Bacterial Sensor Triggering Receptor Expressed on Myeloid Cells-2 Regulates the Mucosal Inflammatory Response

GASTROENTEROLOGY 2013;144:346 –356

BASIC AND TRANSLATIONAL—ALIMENTARY TRACT Bacterial Sensor Triggering Receptor Expressed on Myeloid Cells-2 Regulates the Mucosal Inflammatory Response CARMEN CORREALE,1,* MARCO GENUA,1,* STEFANIA VETRANO,1,* ELISA MAZZINI,2 CHIARA MARTINOLI,2 ANTONINO SPINELLI,1 VINCENZO ARENA,3 LAURENT PEYRIN–BIROULET,4 FLAVIO CAPRIOLI,5 NADIA PASSINI,6 PAOLA PANINA–BORDIGNON,7 ALESSANDRO REPICI,1 ALBERTO MALESCI,1 SERGIO RUTELLA,8 MARIA RESCIGNO,2 and SILVIO DANESE1 1

Division of Gastroenterology, Humanitas Clinical and Research Center, Rozzano, Milan, Italy; 2Department of Experimental Oncology, European Institute of Oncology, Milan, Italy; 3Department of Pathology, Catholic University Medical School, Rome, Italy; 4Department of Gastroenterology, Inserm U954, University Hospital of Nancy, Henri Poincaré University, Vandœuvre-lès-Nancy, France; 5U.O. Gastroenterologia 2, Fondazione IRCCS Ca’ Granda Ospedale Policlinico di Milano, Milan, Italy; 6 Bioxell, Milan, Italy; 7Reproductive Sciences Lab, San Raffaele Scientific Institute, Milan, Italy; and 8Department of Pediatric Hematology/Oncology, IRCCS Bambino Gesù Children’s Hospital, Rome, Italy

BASIC AND TRANSLATIONAL AT

BACKGROUND AND AIMS: Triggering receptor expressed on myeloid cells (TREM)–2 is a surface receptor detected on macrophages, dendritic cells, and microglia that binds repeated anionic motifs on yeast and Gram-positive and Gram-negative bacteria. Little is known about TREM-2 expression and function in the intestine or its role in inflammatory bowel disease (IBD). We investigated the expression of TREM-2 in the intestinal lamina propria and its role in the development of colonic inflammation. METHODS: We measured levels of TREM-2 in lamina propria mononuclear cells from surgical specimens collected from patients with IBD or cancer (controls). We analyzed the development of colitis in TREM-2 knockout and wild-type mice. Colon samples were isolated from mice and analyzed for cytokine expression, phagocytosis of bacteria, proliferation in colonic crypts, lamina propria mononuclear cell function, and T-cell activation by ovalbumin. RESULTS: TREM-2 was virtually absent from colon samples of control patients, but levels were significantly higher in within the inflamed mucosa of patients with IBD; it was mainly expressed by CD11c⫹ cells. Levels of TREM-2 increased as acute or chronic colitis was induced in mice. TREM-2 knockout mice developed less severe colitis than wild-type mice; the knockout mice lost less body weight, had a lower disease activity index, and had smaller mucosal lesions in endoscopic analysis. Colon dendritic cells from TREM-2 knockout mice produced lower levels of inflammatory cytokines and had reduced levels of bacterial killing and T-cell activation than cells from wild-type mice. CONCLUSIONS: TREM-2 contributes to mucosal inflammation during development of colitis in mice. Levels of TREM-2 are increased within the inflamed mucosa of patients with IBD, indicating its potential as a therapeutic target. Keywords: Crohn’s Disease; Ulcerative Colitis; Mouse Model; Immune Response.

T

riggering receptor expressed on myeloid cells (TREM) proteins are a recently discovered family of cell surface receptors broadly expressed on myeloid cells of hu-

man and mouse origin.1–3 TREM-2 has been detected on the surface of immature monocyte-derived dendritic cells (DCs),1 macrophages,4 osteoclasts,5 and microglia.6 –9 TREM-2 ligands have not yet been fully characterized, although it was found that TREM-2 binds several Grampositive and Gram-negative bacteria, as well as yeast, by interacting with surface repeated anionic motifs (lipopolysaccharide [LPS], lipoteichoic acid, peptidoglycan).10 Several biological pathways were shown to be positively regulated by TREM-2, including phagocytosis in both DC and macrophages,11 osteoclast, and macrophage fusion,12 and formation of multinucleated giant cells.13 Studies performed on microglia cells suggested that TREM-2 can act as a negative regulator of inflammation.7,9 For example, TREM-2 neutralization in experimental autoimmune encephalomyelitis exacerbated the disease and caused more diffuse and severe inflammation. Also, in mouse macrophages, TREM-2 negatively regulates inflammation.4,14 TREM-2 deficiency on bone marrow (BM)– derived macrophages is associated with an increase in proinflammatory cytokine secretion after stimulation with Toll-like receptor (TLR) ligands. In humans, TREM-2 is not expressed by monocytes and circulating DC precursors, but it can be detected on the surface of monocyte-derived DCs.15 DC maturation induced by several TLR ligands and proinflammatory cytokines induces the complete down-regulation of TREM-2. *Authors

share co-first authorship.

Abbreviations used in this paper: BM, bone marrow; CD, Crohn’s disease; CFSE, carboxyfluorescein succinimidyl ester; CFU, colony-forming unit; DC, dendritic cell; DSS, dextran sodium sulfate; IBD, inflammatory bowel disease; IL, interleukin; KO, knockout; LP, lamina propria; LPS, lipopolysaccharide; MMP, matrix metalloproteinase; mRNA, messenger RNA; NLR, NOD-like receptor; OVA, ovalbumin; TLR, Toll-like receptor; TNBS, 2,4,6-trinitrobenzene sulfonic acid; TNF, tumor necrosis factor; TREM, triggering receptor expressed on myeloid cells; UC, ulcerative colitis; WT, wild-type. © 2013 by the AGA Institute 0016-5085/$36.00 http://dx.doi.org/10.1053/j.gastro.2012.10.040

February 2013

Materials and Methods Human Tissues Actively inflamed intestinal specimens and normal tissue specimens were obtained from CD and UC patients undergoing surgery or colonoscopy and from patients undergoing abdominal surgery for cancer. Clinical disease activity was assessed by the Harvey-Bradshaw Activity Index and the Colitis Activity Index, as reported previously.20 All diagnoses were confirmed by clinical, radiologic, endoscopic, and histological criteria. Tissues were embedded in optimum cutting temperature compound, frozen on dry ice, and stored at ⫺80°C for additional analysis, or readily used for isolation of LP mononuclear cells. Human studies were approved by the ethical committee of Istituto Clinico Humanitas, Milan, Italy.

Mice C57BL/6 TREM-2 knockout (KO) mice were provided by Bioxell-Cosmo Pharmaceutical (Milan, Italy). Wild-type littermates were used as controls in all the experiments. Six- to 10-week-old, age- and sex-matched wild-type (WT) and TREM-2 KO mice were used for all the experiments. Typically, 4 to 6 mice were used for each treatment group and experiments were repeated at least twice. Animal studies adhered to the requirements of the Commission Directive 86/609/EEC and to the Italian legislation (Decreto Legislativo 116; January 27, 1992). Studies were approved by the Animal Care and Use Committee (authorization no. 192/2008B, Istituto Clinico Humanitas, Milan, Italy).

