IBD: Inflammatory Bowel Diseases; CD: Crohn’s disease; UC: Ulcerative Colitis; TGF: Transforming Growth Factor; TNF: Tumor Necrosis factor; TGFβRII: Type II TGF-β receptor; TGFβRI: Type I TGF-β receptor; LPMC: Lamina Propria Mononuclear Cells; NF-kB: Nuclear Factor Kappa B; TNBS: Trinitrobenzene Sulfonic Acid

Crohn’s disease (CD) and ulcerative colitis (UC) are the principal types of inflammatory bowel diseases (IBD) in human beings. CD is characterized by a transmural and segmental inflammation that can involve any part of the alimentary tract, even though lesions are more common in the terminal ileum and right colon. In contrast, in UC, the inflammatory process involves the mucosa of the rectum and can extend proximally to the colon [1]. The natural history of both CD and UC patients can be also marked by local complications and extra-intestinal manifestations. IBD arise as a result of the interaction between environmental and genetic factors, which leads to an excessive immune response directed against the components of the intestinal microflora [2,3]. Inflamed tissue of IBD patients is massively infiltrated with many immune cell types, which exhibit phenotypic features of activated cells and have the ability to produce huge amounts of inflammatory molecules [4,5]. Identification of the major inflammatory pathways leading to the IBD-associated tissue damage has contributed to the development of various immunomodulating drugs. For instance, different strategies have been explored to inhibit tumor necrosis factor (TNF), since this cytokine regulates positively multiple detrimental signals in the gut. Clinical trials and data emerging from the real-life indicate that monoclonal anti-human TNF antibodies (e.g. infliximab, adalimumab) are useful to induce and maintain clinical, endoscopic and histological remission in IBD thereby reducing the risk of complications, hospitalization and surgery [6-9]. More recently, monoclonal antibodies targeting integrins or other inflammatory cytokines have been used with success in IBD patients, reinforcing the notion that targeting therapy can help dampen the tissue-damaging immune response in these disorders [10,11]. However, not all the patients respond to or tolerate such therapies thereby indicating the need for additional effective and safe drugs.

In the last decade, evidence has been accumulated to suggest that IBD-related inflammation is also sustained by defects in counterregulatory mechanisms. One such a mechanism involves transforming growth factor (TGF)-β1, a multifunctional cytokine that is produced by many intestinal cell types and controls negatively multiple inflammatory signals [2,12]. Indeed, mice deficient in TGF-β1 expression and/or activity develop intestinal inflammation while induction of TGF-β1 production in mice associates with diminished severity of colitis [13-15]. Along the same line is the demonstration that human IBD are marked by defective TGF-β1 activity due to elevated levels of Smad7, an intracellular inhibitor of TGF-β1 signaling [16]. We here review the available data supporting the pathogenic role of Smad7 in the gut.

Binding of TGF-β1 to the type II subunit of TGF-β receptor (TGFβRII) leads to phosphorylation and activation of TGFβR type I subunit (TGFβRI) [17,18], which in turns promotes phosphorylation of Smad2 and Smad3, two intracellular proteins. Once activated, Smad2 and Smad3 associate with Smad4 and translocate to the nucleus, where Smad protein complexes suppress the transcription of many inflammatory genes [19,20]. Several pieces of evidence support the importance of TGF-β1-associated Smad2/3 signaling in the control of gut homeostasis. For instance, lamina propria mononuclear cells (LPMC) and biopsies taken from the small intestine and colon of healthy individuals and colonic specimens of normal mice express constitutively high levels of phosphorylated Smad3 [16,21]. Normal LPMC responds to exogenous TGF-β1 with enhanced Smad3 phosphorylation, inhibition of nuclear factor kappa B (NF-kB) activation and suppression of inflammatory molecule synthesis [22]. Mice lacking Smad3 have an infiltration of T cells and pyogenic abscess formation in the intestine [23].

