Gastroenterology
Volume 132, Issue 3 , Pages 994-1008, March 2007

Inhibition of TGF-β Signaling by IL-15: A New Role for IL-15 in the Loss of Immune Homeostasis in Celiac Disease

  • Mélika Benahmed

      Affiliations

    • INSERM U793, Paris, France
    • Université Paris Descartes, Faculté de Médecine René Descartes, Paris, France
    • M. Ben Ahmed was supported by INSERM, IRMAD, and AFDIAG (Association Française des Intolérants au Gluten).
    • ,
  • Bertrand Meresse

      Affiliations

    • INSERM U793, Paris, France
    • Université Paris Descartes, Faculté de Médecine René Descartes, Paris, France
    • ,
  • Bertrand Arnulf

      Affiliations

    • Université Paris Descartes, Faculté de Médecine René Descartes, Paris, France
    • CNRS UMR 8147, Paris, France
    • Department of Hematology, AP-HP, Hôpital Necker-Enfants Malades, Paris, France
    • ,
  • Ullah Barbe

      Affiliations

    • INSERM U793, Paris, France
    • Université Paris Descartes, Faculté de Médecine René Descartes, Paris, France
    • ,
  • Jean–Jacques Mention

      Affiliations

    • INSERM U793, Paris, France
    • Université Paris Descartes, Faculté de Médecine René Descartes, Paris, France
    • ,
  • Virginie Verkarre

      Affiliations

    • INSERM U793, Paris, France
    • Université Paris Descartes, Faculté de Médecine René Descartes, Paris, France
    • Department of Pathology, AP-HP, Hôpital Necker-Enfants Malades, Paris, France
    • ,
  • Matthieu Allez

      Affiliations

    • Department of Gastroenterology, AP-HP, Hôpital Saint Louis, Paris, France
    • ,
  • Christophe Cellier

      Affiliations

    • INSERM U793, Paris, France
    • Université Paris Descartes, Faculté de Médecine René Descartes, Paris, France
    • Department of Gastroenterology, AP-HP, Hôpital Européen Georges Pompidou, Paris, France
    • ,
  • Olivier Hermine

      Affiliations

    • Université Paris Descartes, Faculté de Médecine René Descartes, Paris, France
    • CNRS UMR 8147, Paris, France
    • Department of Hematology, AP-HP, Hôpital Necker-Enfants Malades, Paris, France
    • ,
  • Nadine Cerf–Bensussan

      Affiliations

    • INSERM U793, Paris, France
    • Université Paris Descartes, Faculté de Médecine René Descartes, Paris, France
    • Corresponding Author InformationAddress requests for reprints to: Nadine Cerf-Bensussan, MD, PhD, INSERM U793, Faculté de Médecine René Descartes-Paris 5, 156 rue de Vaugirard, 75730 Paris Cedex 15, France. fax: (33) 1-40-61-56-38.

Received 15 November 2005; accepted 27 November 2006. published online 19 December 2006.

Article Outline

Background & Aims: Interleukin (IL)-15 delivers signals that drive chronic inflammation in several diseases, including celiac disease. Smad3–transforming growth factor-beta (TGF-β) signaling is instrumental to counteract proinflammatory signals and maintain immune homeostasis. Our goal has been to investigate why the proinflammatory effects of IL-15 cannot be efficiently controlled by TGF-β in celiac disease. Methods: The impact of IL-15 on TGF-β signaling in T cells and in the intestinal mucosa of celiac disease patients was analyzed by combining cell and organ cultures, immunohistochemistry, flow cytometry, real-time polymerase chain reaction, electromobility gel shift, and Western blot. Results: IL-15 impaired Smad3-dependent TGF-β signaling in human T lymphocytes downstream from Smad3 nuclear translocation. IL-15-mediated inhibition was associated with a long-lasting activation of c-jun-N-terminal kinase and reversed by c-jun antisense oligonucleotides, consistent with the demonstrated inhibitory effect of phospho-c-jun on the formation of Smad3–DNA complexes. In active celiac disease, intestinal lymphocytes showed impaired TGF-β–Smad3-dependent transcriptional responses and up-regulation of phospho-c-jun. Anti-IL-15 antibody and c-jun antisense both downmodulated phospho-c-jun expression and restored TGF-β–Smad-dependent transcription in biopsies of active celiac disease. c-jun antisense decreased interferon gamma transcription. Conclusions: Impairment of TGF-β-mediated signaling by IL-15 might promote and sustain intestinal inflammation in celiac disease. More generally, our data provide a new rationale for the potent proinflammatory effects of IL-15, and further support the concept that IL-15 is a meaningful therapeutic target in inflammatory diseases associated with irreducible elevation of IL-15.

Abbreviations used in this paper: CD, celiac disease, HLA, human leukocyte antigen, IEL, intraepithelial lymphocytes, IL, interleukin, JNK, c-jun-N-terminal kinase, LPL, lamina propria lymphocytes, NK, natural killer, PAI, plasminogen activator inhibitor, PBMC, peripheral blood mononuclear cells, SDS, sodium dodecyl sulfate, TGIF, TG-interacting factor, TNF-alpha, tumor necrosis factor-α

 

See editorial on page 1174.

Interleukin (IL)-15, a cytokine produced by many cell types, plays pleiotropic functions at the interface between innate and adaptive immunity. IL-15 is critical for the differentiation and/or homeostasis of several murine innate immune cell subsets, including natural killer (NK), NK/T, and CD8αα intraepithelial lymphocytes (IEL), as well as for the generation and maintenance of specific memory CD8 TCRαβ cells. In addition, IL-15 plays redundant functions with other cytokines to promote maturation of dendritic cells, proliferation of T and B cells, cytotoxicity of NK and CD8+ T cells, and production of proinflammatory cytokines.1 IL-15 may thus play an important role in the triggering and maintenance of immune responses to invading pathogens and in tumor immunosurveillance. Yet, due to its potent antiapoptotic and proinflammatory properties, its over expression at different sites in transgenic mice has severe deleterious consequences, causing benign and malignant proliferation of NK and CD8 T cells, multiorgan lymphocyte infiltration, and/or severe inflammatory lesions of the skin or the small intestine.2, 3 Multifaceted regulatory checkpoints, particularly at the level of IL-15 message translation, allow a tight control of IL-15 expression.4 Yet, disorders of IL-15 have been observed in association with several proinflammatory or autoimmune-related diseases in humans, including psoriasis and rheumatoid arthritis (where IL-15 was recently shown to be a relatable therapeutic target),5, 6 and celiac disease.

Celiac disease (CD) is a small intestinal enteropathy induced by cereal-derived prolamins (gluten) in genetically susceptible individuals. In the current view of the pathogenesis of CD, adaptive immunity plays a key role, accounting for the interplay between the triggering environmental factor, prolamins, and HLA-DQ2/8 haplotypes, the major genetic risk factor. Due to their content in proline and glutamine, some gluten peptides can be deamidated by tissue transglutaminase, the autoantibody target, and adopt a configuration that enables their binding into the peptide pocket of HLA-DQ2/8 molecules.7 Gluten peptides can subsequently be presented to lamina propria CD4+ T cells, triggering their activation and the release of interferon gamma (IFN-γ).8 Although this attractive scheme has considerably improved our understanding of CD pathogenesis, the fact that only a small subset of HLA-DQ2/8 individuals develop CD points to additional factors mandatory for disease development. We and others have provided evidence of the complementary role of IL-15. IL-15 is induced in the intestinal mucosa of CD patients by a peptide distinct from T cell epitopes, the peptide 31–43/49 common to the N-terminus of α-gliadins.9, 10 IL-15 orchestrates within the epithelium an abnormal immune response that induces the expansion of IEL and their cytotoxicity against enterocytes via interaction of the NKG2D innate receptor with its major histocompatibility complex class I-related epithelial ligands.9, 11, 12, 13 IL-15 may also act on dendritic cells and enhance presentation of gliadin epitopes to CD4+ T cells.10 However, it remains unclear why the potent immunoregulatory mechanisms that control intestinal immune responses to intraluminal antigens cannot prevent the inappropriate activation of lamina propria lymphocytes (LPL) and IEL in celiac patients. Notably, transgenic mice expressing human CD4 and HLA-DQ8 fail to develop enteropathy upon oral challenge with gliadins, despite a strong specific CD4+ T cell response.14 A striking feature of this mouse model is the high production of transforming growth factor-beta (TGF-β1) induced by gluten exposure, suggesting that this cytokine may avoid proinflammatory intestinal responses to gluten.

