Abstract
Loss of function in the NOD2 gene is associated with a higher risk of developing Crohn’s disease (CD). CD is characterized by activation of T cells and activated T cells are involved in mucosal inflammation and mucosal damage. We found that acute T cell activation with anti-CD3 mAb induced stronger small intestinal mucosal damage in NOD2−/− mice compared with wild-type mice. This enhanced mucosal damage was characterized by loss of crypt architecture, increased epithelial cell apoptosis, delayed epithelial regeneration and an accumulation of inflammatory cytokines and Th17 cells in the small intestine. Partial microbiota depletion with antibiotics did not decrease mucosal damage 1 d after anti-CD3 mAb injection, but it significantly reduced crypt damage and inflammatory cytokine secretion in NOD2−/− mice 3 d after anti-CD3 mAb injection, indicating that microbial sensing by Nod2 was important to control mucosal damage and epithelial regeneration after anti-CD3 mAb injection. To determine which cells play a key role in microbial sensing and regulation of mucosal damage, we engineered mice carrying a cell-specific deletion of Nod2 in villin and Lyz2-expressing cells. T cell activation did not worsen crypt damage in mice carrying either cell-specific deletion of Nod2 compared with wild-type mice. However, increased numbers of apoptotic epithelial cells and higher expression of TNF-α and IL-22 were observed in mice carrying a deletion of Nod2 in Lyz2-expressing cells. Taken together, our results demonstrate that microbial sensing by Nod2 is an important mechanism to regulate small intestinal mucosal damage following acute T cell activation.
This article is featured in In This Issue, p.1
Introduction
In the gastrointestinal tract, regulation of host–microbiome interactions is crucial to maintain intestinal homeostasis (1). Nod2, a cytosolic sensor of peptidoglycan, is one of the important regulators of this interaction. Nod2 is expressed in a variety of cell types from hematopoietic and nonhematopoietic origins, including myeloid and lymphoid cells, stromal cells, intestinal epithelial cells (IEC), and Paneth cells (2–6), yet it is unclear which Nod2-expressing cell is key for a balanced host–microbial relationship. After stimulation with Nod2-specific ligand muramyl dipeptide (MDP), Nod2 signaling regulates gene expression of a number of mediators thought to promote maintenance of intestinal barrier integrity, the activation of antimicrobial function and autophagy, and the regulation of immune homeostasis (7). More recently, a crucial role of Nod2 in intestinal epithelial stem cell survival was reported, highlighting the importance of Nod2 in intestinal epithelial regeneration (8).
Loss-of-function mutations in the NOD2 gene were the first defined genetic risk factor identified for Crohn’s disease (CD) (9, 10), and carriage of NOD2 risk alleles is associated with ileal disease location (11). Although the etiology of CD remains unclear, the high numbers of T cells in the inflamed intestinal mucosa, the secretion of large amounts of T cell–derived proinflammatory cytokines, and the requirement for T cells in various animal models of chronic intestinal inflammation strongly suggest a role of T cells in the pathogenesis of CD (12). Our group previously showed that, although Nod2 is expressed and functionally active in murine CD4+ T cell subsets, the expression of Nod2 is not required for the development or the prevention of the T cell transfer model of colitis in Rag-deficient mice (6). It remains possible that Nod2 modulates immune homeostasis in the small intestine in response to acute T cell activation.
In this study, we used a model of T cell–induced enteropathy triggered by i.p. injection with T cell activating monoclonal anti-CD3 Ab (anti-CD3 mAb) (13) and investigated the role of Nod2 on T cell activation, T cell–induced mucosal damage and recovery, specifically within the small intestine. Our results showed that acute T cell activation led to more severe crypt damage, increased crypt IEC apoptosis, and delayed epithelial regeneration in Nod2-deficient mice. Microbial sensing by Nod2 was important for the regulation of mucosal damage and epithelial regeneration. Deletion of Nod2 in intestinal epithelial cells and in Lyz2-expressing phagocytes was not sufficient to recapitulate the crypt damage observed in NOD2−/− mice. However, increased numbers of apoptotic epithelial cells and higher expression of TNF-α and IL-22 were observed in mice carrying a deletion of Nod2 in Lyz2-expressing cells, indicating that Nod2 expression in phagocytes can alter the immune response and the mucosal damage induced by anti-CD3 mAb injection.
Materials and Methods
Mice
C57BL/6, NOD2−/−, NOD2flox/flox, Villin-Cre, and LyzM-Cre mice were maintained under standard pathogen-free conditions at the University of Toronto animal facility. Wild-type (WT) and NOD2−/− mice were used as separate strains and were crossed to generate littermate control mice. NOD2−/− mice were obtained from Dr. J.-P. Hugot (Hôpital Robert Debré, Université Paris Diderot, Paris, France) (14), Villin-Cre (B6.Cg-Tg(Vil-cre)997Gum/J) and LyzM-Cre (B6.129P2-Lyz2tm1(cre)Ifo/J) mice were purchased from The Jackson Laboratory (Bar Harbor, ME), and NOD2flox/flox mice were obtained from Dr. P. Rosenstiel (Institute for Clinical Molecular Biology, Christian-Albrechts-University, Kiel, Germany) (15).
NOD2flox/flox mice were generated by genOway (Lyon, France). The loxP sites were inserted after exons 1 and 3 of the NOD2 gene in 129SV mouse embryonic stem cells. The distal loxP site was inserted together with a FRT-neomycin selection cassette. The engineered mouse line was bred with FLP deleter-mice allowing the deletion of the FRT-flanked neomycin selection cassette. The resultant NOD2flox/+ mice expressed two loxP sites in a genomic region including exons 2 and 3. Mice were backcrossed onto C57BL/6 background for 10 generations.
Mice expressing a conditional deletion of NOD2 in villin-expressing cells (NOD2ΔIEC) and in lysozyme 2–expressing cells (NOD2ΔLyz2) were obtained by crossing NOD2flox/flox mice and Villin-Cre or LyzM-Cre mice. Expression of the Cre recombinase resulted in the deletion of both exons 2 and 3.
