed on most human dendritic cells, monocytes, and macrophages. CD11c⁺ cells respond to stimuli, as well as recruit and activate T cells.^28,29 CD11b⁺/CD11c⁺ (Figure 1, A and C) and CD103⁺/CD11c⁺ populations in B. ovatus mice were larger compared with those of germ-free controls (Figure 1, A and D). Interestingly, no differences were observed in CD4⁺CD8⁺ or CD4⁺Fox3p⁺ T-cell populations (data not shown).

Figure 1.

Immune profiling of germ-free and Bacteroides ovatus ATCC 8483 mono-associated mouse colon. A: Representative images of flow cytometry gating for immune cells isolated from the colons of germ-free mice that received phosphate-buffered saline or B. ovatus. Gating was performed with CD19⁺, CD11b⁺, CD103⁺, and CD11c⁺ markers. Immune cells were gated for dendritic cell (DC)/macrophage/monocyte (CD11c⁺, CD11b⁺, and CD103⁺) populations. B–D: Quantitation of the percentage of CD11c⁺ immune cell populations (B), CD11b⁺/CD11c⁺ populations (C), and CD103⁺/CD11c⁺ populations (D) in the colons of germ-free and B. ovatus treated mice. n = 5 mice per group (B–D). ∗P < 0.05 (t-test).

qPCR was used to further assess protective epithelial and immune responses. Expression of barrier function-related genes (mucins and tight junctions) and cytokine genes was analyzed in colonic tissue of germ-free and B. ovatus mono-associated mice (Figure 2A). No differences were observed in secreted mucin (MUC2), adherent mucin (MUC1 and MUC3), tight junction (claudin1,2,3,4,6, occludin, and ZO-1), or adherin (Ecadherin) gene expression between germ-free and B. ovatus mono-associated mice. Similarly, there were no differences in the expression of cytokines IL-12, IL-13, IL-33, interferon-γ, IL-6, keratinocytes-derived chemokine (KC; IL-8), or tumor necrosis factor. However, there was an approximately 30-fold increased expression of colonic IL-22 in B. ovatus mono-associated mice compared with germ-free mice (Figure 2A). IL-22 is primarily generated by dendritic cells, T cells, and innate lymphoid cells and has been associated with epithelial homeostasis and alleviation of inflammation.30, 31, 32, 33, 34, 35, 36, 37 The flow cytometry and qPCR data support the immunomodulatory role of B. ovatus in the gastrointestinal tract.

Figure 2.

Bacteroides ovatus ATCC 8483 secretes indole-3-acetic acid and up-regulates IL-22. A: Quantitative real-time PCR analysis of epithelial and immune-related genes in germ-free and B. ovatus mono-associated mice. B: Tryptophan pathway illustrating downstream products. C: Liquid chromatography–tandem mass spectrometry (LC-MS/MS) analysis of B. ovatus conditioned chemically defined minimal media (CDMM) for tryptophan metabolites (ng/mL) depicted as heat maps. D: LC-MS/MS analysis of cecal and fecal contents from germ-free and B. ovatus mono-associated mice for tryptophan metabolites (ng/mg wet weight). Concentrations of metabolites are indicated in gradations of green, with black indicating 0 ng/mL or 0 ng/mg wet weight. n = 6 mice per group (A and D). ∗P < 0.05 (analysis of variance). 5-HIAA, 5-hydroxyindoleacetic acid; 5-HT, 5-hydroxytryptamine; 5-HTP, 5-hydroxytryptophan.

B. ovatus Secretes Indole-3-Acetic Acid, a Known Driver of IL-22 Production

The B. ovatus ATCC 8483 genome contains the tnaA1 gene (tryptophanase; EC:4.1.99.1) that can convert tryptophan into indole. Indoles have been shown to activate aryl hydrocarbon receptors (AhRs) on immune cells and stimulate IL-22 secretion.38, 39, 40, 41 Thus, LC-MS/MS was used to examine B. ovatus–derived tryptophan pathway metabolites in a fully defined media (CDMM) (Figure 2B). Tryptophan and tryptamine concentrations increased above the CDMM medium baseline in B. ovatus conditioned media (Figure 2C). Moreover, indole-3-acetic acid (48.1 ± 11.6 ng/mL) in B. ovatus cultured supernatant, indicated that B. ovatus can generate indole-3-acetic acid in vitro. Metabolic profiles of cecal contents and feces from germ-free and B. ovatus mono-associated mice by LC-MS/MS were compared to determine if B. ovatus ATCC 8483 could generate these metabolites in vivo (Figure 2D). A decrease in tryptophan in samples from B. ovatus mono-associated mice compared with germ-free controls indicated consumption of tryptophan by B. ovatus in the gut. There were no statistically significant differences in other tryptophan metabolites, such as serotonin, melatonin, tryptamine, or 5-hydroxy-indoleacetic acid. However, there was an approximately fivefold increase in indole-3-acetic acid concentrations in the cecum and an approximately fourfold increase of indole-3-acetic acid in feces of mice mono-associated with B. ovatus compared with germ-free controls (Figure 2D). These findings suggest that B. ovatus can metabolize tryptophan to produce indole-3-acetic acid, which has been shown to elicit IL-22 production.⁴¹