Statistical Analysis Statistical significance was evaluated by the nonparametric, 2-tailed Mann-Whitney U test for the analysis of variables that were not normally distributed. Statistical significance was defined at P ⬍ .05.

347

Results TREM-2 Is Up-Regulated in LP-DCs of IBD Patients and During Experimental Colitis To investigate the expression levels of TREM-2 in normal and IBD-involved mucosa, we performed confocal fluorescence analysis of 14 histologically normal mucosa and 15 active CD- and 13 UC-involved mucosa. Normal colonic mucosa showed no expression of TREM-2 (Figure 1A). On the contrary, in the mucosa of actively inflamed CD or UC patients, TREM-2–positive cells were readily detected (7 ⫾ 2 positive cells/field). Co-staining with CD11c indicated that TREM-2 was mainly expressed on DC (Figure 1A) rather than a macrophage population (Supplementary Figure 1A). In order to accurately quantify TREM-2 expression on LP-DCs at the single-cell level, we isolated mononuclear cells from the LP of both IBD patients and controls and analyzed the expression of TREM-2 through flow cytometry (Figure 1B). DCs were defined as CD45⫹CD11c⫹CD3⫺ cells. TREM2 was expressed on a median of 2.3% (range, 0.2%–10.4%) of LP-DCs obtained from colonic specimens of healthy controls, and significantly more LP-DCs expressed TREM-2 in UC (14.5%; range, 1%–30.4%; P ⬍ .001) and in CD (13.1%; range, 0.3%–21.3%; P ⬍ .05). No differences were found when comparing noninflamed CD and UC patients with controls (data not shown). TREM-2⫹ DC did not share surface markers with plasmacytoid DC, such as BDCA-2 (data not shown), suggesting that they might be of myeloid origin. We next assessed whether different stimuli had the ability to modulate TREM-2 expression on LP-DCs. LPS, interleukin (IL)-4, tumor necrosis factor (TNF)–␣, IL-13, interferon gamma, and IL-10 failed to up-regulate TREM2 on LP-DCs (Supplementary Figure 1B). The up-regulation of TREM-2 in the tissues and LP-DC of IBD patients suggested that TREM-2 might have a functional role in IBD pathogenesis. To address this hypothesis, we investigated the expression pattern of TREM-2 in dextran sodium sulfate (DSS) and 2,4,6-trinitrobenzene sulfonic acid (TNBS)–induced colitis, 2 widely used models of acute intestinal inflammation. As expected, the development of colitis was associated with variable degrees of body weight loss (Figure 2A, B). Colonic samples were collected at different time points after disease induction, and analyzed for TREM-2 messenger RNA (mRNA) and protein expression. Very low levels of TREM-2 mRNA and protein were detected in healthy untreated mice or vehicle-treated mice (Figure 2C–F). By contrast, TREM-2 expression significantly increased on colitis induction, that is, 6 days after DSS administration and 24 hours after TNBS injection (Figure 2C–F). Consistent with the human findings, DC infiltrating the inflamed mucosa of colitic mice expressed TREM-2 protein in the TNBS (Supplementary Figure 1C) and DSS disease models (not shown).

TREM-2 Is Implicated in Colitis Induction To determine whether and how TREM-2 contributes to the development of colitis, we explored whether

BASIC AND TRANSLATIONAL AT

Abnormal responses to the gut flora and excessive activation of myeloid cells, such as DCs, in response to bacterial components occur in Crohn’s disease (CD) and ulcerative colitis (UC), the 2 major forms of inflammatory bowel disease (IBD), leading to the secretion of high amounts of inflammatory mediators and ultimately to tissue damage.16–19 The involvement of TREM-2 in bacterial phagocytosis and in regulating inflammatory responses suggests that TREM-2 might play a prominent role also in the intestine. However, expression of TREM-2 in the gut and its role in intestinal inflammation have not been investigated thoroughly. The aims of this study were to evaluate TREM-2 expression and function in the normal and inflamed gut, and to determine whether TREM-2 plays a role in intestinal inflammation. We present evidence that TREM-2 is up-regulated in lamina propria (LP)-DCs in human IBD and during the course of experimental colitis in mice. Genetic deletion of TREM-2 conferred protection against colitis, with reduced secretion of proinflammatory cytokines and expression of matrix metalloproteinases. In addition, TREM-2–deficient DCs manifested a reduced capacity to produce proinflammatory cytokines, kill bacteria, and activate T cells in response to bacteria-associated antigens. We propose that, in the gut microenvironment, TREM-2 might be an amplifier of inflammation and a potential target for treatment in IBD.

TREM-2 IN IBD PATHOGENESIS

348

CORREALE ET AL

GASTROENTEROLOGY Vol. 144, No. 2

BASIC AND TRANSLATIONAL AT

Figure 1. TREM-2 expression in the IBD mucosa. (A) Representative pictures of immunefluorescence analysis of colonic mucosa in normal (NL) and actively inflamed CD and UC patients. CD11c⫹ cells are depicted in green, TREM-2⫹ cells in red, and merge picture shows antigen colocalization in yellow. Panels are representative of 14 healthy controls, 15 patients with CD, and 13 patients with UC. Images were acquired with an oil immersion objective (60⫻, 1.4 NA Plan-Apochromat; Olympus). Scale bar ⫽ 40 ␮m. The histogram graph is representative of 10 different cell counts (mean ⫾ standard deviation). (B) Mononuclear cells were isolated from the lamina propria of healthy controls and IBD patients, and then analyzed for TREM-2 expression on DCs with flow cytometry. The percentages of TREM2– expressing LP-DCs in 12 healthy subjects (NL), 10 patients with CD and 10 patients with UC are plotted; *P ⬍ .05; **P ⬍ .01.

the loss of TREM-2 could induce changes in leukocyte populations, affecting the inflammatory response. To this end, we first analyzed the expression of CD11c, CD3, CD4, CD8, and F4/80 by flow-cytometry in intestinal mucosa, mesenteric lymph nodes, and spleen. No differences in the relative percentage of each cell population were found, suggesting that the proportions of leukocyte populations in the gut-associated lymphoid tissue of TREM-2 KO mice were similar to WT mice (data not shown). Next, we compared the susceptibility of TREM-2 KO and WT mice to the induction of DSS and TNBS colitis. In contrast to WT mice, TREM-2 KO mice were protected from DSS-induced colitis, showing less severe disease and weight loss, associated with lower disease activity index (Figure 3A, B). In addition, TREM-2 KO mice displayed

significantly milder mucosal lesions at endoscopy compared with WT mice (Figure 3C, D), but no differences were observed between WT and TREM-2 KO mice at baseline (Figure 3C, D). At time of sacrifice, TREM-2 KO mice had reduced colon shortening (data not shown) and colons displayed fewer signs of inflammation and less severe histopathological alterations (Figure 3C, D). Similarly, when we compared the susceptibility to TNBS-induced colitis in TREM-2 KO and WT mice, we found that KO mice experienced less severe disease, with significantly reduced body weight loss (Figure 4A). In addition, 3 days after TNBS injection, TREM-2 KO mice started recovering from the disease and WT mice were still losing weight (Figure 4A). Consistent with less clinical disease, TREM-2 KO mice also had significantly fewer endoscopic and histopathological abnormalities (Figure 4B, C). To test

February 2013

TREM-2 IN IBD PATHOGENESIS

349

whether the loss of TREM-2 conferred protection from chronic colitis, WT and TREM-2 mice were subjected to a chronic colitis model. Because repeated cycles of 3% DSS administration were not well-tolerated by WT mice, resulting in high mortality rates (data not shown), mice received 3 cycles of 2% DSS administration, each consisting of 5 days of DSS treatment, followed by 10 days of regular water. In line with data generated in the acute model, TREM-2 KO mice were significantly less susceptible to chronic colitis compared with WT mice, as suggested by lower body weight loss and disease activity index, coupled with milder leukocyte infiltrates, reduced histological damage score, and lower amounts of mucosal TNF-␣, IL-1␤, and IL-10 production (Supplementary Figure 2A–D). Taken together, these findings indicate that loss of TREM-2 confers protection against the development of acute and chronic experimental colitis and suggest that TREM-2 acts as an amplifier of gut inflammation.