All mouse experiments were conducted as approved by the University of Toronto animal care committee in accordance with the regulations of the Canadian Council on animal care (University of Toronto approved protocol number 20010966).
Material and reagents
The following Abs were used for the experiments: anti-CD3ε (145-2C11; BioLegend, San Diego, CA), anti–CD3ε-PE.Cy5 (145-2C11; eBioscience, San Diego, CA), anti–CD4-A780 (GK1.5; eBioscience), anti–Foxp3-PE (FJK-16s; eBioscience), anti–IL-17A-Brilliant Violet (TC11-18H10.1; eBioscience), anti–IFN-γ-FITC (XMG1.2; eBioscience), anti-CD11b (M1/70; eBioscience), anti–Gr-1 (RB6-8C5; eBioscience), brefeldin A, and Foxp3 Fixation/Permeabilization Kit (eBioscience). PMA ( P8139) and ionomycin (I0634) were purchased from Sigma-Aldrich (St. Louis, MO). Collagenase D was purchased from Roche Applied Science (Penzberg, Germany).
T cell–induced enteropathy
Mice received an i.p. injection of 50 μg monoclonal anti-CD3ε Ab (145-2C11; BioLegend). Mice were assessed for clinical signs of enteropathy, including activity level, appearance of the coat, perianal state and behavior. Mice were sacrificed at days 1, 3, and 5 postinjection.
Antibiotic treatment consisted of administration of ampicillin, neomycin, and metronidazole each at 1 g/l, in drinking water provided ad libitum from 10 d prior to anti-CD3 mAb injection until sacrifice, 1 and 3 d postinjection.
Histological examination
Small-bowel tissues were fixed in 10% buffered formalin and embedded in paraffin sections. H&E, TUNEL, and Ki67 staining were performed by the Pathology Core, Center for Phenogenomics (Toronto, ON, Canada). Small bowels were graded for severity of crypt damage by an experienced pathologist (C.S.) blinded to the experimental set-up. Each segment was graded on a scale from 1 (no change) to 4 (severe damage and loss of crypts).
Analysis of mRNA relative expression using quantitative real-time PCR
Total RNA was isolated from T cells using RNeasy Mini Kit (Qiagen, Germantown, MD). RNA samples were treated with DNAse Amp I Grade (Invitrogen) and cDNA were synthesized using Mu-MLV reverse transcriptase (Eurogentec, Liège, Belgium). For RT-PCR, cDNA samples were combined with primer/probe sets and Power SYBR Green Master Mix (Applied Biosystems, Warrington, U.K.), according to the manufacturer’s recommendations. Samples were normalized internally using the average cycle quantification of GUSB, HMBS, and TBP simultaneously. The primer sequences were as follows: GUSB sense 5′-AGC-CCT-TCG-GGA-CTT-TAT-TG-3′, GUSB antisense 5′-AAT-GGG-CAC-TGT-TGA-TCC-TC-3′, HMBS sense 5′-GTA-CCC-TGG-CAT-ACA-GTT-TG-3′, HMBS antisense 5′-CCT-TGG-TAA-ACA-GGC-TCT-TC-3′, TBP sense 5′-GCA-ACA-GCA-GCA-GCA-ACA-AC-3′, TBP antisense 5′-CAA-CGG-TGC-AGT-GGT-CAG-AG-3′, CCL20 sense 5′-TTG-CTT-TGG-CAT-GGG-TAC-TG-3′, CCL20 antisense 5′-TTC-ATC-GGC-CAT-CTG-TCT-TG-3′, CXCL10 sense 5′-CAT-CCT-GCT-GGG-TCT-GAG-TG-3′, CXCL10 antisense 5′-AGG-ATA-GGC-TCG-CAG-GGA-TG-3′, IL-1β sense 5′-TGT-CTT-TCC-CGT-GGA-CCT-TC-3′, IL-1β antisense 5′-TCA-TCT-CGG-AGC-CTG-TAG-TG-3′, IL-6 sense 5′-ACA-AAG-CCA-GAG-TCC-TTC-AG-3′, IL-6 antisense 5′-TGG-ATG-GTC-TTG-GTC-CTT-AG-3′, IL-22 sense 5′-CAC-AGA-TGT-CCG-GCT-CAT-CG-3′, IL-22 antisense 5′-CCT-GCA-TGT-AGG-GCT-GGA-AC-3′, TNF-α sense 5′-GTC-AGC-CGA-TTT-GCT-ATC-TC-3′, TNF-α antisense 5′-AGA-CTC-CTC-CCA-GGT-ATA-TG-3′, LGR5 sense 5′-AGA-CTA-CGC-CTT-TGG-AAA-CC-3′, LGR5 antisense 5′-TGG-AGA-GTG-TCT-TGA-TTG-CAG-3′, OLFM4 sense 5′-CAA-AAG-TGA-CCT-TGT-GCC-TG-3′, and OLFM4 antisense 5′-ACC-ATG-ACT-ACA-GCT-TCC-AAG-3′. Real-time assays were run on a Bio-Rad C1000 Touch Thermal Cycler (Bio-Rad, Hercules, CA). Expression data are expressed as relative values after Genex macro analysis (Bio-Rad).
Cytokine and myeloperoxidase quantification
For quantification of cytokine level, one centimeter of medial small-bowel tissue was crushed in tubes containing ceramic beads and 500 μl HBSS-1% protease inhibitors (Sigma-Aldrich). The level of IFN-γ, IL-10, IL-17A, and myeloperoxidase (MPO) were quantified by ELISA following the manufacturer recommendations (R&D Systems, Minneapolis, MN).