B. ovatus Stimulates Dendritic Cell Maturation and IL-22 Production

Dendritic cells harbor AhRs that can recognize microbially produced indole metabolites and generate IL-22.^36,42,43 To address whether metabolites produced by B. ovatus ATCC 8483 are able to modulate dendritic cells, bone marrow–derived dendritic cells were isolated from germ-free mice and SPF conventional mice and differentiated for 6 days with IL-4 and granulocyte-macrophage colony-stimulating factor. Immature bone marrow–derived dendritic cells were stimulated for 16 hours with either B. ovatus ATCC 8483 CDMM-conditioned media, which contained indole-3-acetic-acid (Figure 2C), or the known maturation factor for dendritic cells, E. coli LPS. Quantitative protein data revealed that lipopolysaccharide-supplemented media and B. ovatus–conditioned CDMM media resulted in dendritic cell maturation, as measured by the increased release of granulocyte-macrophage colony-stimulating factor, interferon-γ, IL-1α, IL-1β, IL-6, tumor necrosis factor, and IL-10 (Figure 3). This effect was observed in bone marrow–derived dendritic cells isolated from germ-free mice (Figure 3A), as well as SPF conventionalized mice (Figure 3B). Moreover, application of B. ovatus–conditioned CDMM media to bone marrow–derived dendritic cells from SPF conventional mice resulted in enhanced release of IL-22 (Figure 3C). These data indicate that B. ovatus metabolites induce dendritic cell maturation and promote IL-22 secretion.

Figure 3.

Modulation of mouse bone marrow–derived dendritic cells by Bacteroides ovatus ATCC 8483 metabolites. A: Multiplexed cytokine analysis of germ-free derived bone marrow–derived dendritic cell supernatant after overnight incubation with media (RPMI 1640), 100 ng/mL Escherichia coli lipopolysaccharide (LPS), or 25% B. ovatus chemically defined minimal media (CDMM) conditioned media (B. ovatus). B: Multiplexed cytokine analysis of specific pathogen-free (SPF) derived bone marrow dendritic cell supernatants after overnight incubation with media (RPMI 1640), 100 ng/mL E. coli LPS (LPS) in RPMI 1640, or 25% B. ovatus CDMM-conditioned media (B. ovatus) in RPMI 1640. C: Analysis of IL-22 (pg/mL) in SPF dendritic cells treated with media (RPMI 1460) or B. ovatus CDMM conditioned media (B. ovatus) by enzyme-linked immunosorbent assay. n = 4 replicates, repeated in 2 independent experiments (A and B); n = 5 replicates, repeated in 2 independent experiments (C). ∗P < 0.05 (analysis of variance). GM-CSF, granulocyte-macrophage colony-stimulating factor; IFN-γ, interferon-γ; TNF, tumor necrosis factor.