TGF-β1 is a multifunctional cytokine that plays a critical role in controlling autoimmune conditions, as well as intestinal inflammatory responses.15 TGF-β1 is also a key cytokine in the establishment of oral tolerance.16 TGF-β1 binds to a complex of transmembrane serine–threonine kinases type I and type II receptors that trigger several intracellular pathways. The main pathway entails phosphorylation of receptor-regulated Smad2 and Smad3, which then associate with the common mediator Smad4 and translocate into the nucleus where the Smad2/3/4 complexes act as transcription factors regulating gene expression.17 This pathway is instrumental for the regulatory effects of TGF-β1, particularly in the mucosal compartments. Thus, disruption of Smad3 results in inflammatory T-cell infiltrates almost exclusively localized to mucosal surfaces and chronic intestinal inflammation.18 In active CD, large amounts of TGF-β1 have been observed in the intestinal mucosa,19 pleading against a quantitative defect, but reminiscent of the situation in other human inflammatory bowel diseases. In the latter diseases, TGF-β1 is present but the Smad3 signaling pathway is severely impaired due to up-regulation of the inhibitor Smad7 by as-yet unidentified mucosal proinflammatory factors.20, 21, 22

The massive increase of the proinflammatory cytokine IL-15 in CD led us to investigate whether and how IL-15 might interfere with the Smad signaling pathway of TGF-β. We then sought evidence that the inhibitory pathway elicited by IL-15 is operative in CD, and may represent a novel mechanism by which IL-15 contributes to the loss of intestinal homeostasis, and more generally promotes chronic inflammation.

Back to Article Outline

Materials and Methods 

CD Patients and Controls 

Peripheral lymphocytes were from healthy volunteers. Histologically normal small intestinal samples were from 16 adults (age, 33–80 years; mean, 55 years) undergoing intestinal surgery for morbid obesity or pancreatic cancer. Duodenal biopsies were from 25 active CD adults (age, 18–70 years; mean age, 39 years) with partial to subtotal villous atrophy and positive serology for antiendomysium antibodies. This study was approved by the local ethics committee.

Lymphocyte Isolation and Cell Culture 

Peripheral blood mononuclear cells (PBMC) were isolated on Ficoll-Hypaque gradient and lymphocyte subsets were separated using magnetic beads (CD3 microbeads, CD4+ or CD8+ T cell isolation kits II, Miltenyi Biotec, Paris, France). Purity ranged from 85% to 98%. IEL and LPLs were isolated from endoscopic or surgical samples as described.23, 24 Yield from biopsies was 0.4–0.6 × 106 for IEL and 0.6–0.8 × 106 for LPL. Contamination by live epithelial cells evaluated after 2-hour incubation at 37°C was <5%. Cells were cultured in medium11 supplemented or not with human recombinant cytokines IL-2, IL-15, TGF-β1 (R&D Systems, Abingdon, Scotland), or of humanized antihuman tumor necrosis factor alpha (TNF-α) (Remicade, Schering-Plough, Hérouville Saint Clair, France). TGF-β1 was used in all studies at 10 ng/mL, as this concentration was optimal to induce Smad3-dependent activation and to block lymphocyte proliferation25 (and not shown). For proliferation studies, lymphocytes were cultured in 96-well plates (105/well) and uptake of [3H]thymidine (Amersham Biosciences, Saclay, France) was measured 18 hours after adding 0.4 μCi/well. For c-jun antisense treatment, 25 μg/mL phosphorothioate single-stranded oligonucleotides to c-jun (antisense 5′- TTC CAT CTT TGC AGT CAT-3′; sense 5′-ATG ACT GCA AAG ATG GAA-3′) (Invitrogen, Cergy Pontoise, France) were added to lymphocyte cultures 24 hours after stimulation and every 24 hours. For organ culture, duodenal biopsies were cultured in 95% oxygen–5% carbon dioxide at 37°C for 24 to 30 hours on a stainless steel mesh in an organ culture dish (BD Biosciences, Le Pont de Claix, France) containing DMEM-F12 supplemented with 5% SVF, 1% HEPES, 40 μg/mL gentamycine, 2.5 μg/mL amphotericin B (Invitrogen) added or not with 50 μg/mL of sense or antisense phospho-c-jun oligonucleotides, 25 μg/mL of blocking antihuman IL-15 mouse mAbs or control IgG1 isotype (R&D Systems) and/or 10 ng/mL TGF-β1.

Flow Cytometry 

Lymphocytes (105) were incubated with phycoerythrin-, fluorescein isothiocyanate-, APC-Cy-chrome-conjugated mAbs to human, CD3, CD4, CD8, CD45, (BD Biosciences, Le Pont de Claix, France), TGF-βRII, or control isotypes (R&D Systems) for 30 minutes at 4°C. For intracellular phospho-c-jun detection, cells first stained with fluorescein isothiocyanate and APC-conjugated mAbs to CD3, CD4, or CD8, were fixed, permeabilized using DAKO intrastain kit (Dakocytomation, Trappes, France), and labeled with 10 μg/mL phycoerythrin-conjugated mAb KM1 to human phospho-c-jun (Santa Cruz Biotechnology, Santa Cruz, CA) or control isotype (R&D Systems). For intracellular TNF-α labeling, cells were treated with Golgistop (BD Biosciences) for the last 6 hours of culture. They were then fixed and permeabilized using BD Cytoperm/cytofix plus kit and labeled with phycoerythrin-conjugated mouse antihuman TNF-α or control isotype (BD Biosciences) according to manufacturer’s instructions. Analyses were performed with a BD-LSR using the CELLQuest software (BD Biosciences).

Immunochemistry 

Deparaffinized sections (5 μm) of duodenal formol fixed biopsies were incubated overnight with 10 μg/mL of mouse mAb against human phospho-c-jun or human TGF-β1 or control mouse IgG1 (R&D Systems), or with a 1:2000 dilution of anti-phospho-Smad2/3 antibody, a polyclonal rabbit antibody that detects Smad2 only when dually phosphorylated at serine residues 463 and 465, and that can react with phospho-Smad3 but not other Smad-related proteins (Cell Signaling Technology, Beverly, MA). Labeling was revealed by the R.T.U Vectastain universal Elite ABC kit using either diaminobenzidine or VIP (Vector Laboratories, Burlingame, CA). To study nuclear translocation of Smad3, cytospins of peripheral CD3+ T cells cultured in medium added or not with IL-15 for 5 days and stimulated overnight with 10 ng/mL of TGF-β, were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100 (Sigma, St. Louis, MO), incubated with 5 μg/mL of rabbit antihuman Smad3 (Zymed Laboratories, South San Francisco, CA) for 2 hours, and revealed by the R.T.U Vectastain universal Elite ABC kit using VIP and methyl green counterstain. Percentages of cells containing nuclear Smad3 were determined at ×40 on a Leitz Dialux 20 microscope.

Western-Blot Analysis 

Whole-cell extracts were obtained from intestinal biopsies of healthy controls and active CD patients using the Nuclear extract kit (Active motif, Rixensart, Belgium). Proteins (20 μg) were separated on a 10% sodium dodecyl sulfate (SDS)-PAGE gel, transferred to nitrocellulose membranes, and labeled with monoclonal antibody against human TGF-β1 (R&D Systems), or polyclonal rabbit antibodies against human Smad7 (Santa Cruz Biotechnology) and phospho Smad2/3 (Cell Signaling Technology) followed by horseradish peroxidase-conjugated antimouse or antirabbit antibodies (Cell Signaling Technology). Membranes were then stripped and analyzed with monoclonal antibody against human β-actin (Santa-Cruz Biotechnology) and horseradish peroxidase-conjugated antimouse antibody. Visualization was performed using the enhanced chemiluminescence system (ECL plus, Amersham Biosciences).

Electromobility Shift Assay 

Electromobility shift assay was performed with nuclear extracts from lymphocytes treated or not with TGF-β (10 ng/mL) for 45 minutes using double-stranded oligonucleotide probes plasminogen activator inhibitor (PAI) probe: 5′-TCG AGA GCC AGA CAA GGA GCC AGA CAA GCA GCC AGA CAC-3′, SBE probe, 5′-CTCTATCAATTGGTCTAGACTTAACCGGA-3′) end-labeled with [γ-32P] dCTP as described.25 Protein–DNA complexes were resolved in a 5% polyacrylamide gel. Binding specificity was checked using a 50 molar excess of nonradiolabeled PAI-1 promoter as a specific competitor.