Cell sorting and flow cytometry
The small bowel was extracted, washed in PBS, and incubated (37°C, 20 min) in stripping buffer (PBS, 5 mM EDTA). After stripping, the small bowel was cut in 0.5-cm pieces and digested in buffer containing RPMI 1640 medium, 20% FBS, and 2 mg/ml collagenase D for 1 h at 37°C. Digested material was passed through 100- and 70-μm cell strainers, and the cells were collected by centrifugation.
For detection of intracellular cytokines, cells were cultured in 1 ml RPMI 1640 medium-20% FBS containing 50 ng/ml PMA, 1000 ng/ml ionomycin, and 1 μl brefeldin A for 4 h at 37°C. For detection of Foxp3, cells were not simulated and were directly stained with anti-CD3 and anti-CD4 fluorescently labeled mAbs for 30 min. Cells were permeabilized with Foxp3 Fixation/Permeabilization Kit (eBioscience) and then incubated with anti–IL-17A, anti–IFN-γ, or anti-Foxp3 fluorescently labeled mAb for 30 min. Data were acquired on a FACSFortessa (BD Biosciences, Franklin Lakes, NJ) and analyzed using FlowJo Software (Tree Star, Ashland, OR).
Assessment of NOD2 deletion in NOD2ΔIEC and NOD2ΔLyz2 mice
To quantify NOD2 deletion in IEC, villi of the small intestine were removed by scraping and washed with cold PBS. Small intestinal tissue was cut into 2- to 3-mm pieces, carefully rocked in PBS containing 2 mM EDTA for 30 min at 4°C, and then vigorously washed in PBS. The crypt-enriched supernatant fractions were filtered through a 70-μm cell strainer, centrifuged at 300 × g for 5 min at 4°C and resuspended in 50 μl Matrigel (Corning, New York, NY). The crypt-containing organoid cultures were plated onto a 24-well plate and maintained in 500 μl crypt culture medium containing growth factors (EGF at 0.05 μg/ml, R-spondin 1, and Noggin at 10% of the conditioned medium). Organoids were maintained for 7 d in culture, and the medium was changed once a week.
Deletion of NOD2 in Lyz2-expressing cells was assessed after i.p. injection of 50 μg MDP or vehicle (PBS) in NOD2ΔLyz2 and WT littermate mice. The peritoneal cavity was lavaged 24 h later with 5 ml cold PBS. The resulting cellular exudate was collected and incubated with anti-CD11b and anti–Gr-1 fluorescently labeled mAb to assess neutrophil recruitment.
Statistical analysis
The results are depicted as mean ± SEM. Statistical analysis was performed using nonparametric Mann–Whitney U test. The differences between two groups were considered significant when p ≤ 0.05. All calculations were performed using GraphPad Prism software (GraphPad, La Jolla, CA). Results are representative of ≥3 experiments.
Results
Nod2 regulates mucosal damage and small intestinal crypt architecture after acute T cell activation
To investigate whether Nod2 regulates acute T cell activation and T cell–induced mucosal damage in the small intestine, we treated WT and NOD2−/− mice with a single i.p. injection of 50 μg anti-CD3 mAb. One day after T cell activation, mucosal damage in both WT and NOD2−/− mice was characterized by increased IEC apoptosis and villus blunting (Fig. 1A). Three days after anti-CD3 mAb injection, WT mice began to recover as shown by crypt hyperplasia and villus regeneration, with full recovery achieved by day 5 (Fig. 1A). In contrast, 3 d after anti-CD3 mAb injection, NOD2−/− mice continued to show severe mucosal damage. This was characterized not only by increased cell infiltration in the lamina propria and villous blunting but also with significant shortening and loss of small intestinal crypts compared with WT mice (Fig. 1). NOD2−/− mice started to recover 5 d after anti-CD3 mAb injection as observed by crypt hyperplasia and villus regeneration (Fig. 1). However, NOD2−/− mice did not restore a normal crypt architecture compared with WT mice at day 5 (Fig. 1).
Increase of mucosal damage and loss of small intestinal crypts in NOD2−/− mice after acute T cell activation. WT and NOD2−/− mice were i.p. injected with anti-CD3 mAb and were sacrificed at various time points. (A) Representative photographs at original magnification ×10 of H&E-stained sections from the small bowel at days 0, 1, 3, and 5 after anti-CD3 mAb. (B) Depicted are scores for small intestinal crypt damage at days 0, 1, 3, and 5 after anti-CD3 mAb (n = 8–14 mice). (C) Depicted are measures for villus height at days 0, 1, 3, and 5 after anti-CD3 mAb (n = 8–14 mice). *p ≤ 0.05, **p ≤ 0.01.
Increase of mucosal damage and loss of small intestinal crypts in NOD2−/− mice after acute T cell activation. WT and NOD2−/− mice were i.p. injected with anti-CD3 mAb and were sacrificed at various time points. (A) Representative photographs at original magnification ×10 of H&E-stained sections from the small bowel at days 0, 1, 3, and 5 after anti-CD3 mAb. (B) Depicted are scores for small intestinal crypt damage at days 0, 1, 3, and 5 after anti-CD3 mAb (n = 8–14 mice). (C) Depicted are measures for villus height at days 0, 1, 3, and 5 after anti-CD3 mAb (n = 8–14 mice). *p ≤ 0.05, **p ≤ 0.01.