IL-22 Enhances Repair within the Intestinal Epithelium

IL-22 promotes wound repair in multiple models.30, 31, 32, 33, 34, 35, 36, 37 To confirm the ability of IL-22 to drive repair in the intestinal epithelium, HIEs (alias intestinal organoids) were transduced to stably express a fluorescent actin label (LifeAct) (Figure 4). HIE two-dimensional monolayers were scratched to generate a wound in the presence or absence of recombinant human IL-22 (10 ng/mL) and imaged over time to assess repair. Live cell imaging revealed that media control HIEs reached >50% wound closure at 9.5 hours after wounding (Figure 4, A and B). In contrast, IL-22-treated HIEs reached >50% wound closure at 5.5 hours, indicating a significant enhancement in wound closure consistent with area under the curve analysis (Figure 4C). In a complementary approach, mouse colonic CMT93 cells were used for wound repair analysis. CMT93 cells were treated with DMEM (control), DMEM containing 25% B. ovatus–conditioned CDMM media, supernatant from SPF bone marrow–derived dendritic cells, or supernatant from SPF bone marrow–derived dendritic cells stimulated with 25% B. ovatus conditioned CDMM media (Figure 5). Control DMEM and 25% B. ovatus conditioned media-treated CMT93 cells exhibited a similar repair rate (Figure 5, A and B), suggesting that B. ovatus metabolites did not enhance wound repair directly through the epithelium. Addition of supernatant from SPF bone marrow–derived dendritic cells and to a greater degree, supernatant from SPF bone marrow–derived dendritic cells stimulated with 25% B. ovatus conditioned media, which harbors IL-22, increased the rate of wound repair (Figure 5, A and B). There was a complete wound closure of CMT93 cells after 24 hours of treatment with supernatant from SPF bone marrow–derived dendritic cells stimulated with 25% B. ovatus conditioned media (Figure 5C). Thus, IL-22 from immune cells may contribute to wound repair and may mediate the enhanced wound healing observed in DSS colitis mice treated with B. ovatus.

Figure 4.

IL-22 promotes epithelial repair in human intestinal enteroids (HIEs). HIEs were transduced with the LifeActRuby sensor to label F-actin (red). Cell migration was visualized by live cell microscopy on a Nikon TiE with 20× Plan Apo (numerical aperture, 0.75) differential interference contrast objective, using a SPECTRA X LED light source and ORCA-Flash 4.0 sCMOS camera. A: Representative images of LifeActRuby HIEs over time (0 and 12 hours) after scratching and exposure to media control or 100 ng/mL recombinant IL-22. Dotted black sides highlight wound edge. B: FIJI software was used to calculate wound distance over time. C: Area under the curve calculations of media control and IL-22 treated HIE wound repair. n = 3 replicates per experiment, repeated 3 independent times (A–C). ∗P < 0.05 (one-way analysis of variance). Scale bars = 100 μm (A).

Figure 5.

Supernatant from specific pathogen-free (SPF) bone marrow–derived dendritic cells (DCs) treated with Bacteroides ovatus ATCC 8483 conditioned medium (CM) promotes epithelial repair in mouse colon CMT93 cells. Mouse colonic CMT93 cells were treated with media (control), 25% B. ovatus conditioned media, supernatant from dendritic cells treated with media, or supernatant from dendritic cells treated with 25% B. ovatus conditioned media. Cell migration of CMT93 cells was visualized by microscopy every 6 hours for 24 hours on a Nikon TiE with 20× Plan Apo (numerical aperture, 0.75) differential interference contrast objective, using a SPECTRA X LED light source and ORCA-Flash 4.0 sCMOS camera. A: FIJI software was used to calculate CMT93 wound distance over time. B: Area under the curve calculations of CMT93 wound repair. C: Representative images of CMT93 cells treated with supernatant from SPF bone marrow–derived dendritic cells stimulated with B. ovatus CM at time 0 and 24 hours. Dotted black sides highlight wound edge. n = 2 replicates per experiment, repeated 3 independent times (A–C). ∗P < 0.05 versus control [repeated-measures analysis of variance (ANOVA)]; ^†P < 0.05 (one-way ANOVA). Scale bars = 100 μm (C).

B. ovatus ATCC 8483 Reduces Intestinal Inflammation and Ameliorates TNBS-Induced Colitis

TNBS causes acute colitis with similarities to IBD in BALB/c mice.⁴⁴ To determine if administration of B. ovatus could ameliorate TNBS-associated colitis, BALB/c mice were administered B. ovatus or PBS (as a control) by oral gavage daily for 7 days, followed by rectal administration of TNBS. Within the TNBS model, B. ovatus–treated mice exhibited improved histology by H&E staining, with retention of goblet cells, maintenance of crypt architecture, and decreased edema in the lamina propria compared with PBS-treated mice (Figure 6A). These findings mirrored the reduced histologic scores, which demonstrated decreased severity of gut inflammation in B. ovatus–treated mice compared with PBS-treated mice (Figure 6B). Consistent with the histologic scores, B. ovatus treatment decreased serum proinflammatory cytokines KC (IL-8), monocyte chemoattractant protein-1 (MCP-1), tumor necrosis factor (TNF), IL-6, and IL-1α compared with PBS-treated animals, indicating reductions in cytokines contributing to intestinal inflammation (Figure 6C). Similar to the gnotobiotic mouse model, an increased colonic Il-22 expression was observed in the colons of B. ovatus–treated animals compared with PBS-treated mice (Figure 6D).