Real-Time Reverse-Transcription Polymerase Chain Reaction 

Total RNAs extracted with RNeasy Mini Kit (Qiagen, Courtaboeuf, France) from intestinal tissues or lymphocytes, were reverse transcribed as described.11 TGF-β-interacting factor (TGIF) and IFN-γ mRNAs were quantified by real-time polymerase chain reaction (PCR) using SYBR Green PCR Master Mix (Applied Biosystems) and 300 nmol/L of the corresponding primers, and data normalized referring to expression of glyceraldehyde-3-phosphate dehydrogenase. Primers for IFN-γ and glyceraldehyde-3-phosphate dehydrogenase have been previously published.11 Upper and lower primers for TGIF were, respectively, 5′-AGCAAACACACCTGTCTACGCTAC-3′ and 5′-GGCGGGAAATTGTGAACTGAT-3′. TTP (Tristetraprolin) and Smad7 mRNAs were quantified by real-time PCR using available Assay-on-demand and Taqman PCR Master Mix (Applied Biosystems) and data normalized referring to expression of ribosomal Protein, Large, PO. For PCR, 40 cycles were performed as followed: denaturation at 95°C for 15 seconds and annealing and extension at 60°C for 1 minute, using an ABI PRISM 7700 sequence detection system (software version 1.6). TGF-β-dependent transcription was monitored in biopsies using Smad7 and TTP genes, and Smad7 and TGIF genes in isolated lymphocytes.

GST-c-Jun(1-79) Binding/Protein Kinase Assay 

Ficoll-isolated human PBMC were stimulated with 10 ng/mL of IL-15, 300 U/ml IL-2, or medium alone. Cells were lysed in 20 mmol/L Tris–HCl (pH 7.5), 150 mmol/L NaCl, 1 mmol/L Na2EDTA, 1 mmol/L EGTA, 1% Triton X-100, 25 mmol/L sodium pyrophosphate, 1 mmol/L β-glycerophosphate, 1 mmol/L Na3VO4, 1 μg/mL leupeptin, 1 mmol/L phenylmethylsulfonyl fluoride. Protein extracts (200 μg) were incubated overnight at 4°C with GST-c-Jun (1-79) (a kind gift from J. Raingeaud, Chatenay-Malabris) immobilized to glutathione-agarose (10 μL of packed beads per sample containing 10 μg of protein). After washing, the c-jun-N-terminal kinase (JNK)-GST-c-Jun (1–79)–agarose complexes were incubated for 1 hour at 30°C in 40 μL of kinase buffer containing 25 mmol/L Tris–HCl (pH 7.5), 5 mmol/L β-glycerophosphate, 2 mmol/L DTT, 0.1 mmol/L Na3VO4, 10 mmol/L MgCl2, 20 μmol/L cold ATP, and 5 μCi [-32P] ATP. Reactions were stopped by adding SDS sample buffer and boiling. Samples were loaded and resolved on 10% SDS-polyacrylamide gels. Radioactivity incorporated into the GST fusion proteins was visualized using a STORM 840 phospho-imager (Molecular Dynamics, Sunnyvale, CA).

Statistical Analysis 

Values obtained in the different study groups were compared by the nonparametric Mann–Whitney U test. Statistical significance was assigned to a value of P < .05.

Back to Article Outline

Results 

IL-15 Inhibits Smad3-Dependent TGF-β Signaling in Human T Lymphocytes 

TGF-β1 plays a role in the negative regulation of the immune response in part by inhibiting normal T-cell proliferation after stimulation. Notably, TGF-β1, via activation of the Smad pathway, efficiently inhibits the proliferative response of T cells to IL-2,26 a cytokine that shares its receptor β and γ chains with IL-15.1 In agreement with previous reports,25, 27 TGF-β1 inhibited IL-2 induced proliferation of PBMC (Figure 1A). The inhibitory effect of TGF-β1, absent during the early phase of stimulation, progressively increased after 48 hours and became maximal at 96 hours (59% ± 6% inhibition, P < .05). In contrast, TGF-β1 had no detectable effect on IL-15-induced proliferation at all time points tested (P > .05) (Figure 1A). Furthermore, although TGF-β inhibited the IL-2-induced proliferative response of IEL (90% ± 5% CD8+) and LPL (68% ± 6% CD4+) from healthy controls, it failed to block IL-15-induced proliferation in both intestinal T-cell subsets (Figure 1B). Although these results might reflect the lack of impact of TGF-β on the IL-15 signaling pathway leading to proliferation, they also raised the possibility that IL-15 inhibited Smad-mediated TGF-β signaling.

  • View full-size image.
  • Figure 1. 

    IL-15 inhibits TGF-β signaling in human T-cell lymphocytes. (A, B) Lymphocytes were stimulated with IL-2 (300 IU/mL) or IL-15 (10 ng/mL) in the absence or presence of TGF-β1 (10 ng/mL). Proliferative responses were assessed by [3H] thymidine uptake in triplicate. (A) PBMC cultured for 24 hours to 96 hours (results representative of 5 independent experiments). (B) PBMC, IEL, and LPL from controls stimulated for 96 hours (results of 3 to 4 independent experiments). Bars indicate means ± SD. *P < .05 when compared with the condition without TGF-β. (C) PBMC cultured in the presence or not of IL-15 (10 ng/mL) for 24 hours (upper panel) or 96 hours (lower panel) were stimulated by TGF-β1 (10 ng/mL) for 45 minutes. Nuclear extracts were analyzed by electromobility shift assay using a 32P-labeled probe derived from PAI-1 promoter. Fifty molar excess of nonradiolabeled PAI-1 promoter was added as a specific competitor (cold probe). (D) Peripheral CD3+ and CD8+ lymphocytes from controls were cultured in medium added or not with IL-15 (10 ng/mL) for 1, 2, 6, and 24 hours and then stimulated with 10 ng/mL TGF-β1 for 2 hours. TGIF mRNA levels were quantified by real-time PCR. Results are expressed in arbitrary units and are representative of 3 independent experiments. (Asterisks indicate that inhibition is significant at 1,2,6, and 24 hours). (E) Peripheral CD3+, CD4+, and CD8+ lymphocytes, IEL, and LPL from controls were cultured in medium added or not with IL-15 (10 ng/mL) for 24 hours and/or 96 hours and then stimulated with 10 ng/mL TGF-β1 for 2 hours. TGIF mRNA levels were quantified as in D. Results are representative of 3 independent experiments.

Following activation by TGF-β, Smad2/3/4 proteins assemble into a complex that translocates into the nucleus, binds to the promoters of TGF-β1-target genes, and induces their transcription. The inhibitory effect of IL-15 on Smad-dependent TGF-β-signaling was first studied by investigating the impact of IL-15 on the binding of Smad complexes to the promoters of TGF-β-target genes, using an electromobility shift assay and a probe containing a 3-CAGAC box derived from the PAI promoter, as described.28 IL-15 pretreatment of PBMCs for 24 hours impaired the formation of Smad3–DNA complexes in response to a 45-minute stimulation with TGF-β. A comparable inhibitory effect was observed after 5 days (Figure 1C), indicating a long-lasting effect of IL-15 compatible with the lack of inhibition of TGF-β on IL-15-induced proliferation at 96 hours. Similar results were obtained with another synthetic probe that contains a palindromic Smad3 specific target sequence CAGATCTG (data not shown).

The inhibitory effect of IL-15 on Smad-mediated TGF-β signaling in T lymphocytes was further tested by monitoring the transcription of TGIF, a target gene of TGF-β with a promoter activated by Smads proteins.29 This gene was chosen as expressed in T cells and not modulated directly by IL-15 in peripheral or intestinal T cells (Figure 4C and supplemental Figure 1 [See supplemental material online at www.gastrojournal.org]). TGF-β-dependent gene transcription was studied after a 2-hour stimulation with TGF-β, using real-time RT-PCR. A 1-, 2-, 6-, and 24-hour pretreatment with IL-15 inhibited TGF-β-induced transcription of TGIF in CD3+ and CD8+ peripheral T cells (Figure 1D and E). The inhibitory effect of IL-15 persisted after a 5-day incubation in peripheral T cells, particularly in the CD8+ subset (Figure 1E), a finding perhaps ascribable to the known preferential effect of IL-15 on CD8+ peripheral T cells.30 Noticeably, IL-15 also reduced TGF-β induction of TGIF in IEL and LPL by approximately 50% (P < .05) on day 4 (Figure 1E). By comparison, IL-2 exerted a transient inhibitory effect on Smad3-dependent transcription that was detected after a 24-hour incubation, but was much less pronounced at 96 hours. The distinct effects of IL-2 and IL-15 on Smad3-dependent transcription at late time points were compatible with the distinct effects of TGF-β on IL-2 and IL-15-induced T-cell proliferation (Figure 1B). These results led us to investigate the mechanism by which IL-15 inhibits Smad-dependent TGF-β-signaling.

  • View full-size image.
  • Figure 4. 

    Antisense oligonucleotides for c-jun restore TGF-β effects in lymphocytes cultured with IL-15. Lymphocytes were cultured in the presence of cytokines for 96 hours. Sense or antisense c-jun oligonucleotides (25 μg/mL) were added every 24 hours to the cultures. (A) Histogram analysis of intracellular phospho-c-jun (p-c-jun) expression in PBMC at 96 hours. (B) Proliferative responses of PBMC after a 96-hour culture in the presence of the indicated reagents. Oligonucleotides were added at 24 hours. Results are expressed as mean cpm of 3H thymidine incorporation (±SD) and are representative of 3 independent experiments. (C) TGIF and Smad7 mRNA induction was quantified by real-time PCR in peripheral CD8+ lymphocytes stimulated or not with TGF-β1 (10 ng/mL) for 2 hours after a 5-day culture in the indicated conditions. Results, expressed in arbitrary units, are from 2 independent experiments. *P < .05 when compared with the condition without IL-15.