Nod2 regulates intestinal epithelial cell regeneration in response to T cell–induced enteropathy
Because NOD2−/− mice developed more severe mucosal injury with a delayed recovery phase after acute T cell activation, we explored the factors that contributed to epithelial regeneration. In order to quantify the degree of apoptosis after mucosal injury and repair, we used an immunohistochemical TUNEL assay to identify apoptotic bodies and found an equal number in the mucosa of WT and NOD2−/− mice 1 d after anti-CD3 mAb injection (Fig. 2A, 2B). The number of mucosal apoptotic bodies started to decrease in WT mice 3 d after T cell activation, whereas it remained as high as on day 1 in NOD2−/− mice (Fig. 2A, 2B). Five days after anti-CD3 mAb injection, the number of apoptotic bodies was significantly higher in NOD2−/− mice as compared with WT mice (Fig. 2A, 2B). In addition, these apoptotic bodies were specifically located in the epithelium of small intestinal crypts in NOD2−/− mice, as opposed to the lamina propria (Fig. 2C, 2D). Paneth cells and intestinal stem cells (ISC) are located in the crypt epithelium; therefore, we counted the number of granulated Paneth cells per crypt and assessed the expression of LGR5 and OLFM4, ISC markers in the small intestine. One and three days after T cell activation, the number of Paneth cells significantly decreased in both WT and NOD2−/− mice (Fig. 2E). Five days after anti-CD3 mAb injection, the number of Paneth cells recovered in WT mice but not in NOD2−/− mice (Fig. 2E). Our data also showed that the mRNA expression levels of LGR5 and OLFM4 were significantly decreased in both WT and NOD2−/− mice after T cell activation (Fig. 2F). LGR5 level was not different between WT and NOD2−/− mice at days 1 and 3 after anti-CD3 mAb injection. On day 5, LGR5 expression was upregulated in both WT and NOD2−/− mice, but this upregulation was significantly greater in WT mice (Fig. 2F). OLFM4 level was significantly decreased in both WT and NOD2−/− mice on day 1, and this decrease was more pronounced in NOD2−/− mice (Fig. 2F). On day 3, OLFM4 level recovered in WT mice but not in NOD2−/− mice, who showed recovery only by day 5 after anti-CD3 mAb injection (Fig. 2F).
Increase of apoptosis in NOD2−/− mice after acute T cell activation. WT and NOD2−/− mice were i.p. injected with anti-CD3 mAb and were sacrificed at various time points. (A) Representative photographs (original magnification ×20 and ×40 [inset]) of TUNEL-stained sections from the small bowel at days 1, 3, and 5 after anti-CD3 mAb. (B) Average number of mucosal apoptotic bodies in 10 fields representing five villus per crypt units at days 0, 1, 3, and 5 after anti-CD3 mAb (n = 8–14 mice). Average number of lamina propria (C) and epithelium (D) apoptotic bodies in 10 fields representing five villus per crypt units at day 5 after anti-CD3 mAb (n = 3 control mice and n = 8 anti-CD3 mAb-injected mice). (E) Average number of granulated Paneth cells in 10 fields representing five villus per crypt units (n = 8–14 mice). (F) mRNA relative expression of LGR5 and OLFM4 ISC markers (n = 8–14 mice). *p ≤ 0.05, **p ≤ 0.01.
Increase of apoptosis in NOD2−/− mice after acute T cell activation. WT and NOD2−/− mice were i.p. injected with anti-CD3 mAb and were sacrificed at various time points. (A) Representative photographs (original magnification ×20 and ×40 [inset]) of TUNEL-stained sections from the small bowel at days 1, 3, and 5 after anti-CD3 mAb. (B) Average number of mucosal apoptotic bodies in 10 fields representing five villus per crypt units at days 0, 1, 3, and 5 after anti-CD3 mAb (n = 8–14 mice). Average number of lamina propria (C) and epithelium (D) apoptotic bodies in 10 fields representing five villus per crypt units at day 5 after anti-CD3 mAb (n = 3 control mice and n = 8 anti-CD3 mAb-injected mice). (E) Average number of granulated Paneth cells in 10 fields representing five villus per crypt units (n = 8–14 mice). (F) mRNA relative expression of LGR5 and OLFM4 ISC markers (n = 8–14 mice). *p ≤ 0.05, **p ≤ 0.01.
To assess the degree of epithelial cell proliferation during the recovery phase, we used Ki67 staining. In WT mice, 3 d after anti-CD3 mAb injection, most of the Ki67+ cells were located in the hyperplastic small intestinal crypts, corresponding to the normal proliferative zone of the small intestine, whereas few Ki67+ cells were found in the shortened crypts of NOD2−/− mice (Fig. 3). Moreover, we observed an atypical pattern of Ki67+ within epithelial cells extending along the villus in NOD2−/− mice (Fig. 3A). Five days after anti-CD3 mAb injection, Ki67+ cells were confined to the crypt epithelium in WT mice, whereas a high number of Ki67+ cells were located in the hyperplastic crypts and along the villus in NOD2−/− mice (Fig. 3). Taken together, these data indicate that acute T cell activation with anti-CD3 mAb induces stronger apoptosis of small intestinal mucosal cells, particularly crypt epithelial cells of NOD2−/− mice compared with WT mice. The higher number of apoptotic cells and the delayed epithelial regeneration in the small intestine of NOD2−/− mice contribute to the severe mucosal injury, the loss of crypt architecture, and the delayed recovery phase.
Delayed epithelial regeneration in NOD2−/− mice after acute T cell activation. (A) Representative photographs at original magnification ×10 of Ki67+-stained sections from the small bowel at days 0, 3, and 5 after anti-CD3 mAb. (B) Average number of Ki67+ epithelial cells in 10 crypts at days 0, 3, and 5 after anti-CD3 mAb (n = 8–14 mice). **p ≤ 0.01, ***p ≤ 0.001.
Delayed epithelial regeneration in NOD2−/− mice after acute T cell activation. (A) Representative photographs at original magnification ×10 of Ki67+-stained sections from the small bowel at days 0, 3, and 5 after anti-CD3 mAb. (B) Average number of Ki67+ epithelial cells in 10 crypts at days 0, 3, and 5 after anti-CD3 mAb (n = 8–14 mice). **p ≤ 0.01, ***p ≤ 0.001.