Figure 6.

Bacteroides ovatus ATCC 8483 up-regulates IL-22 and ameliorates inflammation in a trinitrobenzene sulfonic acid (TNBS) model of colitis. A: Representative hematoxylin and eosin stains of mouse colon 3 days following TNBS administration in mice treated with either phosphate-buffered saline (PBS; control) or 2 × 10⁸ colony-forming units B. ovatus. B: Histologic scores of TNBS mouse colon from PBS or B. ovatus treated mice, as assessed by a blinded board-certified pathologist (D.A.S.). C: Heat maps of multiplexed cytokine analysis of sera from TNBS mice treated with PBS or B. ovatus. Proinflammatory cytokines keratinocytes-derived chemokine (KC, homolog of human IL-8), monocyte chemoattractant protein-1 (MCP-1), tumor necrosis factor (TNF), IL-6, IL-1α, IL-1β, and interferon (IFN)-γ (pg/mL) are depicted as heat maps. Concentrations of metabolites are indicated in gradations of green, with black indicating 0 pg/mL. D: Quantitative real-time PCR analysis of IL-22 from TNBS mouse colons of mice treated with PBS or B. ovatus. n = 8 mice per group (B–D). ∗P < 0.05 (t-test). Scale bars = 100 μm (A).

Finally, flow cytometry was used to analyze colonic immune cells in mice treated with PBS or B. ovatus, in the absence or presence of TNBS colitis (Figure 7). In SPF mice receiving B. ovatus, no significant shifts were observed in immune cell populations compared with SPF mice receiving PBS (Figure 7, A and B), probably due to the presence of a complex community of microbes and the absence of inflammation. However, in mice receiving TNBS, B. ovatus shifted colonic immune cells toward a greater abundance of CD103⁺/CD11c⁺ populations compared with PBS-treated TNBS mice (Figure 7C). CD103⁺/CD11c⁺ immune cells are important for maintaining intestinal immune homeostasis and inducing tolerogenic immune responses.⁴⁵ Once again, no changes were observed in CD4⁺/Foxp3⁺ populations (Figure 7D). These findings expand the characterization of B. ovatus by demonstrating its immunomodulatory effects and potentially identifying a protective mechanism by which B. ovatus aids the host.

Figure 7.

Immune profiling of untreated and trinitrobenzene sulfonic acid (TNBS)–treated mouse colon supplemented with phosphate-buffered saline (PBS) or Bacteroides ovatus ATCC 8483. A: Representative images of flow cytometry gating for immune cells isolated from the colons of untreated mice that received PBS or B. ovatus or TNBS-treated mice that received PBS or B. ovatus (1 day after TNBS). Immune cells were gated for T-cell (CD4⁺, CD8⁺, and forkhead box P3 (Foxp3)⁺CD4⁺ T-regulatory) and dendritic cell (DC)/macrophage/monocyte (CD11c⁺, CD11b⁺, and CD103⁺) populations. B–D: Quantitation of the colonic CD11b⁺/CD11c⁺ populations (B), CD103⁺/CD11c⁺ populations (C), and Fox3p⁺/CD4⁺ populations (D). n = 5 mice per group (A–D). ∗P < 0.05 (analysis of variance).

Discussion

These data demonstrate that members of the phylum Bacteroidetes, and specifically B. ovatus ATCC 8384, can generate indole-3-acetic acid, influence CD11b⁺/CD11c⁺ and CD103⁺/CD11c⁺ immune populations, and stimulate IL-22 in vitro and in vivo. IL-22 promoted wound repair in HIE monolayers. In the TNBS model of colitis, B. ovatus shifted the CD103⁺/CD11c⁺ immune populations and enhanced colonic IL-22 production, which likely promoted epithelial integrity and dampened inflammation B. ovatus beneficially modulated the inflamed intestinal tract. Before this work, B. ovatus' therapeutic potential had not been thoroughly investigated. In this study, we defined mechanism(s) behind the therapeutic effects of B. ovatus and possibly other commensal Bacteroides species. This work provides the first evidence for B. ovatus' direct effects on the IL-22 pathway to suppress colitis and enhance wound repair (Figure 8).

Figure 8.

Proposed model for Bacteroides ovatus immunomodulation. B. ovatus generates indole-3-acetic acid, which then binds to aryl hydrocarbon receptors (AhRs) on dendritic cells and drives IL-22 production. IL-22, in turn, promotes epithelial repair and alleviates inflammation, thereby promoting intestinal homeostasis.