BASIC AND TRANSLATIONAL AT

Figure 2. TREM-2 expression is up-regulated during experimental colitis. (A, B) C57BL/6 mice received DSS in drinking water or TNBS intra-rectally. Mice were weighed daily and body weight loss was calculated. At the indicated time points, mice were sacrificed and the colon excised in order to obtain both mRNA and protein, as described in Material and Methods. TREM2 expression was estimated by quantitative real-time polymerase chain reaction (C, D) and by Western blot analyses (E, F). Data are representative of 3 different experiments; *P ⬍ .05; **P ⬍ .01.

TREM-2 Modulates Proinflammatory Cytokine Secretion and Expression of Matrix Metalloproteinase in Experimental Colitis Because the DSS and TNBS models of colitis are associated with an unbalanced cytokine profile, we evaluated whether loss of TREM-2 could affect the secretion of proinflammatory and anti-inflammatory cytokines. We measured the amounts of IL-1␤, TNF-␣, and IL-10 produced by colonic tissues under steady-state conditions (before colitis induction) and after intestinal injury. Comparable levels of cytokines were found in healthy WT and TREM-2 KO mice (Figure 5A–C). By contrast, treatment with either DSS or TNBS induced a greater increase of IL-1␤ and TNF-␣ secretion in WT mice than in TREM-2 KO mice (Figure 5A, B), and no differences in IL-10 production were evident (Figure 5C). Because matrix metalloproteinase (MMPs) are involved in inflammation and tissue remodeling, we investigated

350

CORREALE ET AL

GASTROENTEROLOGY Vol. 144, No. 2

BASIC AND TRANSLATIONAL AT

Figure 3. TREM-2 KO mice display resistance to DSS-induced colitis. Evaluation of colitis by (A) body weight loss expressed as a percentage of the initial weight, (B) clinical disease activity index (DAI), (C) endoscopic and histological images of mucosal damage in the colon of WT and TREM-2 mice before and after 8 days of 3% DSS, and (D) quantification of mucosal injury by endoscopic and histological score before and after treatment. Data are representative of 3 independent experiments; *P ⬍ .05; **P ⬍ .01.

whether the protection from colitis observed in TREM-2 KO mice was associated with a down-regulation of MMP expression. To accomplish this goal, colons from mice that were treated with TNBS or DSS were analyzed for MMP-3, MMP-9, and MMP-14 mRNA expression. In both models of colitis, we found no differences between WT and KO mice at baseline (Figure 5D–F). After DSS treatment, MMP-3 and MMP-9 expression were up-regulated in WT compared with TREM-2 KO mice (4-fold and 2-fold increase, respectively), and MMP-14 expression remained unaltered. After TNBS instillation, both MMP-3 and MMP-9 expression significantly increased in WT mice compared with TREM-2 KO mice (3-fold and 2-fold, respectively), whereas MMP-14 expression was slightly higher in WT mice (Figure 5D–F). Taken together, these data show that both the release of proinflammatory cytokines and expression of MMPs are strongly reduced in TREM-2 KO mice during colitis induction.

TREM-2 Does Not Directly Regulate Epithelial Barrier Function To address the hypothesis that a defect in the epithelial barrier could be linked to decreased susceptibility to colonic injury in TREM-2– deficient mice, we examined the proliferative state of colonic crypts of WT and TREM-2 KO mice before and after TNBS-induced colitis. Analysis of Ki-67⫹ intestinal epithelial cells revealed no differences between the 2 groups at baseline (data not shown) and after 72 hours in the presence of vehicle only (Supplementary Figure 3A). By contrast, a significantly increased number of proliferating cells was observed in the crypts of WT compared with TREM-2 KO mice after 72 hours of TNBS treatment (Supplementary Figure 3A). In addition, the apoptotic rate of epithelial cells evaluated by the terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling assay was

February 2013

TREM-2 IN IBD PATHOGENESIS

351

responsible for increased barrier permeability during colitis, via the amplification of inflammatory responses.

Figure 4. TREM-2 KO display resistance to TNBS-induced colitis. Monitoring of TNBS-induced colitis by (A) body weight loss, (B) endoscopic and histological analyses before and after colitis induction in WT and TREM-2 KO mice, (C) quantification of mucosal injury by endoscopic and histological score. Data are representative of 3 independent experiments; *P ⬍ .05; **P ⬍ .01.

similar between TREM-2 KO and WT mice at baseline or after vehicle injection, and it significantly increased by 2-fold in WT compared with TREM-2 KO mice after TNBS treatment (Supplementary Figure 3B). Similar findings were detected in DSS-induced colitis (data not shown). In order to test whether the protective effect observed in TREM-2 KO mice was mediated through the enhancement of intestinal epithelial barrier, we used the Evans blue test to measure intestinal permeability in WT and TREM-2 KO mice before and after TNBS-induced colitis. At baseline, no alterations in barrier function were observed in the 2 groups. However, after induction of colitis, a significant increase in intestinal permeability was observed (24 hours and 72 hours) in WT compared with TREM-2 KO mice (Supplementary Figure 3C). Altogether, these data suggest that TREM-2 does not directly regulate intestinal epithelial turnover and barrier function under steady-state conditions, but that it might be indirectly

Because DCs are the main cell type expressing TREM-2 in the inflamed gut, we reasoned that both protection from colitis and impaired production of inflammatory cytokines at mucosal sites might be attributed to DC dysfunction in TREM-2 KO mice. It has recently been shown that TREM-2 regulates TLR-induced inflammatory cytokine production in DCs.21 To assess the role of TREM-2 in cytokine production by DC, we generated DCs from either the BM or LP of WT and TREM-2 KO mice. LP-DCs were stimulated with either LPS (a TLR4 ligand), CpG (a TLR9 ligand), or muramyl dipeptide (a NOD2 ligand). Interestingly, no difference in TNF-␣ production was found in WT and TREM-2 KO LP-DC in response to LPS (Figure 6A). By contrast, LPS significantly stimulated the secretion of IL-1␤, IL-6, and IL-12p70 in WT compared with TREM-2 KO LP-DC (Figure 6B–D). Similarly, stimulation with CpG significantly up-regulated IL-1␤, IL-6, and IL-12p70 production, and TNF-␣ release was comparable in WT and TREM-2 KO cells (Figure 6A–D). In addition, in the absence of TREM-2, NOD2 stimulation did not translate into increases in TNF-␣, IL-12p70, and IL-6 (Figure 6A–D). We then evaluated the effect of additional TLR and NOD-like receptor (NLR) ligands on LP-DCs by stimulating with Pam3CSK4 (a TLR1/2 ligand), poly (I:C) (a TLR3 ligand), flagellin (a TLR5 ligand), and Pam2CSK4 (a TLR2/6 ligand), but no major differences in cytokine production profile were observed (data not shown). Finally, we extended our analyses to BM-DCs that were stimulated with different concentrations of LPS, CpG, Pam3CSK4, poly (I:C), flagellin, Pam2CSK4, and muramyl dipeptide. Notably, according to the ligand and cytokine analyzed, differences in levels of IL-1␤, IL-6, IL-12p70, and TNF-␣ were detected between BM-DCs from WT and TREM-2 KO mice after stimulation, confirming the observation made in LP-DCs (Supplementary Figure 4A–D). Altogether, these data suggest that TREM-2 acts by amplifying cytokine induction by selected TLR and NOD2 ligands.