Small intestinal mucosal damage is associated with stronger immune activation and accumulation of Th17 cells in NOD2−/− mice
Repeated low-dose injections (20 μg) of anti-CD3 mAb have been shown to induce IL-6–dependent generation and CCL20-dependent accumulation of Th17 cells in the small intestine of mice (16). To explore the mechanisms by which acute T cell activation induced more severe small intestinal crypt damage in NOD2−/− mice, we examined the expression of T cell–associated cytokines and the accumulation of T cell subsets within the small intestinal lamina propria of WT and NOD2−/− mice. Our results showed that anti-CD3 mAb injection significantly upregulated the level of IL-17A, IFN-γ, and IL-10 cytokines in NOD2−/− mice compared with WT mice (Fig. 4A). IL-17A level was significantly higher in NOD2−/− mice at days 3 and 5 after anti-CD3 mAb injection, whereas IFN-γ and IL-10 levels were significantly higher only at day 3 (Fig. 3A). In parallel, we found that IL-1β, IL-6, and CCL20 transcripts were significantly upregulated in NOD2−/− mice (Fig. 4B), suggesting that these factors may contribute to the generation and the accumulation of Th17 cells in the lamina propria of NOD2−/− mice. The expression of IL-22, TNF-α, and CXCL10 (also known as IFN-γ–induced protein 10) were also increased in NOD2−/− mice. Taken together, the increased secretion and expression in these cytokines indicated a stronger inflammatory response in NOD2−/− mice (Fig. 4B).
Acute T cell activation leads to stronger immune activation and accumulation of IL-17A–expressing CD4+ T cells in NOD2−/− mice. WT and NOD2−/− mice were i.p. injected with anti-CD3 mAb and were sacrificed at various time points. (A) Levels of IL-17A, IFN-γ, and IL-10 cytokines in small-bowel tissue were quantified by ELISA at days 0, 1, 3, and 5 after anti-CD3 mAb (n = 8–14 mice). (B) mRNA expression of IL-1β, IL-6, IL-22, TNF-α, CCL20, and CXCL10 in small-bowel tissue was quantified by RT-quantitative PCR at days 0, 1, 3, and 5 after anti-CD3 mAb (n = 8–14 mice). (C) Depicted are representative dot plots of lamina propria CD3+CD4+ T cells in WT and NOD2−/− mice for expression of CD3, CD4, Foxp3, IL-17A, and IFN-γ. (D) Columns represent the number of CD4+ T cells recovered from the small-bowel lamina propria of mice and the number of IL-17A, Foxp3, and IFN-γ–expressing CD4+ T cells per 100,000 CD4+ T cells at day 3 after anti-CD3 mAb (n = 3 control mice and n = 6 anti-CD3 mAb-injected mice). *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.
Acute T cell activation leads to stronger immune activation and accumulation of IL-17A–expressing CD4+ T cells in NOD2−/− mice. WT and NOD2−/− mice were i.p. injected with anti-CD3 mAb and were sacrificed at various time points. (A) Levels of IL-17A, IFN-γ, and IL-10 cytokines in small-bowel tissue were quantified by ELISA at days 0, 1, 3, and 5 after anti-CD3 mAb (n = 8–14 mice). (B) mRNA expression of IL-1β, IL-6, IL-22, TNF-α, CCL20, and CXCL10 in small-bowel tissue was quantified by RT-quantitative PCR at days 0, 1, 3, and 5 after anti-CD3 mAb (n = 8–14 mice). (C) Depicted are representative dot plots of lamina propria CD3+CD4+ T cells in WT and NOD2−/− mice for expression of CD3, CD4, Foxp3, IL-17A, and IFN-γ. (D) Columns represent the number of CD4+ T cells recovered from the small-bowel lamina propria of mice and the number of IL-17A, Foxp3, and IFN-γ–expressing CD4+ T cells per 100,000 CD4+ T cells at day 3 after anti-CD3 mAb (n = 3 control mice and n = 6 anti-CD3 mAb-injected mice). *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.
We then examined lamina propria lymphocytes from WT and NOD2−/− mice and found that total CD4+ T cell number, IL-17A–expressing T cells, IFN-γ–expressing T cells, and Foxp3+ regulatory T (Treg) cell numbers were no different in untreated WT and NOD2−/− mice (Fig. 4D). We observed that anti-CD3 mAb injection reduced total CD4+ T cell numbers within the lamina propria of both WT and NOD2−/− mice, but these numbers were not different, suggesting that WT and Nod2-deficient CD4+ T cells presented the same sensitivity to apoptosis (Fig. 4C, 4D). Three days after anti-CD3 mAb, we found a significant increase in IL-17A–expressing T cells within the lamina propria of NOD2−/− mice compared with WT mice (Fig. 4C, 4D). In contrast, Foxp3+ Treg cell number was lower in NOD2−/− mice compared with WT mice, suggesting that the Th17/Treg ratio may in part explain the severe mucosal damage in NOD2−/− mice (Fig. 4C, 4D). Last, we found that the number of IFN-γ–expressing T cells was not different between WT and NOD2−/− mice 3 d after anti-CD3 mAb injection (Fig. 4C, 4D). Taken together, these data indicate that small intestinal mucosal damage is associated with accumulation of IL-17A–expressing T cells in NOD2−/− mice.
Sensing of gut microbiota by Nod2 is important for the regulation of T cell–induced enteropathy
To rule out the possibility that NOD2−/− mice are more sensitive to T cell–induced enteropathy because of major differences in microbiota composition with WT mice, we crossed WT and NOD2−/− mice to generate littermate controls. NOD2+/+and NOD2−/− littermate control mice were sacrificed 3 d after anti-CD3 mAb injection because the strongest difference in mucosal damage was shown at that time point. As previously described in nonlittermate mice, we found that WT littermate control mice showed robust crypt hyperplasia and villus regeneration, whereas NOD2−/− mice presented a stronger mucosal damage characterized by significant shortening and loss of small intestinal crypts and an increase of IL-17A expression in small-bowel tissue (Fig. 5A–C). To support this finding that NOD2−/− littermate control mice developed a severe acute inflammation, we quantified MPO level in small intestinal tissue and observed a significant increase in these NOD2−/− littermate mice (Fig. 5D).