IL-15 Exerts Its Inhibitory Effect on TGF-β Signaling Downstream of Nuclear Translocation of Smad3 

Several checkpoints regulate the Smad-dependent TGF-β signaling pathway. First, surface expression of TGF-β receptor II (TGF-β RII) can be downmodulated by inflammatory cytokines.31 No significant variation of TGF-β RII levels was noticed, however, upon IL-15 stimulation (Figure 2A).

  • View full-size image.
  • Figure 2. 

    IL-15 acts downstream from Smad3 nuclear translocation. (A) PBMC were cultured for 24 hours to 96 hours in medium added or not with IL-15 (10 ng/mL). Membrane expression of TGF-β RII was monitored by flow cytometry. Results are representative of 3 independent experiments. (B) Messenger RNA expression of Smad7 was monitored by real-time PCR in PBMC cultured for 2 hours to 96 hours in the presence or not of IL-15 (10 ng/mL) or TGF-β1 (10 ng/mL). Results are representative of 2 independent experiments. (C) Peripheral CD3+, CD4+, and CD8+ lymphocytes, IEL, and LPL from controls were cultured in medium added or not with IL-15 (10 ng/mL) for 24 hours and/or 96 hours and then stimulated with 10 ng/mL TGF-β1 for 2 hours. Smad7 mRNA levels were quantified by real-time PCR. Results expressed are representative of 3 independent experiments. (D) Immunocytochemical staining of Smad3 in peripheral CD3+ T cells stimulated overnight with TGF-β1 (10 ng/mL) after a 5-day culture in the presence or not of IL-15 (10 ng/mL). Percentage of cells with nuclear staining was determined among 500 cells. In 3 independent experiments, TGF-β-induced nuclear translocation of Smad3 was comparable in lymphocytes cultured in medium alone or in the presence of IL-15 (85%–99%). Nuclear translocation was observed in <2% of lymphocytes cultured in medium alone but in 25%–35% of lymphocytes cultured in IL-15, likely due to the induction of an endogenous production of TGF-β in the latter cells (data not shown). Note that IL-15-activated T cells become enlarged blast cells, while T cells cultured in control medium remain small and round.

A second important checkpoint lies at the level of Smad2 and Smad3 phosphorylation by the kinase activity of the TGF-β RII. The inhibitory factor, Smad7, not only promotes the degradation of the TGF-β receptor but also competitively blocks Smad2 and Smad3 phosphorylation and the subsequent signaling cascade (reviewed in Javelaud et al17). Because Smad7 is a target gene of the Smad3–TGF-β pathway, its primary role may be to provide a feedback mechanism modulating TGF-β signaling. Smad7 transcription can also be induced by several proinflammatory cytokines including IFN-γ,32 IL-1β, and TNF-α,33 an effect that underlies the inhibitory effect of these cytokines on TGF-β signaling. Smad7 mRNA expression was therefore monitored by real-time quantitative RT-PCR in peripheral T lymphocytes cultured for 2 hours to 96 hours with IL-15 or TGF-β as a positive control. Although TGF-β, as expected, had a strong activating effect on Smad7 transcription, IL-15 had no detectable direct impact on Smad7 RNA levels neither in peripheral (Figure 2B) nor in intestinal T cells (Supplemental Figure 1 [See supplemental material online at www.gastrojournal.org]). IL-15, however, inhibited TGF-β-induced transcription of Smad7 in peripheral and intestinal T cells (Figure 2C), confirming results obtained with TGIF (Figure 1C), showing that IL-15 impairs Smad3-dependent transcription. Altogether these findings suggested that IL-15 does not require Smad7 to block the TGF-β-Smad3 signaling cascade but acts after the steps regulated by Smad7. Therefore, we investigated whether IL-15 acts upstream or downstream from Smad3 nuclear translocation. As shown in Figure 2D, TGF-β induced a massive nuclear translocation of Smad3 in peripheral T lymphocytes either untreated or cultured in the presence of IL-15, indicating that IL-15 blockade takes place downstream from Smad3 nuclear translocation.

IL-15 Inhibition of TGF-β Signaling in T Lymphocytes Is Mediated by Phospho-c-jun Up-regulation 

Recent work has demonstrated that the JNK pathway can repress TGF-β signaling following Smad3 nuclear translocation via the inhibitory effect of phospho-c-jun on the formation of Smad–DNA complexes.25, 34 Previous studies in the Ba/F3 cell line have shown that IL-2 can activate JNK through recruitment of Shc to the IL-2Rβ chain that is shared by IL-2 and IL-15 receptors.35 Using a sensitive JNK assay, activation of JNK was detected in PBMC 30 minutes after stimulation with IL-15 (Figure 3A) and IL-2 (data not shown). The signal, strong at 24 hours with both cytokines, remained comparable at 48 hours and 96 hours in the presence of IL-15 (Figure 3A) but, as previously observed,25 decreased at 48 hours in the presence of IL-2. Intracellular expression of phospho-c-jun was next monitored by flow cytometry in peripheral and intestinal lymphocytes cultured or not in the presence of IL-2 or IL-15 for 1 to 10 days. Staining was specifically absent or markedly inhibited when stimulation was performed in the presence of c-jun antisense oligonucleotide (Figure 4A) or of SP600, a specific inhibitor of JNK (data not shown). In PBMC, induction of phospho-c-jun by IL-2 and IL-15, visible at 24 hours (data not shown), was more obvious at 48 hours (Figure 3B). C-jun phosphorylation remained high in peripheral lymphocytes after 5 or 10 days in the presence of IL-15, while it markedly decreased in cells stimulated by IL-2 (Figure 3B and C). Simultaneous membrane labeling with anti-CD3, -CD4, and/or -CD8 antibodies revealed higher phospho-c-jun up-regulation in CD8+ than in CD4+ peripheral T lymphocytes (Figure 3D). Phospho-c-jun up-regulation was even more striking in IEL and LPL: by 96 hours, phospho-c-jun median fluorescence was 5-fold increased in both subsets in response to IL-15, but only 2-fold increased in response to IL-2 (Figure 3D). In LPL, the effects of IL-15 were identical in either CD4+ or CD8+ subsets (data not shown). The higher induction of phospho-c-jun in intestinal T cells might be related to their activation state,36 because T-cell receptor ligation promotes expression of JNK.37

  • View full-size image.
  • Figure 3. 

    IL-15 induces a long-lasting expression of phospho-c-jun in peripheral and intestinal lymphocytes. (A) Peripheral lymphocytes were cultured with IL-15 (10 ng/mL) or IL-2 (300 IU/mL) for 30 minutes to 96 hours. JNK kinase assay was performed on cell extracts precipitated with GST-c-jun as described in Materials and Methods, and proteins were separated by 10% SDS-PAGE (upper panel). Loading control was obtained by running an identically loaded SDS-polyacrylamide gel that was stained by Coomassie blue (lower panel). (B–D) Peripheral and intestinal lymphocytes were cultured with IL-15 (10 ng/mL) or IL-2 (300 IU/mL) for 10 days. Intracellular staining of phospho-c-jun (p-c-jun) was monitored by flow cytometry. (B) Histogram analysis of intracellular staining of phospho-c-jun in the lymphocyte gate at the indicated time points. Isotype control was identical to unstimulated PBL (data not shown). (C) Representation of the increase of phospho-c-jun median fluorescence in PBMC stimulated for 10 days. Results were pooled from 3 independent experiments. Bars indicate mean ±SD. (D) Increase of phospho-c-jun median fluorescence after a 96-hour culture is compared in different subsets of peripheral and intestinal lymphocytes. Results are the mean value (±SD) of 3 independent experiments.

C-jun antisense treatment was used to investigate whether blockade of TGF-β signaling by IL-15 in T lymphocytes depends on up-regulation of phospho-c-jun. Thus, c-jun antisense but not sense oligonucleotide efficiently inhibited phospho-c-jun induction in T lymphocytes stimulated by IL-15 (Figure 4A). The capacity of c-jun antisense to restore the inhibitory effect of TGF-β on IL-15-induced proliferation was tested by adding the oligonucleotides 24 hours after initiation of the culture. Indeed, preliminary experiments indicated that JNK activation was necessary to initiate IL-15 and IL-2-induced proliferation. In contrast, after 24 hours, cytokine-induced proliferation was only reduced by approximately 40% in the presence of c-jun antisense, allowing to investigate the specific effect of c-jun blockade on TGF-β signaling (Figure 4B). In these conditions, adding c-jun antisense but not sense oligonucleotides restored the inhibitory effect of TGF-β on IL-15-induced proliferation (Figure 4B). This effect was not associated with any increased apoptosis (data not shown). Furthermore, c-jun antisense treatment restored a normal level of TGIF and Smad7 transcription in response to TGF-β in peripheral CD8+ T lymphocytes cultured with IL-15 for 5 days (Figure 4C). These data demonstrate that IL-15 prevents TGF-β signaling in T lymphocytes by up-regulation of phospho-c-jun.