TREM-2 Impacts on Bacterial Handling by DCs and on T-Cell Activation in Response to Bacteria-Associated Antigens It has been shown that TREM-2 is involved in the recognition and phagocytosis of several species of bacteria.11 To investigate whether the loss of TREM-2 affects these functions, we determined the levels of bacteria in the spleen, mesenteric lymph nodes, and colon of WT and TREM-2 KO mice before and after colitis induction. The loss of TREM-2 significantly reduced the number of colony-forming units (CFUs) in the mesenteric lymph nodes, spleen, and colon with respect to WT mice after DSS induction, and CFUs were undetectable in both groups at

BASIC AND TRANSLATIONAL AT

TREM-2 Amplifies the Inflammatory Response Induced by TLR and NOD-Like Receptor Ligands

352

CORREALE ET AL

GASTROENTEROLOGY Vol. 144, No. 2

BASIC AND TRANSLATIONAL AT

Figure 5. TREM-2 KO mice produce low levels of metalloproteinase and proinflammatory cytokines upon DSS- and TNBS-induced colitis. WT littermates (black columns) and TREM-2 KO mice (gray columns) before and after DSS and TNBS treatment were tested for mucosal production of (A) IL-1␤, (B) TNF-␣, and (C) IL-10 by enzyme-linked immunosorbent assay. (D) MMP-3, (E) MMP-9, and (F) MMP-14 expressions were estimated by quantitative real-time polymerase chain reaction. Untreated mice for the DSS model and vehicle-treated mice for the TNBS model are indicated as NT and EtOH, respectively. *P ⬍ .05.

baseline (Supplementary Figure 5A–C). In addition, fluorescence in situ hybridization showed that the presence of bacteria within the colonic mucosa was comparable in WT and TREM-2 KO mice, not only at baseline but also after colitis induction, indicating that both the epithelial barrier function and bacterial uptake were intact in TREM-2 KO mice (Supplementary Figure 5D). To dissect the role of TREM-2 in bacterial handling, LP-derived DCs were infected for 1 hour with metabolically defective invasive and noninvasive strains of Salmonella Typhimurium and, after gentamycin treatment, viable intracellular bacteria CFUs were counted after 2 and 24 hours to evaluate bacterial phagocytosis and killing, respectively. We found that WT and TREM-2 KO LP-DCs did not differ in their ability to phagocytose invasive or noninvasive strains of Salmonella (Figure 7A), but they differed in their capacity to kill bacteria. Indeed, TREM-2 KO LP-DCs showed an

impaired killing of the invasive intracellular bacteria and this was reflected also by a decrease in IL-1␤ levels in the supernatant (Figure 7B). Next, we investigated whether the defect in bacterial killing of TREM-2 KO LP-DCs could affect their ability to present bacteria-associated antigens. To address this point, we first evaluated the ability of WT and TREM-2 KO LP-DCs to prime OT-II T cells in an antigen-specific T-cell response. WT and TREM-2 KO LP-DCs were loaded with ovalbumin (OVA) and then cocultured with antigen-specific, carboxyfluorescein succinimidyl ester (CFSE)-loaded OT-II naïve T cells for 4 days. OT-II T-cell priming and proliferation were evaluated by flow cytometry analysis in terms of CFSE dilution. As shown in Figure 7C, OVA-loaded WT and TREM-2 KO LP-DCs were equally able to prime OT-II T cells, as reflected by comparable percentages of CFSElow CD4⫹ T cells. Antigen presentation itself was apparently

TREM-2 IN IBD PATHOGENESIS

353

Figure 6. TREM-2 amplifies the inflammatory responses induced by TLR and NLR ligands. LP-DCs generated from WT (black columns) and TREM-2 KO (gray columns) mice were seeded at a density of 2.5 ⫻ 104/well and stimulated with the indicated concentrations of LPS, CpG, or muramyl dipeptide for 24 hours. Thereafter, levels of (A) TNF-␣, (B) IL-1␤, (C) IL-6, and (D) IL-12p70 were measured by enzyme-linked immunosorbent assay. Data are representative of 3 independent experiments and are shown as mean ⫾ standard deviation; *P ⬍ .05; **P ⬍ .01.

not altered in TREM-2 KO DCs. Subsequently, WT and TREM-2 KO LP-DCs were infected with invasive or noninvasive S typhimurium expressing the glutathione S-transferase–OVA fusion protein (SL-pOVA), or with the same strains expressing only glutathione S-transferase (SLpGEX), as a control culture condition (data not shown). CFSE-loaded OT-II naïve T cells were then added for 4 days. Infected TREM-2 KO LP-DCs displayed a reduced ability to prime OT-II T cells when compared with WT LP-DCs, both when loaded with invasive and noninvasive bacterial strains (Figure 7C), in good agreement with the defect in bacterial killing and capacity to process the relevant bacterial antigen. Similar results were obtained with BM-DCs (Supplementary Figure 6). Finally, WT and TREM-2 KO DCs were infected with invasive or noninvasive S typhimurium and CD86 expression was measured by flow cytometry. TREM-2 KO BMDCs expressed constitutively lower levels of CD86 compared with WT. This difference, although not statistically significant, was maintained also after salmonella infection (data not shown). In line with BM-DCs, TREM-2 KO LP-DCs showed lower constitutive CD86 expression compared with WT, although this failed to achieve statistical significance. Interestingly, both KO and WT LP-DCs responded to salmonella infection by maturing into 2 CD86 subpopulations. Although the percentage of CD86dim/int cells was

similar between the 2 groups, TREM-2 KO LP-DCs displayed a reduced CD86high subpopulation in contrast to WT DCs, indicating that although responsive to salmonella infection, in the absence of TREM-2 LP-DCs undergo less maturation (Supplementary Figure 7). Altogether, these data indicate that the loss of TREM-2 reduces DCs’ ability to induce bacteria-associated, antigen-specific T-cell proliferation, affecting the proinflammatory immune response.

Discussion The pathogenesis of IBD is not completely understood, although it is increasingly appreciated that the innate immune system plays a crucial role.22,23 Loss of tolerance toward the commensal flora appears to be a key pathogenic event, and the molecules involved in sensing the luminal bacterial content, such as NOD-2 and several TLRs, are essential for gut homeostasis. Mutations or polymorphisms in these molecules have been associated with susceptibility to IBD. Other innate immune receptors, such as TREMs, modulate the innate response by either amplifying or dampening TLR-induced signals and orchestrating the fine-tuning of the inflammatory response.24 TREMs are type I membrane proteins with an extracellular immunoglobulin-like domain and a short cytoplasmic tail that has no intrinsic signaling capacity.