Sensing of gut microbiota by Nod2 is important for the regulation of T cell–induced mucosal damage. (A) Representative photographs at original magnification ×10 of H&E-stained sections from the small bowel of NOD2+/+ and NOD2−/− littermate control mice at day 3 after anti-CD3 mAb. (B) Depicted are scores for small intestinal crypt damage of littermate control mice at day 3 after anti-CD3 mAb (n = 9–13 mice). Levels of IL-17A (C) and MPO (D) in the small-bowel tissue of littermate control mice were quantified by ELISA at day 3 after anti-CD3 mAb (n = 9–13 mice). WT and NOD2−/− mice were treated with water or a mixture of ampicillin-neomycin-metronidazole, each at 1 g/l (n = 8–12 mice). (E) Representative photographs at original magnification ×10 of H&E-stained sections from the small bowel of antibiotic-treated mice at day 1 after anti-CD3 mAb. (F) DNA concentration (nanograms per microliter) per 10 mg feces in water or antibiotic-treated mice. (G) Average number of mucosal apoptotic bodies in 10 fields representing five villus per crypt units of antibiotic-treated mice at day 1 after anti-CD3 mAb. (H) Representative photographs at original magnification ×10 of H&E-stained sections from the small bowel of water and antibiotic-treated mice at day 3 after anti-CD3 mAb. (I) Depicted are scores for small intestinal crypt damage of water and antibiotic-treated mice at day 3 post anti-CD3 mAb. (J) Average number of mucosal apoptotic bodies in 10 fields representing five villus per crypt units of water and antibiotic-treated mice at day 3 after anti-CD3 mAb. (K) Levels of IL-17A, IFN-γ, and MPO in the small-bowel tissue of water and antibiotic-treated mice at day 3 after anti-CD3 mAb. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.
Sensing of gut microbiota by Nod2 is important for the regulation of T cell–induced mucosal damage. (A) Representative photographs at original magnification ×10 of H&E-stained sections from the small bowel of NOD2+/+ and NOD2−/− littermate control mice at day 3 after anti-CD3 mAb. (B) Depicted are scores for small intestinal crypt damage of littermate control mice at day 3 after anti-CD3 mAb (n = 9–13 mice). Levels of IL-17A (C) and MPO (D) in the small-bowel tissue of littermate control mice were quantified by ELISA at day 3 after anti-CD3 mAb (n = 9–13 mice). WT and NOD2−/− mice were treated with water or a mixture of ampicillin-neomycin-metronidazole, each at 1 g/l (n = 8–12 mice). (E) Representative photographs at original magnification ×10 of H&E-stained sections from the small bowel of antibiotic-treated mice at day 1 after anti-CD3 mAb. (F) DNA concentration (nanograms per microliter) per 10 mg feces in water or antibiotic-treated mice. (G) Average number of mucosal apoptotic bodies in 10 fields representing five villus per crypt units of antibiotic-treated mice at day 1 after anti-CD3 mAb. (H) Representative photographs at original magnification ×10 of H&E-stained sections from the small bowel of water and antibiotic-treated mice at day 3 after anti-CD3 mAb. (I) Depicted are scores for small intestinal crypt damage of water and antibiotic-treated mice at day 3 post anti-CD3 mAb. (J) Average number of mucosal apoptotic bodies in 10 fields representing five villus per crypt units of water and antibiotic-treated mice at day 3 after anti-CD3 mAb. (K) Levels of IL-17A, IFN-γ, and MPO in the small-bowel tissue of water and antibiotic-treated mice at day 3 after anti-CD3 mAb. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.
Regulation of microbial signals is essential for the maintenance of intestinal homeostasis. To determine whether gut microbiota can alter mucosal damage and recovery in response to acute T cell activation, we treated WT and NOD2−/− mice with a broad-spectrum antibiotic mixture (ampicillin, neomycin, and metronidazole, each at 1 g/l) and analyzed whether the decrease in gut microbiota altered the mucosal damage and recovery induced by anti-CD3 mAb injection. The results showed that both antibiotic-treated WT and NOD2−/− mice developed significant mucosal damage associated with villus blunting and equal number of mucosal apoptotic cells 1 d after anti-CD3 mAb injection (Fig. 5E–G). In addition, the number of mucosal apoptotic cells was comparable to water-treated mice (Fig. 2B), indicating that partial microbiota depletion did not prevent T cell–induced mucosal damage. However, 3 d after anti-CD3 mAb injection, both antibiotic-treated WT and NOD2−/− mice showed significantly reduced mucosal damage compared with water control (Fig. 5H–J). Moreover, IL-17A, IFN-γ, and MPO levels in the small intestine of NOD2−/− mice were reduced to the level of WT mice at day 3 (Fig. 5K). These data collectively indicate that gut microbial load is important in determining the degree of damage and the delay in recovery. In this regard, microbial sensing by Nod2 is a critical mechanism in controlling excessive crypt damage.
Selective deletion of Nod2 in intestinal epithelial cells or in Lyz2-expressing phagocytes is not sufficient to recapitulate the mucosal damage observed in NOD2−/− mice
To determine whether epithelial cell intrinsic Nod2 regulates mucosal damage and epithelial regeneration after T cell-induced enteropathy, we generated mice lacking NOD2 in villin-expressing cells (NOD2∆IEC mice). Absence of NOD2 expression in IEC from NOD2∆IEC mice was controlled by PCR (Supplemental Fig. 1). Our results showed that NOD2∆IEC mice developed comparable crypt damage and IEC apoptosis with NOD2Flox (WT) mice after anti-CD3 treatment (Fig. 6A–C). In addition, the secretion of IL-17A, IFN-γ, and MPO, and the mRNA expression of the ISC markers LGR5 and OLFM4 were not different between NOD2Flox and NOD2∆IEC mice (data not shown). These data indicate that deletion of Nod2 in IEC is not sufficient to increase the sensitivity to T cell–induced enteropathy and delay recovery, suggesting that Nod2 expression outside the epithelium is important in regulating the response to T cell–induced mucosal damage.