In inflamed intestine of IBD patients there are high levels of TGF-β1 RNA transcripts [24] but diminished expression of phosphorylated Smad3 and Smad3/Smad4 complexes as compared with normal intestinal samples [16]. Moreover, in vitro treatment of IBD LPMC with TGF-β1 does not suppress NF-kB activation or production of pro-inflammatory cytokines, suggesting a defect of TGFβ1 signaling [22]. Indeed, IBD mucosal cells express elevated levels of Smad7, an intracellular protein that binds to TGFβRI and prevents Smad2/3 phosphorylation [16]. Knockdown of Smad7 in IBD LPMC and mucosal explants with a specific antisense oligonucleotide restores cell responsiveness to TGF-β1, as indicated by increased level of phosphorylated Smad3 and diminished expression of inflammatory cytokines [16]. The factors/mechanisms that sustain Smad7 expression in IBD remain to be determined. Smad7 is regulated at the post-transcriptional level by enhanced acetylation of the protein on lysine residues. This modification, which is partly mediated by the acetylase p300, prevents Smad7 ubiquitination-mediated proteasomal degradation with the downstream effect of enhancing Smad7 protein stability [25]. We have recently shown that the levels of Sirt1, an enzyme that deacetylates Smad7 lysine residues [26,27], are reduced in inflamed tissue of IBD patients, suggesting a further mechanism by which Smad7 overexpression is sustained in these disorders [28].

Studies in cancer cells have also shown that Smad7 levels are controlled by USP26, a deubiquitinating enzyme, which prevente Smad7 degradation [29]. On the other hand, there are intracellular mechanisms that counteract Smad7 activity. Among these, RNF11, an E3 ligase, has been reported to antagonize Smurf2/Smad7 complex, thus restoring cellular TGFβ signaling [30]. It remains unclear if these mechanisms are also acting in the gut of IBD patients. Mir-195 acts as a negative regulator of Smad7 expression [31] and decreased levels of mir-195 have been described in steroid-resistant UC patients, thus suggesting a link between such a defect and the enhanced expression of Smad7 in UC.

A considerable amount of experimental data has been accumulated to support the pathogenic role of Smad7 in the gut. Studies in mice with trinitrobenzene sulfonic acid (TNBS)-induced colitis, which shows immunological similarities with CD, or oxazolone-induced colitis, which resembles UC, show that colonic inflammation in such animals is characterized by overproduction of TGF-β1, which associates with decreased expression of phosphorylated Smad3 and elevated content of Smad7 [32,33]. Inhibition of Smad7 with an oral antisense oligonucleotide enhances Smad3 phosphorylation, suppresses expression of STAT1 and T-bet, two T helper (Th)1-related transcription factors, and Th1 cytokines in mice with TNBS-induced colitis and interleukin (IL)-4 production in mice with oxazolone induced colitis, thus leading to attenuation of colitis [33,34].

Analysis of Smad7-expressing cells in the gut revealed that such a protein is overexpressed in both T cells and non-T cells [16]. To examine the role of Smad7 in T cells, we developed a transgenic mouse overexpressing Smad7 in this cell type. The transgenic mice do not spontaneously develop intestinal inflammation but show increased mucosal production of inflammatory cytokines and more severe colitis than wild-type mice following oral ingestion of dextran sulfate sodium [35]. Interestingly, Smad7-overexpressing T cells transferred into immunodeficient mice cause a severe colitis that is resistant to regulatory T-cell (Tregs)-mediated immunosuppression [36]. This finding is consistent with the demonstration that CD mucosal T cells are not responsive to Tregs-mediated immunosuppression, a phenomenon that is reverted by Smad7 knockdown. T cells of Smad7- transgenic mice also exhibit a defective expression of aryl hydrocarbon receptor (AhR), a transcription factor that promotes IL-22 production thus delivering protective signals in the gut [37]. In immunodeficient mice, 6-formylindolo[3,2-b]carbazole, an activator of AhR, ameliorates colitis induced by wild-type T cells but does not affect colitis induced by transfer of Smad7-overexpressing T cells [37]. CD mucosal cells have reduced levels of AhR but expression of such a protein is increased by TGF-1 stimulation following Smad7 knockdown [38].

Treatment of IBD mucosal cells with Smad7 antisense oligonucleotide also increases production of both interleukin 25 [39], a cytokine that negatively regulates Th1 and Th17 inflammatory responses in the gut, and tissue inhibitor of matrix metalloproteinase-3, an enzyme that inhibits multiple tissue degrading enzymes [40].