IL-22 is a cytokine produced by immune cells, and plays an important role in regulation of intestinal inflammation in the context of IBD.⁴⁶ IL-22 receptor activation induces the expression of genes involved in immunosurveillance, epithelial barrier function, inflammation, and homeostasis.47, 48, 49, 50, 51 IL-22–mediated STAT3 activation leads to colitis suppression and epithelial repair in murine models of IBD.^36,52 We demonstrated up-regulation of colonic IL-22 in B. ovatus mono-associated mice compared with germ-free mice. We also showed that B. ovatus treatment during TNBS colitis leads to increased expression of colonic IL-22. These data suggest that B. ovatus–mediated protection of the intestinal mucosa may involve immune cell-driven up-regulation of the IL-22–STAT3 pathway, with resultant colitis suppression and enhanced epithelial wound repair.

Although this work focused on dendritic cells, it is possible that IL-22 may be produced by multiple cell types in vivo. CD4⁺ T cells,^53,54 natural killer cells,54, 55, 56, 57 type 17 helper T cells,^55,57 neutrophils,^34,58 and T-cell receptor (TCR)γδ T cells⁵⁹ have all been shown to secrete IL-22. Interestingly, in previous DSS mouse model experiments, dendritic cells,³⁶ neutrophils,³⁴ and TCRγδ T cells⁵⁹ appear to be the major sources of IL-22. Dendritic cells have also been suggested as a major source of IL-22 in an infectious (Citrobacter rodentium) model of colitis.⁵¹ In addition to these other cell types, multiple groups have also identified RAR-related orphan receptorγt–dependent innate lymphoid cell (ILC)3 as a major source of IL-22 in vivo.60, 61, 62 Although ILCs only constitute approximately 5% of lymphocytes in human colon and small intestines^56,63 and 1% to 2% of cells in mouse colon and small intestines,^64,65 ILC production of IL-22 is considered essential for gut homeostasis. ILCs express AhR,^62,66, 67, 68 and microbial tryptophan metabolites, including indole, can activate ILC3s to produce IL-22.^38,69,70 These studies suggest that ILC3 could also be a major contributor to IL-22 production in our animal model, and we are interested in pursuing this question in future studies.

Tryptophanases are prominent in Bacteroides species,^71,72 and multiple groups have identified Bacteroides-produced indole.72, 73, 74 On the basis of the prevalence of tryptophanases, it has been speculated that Bacteroides species are prominent producers of indole in the gut. The ability of indole-producing B. ovatus to stimulate immune cells to generate IL-22 is consistent with another study that identified that B. fragilis, another indole-producing Bacteroides species, could stimulate peripheral blood mononuclear cells to secrete IL-22.⁷⁵ Although Bacteroides species are known to produce indole which can activate AhRs on immune cells to generate IL-22, there have been no definitive studies linking Bacteroides, indole production, and IL-22 secretion. This gap could be due to prior studies focusing on other cytokines and aspects of immunity. Herein, we provide a novel link between the intestinal microbe B. ovatus, the production of the microbial metabolite indole-3-acetic-acid, and the generation of IL-22 by innate immune cells in the intestinal mucosa. We speculate that indole-3-acetic acid production by other Bacteroides species may also activate IL-22 and diminish colitis.

Depletion of tryptophan and increased levels of downstream tryptophan metabolites, such as indole, can confer protection against inflammation.⁷⁶ In our germ-free mice, B. ovatus mono-association correlated with decreased tryptophan levels and increased indole-3-acetic acid production. It is possible that both decreased tryptophan and increased indole contribute to the anti-inflammatory effects observed in B. ovatus–treated mice receiving TNBS. Further experiments are warranted to identify whether B. ovatus–produced indole-3-acetic acid alone drives IL-22 production.

In contrast to Bacteroides, many bacterial species are not capable of producing indole compounds. Another human gut commensal, Bifidobacterium dentium, does not harbor the enzymatic machinery to produce indole-3-acetic acid (Kyoto Encyclopedia of Genes and Genomes). Consistent with the genome, analysis of B. dentium metabolites in a fully defined media by LC-MS/MS revealed no detectable levels of indole-3-acetic acid (data not shown). In vivo, B. dentium mono-associated feces had comparable indole-3-acetic acid levels with germ-free mice, and colonic IL-22 gene expression was similar between B. dentium mono-associated mice and germ-free controls.⁷⁷ These findings indirectly provide support for the notion that commensal indole-3-acetic acid can drive IL-22 production.