BASIC AND TRANSLATIONAL AT

February 2013

354

CORREALE ET AL

BASIC AND TRANSLATIONAL AT

Figure 7. TREM-2 KO DCs are defective in bacterial killing and in CD4⫹ T-cell priming after bacterial infection. LP-DCs generated from WT (black columns) and TREM2 KO mice (gray columns) were seeded at a density of 2.5 ⫻ 104/well and infected with invasive or noninvasive S Typhimurium for 1 hour and subsequently incubated in gentamycin-supplemented medium for 2 and 24 hours to evaluate bacterial phagocytosis and killing, respectively. Cells were lysed and seeded on Luria broth agar plates and CFUs were counted (A). After 24 hours, supernatants were collected and IL-1␤ was measured by enzyme-linked immunosorbent assay (B). Antigen presentation and CD4⫹ T-cell stimulation were evaluated (C). CD11c⫹ LP-derived DC from WT and TREM-2 KO mice were cocultured for 4 days with CFSE-labeled CD4⫹ naïve OT-II T cells in the presence or in absence of OVA, and with or without infection with invasive or noninvasive S Typhimurium. After coculture, CFSE dilution in CD4⫹ OT-II T cells was detected by flow cytometry. Data are representative of 3 independent experiments and are shown as mean ⫾ standard deviation; *P ⬍ .05.

TREM signaling relies on the association with DAP12, a cytosolic adapter that also associates with other receptors, and leads to an increase of intracellular calcium and

GASTROENTEROLOGY Vol. 144, No. 2

phosphorylation of ERK1/2. To date, TREM-2 ligand(s) have not been identified thoroughly.24 It has been shown that TREM-2 binds to a wide variety of bacteria10 and that TREM-2-DAP12 functions as a phagocytic receptor for bacteria.11 Interestingly, DAP12⫺/⫺ macrophages display a reduced internalization of bacteria by both nonphagocytic cells and professional phagocytes, but not a reduced binding.11 In addition, TREM-2 was still expressed in the DAP12⫺/⫺ macrophages and reintroduction of either DAP12 alone or TREM-2-DAP12 into DAP12⫺/⫺ macrophages restored phagocytosis. Cross-linking of TREM-2 on immature DC translates into the up-regulation of costimulatory molecules as well as the chemokine receptor CCR7.15 Previous studies have suggested that TREM-2 might be involved in the regulation of inflammatory and immune responses.25 In experimental autoimmune encephalomyelitis, a murine model of human multiple sclerosis, TREM-2 expression is increased and its neutralization worsens the disease.7 In addition, TREM-2 and its adaptor DAP-12 have been shown to regulate alveolar macrophage chemotaxis and recruitment to the lung, both physiologically and in disease states such as emphysema. In IBD, TREM-2 expression and function had not been investigated yet. We first assessed the expression of TREM-2 in mucosal sections of patients with CD and UC and in healthy controls. Consistent with other forms of tissue chronic inflammation, we found that TREM-2 is up-regulated in human IBD. The cell types responsible for TREM-2 expression in the LP of both CD and UC patients were mainly DC, as indicated by confocal microscopy and flow cytometry. Growing evidence now supports the view of a pathogenic role of DCs, either in the establishment or in the maintenance of colitis.16,23 However, data on human DC in the IBD setting are still very limited. In this study, we showed that DC infiltrating the inflamed mucosa of IBD patients express high levels of TREM-2, suggesting that they might play a role in IBD pathogenesis. The molecular determinants of TREM-2 induction are presently unknown. We tested whether mediators that positively or negatively affect inflammatory responses, such as TNF-␣, IL-10, LPS, IL-4, IL-13, and interferon gamma, can affect the expression of TREM-2. However, none of these stimuli were capable of inducing TREM-2 expression, at least in vitro. Consistent with the ex vivo findings, TREM-2 expression increased on colitis induction and, surprisingly, TREM-2 up-regulation correlated with the onset of disease. Conceivably, TREM-2 induction might represent a physiological response aimed at reducing inflammation in the gut. Alternatively, TREM-2 overexpression could be part of the disease itself, contributing to pathology establishment. In order to answer this fundamental question and to gain insights into the functional role of TREM-2 in intestinal inflammation, we took advantage of TREM-2 KO mice. We found that, compared with WT animals, TREM-2 KO mice were protected from acute and chronic DSS or TNBS-induced colitis, showing only minimal endoscopic and histological inflammation,

February 2013

355

be beneficial to dampen intestinal inflammation through effects on the innate immune response.

Supplementary Materials Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at www.gastrojournal.org, and at http:// dx.doi.org/10.1053/j.gastro.2012.10.040. References 1. Bouchon A, Dietrich J, Colonna M. Cutting edge: inflammatory responses can be triggered by TREM-1, a novel receptor expressed on neutrophils and monocytes. J Immunol 2000;164:4991– 4995. 2. Klesney-Tait J, Turnbull IR, Colonna M. The TREM receptor family and signal integration. Nat Immunol 2006;7:1266 –1273. 3. Daws MR, Lanier LL, Seaman WE, et al. Cloning and characterization of a novel mouse myeloid DAP12-associated receptor family. Eur J Immunol 2001;31:783–791. 4. Turnbull IR, Gilfillan S, Cella M, et al. Cutting edge: TREM-2 attenuates macrophage activation. J Immunol 2006;177:3520 –3524. 5. Cella M, Buonsanti C, Strader C, et al. Impaired differentiation of osteoclasts in TREM-2-deficient individuals. J Exp Med 2003;198: 645– 651. 6. Kiialainen A, Hovanes K, Paloneva J, et al. Dap12 and Trem2, molecules involved in innate immunity and neurodegeneration, are co-expressed in the CNS. Neurobiol Dis 2005;18:314 –322. 7. Piccio L, Buonsanti C, Mariani M, et al. Blockade of TREM-2 exacerbates experimental autoimmune encephalomyelitis. Eur J Immunol 2007;37:1290 –1301. 8. Schmid CD, Sautkulis LN, Danielson PE, et al. Heterogeneous expression of the triggering receptor expressed on myeloid cells-2 on adult murine microglia. J Neurochem 2002;83:1309 –1320. 9. Takahashi K, Rochford CD, Neumann H. Clearance of apoptotic neurons without inflammation by microglial triggering receptor expressed on myeloid cells-2. J Exp Med 2005;201:647– 657. 10. Daws MR, Sullam PM, Niemi EC, et al. Pattern recognition by TREM-2: binding of anionic ligands. J Immunol 2003;171:594 –599. 11. N’Diaye EN, Branda CS, Branda SS, et al. TREM-2 (triggering receptor expressed on myeloid cells 2) is a phagocytic receptor for bacteria. J Cell Biol 2009;184:215–223. 12. Paloneva J, Mandelin J, Kiialainen A, et al. DAP12/TREM2 deficiency results in impaired osteoclast differentiation and osteoporotic features. J Exp Med 2003;198:669 – 675. 13. Helming L, Tomasello E, Kyriakides TR, et al. Essential role of DAP12 signaling in macrophage programming into a fusion-competent state. Sci Signal 2008;1:ra11. 14. Hamerman JA, Jarjoura JR, Humphrey MB, et al. Cutting edge: inhibition of TLR and FcR responses in macrophages by triggering receptor expressed on myeloid cells (TREM)-2 and DAP12. J Immunol 2006;177:2051–2055. 15. Bouchon A, Hernandez-Munain C, Cella M, et al. A DAP12-mediated pathway regulates expression of CC chemokine receptor 7 and maturation of human dendritic cells. J Exp Med 2001;194:1111–1122. 16. Rescigno M, Di Sabatino A. Dendritic cells in intestinal homeostasis and disease. J Clin Invest 2009;119:2441–2450. 17. Strober W, Kitani A, Fuss I, et al. The molecular basis of NOD2 susceptibility mutations in Crohn’s disease. Mucosal Immunol 2008;1(Suppl 1):S5–S9. 18. Xavier RJ, Podolsky DK. Unravelling the pathogenesis of inflammatory bowel disease. Nature 2007;448:427– 434. 19. Zhang H, Massey D, Tremelling M, et al. Genetics of inflammatory bowel disease: clues to pathogenesis. Br Med Bull 2008; 87:17–30. 20. Danese S, Sans M, de la Motte C, et al. Angiogenesis as a novel component of inflammatory bowel disease pathogenesis. Gastroenterology 2006;130:2060 –2073.