Previous studies have shown that TNF-α can prevent TGF-β signaling via JNK-mediated c-jun phosphorylation.34, 38 Because this cytokine can be induced by IL-15 in T cells activated by the anti-CD3 antibody,39 we tested whether the effect of IL-15 on phospho-c-jun up-regulation depended or not on TNF-α production in lymphocytes. Phospho-c-jun induction by IL-15 in CD3+ peripheral and intestinal T lymphocytes was not modified in the presence of a blocking anti-TNF-α antibody (Supplemental Figure 1A, [See supplemental material online at www.gastrojournal.org] and not shown). In addition, IL-15 alone failed to induce significant amounts of TNF-α in all tested T lymphocyte subsets (Supplemental Figure 1B [See supplemental material online at www.gastrojournal.org]).

TGF-β Signaling Is Impaired in the Mucosa of Patients With Active CD Secondary to Phospho-c-jun Up-Regulation by IL-15 

Confirming a previous report,19 immunohistochemical staining of intestinal biopsies revealed intense TGF-β labeling in epithelium and lamina propria of patients with active CD, comparable with controls (Figure 5A). Immunoblotting confirmed that amounts of TGF-β were comparable in intestinal biopsies from patients with active CD and from controls (Figure 5A). TGF-β signaling in intestinal biopsies was first assessed by measuring the transcription levels of TTP (tristetraprolin), a TGF-β target gene under the control of the Smad3 pathway,40 which is easily detectable in tissue specimens. Contrasting with the substantial in situ synthesis of TGF-β protein, transcription levels of TTP were approximately decreased 3-fold in duodenal biopsies from patients with active CD compared with controls (Figure 5B). Impaired response to TGF-β was confirmed in isolated intestinal T lymphocytes using TGIF and Smad7, 2 reliable probes to monitor TGF-β/Smad3-induced transcription in T cells. Although incubation in the presence of TGF-β for 2 hours induced transcription of TGIF and Smad7 in IEL and LPL isolated from normal controls, it had no effect on the transcription of these 2 genes in IEL and LPL from active CD (Figure 5C).

  • View full-size image.
  • Figure 5. 

    Intestinal lymphocytes from active CD have defective Smad-dependent response to TGF-β. (A) Expression of TGF-β1 was analyzed by immunohistochemistry (left panel) and immunoblotting (right panel) in duodenal biopsies of 5 patients with active CD (ACD) and of 5 controls. Results of the immunoblots are shown in the upper right panel for 2 active CD patients and 2 controls. The ratios of TGF-β1 to β-actin assessed by densitometry were comparable between patients and controls (bars indicate the median value). (B) Messenger RNA expression of TTP was quantified by real-time PCR in duodenal samples of 13 active CD and 8 controls. (C) IEL and LPL were isolated from active CD and controls and stimulated for 2 hours with TGF-β1 (10 ng/mL). TGIF and Smad7 mRNA levels were quantified by real-time PCR. Results are expressed in arbitrary units. Bars indicate the median values. *P < .05 compared with controls.

Impairment of TGF-β signaling despite strong TGF-β expression has been observed in other inflammatory bowel diseases and ascribed to Smad7 up-regulation.20, 21 Smad7 was first assessed at the transcription level. Smad7 mRNA levels were significantly less in the biopsies of patients with active CD compared with controls (Figure 6A). This result is compatible with inhibition of Smad3-dependent transcription in the intestine of patients with active CD (Figure 5B and C). Smad7 is, however, not only regulated at the transcriptional but also at the posttranscriptional levels by antagonist mechanisms that modulate its degradation by the proteasome.41 Protein levels were therefore assessed by immunoblotting. Comparable amounts of Smad7 protein were detected in the intestinal biopsies of patients with active CD and in controls (Figure 6B). To define a possible impact of Smad7 on TGF-β signaling in active CD, the levels of Smad2/3 phosphorylation were compared in patients with active CD and controls. No difference was demonstrated by immunoblotting with a polyclonal antibody that detects phosphorylated Smad2/3 (Figure 6C). Furthermore, immunostaining of duodenal biopsies from patients with active CD with the same antibody revealed extensive nuclear labeling (Figure 6D). The latter results indicated that the initial steps of TGF-β signaling were essentially preserved in patients with active CD, pleading against a major inhibitory effect of Smad7 on Smad3-dependent transcription and suggesting that blockade of TGF-β signaling in CD takes place downstream nuclear translocation of Smad2/3.

  • View full-size image.
  • Figure 6. 

    The early steps of TGF-β-mediated signaling are not impaired in situ in active CD. (A) Messenger RNA expression of Smad7 was quantified by real-time PCR in duodenal samples of 13 active CD (ACD) and 8 controls and expressed as arbitrary units (AU). *P < .05 when compared with controls. (B) Expression of Smad7 was analyzed by immunoblotting in total extracts from 5 active CD patients and 4 controls. The ratios of Smad7 to β-actin assessed by densitometry (right panel) were comparable between patients and controls (bars indicate the median value of arbitrary units). Results of the immunoblot are shown in the left panel for 4 active CD patients and 3 controls. (C) Expression of phospho-Smad2/3 was analyzed by immunoblotting in total extracts from 4 active CD patients and 4 controls. The ratios of phospho-Smad2/3 to β-actin assessed by densitometry (right panel) were comparable between patients and controls. (D) Phospho-Smad2/3 distribution was studied by immunohistochemistry in intestinal biopsies of 4 patients with active CD. Staining is shown at 100× (left panel) and 200× (right panel) magnifications. Arrows show positive nuclei in lamina propria and epithelium.

The latter finding, compatible with an inhibitory effect of IL-15, a cytokine strongly expressed in the mucosa of patients with active CD11 (and not shown), led us to investigate phospho-c-jun expression in active CD. Immunohistochemical staining of intestinal biopsies showed phospho-c-jun labeling in a small number of nuclei in the villous epithelium of histologically normal controls, while in active CD patients, almost all IEL, LPL, and most epithelial cells were intensively positive with both nuclear and cytoplasmic staining (Figure 7A). Up-regulation of phospho-c-jun was confirmed in freshly isolated intestinal lymphocytes by flow cytometry: while IEL and LPL from controls were negative, phospho-c-jun was strongly expressed in IEL and LPL from patients with active CD (Figure 7B). The role of phospho-c-jun overexpression in the impairment of Smad3-dependent transcription in active CD was then assessed in 24-hour organ cultures by antagonizing c-jun expression with antisense oligonucleotides. c-jun antisense oligonucleotide abolished phospho-c-jun staining (Figure 8A). Concurrent to inhibition of phospho-c-jun expression, c-jun-antisense induced a small but reproducible increase in both TTP and Smad7 mRNA levels (P < .05) (Figure 8B). Comparable effects were observed with and without addition of exogenous TGF-β (Figure 8B). Notably, increase in Smad3-dependent transcription was associated with a decrease in IFN-γ mRNA expression, pointing to a possible restoration of the regulatory function of TGF-β. An effect via JNK blockade of TCR-42, 43 and NKG2D-mediated12 activation could not, however, be excluded.

  • View full-size image.
  • Figure 7. 

    Phospho-c-jun is up-regulated in intestinal lymphocytes in active celiac disease. (A) Small intestinal sections from 4 active CD (ACD) and 4 controls were stained with antiphospho-c-jun antibody (10 μg/mL). Immunohistochemical staining is shown at 100× (left panel) and at 400× magnification (right panel). (B) IEL and LPL were isolated from active CD and controls. Intracellular staining of phospho-c-jun (p-c-jun) was studied by flow cytometry (results representative of 3–4 experiments).

  • View full-size image.
  • Figure 8. 

    Effect of c-jun antisense and anti-IL-15 antibody ex vivo in duodenal biopsies from patients with active CD. Duodenal biopsies were cultured for 24–36 hours with medium containing sense or antisense c-jun oligonucleotides (50 μg/mL), monoclonal anti-IL-15 antibody or control mouse IgG1 (20 μg/mL) in the absence or presence of 10 ng/mL exogenous TGF-β added 3 hours before the end of the culture. (A) Immunohistochemical staining of organ cultures (OC) with antiphospho-c-jun antibody. (B) Quantification by real-time PCR of TTP, Smad7, or IFN-γ mRNAs expression in duplicate biopsies after incubation in the presence of the indicated reagents. Results, expressed as relative induction, are means (±SD) of 3 to 4 independent experiments. *P < .05 when compared with the control condition.