In addition to indole, Bacteroides species are known to secrete other immunomodulatory metabolites and subcellular structures. For example, B. fragilis secretes polysaccharide A packaged in outer membrane vesicles, which can then be delivered to dendritic cells, leading to enhanced numbers of regulatory T cells and anti-inflammatory cytokine production.78, 79, 80 Likewise, B. thetaiotaomicron– and B. vulgatus–derived outer membrane vesicles regulate dendritic cell responses.^81,82 B. ovatus mono-associated animals had increased abundances of CD11c⁺ populations, which include dendritic cells. Thus, similar to other strains of Bacteroides, B. ovatus could secrete outer membrane vesicles, in addition to indole-3-acetic acid, synergistically contributing to immune modulation. The genus Bacteroides represents a prominent group of commensal bacteria in the human gastrointestinal tract. Several Bacteroides species may possess beneficial and symbiotic features benefitting the human host, including the ability to suppress colitis. Bacteroides thetaiotaomicron can suppress inflammation in multiple models of murine colitis.⁷ Similarly, B. fragilis can ameliorate DSS colitis in mice by inhibiting the expression of proinflammatory cytokines.⁸ However, some Bacteroides species can assume pathogenic roles in the human host as well. Bacteroides thetaiotaomicron has been directly implicated in bacteremia-related mortality, and B. fragilis has potent virulence factors that can contribute to tissue destruction and abscess formation in the human host.¹² More importantly, these virulence factors appear to be absent in B. ovatus, making this commensal microbe an ideal candidate as a next-generation probiotic.

We have shown in two different mouse colitis models (DSS¹¹ and TNBS) that B. ovatus reduces inflammation. However, the exact role of B. ovatus–induced IL-22 in the setting of colitis is unclear. Future studies using IL-22 knockout, STAT-3 knockout, and AhR knockout mice would help identify the specific roles of indole-3-acetic acid and IL-22 and mechanisms of intestinal disease modulation in colitis models. On the basis of existing data, we propose that B. ovatus could be an attractive candidate as a next-generation probiotic for the suppression of inflammation, particularly in the setting of IBD. Many patients are afflicted with IBD refractory to current standard therapies. Although probiotics have been studied for the treatment of IBD, these studies have yielded conflicting results. Some studies indicate probiotics can promote resolution of inflammation, whereas other studies showed no benefit. However, these studies have been mainly limited to strains of Lactobacillus, Bifidobacterium, E. coli, Streptococcus, and Saccharomyces.83, 84, 85, 86, 87 Bacteroidetes is not currently used as a source for probiotics, despite being a dominant bacterial phylum in the healthy human intestine. At present, B. ovatus has not been studied as an adjuvant treatment option for IBD. This work supports the potential for B. ovatus (and possibly other Bacteroides species) to be a therapeutic microbe for IBD based on its capacity to beneficially modulate the intestinal immune system.

Author Contributions

F.D.I. and J.V. developed the concept and design; F.D.I., M.A.E., W.R., A.H., J.M.H., and J.V. made intellectual contributions; F.D.I., M.A.E., W.R., K.A.E., Z.S., A.C.G., J.F., E.S.P., S.V., T.D.H., S.J.H., A.M.H., D.A.S., J.K.S., and Y.L. performed data acquisition; F.D.I., M.A.E., W.R., T.D.H., S.J.H., J.K.S., and Y.L. performed data analysis, statistical analysis, and interpretation; D.A.S. performed histologic scoring; F.D.I., M.A.E., W.R., K.A.E., Z.S., A.C.G., S.V., T.D.H., S.J.H., A.M.H., D.A.S., J.M.H., J.K.S., and J.V. performed drafting and editing manuscript; F.D.I., J.M.H., and J.V. obtained funding.

Supplemental Data

Supplemental Figure S1.

Bacteroides ovatus ATCC 8483 colonization does not influence intestinal morphology. A: Representative images of hematoxylin and eosin stains of the ileum and colon in germ-free mice that received phosphate-buffered saline vehicle control or B. ovatus mono-associated mice (2 × 10⁸B. ovatus). B: Quantitation of villi length and crypt depth in the ileum of germ-free and B. ovatus treated mice, as assessed by FIJI software; analysis of variance was performed. C: Quantitation of crypt depth in the colon of germ-free and B. ovatus treated mice, as assessed by FIJI software; t-test was performed. n = 6 mice per group (A–C). Scale bars = 100 μm (A).