The demonstration that inhibition of Smad7 restores a TGF-β- dependent counter-regulatory mechanism leads to the development of a pharmaceutical compound, named Mongersen, containing the Smad7 antisense oligonucleotide. This drug is given orally to patients and is protected by an external coating that determines gastro-resistance and allows the compound to be released in the terminal ileum and right colon [41]. A phase 1, clinical, open-label dose-escalating study was conducted in 15 patients with active steroid-dependent or resistant CD. Patients were divided into 3 cohorts and received Mongersen once a day for 7 days at doses of 40, 80, and 160 mg [41]. Treatment with Mongersen was well tolerated and no drug-related adverse event was observed. Measurement of the circulating levels of the drug revealed a very low systemic bioavailability as the compound was detectable at low levels in the plasma of only one patient at a single time point. All CD patients showed a clinical improvement following treatment and more than two thirds of them experienced a clinical remission. Treatment associated with a reduction of the fraction of circulating CCR9-positive T cells producing interferon-γ or IL-17A, a subset of inflammatory T cells with gut-homing properties, which are increased in active CD [41,42]. Since TGF-β1 has pro-fibrogenic effects and CD natural history can be marked by the development of fibrostrictures [43], all the patients receiving Mongersen were monitored for the development of such a complication. No patient developed strictures or obstructive symptoms during the study nor had a significant increase in the serum levels of fibrogenic markers [44]. A subsequent phase II, multicenter, doubleblind placebo-controlled study was conducted to assess the efficacy of Mongersen. One hundred sixty-six, steroid-dependent or resistant, CD patients were enrolled to receive placebo or Mongersen at 10, 40, or 160 mg/d for 2 weeks. Patients receiving the highest doses of the drug have significantly higher rates of remission than those treated with placebo. All patients receiving the drug showed a greater rate of clinical response in comparison with those receiving placebo. No drug-related adverse event was seen [45]. Responders to Mongersen had reduced serum levels of CCL20, a chemokine over-produced in epithelium of CD patients that contributes to recruit immune cells to inflamed gut [46]. This finding is in line with the demonstration that CCL20 production is induced in intestinal epithelial cells by TNF-α through a mechanisms which is negatively regulated by TGF-β1 [46].

The data described in this article highlight the inflammatory role of Smad7 in the gut and support the notion that knockdown of Smad7 with Mongersen can be a promising therapeutic approach for CD patients. However, further studies on larger populations are needed to confirm the clinical benefit seen in phase 1 and phase 2 studies, to examine whether Mongersen promotes also endoscopic/histological remission/improvement and prolongs the remission phases and to identify which subsets of patients could benefit from such treatments. Since phase 1 and phase 2 studies were performed in patients with lesions confined to terminal ileum and/or right colon, it would be relevant to assess whether Mongersen is also useful to control the active phases of the disease in patients with distal colitis. Similarly, further work is needed to evaluate whether Mongersen is effective in UC as this disease is also characterized by elevated mucosal production of Smad7. So far, short-term treatment with Mongersen has been associated with no adverse event perhaps due to the fact that the drug is poorly absorbed following oral administration. However, the safety profile of this compound must be confirmed by long-term studies. In this context, it would be relevant to confirm the preliminary evidence suggesting that inhibition of Smad7 expression by Mongersen is not followed by formation of fibro-strictures.

Studies in other systems have recently shown that expression of Smad7 in antigen presenting cells, such as dendritic cells and macrophages, can facilitate the progression of destructive inflammatory responses and induction of Smad7 seems to rely on bacterial-derived stimuli [47,48]. Therefore, experimental work is still needed to better characterize which cell types over-express Smad7 in the different phases of the natural history of CD and UC and which factors account for such an induction.

The authors have received research funding for the work on Smad7 from Fondazione U. Di Mario (Rome, Italy) and Broad Medical Research Foundation (IBD-0301R). GM has filed a patent related to the treatment of inflammatory bowel diseases with Smad7 antisense oligonucleotides, while the remaining authors have no conflict of interest.