Deletion of Nod2 in intestinal epithelial cells or Lyz2-expressing phagocytes does not recapitulate mucosal damage induced by acute T cell activation in NOD2−/− mice. NOD2Flox, NOD2∆IEC, and NOD2∆Lyz2 mice were i.p. injected with anti-CD3 mAb and were sacrificed at day 3. (A) Representative photographs at original magnification ×10 of H&E-stained sections from the small bowel of NOD2Flox and NOD2∆IEC. (B) Depicted are scores for small intestinal crypt damage (n = 6 mice). (C) Average number of epithelial apoptotic bodies in 10 fields representing five villus per crypt units (n = 6 mice). (D) Representative photographs at original magnification ×10 of H&E-stained sections from the small bowel of NOD2Flox and NOD2∆Lyz2. (E) Depicted are scores for small intestinal crypt damage (n = 10–11 mice). (F) Average number of epithelial apoptotic bodies in 10 fields representing five villus per crypt units (n = 10–11 mice). (G) mRNA expression of TNF-α and IL-22 in small-bowel tissue was quantified by RT-quantitative PCR (n = 10–11 mice). *p ≤ 0.05, **p ≤ 0.01.
Deletion of Nod2 in intestinal epithelial cells or Lyz2-expressing phagocytes does not recapitulate mucosal damage induced by acute T cell activation in NOD2−/− mice. NOD2Flox, NOD2∆IEC, and NOD2∆Lyz2 mice were i.p. injected with anti-CD3 mAb and were sacrificed at day 3. (A) Representative photographs at original magnification ×10 of H&E-stained sections from the small bowel of NOD2Flox and NOD2∆IEC. (B) Depicted are scores for small intestinal crypt damage (n = 6 mice). (C) Average number of epithelial apoptotic bodies in 10 fields representing five villus per crypt units (n = 6 mice). (D) Representative photographs at original magnification ×10 of H&E-stained sections from the small bowel of NOD2Flox and NOD2∆Lyz2. (E) Depicted are scores for small intestinal crypt damage (n = 10–11 mice). (F) Average number of epithelial apoptotic bodies in 10 fields representing five villus per crypt units (n = 10–11 mice). (G) mRNA expression of TNF-α and IL-22 in small-bowel tissue was quantified by RT-quantitative PCR (n = 10–11 mice). *p ≤ 0.05, **p ≤ 0.01.
Because of the higher number of apoptotic cells in NOD2−/− mice and the importance of macrophages and neutrophils in phagocytosing dead cells, cellular debris, and invading microorganisms, we hypothesized that Nod2 expression in phagocytes may be important in controlling anti-CD3–induced enteropathy, and we generated mice lacking NOD2 in Lyz2-expressing cells (NOD2∆Lyz2 mice). Deletion of NOD2 in Lyz2-expressing cells was assessed by the decreased recruitment of neutrophils in the peritoneal cavity of NOD2∆Lyz2 mice after i.p. injection of the Nod2 ligand MDP (Supplemental Fig. 1). Although, crypt damage was not different between NOD2Flox and NOD2∆Lyz2 mice, we found higher numbers of apoptotic IEC in NOD2∆Lyz2 mice (Fig. 6D–F). The secretion of IL-17A, IFN-γ, and MPO, and the mRNA expression of LGR5 and OLFM4 was not different between NOD2Flox and NOD2∆Lyz2 mice (data not shown). However, we did find a significant increase of TNF-α and IL-22 expression in NOD2∆Lyz2 mice (Fig. 6G), indicating that Nod2 deletion in phagocytes can affect TNF-α level and IEC viability in the small intestine.
Discussion
The requirement of T cells in various animal models of intestinal inflammation (12) and the increase in T cell–derived inflammatory cytokines in the mucosa of inflammatory bowel disease (IBD) patients (17) suggest that T cells play important roles in the pathogenesis of IBD. Our study demonstrates that following small intestinal mucosal damage induced by T cell activation, microbiota sensing by Nod2 is important for controlling immune activation, crypt damage, and epithelial regeneration.
Looking at a direct role of Nod2 in T cell regulation of colitis, our group previously showed that the expression of Nod2 in murine CD4+ T cells is not required for the development or the prevention of T cell–induced colitis in Rag-deficient mice (6). In addition, NOD2+/+and NOD2−/− littermate control mice have been shown to present the same sensitivity to dextran sulfate sodium–induced colitis (18). These results and the association between Nod2 mutations and ileal CD suggest a more specific role of Nod2 in the regulation of small intestinal homeostasis. The model of anti-CD3 mAb i.p. injection was developed to study T cell driven inflammatory diseases such as IBD, celiac disease, or graft-versus-host disease (13). In mice, acute T cell activation following i.p. injection of T cell–activating anti-CD3 mAb leads to severe small intestinal mucosal damage characterized by villus blunting and apoptosis of IEC (13). Therefore, i.p. injection of anti-CD3 mAb is a helpful model to study the role of Nod2 in regulating T cell–induced small intestinal mucosal damage.
The inability to control inflammation and to regenerate the intestinal epithelial barrier is a common characteristic of IBD. Regeneration of the epithelium is a highly regulated process that is orchestrated by multipotent stem cells located in the bottom of the crypts (19, 20). Study of ISC biology has been facilitated by the identification of ISC markers such as LGR5 and OLFM4 (21, 22). Nod2 is highly expressed in ISC and has an important function in ISC survival (8). Our data showed that the delayed epithelial regeneration in NOD2−/− mice is associated with an increase of crypt IEC apoptosis and a decrease of LGR5 and OLFM4 ISC marker expression, suggesting that Nod2-deficient ISC may be more sensitive to apoptosis induced by anti-CD3 mAb injection. In addition, we found a more severe loss of granulated Paneth cell number in NOD2−/− mice. Nod2 is expressed in Paneth cells (4), and secretion of IFN-γ by immune cells has been shown to induce Paneth cell degranulation and death (23), suggesting that the increase of IFN-γ level in NOD2−/− mice can affect Paneth cell number and function.