Therefore, we next investigated whether IL-15 participated in the up-regulation of phospho-c-jun in CD using 24-hour organ cultures. In 4 patients with active CD, addition of anti-IL-15 antibody markedly reduced phospho-c-jun staining, while control isotype and anti-TNF antibody had no effect (Figure 8A, and not shown), pointing out to the prominent role of IL-15 in the induction of phospho-c-jun. We then tested the capacity of anti-IL-15 antibody to restore TGF-β-induced transcription in intestinal biopsies of patients with CD. There was a marked increase in TTP and Smad7 mRNA levels after a 30-hour culture in the presence of anti-IL-15 antibody (Figure 8B, lower panel). Altogether, these data provide strong evidence that, in active CD, IL-15-induced phospho-c-jun is involved in the local down-regulation of TGF-β signaling.

Back to Article Outline

Discussion 

Our data, showing that IL-15 impedes Smad-dependent signaling of TGF-β, provide a new rationale for the potent proinflammatory effect of this cytokine that may be central to the pathogenesis of CD.

The inhibitory effect of IL-15 on TGF-β1 signaling in T lymphocytes was demonstrated by the negative impact of this cytokine on the formation of Smad–DNA complexes and on Smad3-dependent transcription. Consistent with these results and a previous report,27 TGF-β was unable to inhibit IL-15-induced lymphocyte proliferation. Our results indicate that IL-15 acts late in the Smad cascade via a mechanism that implicates JNK activation. First, IL-15 did not modify TGF-β RII expression or Smad7 transcription, 2 checkpoints important to control the initiation of the cascade. Second, IL-15 did not prevent Smad3 nuclear translocation. Third, IL-15 induced a long lasting activation of JNK and up-regulation of phospho-c-jun. Previous studies have demonstrated that JNK activation impaired TGF-β signaling because phospho-c-jun can, within the nucleus, inhibit Smad3 binding to DNA.25, 34 That IL-15 inhibits the Smad–TGF-β pathway via the latter mechanism was demonstrated using c-jun antisense oligonucleotides because they restored the effects of TGF-β on transcription and cytokine-induced proliferation in lymphocytes. IL-15 belongs to the 4 α-helix bundle cytokine family and shares with the closely related IL-2 cytokine, the receptor β and γc chains. The 2 chains form 1 signaling module able to recruit several pathways including JAK1/3-STAT3/5, PI3–kinase–AKT and RAS–MAP kinase cascades.1 As previously reported,25 IL-2 had only a brief and modest effect on the induction of phospho-c-jun, which paralleled transient refractoriness to TGF-β antiproliferative effects. How IL-15 but not IL-2 can induce a long lasting induction of phospho-c-jun remains to be defined. One possibility might be that IL-15 but not IL-2 induces another cytokine that can activate JNK and take the relay of IL-15. One obvious candidate was TNF-α.34, 38 Yet, our data indicate that the prolonged effect of IL-15 was not associated with the induction of TNF-α (Supplemental Figure 1 [See supplemental material online at www.gastrojournal.org]). Although another as yet unknown relay of IL-15 is not excluded, it is interesting to observe that the private α chain of the IL-15 receptor can recruit the tyrosine kinase syk44 and TRAF2,45 2 proteins able to potentiate JNK activation.46, 47 Yet, the contribution of IL-15 Rα chain to IL-15 signaling in lymphocytes remains controversial.48 Further studies, in progress, will aim to decipher the mechanisms of JNK activation in response to IL-15.

A delicate balance between inflammatory immune responses and tolerance is necessary to permit the eviction of pathogens while avoiding protracted inflammation and preserving immune homeostasis. Smad3–TGF-β signaling plays a central role in tipping the balance toward tolerance.15, 18 The inhibitory effect of IL-15 on this pathway might invert the balance and promote protective immune responses, in particular, against intracellular pathogens.49 Conversely, persistent expression of IL-15, observed in several autoimmune or chronic inflammatory diseases,50 may durably impair the regulatory functions of TGF-β and alter immune homeostasis. Our data in CD support this hypothesis.

A molecular basis to CD pathogenesis has been supplied by deciphering the specific interactions between the peptide pocket of HLA-DQ2/8 molecules and gliadin-derived T-cell epitopes.8 However, why loss of tolerance to gluten and triggering of the intestinal gliadin-specific CD4+ T cell response only occurs in a small subset of patients with the at-risk HLA remain unsolved. Data in transgenic mice expressing both human CD4 and HLA-DQ8 molecules support a role for TGF-β in preventing intestinal inflammation in response to oral challenge by gliadin peptides.14 Our results, showing that Smad–TGF-β-dependent transcription is decreased in the intestinal mucosa of active CD patients and cannot be up-regulated in isolated intestinal lymphocytes or in biopsies by exogenous TGF-β, indicate that Smad-dependent TGF-β signaling is impaired in active CD, and may therefore fail to maintain tolerance to gluten. Inhibition of Smad3 signaling has been previously demonstrated in intestinal inflammatory bowel diseases,21, 22 as well as in Helicobacter pylori-induced gastritis.20 In the latter conditions, as-yet unidentified mucosal factors up-regulate Smad7 expression and thereby prevent Smad2/3 phosphorylation. In active CD, however, Smad7 was not up-regulated, and amounts of phosphorylated Smad2/3, as those of TGF-β, were comparable in the mucosa of CD and controls. Altogether, these results pleaded against the prominent role of Smad7, and rather pointed to an inhibitory mechanism operating downstream Smad2/3 nuclear translocation and therefore independent of Smad7 expression.

We have shown that chronic exposure to gluten in CD patients is associated with a massive up-regulation of IL-15 in the intestinal mucosa.11 Consistent with the long-lasting in vitro induction of phospho-c-jun by IL-15, phospho-c-jun was markedly increased in the intestinal mucosa and T lymphocytes of patients with active CD, and downmodulated in duodenal tissue explants treated with blocking anti-IL-15 antibody. Furthermore, both c-jun antisense oligonucleotide and anti-IL-15 antibody increased TGF-β-dependent transcription in the biopsies of patients with active CD. Although of small amplitude, the changes in transcription detectable in organ culture were reproducible for the 2 genes tested. The effect of the antisense was comparable with and without addition of exogenous TGF-β, indicating that sufficient amounts of endogenous TGF-β were present in the biopsies of active CD to induce signaling after downmodulation of phospho-c-jun. Restoration of TTP transcription might be instrumental to downmodulate the local inflammatory reaction. Indeed, TTP is a protein that binds the AU-rich regions in the 3′UTR of mRNA of various proinflammatory molecules, including TNF-α, IL-2, cyclooxygenase 2, IL-1β, and matrix metalloproteases to promote their decay and prevent their translation.51, 52 In line with this hypothesis, TTP−/− mice develop an autoimmune syndrome characterized by arthritis, dermatitis, and autoantibody formation.53 Noticeably, restoration of Smad3-dependent transcription by c-jun antisense oligonucleotide was associated with a decrease in IFN-γ mRNA expression. This finding, coherent with the known role of Smad3 in the control of IFN-γ production,15 points to a possible restoration of the regulatory function of TGF-β. An effect of JNK blockade on TCR-42, 43 and/or NKG2D-mediated12 activation cannot be excluded. Yet, as shown in Figure 8, phospho-c-jun expression was almost entirely abolished by the blocking anti-IL-15 antibody, suggesting a prominent role of IL-15 in the in situ activation of JNK in CD.

Previous studies have substantiated several contributions of IL-15 to the pathogenesis of CD, including activation of lamina propria dendritic cells,10 induction of IEL cytotoxicity against enterocytes via innate immune receptors,9, 11, 12, 13 loss of IEL homeostasis, and emergence of T-cell lymphomas.11 Noticeably, 2 very recent reports have demonstrated that complete ablation of TGF-β signaling in T cells engendered an aggressive, autoimmune disease associated with uncontrolled activation of TH154, 55 and cytolytic differentiation programs in CD4+ and CD8+ peripheral T cells, and strikingly with the appearance of highly cytotoxic CD8 T cells bearing NK receptors,55 the latter finding reminiscent of CD.24, 56 Therefore, our current results suggest an additional scenario where the abnormal and chronic induction of IL-15 by the 31–43/49 peptide of α-gliadins9, 10 also alters local immune regulation by TGF-β, thereby favoring and perpetuating gliadin-specific CD4+ T-cell response in lamina propria and IEL activation.57 Furthermore, inhibition of Smad3-dependent TGF-β signaling may synergize with the potent antiapoptotic effects of IL-15 to promote intestinal T lymphomagenesis,58, 59 a rare but characteristic complication of CD.60 Studies are in progress to demonstrate more directly the consequences of Smad3 blockade on the functions of LPL and IEL in CD.