BASIC AND TRANSLATIONAL AT

coupled with moderate induction of proinflammatory cytokines and reduced levels of MMPs. In addition, the lower levels of proinflammatory cytokines in the mucosa did not affect either barrier permeability or intestinal epithelial turnover. These findings unveil a unique function for TREM-2 in the gut compared with other tissues, and support our contention that TREM-2 can amplify inflammation rather than negatively regulating inflammatory responses, as described. Consistent with this hypothesis, in vitro studies demonstrated that TREM-2 expression on DC, together with TLR and NLR ligation, resulted in the amplification of the inflammatory response. Indeed, the absence of TREM-2 on DC eminently reduced proinflammatory cytokine production in response to several TLR and NLR ligands. This is the first evidence of a synergic interaction between the TREM-2 and NOD2 pathways. Because NOD2 signaling appears to be crucial to the pathogenesis of CD, it is tempting to speculate that, in IBD patients, the inflammatory process ignited by NOD2 dysfunction could be magnified by TREM-2 expression. A recent study reported that TREM-2 acts as a negative regulator of TLR-induced inflammatory cytokine production in DC.21 In particular, cytokine secretion by TREM-2 KO DC was abundant after stimulation with very low doses of ligands for LPS. We also found that LPS doses ⬍1 ng/mL induced a slightly higher release of TNF-␣ by TREM-2 KO DC. However, the major differences were observed at higher doses, which induced a stronger response in WT compared with TREM-2 KO DCs. In addition, it has been shown that TREM-2 is an innate immune receptor involved in recognition and phagocytosis of several bacterial species.11 We observed that loss of TREM-2 did not affect the phagocytic process of invasive and noninvasive Salmonella by DCs, but rather it led to impaired killing of the invasive strain. This correlated with defects in DC maturation and with the induction of antigen-specific T-cell proliferation in response to bacterial antigens. Although previous studies have indicated that depletion of TREM-2 on macrophages inhibited both binding and uptake of bacteria,11 these functions were not investigated on DC. It is likely that TREM-2 plays different functions depending on the phagocyte subset analyzed. Collectively, our data provide evidence that TREM-2 acts as an amplifier of the immune response. Resistance to experimental colitis in TREM-2 KO mice corroborates this multifaceted function of TREM-2 in the intestine. The mechanisms underpinning the protective phenotype in TREM-2 KO mice are related to reduced cytokine induction by TLR and NLR ligation, which ultimately translates into reduced barrier permeability and bacterial translocation from the gut lumen. In addition, TREMdeficient DCs manifested a defect in bacterial killing, with reduced ability to present bacterial antigens and induce antigen-specific T-cell proliferation. This can further reduce gut pathology driven by bacterial triggers because T cells are not activated. The results presented here point to the bacterial sensor TREM-2 as a novel molecule involved in IBD pathogenesis. Targeting TREM-2 in the gut could

TREM-2 IN IBD PATHOGENESIS

356

CORREALE ET AL

21. Ito H, Hamerman JA. TREM-2, triggering receptor expressed on myeloid cell-2, negatively regulates TLR responses in dendritic cells. Eur J Immunol 2012;42:176 –185. 22. Kaser A, Zeissig S, Blumberg RS. Inflammatory bowel disease. Annu Rev Immunol;28:573– 621. 23. Rutella S, Locatelli F. Intestinal dendritic cells in the pathogenesis of inflammatory bowel disease. World J Gastroenterol 2011;17:3761– 3775. 24. Ford JW, McVicar DW. TREM and TREM-like receptors in inflammation and disease. Curr Opin Immunol 2009;21:38 – 46. 25. Sharif O, Knapp S. From expression to signaling: roles of TREM-1 and TREM-2 in innate immunity and bacterial infection. Immunobiology 2008;213:701–713.

GASTROENTEROLOGY Vol. 144, No. 2 Research Center, Via Manzoni 113, 20089, Rozzano, Milan, Italy. e-mail: [email protected]; fax: ⴙ39 02 822 45101. Acknowledgments The authors are indebted to Bioxell-Cosmo Pharmaceuticals (Milan, Italy) for providing TREM-2 KO mice and TREM-2 reagents. The authors thank Achille Anselmo and Chiara Buracchi for providing assistance in revising flow cytometry data, and Monica Rimoldi for her supervision on the experiments with the gentamycin protection assay. Conflicts of interest All authors disclose no conflicts.

Received August 11, 2011. Accepted October 24, 2012. Reprint requests Address requests for reprints to: Silvio Danese, MD, PhD, IBD Center, Division of Gastroenterology, Humanitas Clinical and

Funding Innovative Medicines Initiative IMI-funded project “BeTheCure” (#115142-2) and the Broad Medical Research Program (BMRP-IBD0345R2), The Ely and Edith Broad Foundation.

BASIC AND TRANSLATIONAL AT

February 2013

Supplementary Materials and Methods Isolation of Human LP Mononuclear Cells LP mononuclear cells (MCs) were isolated from intestinal tissues as reported.1 Briefly, strips of mucosa were cut into small pieces and incubated for 30 minutes in constant agitation in Hank’s Balanced Salt Solution without calcium and magnesium, supplemented with 2.5% fetal calf serum (FCS), 0.5 mM EDTA, and antibiotic solution 1⫻ (penicillin, streptomycin, and gentamycin; BioWhittaker, Cambrex, East Rutherford, NJ). After extensive washing, tissues were digested for 60 minutes in RPMI medium supplemented with 5% FCS, 10 mM HEPES, 0.75 mg/mL collagenase type II, and 20 ␮g/mL DNase I (both from Worthington Biochemical Corporation, Lakewood, NJ). DCs were enriched through a Ficoll gradient followed by a 46% Percoll gradient.

Immunofluorescence Staining As reported previously,2 4-␮M frozen sections of human colonic mucosa were fixed in cold acetone (10 minutes at ⫺20°C). Sections were then blocked with phosphate-buffered saline (PBS) containing 2% bovine serum albumin for 60 minutes at room temperature and incubated with the following primary antibodies: antiCD11c (1:50 dilution; BD Biosciences, Mountain View, CA) with anti⫺TREM-2 (1:50 dilution; R&D Systems, Oxon, UK) for 1 hour at room temperature. Alexa Fluor 488⫺conjugated goat anti-rabbit (for TREM-2 detection) and/or Alexa Fluor 594⫺conjugated goat anti-mouse (for CD11c) antibodies (all from BD Biosciences) were used as secondary antibodies (1:1000 dilution; 30 minutes at room temperature), followed by incubation with 0.2 ␮g/mL 4=,6-diamidino-2-phenylindole (Invitrogen, Carlsbad, CA) for 10 minutes at room temperature. Sections were mounted with ProLong antifade medium (Molecular Probes, Grand Island, NY) and analyzed with a laserscanning confocal microscope (FluoView FV1000; Olympus Italia, Milan, Italy). Images were acquired with an oil immersion objective (60⫻, 1.4 NA Plan-Apochromat; Olympus Italia).