Intraperitoneal injection of mice with anti-CD3 mAb has been used to study the role of Nod2 on T cell activation and function. In a murine model of low-dose (15 μg) i.p. injection of anti-CD3 mAb at days 0 and 2, Nod2 deletion has been shown to decrease CXCL9/CXCL10-dependent recruitment of CD8+ T cells in the small intestinal lamina propria and the subsequent expression of IL-10 by CD8+ T cells (24). In addition, the authors found a decrease in IL-17A and IFN-γ mRNA expression in the small intestine of NOD2−/− mice (24). In our study, we found that a single high-dose (50 μg) i.p. injection of anti-CD3 mAb induced higher levels of IL-17A, IFN-γ, IL-10, and CXCL10 in the small intestine of NOD2−/− mice compared with WT mice. These opposite results may indicate that the intensity of T cell activation and the degree of mucosal damage can strongly modulate the immune response. Another possibility is that the use of different NOD2−/− mice strain leads to contradictory results regarding the role of Nod2 (6, 25, 26).
Yet in another model, repeated i.p. injection of 20 μg anti-CD3 mAb promoted generation and accumulation of Th17 cells within the small intestine of WT mice (16). In this model, Th17 cell generation was dependent on IL-6 secretion from macrophages, and their accumulation within the small intestine was triggered by the CCR6–CCL20 axis. In our data, we observed a significant increase of IL-6, CCL20 expression, IL-17A secretion, and IL-17A–expressing T cell accumulation within the small intestine of NOD2−/− mice. After antibiotic treatment, the secretion of IL-17A, IFN-γ, and MPO was decreased to the level of WT mice. In addition, antibiotic-treated NOD2−/− mice did not display the marked loss of small intestinal crypts seen in untreated NOD2−/− mice, highlighting the role of the microbiota in perpetuating mucosal damage induced by acute T cell activation in NOD2−/− mice. Indeed, Nod2-deficient mice show increased sensitivity to several microbial pathogens (27–30). The altered response to injury and infection that is seen in NOD2−/− mice and the development of ileal disease in CD patients carrying NOD2 risk alleles are consistent with our results supporting the importance of microbial sensing by Nod2 in the regulation of small-intestine homeostasis.
CD is associated with structural damage and destruction of the bowel wall (31). Epithelial regeneration is a crucial step in mucosal healing and this mechanism requires the coordinated activity of IEC, Paneth cells, and goblet cells to restore intestinal barrier function (32–34). Mice carrying both Nod1 and Nod2 deletion have decreased paracellular permeability and a higher susceptibility to dextran sulfate sodium–induced colitis, highlighting a potential role of Nod receptors in the maintenance of intestinal barrier function (35). Moreover, another study reported a new function for Nod2 in affording ISC survival and epithelial regeneration (8). The authors showed that stimulation with MDP enhanced in vitro intestinal organoid survival and in vivo stem cell survival following doxorubicin treatment in mice (8). With regards to T cell–induced mucosal damage, our data has identified the importance of Nod2 signaling in sustaining crypt architecture and epithelial regeneration and suggests that apoptotic epithelial cells observed in the small intestinal crypts of NOD2−/− mice might be ISC. Surprisingly, we found that Nod2 deletion in IEC did not exacerbate mucosal damage, suggesting that Nod2 function in other cellular compartments regulates intestinal homeostasis (7, 36–38).
Resolution of acute inflammation requires the elimination of invading microorganisms, apoptotic cells, and cellular debris from the intestinal lamina propria. This mechanism is achieved via the recruitment of professional phagocytes (39, 40). Stimulation of Nod2 in human monocyte–derived macrophages and in mouse intestinal macrophages increases autophagy and bacterial clearance (41). We described that i.p. injection of anti-CD3 mAb induced higher numbers of mucosal apoptotic bodies in the small intestinal lamina propria of NOD2−/− mice. In addition, disruption of the intestinal epithelial barrier was associated with a microbiota-dependent perpetuation of mucosal damage in NOD2−/− mice. These results suggest that Nod2 deletion in macrophages can alter bacterial clearance and consequently the mucosal damage induced by anti-CD3 mAb injection. We found that deletion of Nod2 in Lyz2-expressing cells did not recapitulate the mucosal damage observed in NOD2−/− mice. Indeed, the shortening and the loss of small intestinal crypts were comparable between WT and NOD2∆Lyz2 mice. However, NOD2∆Lyz2 mice showed higher numbers of epithelial apoptotic bodies and increased expression of TNF-α and IL-22 compared with WT mice. TNF-α plays a crucial role in the pathogenesis of CD and high levels of TNF-α can alter intestinal epithelial barrier integrity (42–44). Chronic stimulation of Nod2 in human monocyte–derived macrophages has been associated with a downregulation of TNF-α secretion (45), indicating that expression of Nod2 is important for both the induction and the regulation of this response. These results suggest that deletion of Nod2 in Lyz2-expressing cells can trigger higher levels of TNF-α, thus inducing a stronger IEC apoptosis and the increased expression of IL-22, a cytokine involved in gut barrier integrity and epithelial regeneration (46, 47).
In conclusion, we describe a role for Nod2 in regulating small intestinal crypt damage and epithelial regeneration after acute T cell activation. Understanding the function of Nod2 in mucosal healing will help define the pathophysiology of inflammatory diseases of the gut such as IBD.
Acknowledgements
We thank Dr. Philip Rosenstiel (Institute for Clinical Molecular Biology, Christian-Albrechts-University, Kiel, Germany) for providing NOD2-floxed mice.
Footnotes
This work was supported by the Canadian Association of Gastroenterology, Crohn's and Colitis Canada, the Canadian Foundation of Innovation, the Canadian Institutes of Health Research, and the Mount Sinai Hospital Department of Medicine.
The online version of this article contains supplemental material.
References
Disclosures
The authors have no financial conflicts of interest.