In conclusion, our results delineate a novel mechanism underlying the deleterious effects of chronic overexpression of IL-15. IL-15 was identified as a potential therapeutic target in certain human T-cell malignancies and in several inflammatory and autoimmune disorders.61 A recent report showed that targeting IL-15 with a humanized anti-IL-15 antibody improved patients with rheumatoid arthritis with minimal side effects.5 Our finding of the interplay between IL-15 and the TGF-β–Smads pathway further supports the concept that IL-15 is a meaningful therapeutic target in inflammatory diseases associated with irreducible elevation of IL-15.

Back to Article Outline

Supplementary data 

View full-size image

download image

Supplemental Figure 1. Lack of induction of Smad7 and TGIF by IL-15 in control IEL and LPL. IEL and LPL isolated from controls were left unstimulated or cultured for 5 days with 10 ng/mL of IL-15. TGIF and Smad7 mRNA were quantified by real-time PCR. Results, expressed in arbitrary units, are from 3 independent experiments.

View full-size image

download image

Supplemental Figure 2. Phospho-c-jun up-regulation induced by IL-15 is not mediated by TNF-α secretion. Peripheral and intestinal T lymphocytes were cultured with medium added or not with IL-15 (10 ng/mL) and/or anti-TNF-α (10 μg/mL) for 5 days. (A) Histogram analysis of intracellular staining of phospho-c-jun in peripheral CD3+ lymphocytes on day 5. (B) Intracellular labeling for TNF-α was monitored every day by flow cytometry in peripheral CD3+ lymphocytes, IEL and LPL. Cells induced with PMA (50 ng/mL) and ionomycin (10−6 M) for 24 hours were used as positive controls. Results, obtained on day 1 and day 5 are shown. Results are representative of 2 independent experiments.