Flow Cytometry The following fluorochrome-conjugated antibodies were used for flow cytometry staining: peridinin-chlorophyllprotein⫺conjugated anti-human CD45; phycoerythrin-conjugated anti-human CD3; and allophycocyanin-conjugated antihuman CD11c (all from BD Biosciences). Goat anti-human TREM-2 (R&D Systems) and fluorescein isothiocyanate⫺conjugated anti-goat IgG secondary antibodies were used for the detection of TREM-2. In selected experiments, murine LP-DCs were analyzed for CD86 expression. The following fluorochrome-conjugated antibodies were used for fluorescence-activated cell-sorted staining: peridininchlorophyll-protein⫺conjugated anti-mouse CD45; allophycocyanin-conjugated anti-mouse CD11c; phycoerythrin-

TREM-2 IN IBD PATHOGENESIS

356.e1

conjugated anti-mouse major histocompatibility complex⫺II; and fluorescein isothiocyanate⫺conjugated anti-mouse CD86 (all from BD Bioscences). Fluorochrome-conjugated, isotypematched antibodies were included as negative controls.

Analyses of TREM-2 Modulation in Human LP-MCs LP-MCs were isolated as described here from nonactively and actively inflamed intestinal specimens obtained from UC and CD patients undergoing surgery or colonoscopy. Then 2 ⫻ 106 LP-MCs were plated in RPMI complete medium and incubated for 18 hours with or without the following ligands: 1 ␮g/mL LPS (Sigma Chemical Co., St Louis, MO), IL-4, TNF-␣, IL-13, interferon gamma, and IL-10 (all at 50 ng/mL; purchased from R&D Systems). At the end of treatment, cells were washed and used for flow cytometry analyses. DCs were identified based on their CD45⫹CD3⫺CD11c⫹ phenotype and TREM-2 expression was evaluated as already described.

Mouse Models of Acute and Chronic Colitis For DSS-induced colitis, mice received 3% DSS in drinking water ad libitum for 8 days. For TNBS-induced colitis, mice were anesthetized with intraperitoneal injection of avertin (2.5%) and treated with an intra-rectal injection of 3 mg TNBS in 20% ethanol (vol/vol) or vehicle alone as a control. After instillation, mice were held in an upright position for 60 seconds to avoid reflux. For the chronic colitis model, all mice were treated with 3 cycles of 2% DSS for 5 days, followed by a 10-day recovery period. For all models, weight loss, stool consistency, and presence of fecal occult blood were monitored daily and used to calculate the disease activity index, as described.2 At the indicated time points, mice were sacrificed, colons were excised, and colon length from the end of the cecum to the anus was recorded. Colons of colitic and healthy mice were cut longitudinally into 2 parts after washing with PBS. Half of the colon was weighed and used for further analysis. The other half was fixed in 4% formaldehyde/PBS and included in paraffin and 2-␮M sections were cut and stained with H&E. Degree of inflammatory cell infiltration and mucosal damage was evaluated as described previously.3⫺5

Isolation and Stimulation of Murine DCs Intestinal DC isolation. Briefly, a longitudinally opened small intestine was cut in 1-cm⫺long pieces after removal of Peyer’s Patches. After repeated washings with PBS, tissues were digested in GlutaMAX MEM␣ medium (Invitrogen, Milan, Italy) supplemented with 5% FCS, 0.5 mg/mL collagenase type VIII (Sigma-Aldrich, Milan, Italy), 5 U/mL DNase I (Roche Diagnostics, Milan, Italy), and antibiotic solution 1⫻ for 30 minutes at 37°C by gentle shaking. Cells were resuspended after smashing in

356.e2

CORREALE ET AL

40% Percoll (GE Healthcare, Milan, Italy) and overlaid onto 70% Percoll. MCs were collected at 40%/70% interface and red blood cells lysed. Intestinal DCs were enriched by anti-CD11c MACS beads (Miltenyi Biotec, Bergisch Gladbach, Germany).6 Isolation of BM-DCs. For BM-DC isolation, BM of both WT and KO mice was flushed out and the cell suspension obtained was cultured for 8 days in complete RPMI 1640 medium (GIBCO, Gaithersburg, MD) supplemented with 30% supernatant from granulocyte macrophage colony-stimulating factor⫺producing NIH-373 cells to obtain mature BM-DCs. LP-DCs and BM-DCs were stimulated with different doses of LPS (ng/mL), CpG (␮M), Pam3CSK4 (ng/mL), Pam2CSK4 (ng/mL), poly (IC) (ng/mL), flagellin (ng/mL), and with 50 ␮g/mL muramyl dipeptide. RNA isolation and quantitative real-time polymerase chain reaction. At the indicated time points, mice

were sacrificed and their colons were excised from the end of the cecum to the anus and snap frozen. Total RNA was extracted using RNeasy Mini kit (Qiagen, Milan, Italy), treated with DNase I (Qiagen) and retro-transcribed with Reverse Transcription Reagent (Applied Biosystems, Milan, Italy) and random primers. Real-time polymerase chain reaction analysis was performed using commercially available actin VIC-conjugated probe and TREM-2, MMP-3, MMP-9, and MMP-14 FAM-conjugated probes (all from Applied Biosystems). Real-time polymerase chain reaction was performed with an ABI PRISM 7700 analyzer (Applied Biosystems) following manufacturer’s instruction. Results were normalized vs mouse actin housekeeping gene. Fold differences in gene expression were calculated using the 2⫺⌬Ct method.

Western Blot Studies Protein lysates were obtained from colon of both WT and KO mice, treated as described in Material and Methods. Briefly, colons excised from the end of the cecum to the anus were weighed and resuspended with 20 ␮L/mg precooled lysis buffer (Tris-HCl 10 mM [pH 7.4], EDTA 1 mM [pH 8], NaCl 150 mM, Triton X-100 0.1%). Tissue samples were then disrupted for 2 ⫻ 90 seconds at 25 Hz using the TissueLyser II (Qiagen). Protein were clarified by centrifugation (13,000 rpm for 10 minutes at 4°C) and supernatant quantified by Bradford Assay (Bio-Rad). Normalized proteins were subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis in 10% polyacrylamide gel, transferred onto nitrocellulose membranes and immunoblotted with rat antimouse TREM-2 antibody (R&D Systems). The filter was then stripped with buffer Restore (Pierce, Rockford, IL) and reprobed with an anti–actin antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) for control protein loading. Evaluation of cytokine secretion. One half of the colon was washed in cold PBS supplemented with anti-

GASTROENTEROLOGY Vol. 144, No. 2

biotic solution 1⫻ (BioWhittaker, Cambrex) and incubated in RPMI 1640 medium with 0.1% FCS, and antibiotic solution at 37°C in 5% CO2. After 24 hours, supernatant was collected, centrifuged, and stored at ⫺80°C until analyzed. Supernatants were assayed for TNF-␣, IL-1␤, and IL-10 in duplicate using commercially available enzyme-linked immunosorbent assay kits, as reported previously (R&D Systems). Supernatants from either LP- or BM-DCs stimulated as stated here were evaluated for IL-6, IL-12p70, TNF-␣, and IL-1␤ levels by enzyme-linked immunosorbent assay following the manufacturer’s instructions (R&D Systems).

Bacterial Strains and Gentamycin Protection Assay A metabolically defective invasive strain (aroA) and a metabolically defective noninvasive strain (aroAinvA) of S Typhimurium on SL3261 background were used. Recombinant strains expressing the gene coding for glutathione S-transferase ovalbumin fusion (SL-OVA) or with the glutathione S-transferase alone (SL-pGEX) were generated. Single colonies were grown overnight at 37°C in Luria broth supplemented with 100 ␮g/mL ampicillin diluted the next day at 1:10 of the original volume and grown until they reached OD600 ⫽ 0.6, which corresponds to 8 ⫻ 108 colony-forming units/mL. OVA expression was induced by the addition of 0.1 mmol/L isopropyl-L-thio-B-Dgalactopyranoside. For the Gentamycin Protection Assay, 2.5 ⫻ 104 WT or TREM2 KO LP- or BM-DCs were infected at multiplicity of infection 1:50 for 1 hour in RPMI 1640 medium (supplemented with 10% FCS, 1% glutamine without antibiotics) and then incubated with 100 ␮g/mL gentamycin for 2 and 24 hours (when supernatant was also collected). Cells were subsequently lysed in PBS containing 0.5% sodium deoxycholate and seeded on Luria broth agar plates. After overnight incubation at 37°C, bacterial colonies (colony-forming units) were counted as a measure of intracellular bacteria.

DC-T Cells Coculture Then 2.5 ⫻ 104 WT or TREM2 KO LP- or BMDCs were loaded with 250 ␮g/mL OVA (Sigma-Aldrich) or infected with the invasive or noninvasive strains of S Typhimurium, OVA- (SL-OVA) or glutathione S-transferase⫺expressing (SL-pGEX) at MOI 1:50 for 1 hour. LP- or BM-DCs were then incubated in complete RPMI 1640 medium supplemented with 100 ␮g/mL gentamycin for 10 hours and subsequently cocultured with CFSE-labeled CD4⫹ CD25⫺ T cells purified according to manufacturer’s instructions (Miltenyi Biotec) from spleens of OTII mice. After 4 days, cells were stained and T cell proliferation was evaluated as CFSE dilution.

February 2013

Evaluation of Proliferation and Apoptosis Index To evaluate proliferation, 4-␮M paraffin sections were stained with anti-mouse Ki-67 polyclonal antibodies (1:1000; Abcam) at room temperature for 1 hour. Ki67⫺positive cells were detected using a horseradish peroxidase⫺conjugated polymer system. 3,3=-diaminobenzidine tetra hydrochloride was used as the chromogen. Colonic sections from both WT and TREM-2 KO mice were counterstained with hematoxylin and mounted. Proliferation was evaluated as the number of Ki-67⫹ cells/crypt. For the analyses of the apoptosis index, the terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling assay was performed on 4-␮M cryostatic sections using the In Situ Cell Death detection Kit Fluorescein (Roche, Mannheim, Germany) according to manufacturer’s instructions. For the detection of nuclei, 4=,6-diamidino-2-phenylindole counterstaining was performed in parallel to terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling, as reported previously. Apoptosis was quantified as the number of apoptotic cells over 60 different fields.

Analysis of Colonic Epithelial Permeability by Evans Blue Assay Epithelial permeability was evaluated in colon tissue by penetrance of Evans blue from the lumen into the wall of the colon. Briefly, mice were anesthetized and subjected to surgical procedures. A fine catheter was introduced in the proximal colon, which was incised at the junction with the cecum and into the distal rectum via the anus. The proximal colon and the distal rectum were then ligated and washed with saline and with 6 mM N-acetylcysteine at 0.5 mL/min. A total volume of 5 mL Evans Blue (0.4%) was then perfused from the cecum to the anus in 10 to 15 minutes. Mice were then scarified and colon excised rapidly and weighted. Colons were washed with saline and placed in 15 mL dimethyl-formamide at 25°C for 18 hours. Gut permeability was expressed as absorbance at 620 nm/g tissue.

Fluorescence In Situ Hybridization Probe EUB 338, which is complementary to a portion of the 16S ribosomal RNA gene conserved in the domain bacteria, was used to visualize the entire bacterial population in the colon specimens. The probe was synthesized commercially and 5= end-labeled with fluorochrome Cy3 (Invitrogen), giving a bright orange signal. The hybridization buffer contained 0.9 M NaCl, 20 mM

TREM-2 IN IBD PATHOGENESIS

356.e3

Tris/HCl (pH 7.3), and 0.01% sodium dodecyl sulfate. Prewarmed hybridization buffer (20 ␮L) was mixed with approximately 5 pmol of the oligonucleotide probe and carefully applied to the tissue sections. After incubation for 3 to 5 hours in a dark humid chamber at 46°C, each of the slides was rinsed, counterstained with 4=,6-diamidino-2-phenylindole, and mounted. Assessment of bacterial growth from mesenteric lymph nodes, spleens, and colons. Mesenteric lymph

nodes, spleens, and colons were collected from WT and KO mice after 8 days of 3% DSS treatment (mesenteric lymph nodes, spleens, and colons were also collected from untreated mice) and mechanically disaggregated over a 70-␮M pore nylon mesh cell strainer (BD) using a 5-mL syringe pestle to obtain single-cell suspensions. Cell strainers were washed twice with Iscove’s modified Dulbecco’s medium and cell suspensions were centrifuged and cells were counted using Trypan blue. Cells were resuspended in PBS containing 0.5% sodium deoxycholate and seeded on Luria broth agar plates in the absence of antibiotics at escalating numbers (from 2 to 0.25 ⫻ 106). Plates were incubated overnight at 37°C and bacterial colony growth was measured (CFUs).7 References 1. Danese S, Sans M, Scaldaferri F, et al. TNF-alpha blockade downregulates the CD40/CD40L pathway in the mucosal microcirculation: a novel anti-inflammatory mechanism of infliximab in Crohn’s disease. J Immunol 2006;176:2617–2624. 2. Vetrano S, Rescigno M, Cera MR, et al. Unique role of junctional adhesion molecule—a in maintaining mucosal homeostasis in inflammatory bowel disease. Gastroenterology 2008;135:173– 184. 3. Danese S, Scaldaferri F, Vetrano S, et al. Critical role of the CD40 CD40-ligand pathway in regulating mucosal inflammation-driven angiogenesis in inflammatory bowel disease. Gut 2007;56:1248 – 1256. 4. Dieleman LA, Palmen MJ, Akol H, et al. Chronic experimental colitis induced by dextran sulphate sodium (DSS) is characterized by Th1 and Th2 cytokines. Clin Exp Immunol 1998;114:385–391. 5. McCafferty DM, Miampamba M, Sihota E, et al. Role of inducible nitric oxide synthase in trinitrobenzene sulphonic acid induced colitis in mice. Gut 1999;45:864 – 873. 6. Matteoli G, Mazzini E, Iliev ID, et al. Gut CD103⫹ dendritic cells express indoleamine 2,3-dioxygenase which influences T regulatory/T effector cell balance and oral tolerance induction. Gut 2010; 59:595– 604. 7. Rotta G, Matteoli G, Mazzini E, et al. Contrasting roles of SPARCrelated granuloma in bacterial containment and in the induction of anti-Salmonella typhimurium immunity. J Exp Med 2008;205:657– 667.