Back to Article Outline

References 

  1. Fehniger TA, Caligiuri MA. Interleukin 15: biology and relevance to human disease. Blood. 2001;97:14–32
  2. Fehniger T, Suzuki K, Ponnapan A, Van Deusen J, Cooper A, Florea S, et al. Fatal leukemia in interleukin15 transgenic mice follows early expansion in natural killer and memory phenotype CD8+ T cells. J Exp Med. 2001;193:219–231
  3. Ohta N, Hiroi T, Kweon M, Kinoshita N, Jang M, Mashimo T, et al. IL-15-dependent activation-induced cell death-resistant Th1 type CD8alphabeta(+)NK1.1(+) T cells for the development of small intestinal inflammation. J Immunol. 2002;169:460–468
  4. Bamford RN, DeFilippis AP, Azimi N, Kurys G, Waldmann TA. The 5′ untranslated region, signal peptide, and the coding sequence of the carboxyl terminus of IL-15 participate in its multifaceted translational control. J Immunol. 1998;160:4418–4426
  5. Baslund B, Tvede N, Danneskiold-Samsoe B, Larsson P, Panayi G, Petersen J, et al. Targeting interleukin-15 in patients with rheumatoid arthritis: a proof-of-concept study. Arthritis Rheum. 2005;52:2686–2692
  6. Villadsen LS, Schuurman J, Beurskens F, Dam TN, Dagnaes-Hansen F, Skov L, et al. Resolution of psoriasis upon blockade of IL-15 biological activity in a xenograft mouse model. J Clin Invest. 2003;112:1571–1580
  7. Kim CY, Quarsten H, Bergseng E, Khosla C, Sollid LM. Structural basis for HLA-DQ2-mediated presentation of gluten epitopes in celiac disease. Proc Natl Acad Sci U S A. 2004;101:4175–4179
  8. Sollid L. Coeliac disease: dissecting a complex inflammatory disorder. Nat Rev Immunol. 2002;9:647–655
  9. Hue S, Mention JJ, Monteiro RC, Zhang S, Cellier C, Schmitz J, et al. A direct role for NKG2D/MICA interaction in villous atrophy during celiac disease. Immunity. 2004;21:367–377
  10. Maiuri L, Ciacci C, Ricciardelli I, Vacca L, Raia V, Auricchio S, et al. Association between innate response to gliadin and activation of pathogenic T cells in coeliac disease. Lancet. 2003;362:30–37
  11. Mention JJ, Ben Ahmed M, Begue B, Barbe U, Verkarre V, Asnafi V, et al. Interleukin 15: a key to disrupted intraepithelial lymphocyte homeostasis and lymphomagenesis in celiac disease. Gastroenterology. 2003;125:730–745
  12. Meresse B, Chen Z, Ciszewski C, Tretiakova M, Bhagat G, Krausz TN, et al. Coordinated induction by IL15 of a TCR-independent NKG2D signaling pathway converts CTL into lymphokine-activated killer cells in celiac disease. Immunity. 2004;21:357–366
  13. Di Sabatino A, Ciccocioppo R, Cupelli F, Cinque B, Millimaggi D, Clarkson MM, et al. Epithelium derived interleukin 15 regulates intraepithelial lymphocyte Th1 cytokine production, cytotoxicity, and survival in coeliac disease. Gut. 2006;55:469–477
  14. Black K, Murray J, David CS. HLA-DQ determines the response to exogenous wheat proteins: a model of gluten sensitivity in transgenic knockout mice. J Immunol. 2002;169:5595–5600
  15. Aoki CA, Borchers AT, Li M, Flavell RA, Bowlus CL, Ansari AA, et al. Transforming growth factor beta (TGF-beta) and autoimmunity. Autoimmun Rev. 2005;4:450–459
  16. Weiner H. Induction and mechanism of action of transforming growth factor-beta-secreting Th3 regulatory cells. Immunol Rev. 2001;182:207–214
  17. Javelaud D, Mauviel A. Crosstalk mechanisms between the mitogen-activated protein kinase pathways and Smad signaling downstream of TGF-beta: implications for carcinogenesis. Oncogene. 2005;24:5742–5750
  18. Yang X, Letterio JJ, Lechleider RJ, Chen L, Hayman R, Gu H, et al. Targeted disruption of SMAD3 results in impaired mucosal immunity and diminished T cell responsiveness to TGF-beta. EMBO J. 1999;18:1280–1291
  19. Lionetti P, Pazzaglia A, Moriondo M, Azzari C, Resti M, Amorosi A, et al. Differing patterns of transforming growth factor-beta expression in normal intestinal mucosa and in active celiac disease. J Pediatr Gastroenterol Nutr. 1999;29:308–313
  20. Monteleone G, Del Vecchio Blanco G, Palmieri G, Vavassori P, Monteleone I, Colantoni A, et al. Induction and regulation of Smad7 in the gastric mucosa of patients with Helicobacter pylori infection. Gastroenterology. 2004;126:674–682
  21. Monteleone G, Kumberova A, Croft NM, McKenzie C, Steer HW, MacDonald TT. Blocking Smad7 restores TGF-beta1 signaling in chronic inflammatory bowel disease. J Clin Invest. 2001;108:601–609
  22. Monteleone G, Mann J, Monteleone I, Vavassori P, Bremner R, Fantini M, et al. A failure of transforming growth factor-beta1 negative regulation maintains sustained NF-kappaB activation in gut inflammation. J Biol Chem. 2004;279:3925–3932
  23. Cerf-Bensussan N, Guy-Grand D, Griscelli C. Intraepithelial lymphocytes of human gut: isolation, characterisation and study of natural killer activity. Gut. 1985;26:81–88
  24. Jabri B, Patey de Serre N, Cellier C, Evans K, Gache C, Carvalho C, et al. Selective expansion of intraepithelial lymphocytes expressing the HLA-E-specific natural killer receptor CD94 in celiac disease. Gastroenterology. 2000;118:867–879
  25. Arnulf B, Villemain A, Nicot C, Mordelet E, Charneau P, Kersual J, et al. Human T-cell lymphotropic virus oncoprotein Tax represses TGF-beta 1 signaling in human T cells via c-Jun activation: a potential mechanism of HTLV-I leukemogenesis. Blood. 2002;100:4129–4138
  26. Nelson BH, Martyak TP, Thompson LJ, Moon JJ, Wang T. Uncoupling of promitogenic and antiapoptotic functions of IL-2 by Smad-dependent TGF-beta signaling. J Immunol. 2003;170:5563–5570
  27. Campbell JD, Cook G, Robertson SE, Fraser A, Boyd KS, Gracie JA, et al. Suppression of IL-2-induced T cell proliferation and phosphorylation of STAT3 and STAT5 by tumor-derived TGF beta is reversed by IL-15. J Immunol. 2001;167:553–561
  28. Dennler S, Itoh S, Vivien D, ten Dijke P, Huet S, Gauthier JM. Direct binding of Smad3 and Smad4 to critical TGF beta-inducible elements in the promoter of human plasminogen activator inhibitor-type 1 gene. EMBO J. 1998;17:3091–3100
  29. Chen F, Ogawa K, Nagarajan RP, Zhang M, Kuang C, Chen Y. Regulation of TG-interacting factor by transforming growth factor-beta. Biochem J. 2003;371:257–263
  30. Kennedy M, Glaccum M, Brown S, Butz E, Viney J, Embers M, et al. Reversible defects in natural killer and memory CD8 T cell lineages in interleukin 15-deficient mice. J Exp Med. 2000;191:771–780
  31. Yamane K, Ihn H, Asano Y, Jinnin M, Tamaki K. Antagonistic effects of TNF-alpha on TGF-beta signaling through down-regulation of TGF-beta receptor type II in human dermal fibroblasts. J Immunol. 2003;171:3855–3862
  32. Ulloa L, Doody J, Massague J. Inhibition of transforming growth factor-beta/SMAD signalling by the interferon-gamma/STAT pathway. Nature. 1999;397:710–713
  33. Bitzer M, von Gersdorff G, Liang D, Dominguez-Rosales A, Beg AA, Rojkind M, et al. A mechanism of suppression of TGF-beta/SMAD signaling by NF-kappa B/RelA. Genes Dev. 2000;14:187–197
  34. Verrecchia F, Tacheau C, Wagner EF, Mauviel A. A central role for the JNK pathway in mediating the antagonistic activity of pro-inflammatory cytokines against transforming growth factor-beta-driven SMAD3/4-specific gene expression. J Biol Chem. 2003;278:1585–1593
  35. Hunt AE, Lali FV, Lord JD, Nelson BH, Miyazaki T, Tracey KJ, et al. Role of interleukin (IL)-2 receptor beta-chain subdomains and Shc in p38 mitogen-activated protein (MAP) kinase and p54 MAP kinase (stress-activated protein Kinase/c-Jun N-terminal kinase) activation (IL-2-driven proliferation is independent of p38 and p54 MAP kinase activation). J Biol Chem. 1999;274:7591–7597
  36. Guy-Grand D, Griscelli C, Vassalli P. The gut-associated-lymphoid system: nature and properties of the large dividing cells. Eur J Immunol. 1974;4:435–443
  37. Weiss L, Whitmarsh AJ, Yang DD, Rincon M, Davis RJ, Flavell RA. Regulation of c-Jun NH(2)-terminal kinase (Jnk) gene expression during T cell activation. J Exp Med. 2000;191:139–146
  38. Verrecchia F, Pessah M, Atfi A, Mauviel A. Tumor necrosis factor-alpha inhibits transforming growth factor-beta/Smad signaling in human dermal fibroblasts via AP-1 activation. J Biol Chem. 2000;275:30226–30231
  39. Korholz D, Banning U, Bonig H, Grewe M, Schneider M, Mauz-Korholz C, et al. The role of interleukin-10 (IL-10) in IL-15-mediated T-cell responses. Blood. 1997;90:4513–4521
  40. Ogawa K, Chen F, Kim YJ, Chen Y. Transcriptional regulation of tristetraprolin by transforming growth factor-beta in human T cells. J Biol Chem. 2003;278:30373–30381
  41. Gronroos E, Hellman U, Heldin CH, Ericsson J. Control of Smad7 stability by competition between acetylation and ubiquitination. Mol Cell. 2002;10:483–493
  42. Gao Y, Tao J, Li MO, Zhang D, Chi H, Henegariu O, et al. JNK1 is essential for CD8+ T cell-mediated tumor immune surveillance. J Immunol. 2005;175:5783–5789
  43. Yang DD, Conze D, Whitmarsh AJ, Barrett T, Davis RJ, Rincon M, et al. Differentiation of CD4+ T cells to Th1 cells requires MAP kinase JNK2. Immunity. 1998;9:575–585
  44. Bulanova E, Budagian V, Pohl T, Krause H, Durkop H, Paus R, et al. The IL-15Ralpha chain signals through association with Syk in human B cells. J Immunol. 2001;167:6292–6302
  45. Bulfone-Paus S, Bulanova E, Pohl T, Budagian V, Durkop H, Ruckert R, et al. Death deflected: IL-15 inhibits TNF-alpha-mediated apoptosis in fibroblasts by TRAF2 recruitment to the IL-15Ralpha chain. FASEB J. 1999;13:1575–1585
  46. Jacinto E, Werlen G, Karin M. Cooperation between Syk and Rac1 leads to synergistic JNK activation in T lymphocytes. Immunity. 1998;8:31–41
  47. Habelhah H, Frew IJ, Laine A, Janes PW, Relaix F, Sassoon D, et al. Stress-induced decrease in TRAF2 stability is mediated by Siah2. EMBO J. 2002;21:5756–5765
  48. Dubois S, Mariner J, Waldmann TA, Tagaya Y. IL-15Ralpha recycles and presents IL-15 in trans to neighboring cells. Immunity. 2002;17:537–547
  49. Nishimura H, Yajima T, Naiki Y, Tsunobuchi H, Umemura M, Itano K, et al. Differential roles of interleukin 15 mRNA isoforms generated by alternative splicing in immune responses in vivo. J Exp Med. 2000;191:157–170
  50. McInnes IB, Gracie JA. Interleukin-15: a new cytokine target for the treatment of inflammatory diseases. Curr Opin Pharmacol. 2004;4:392–397
  51. Ogilvie RL, Abelson M, Hau HH, Vlasova I, Blackshear PJ, Bohjanen PR. Tristetraprolin down-regulates IL-2 gene expression through AU-rich element-mediated mRNA decay. J Immunol. 2005;174:953–961
  52. Phillips K, Kedersha N, Shen L, Blackshear PJ, Anderson P. Arthritis suppressor genes TIA-1 and TTP dampen the expression of tumor necrosis factor alpha, cyclooxygenase 2, and inflammatory arthritis. Proc Natl Acad Sci U S A. 2004;101:2011–2016
  53. Taylor GA, Carballo E, Lee DM, Lai WS, Thompson MJ, Patel DD, et al. A pathogenetic role for TNF alpha in the syndrome of cachexia, arthritis, and autoimmunity resulting from tristetraprolin (TTP) deficiency. Immunity. 1996;4:445–454
  54. Li MO, Sanjabi S, Flavell RA. Transforming growth factor-beta controls development, homeostasis, and tolerance of T cells by regulatory T cell-dependent and -independent mechanisms. Immunity. 2006;25:455–471
  55. Marie JC, Liggitt D, Rudensky AY. Cellular mechanisms of fatal early-onset autoimmunity in mice with the T cell-specific targeting of transforming growth factor-beta receptor. Immunity. 2006;25:441–454
  56. Meresse B, Curran SA, Ciszewski C, Orbelyan G, Setty M, Bhagat G, et al. Reprogramming of CTLs into natural killer-like cells in celiac disease. J Exp Med. 2006;203:1343–1355
  57. Ebert EC. Inhibitory effects of transforming growth factor-beta (TGF-beta) on certain functions of intraepithelial lymphocytes. Clin Exp Immunol. 1999;115:415–420
  58. Letterio JJ. TGF-beta signaling in T cells: roles in lymphoid and epithelial neoplasia. Oncogene. 2005;24:5701–5712
  59. Wolfraim LA, Fernandez TM, Mamura M, Fuller WL, Kumar R, Cole DE, et al. Loss of Smad3 in acute T-cell lymphoblastic leukemia. N Engl J Med. 2004;351:552–559
  60. Cellier C, Delabesse E, Helmer C, Patey N, Matuchansky C, Jabri B, et al. Refractory sprue, coeliac disease, and enteropathy-associated T-cell lymphoma. [see comments] Lancet. 2000;356:203–208
  61. Waldmann TA. Targeting the interleukin-15 system in rheumatoid arthritis. Arthritis Rheum. 2005;52:2585–2588

 This work was sponsored by INSERM, by ARC Grant 4616, Canceropole Ile de France, and by La Fondation Princesse Grace de Monaco.

 The authors are grateful to members of the GERMC and particularly to Pr. M. Lehman (Hôpital Saint Louis), Dr. T. Matysiak-Budnik, Dr. D. Lamarque (Hôpital Hôtel Dieu), Pr. Chaussade (Hôpital Cochin), and Pr. Cugnenc (Hôpital Georges Pompidou) for providing material and information from their patients. The authors thank Dr. D. Buzoni-Gatel for helpful discussions and Mr. G. Pivert for technical support with histology.

PII: S0016-5085(06)02675-8

doi:10.1053/j.gastro.2006.12.025

Refers to article:

  • Mucosal Inflammation in Celiac Disease: Interleukin-15 Meets Transforming Growth Factor β-1

    Martin F. Kagnoff
    Gastroenterology March 2007 (Vol. 132, Issue 3, Pages 1174-1176)

Gastroenterology
Volume 132, Issue 3 , Pages 994-1008, March 2007