Introduction

Although the pathogenic mechanisms underlying Crohn’s disease (CD) and ulcerative colitis (UC), the two major forms of inflammatory bowel disease (IBD), have not yet been fully clarified, evidence indicates that the chronic and relapsing inflammation derives from the interaction involving genetic predisposition and an abnormal immune response against the intestinal microbiota, resulting from undetermined environmental factors [1, 2]. Nevertheless, in recent years, a role for innate immunity defects has been increasingly considered in the pathogenesis of IBD. In fact, in addition to the early sensing and initiation of an immediate response, the concept of the innate immune system has evolved to include functional integration with adaptive immune responses [3, 4].

In parallel, genome-wide association studies have contributed to the identification of susceptibility genes related to IBD, most of which are involved in microbial pathogen sensing but also in the regulation of cellular stress and autophagy [5–7]. Therefore, in addition to the consequent altered recognition and processing of intracellular bacterial components, it is reasonable to predict that the cell death machinery might fulfill an important role in the pathogenesis of IBD. In this regard, cell death appears to be particularly relevant in epithelial barrier dysfunction, increasing the exposure to luminal contents and leading to the persistent activation of immune and non-immune elements within the intestinal lamina propria [8–10].

Hedgehog signaling constitutes a vital pathway regulating patterning and morphogenesis of the gut early in life [11, 12]. In the adult gastrointestinal tract, Hh signaling plays an important role in homeostasis, as it regulates epithelial cell turnover [13, 14] and mediates epithelial-mesenchymal interactions [15, 16]. In the human intestine, the most relevant Hh ligands, Indian hedgehog (Ihh) and Sonic hedgehog (Shh), are primarily produced in epithelial cells and interact with the receptor Patched 1 (Ptch-1). Ptch-1 is present in the underlying lamina propria cells, preferentially in mesenchymal cells [17]. Once activated, Ptch-1 interacts with a second membrane receptor, Smoothened (Smo), promoting the initiation of an intricate signaling cascade that results in the nuclear translocation of the glioma-associated (Gli) family of zinc finger transcription factors, Gli-1, Gli-2, and Gli-3 [18, 19]. Recently, a germline variation in the transcription factor GLI1 has been associated with IBD disease susceptibility; reduced GLI1 function appears to enhance the inflammatory response within the intestinal mucosa [20].

Currently, few studies have addressed Hedgehog pathway signaling in intestinal disorders including IBD, and those were restricted to experimental models. For example, the results from a study using transgenic mice expressing Hhip, an Hh inhibitor, demonstrated the predominant paracrine nature of Hh signaling within the intestine, characteristically from the epithelial layer to the lamina propria mesenchymal cells and including subepithelial myofibroblasts, and showed that it led to increased epithelial cell proliferation [13]. Another investigation using a transgenic mouse model of chronic Hh inhibition indicated that the modulation of Hh signals can modify inflammatory pathways in the mesenchymal compartment, whereas persistent Hh signaling blockade results in intestinal inflammation [21]. In addition, Hh signaling has been identified as a pathway that is antagonistic to Wnt signaling in the intestine, therefore driving epithelial cell differentiation [22, 23]. In turn, in another model for investigating mechanisms that regulate the adaptive response to intestinal resection, Hh inhibition promoted epithelial cell migration but also apoptosis [14].

It is noticeable that, in the gastrointestinal tract, Hh pathway components have been characteristically detected in the epithelial layer under normal conditions [23, 24]. Moreover, we recently demonstrated that Hh pathway activation in terminally differentiated colon cancer cell lines can dampen inflammatory signals and inhibit cell death by apoptosis [25]. Therefore, in this study, we sought to investigate the expression and function of the Hh signaling pathway in IBD, a disease characterized by chronic mucosal inflammation with abnormal cell death and compromised epithelial integrity.

Materials and methods

Ethics statement

This study was approved by the Ethical Committee of the University Hospital of the Federal University of Rio de Janeiro, and informed consent was obtained from all participants (Approval Number: 092/08).

Study population

Consecutive patients regularly followed up at the outpatients unit for Intestinal Diseases of the Division of Gastroenterology at the University Hospital of the Federal University of Rio de Janeiro, a tertiary referral center, were enrolled in this study. Specimens were obtained from 15 CD, 15 UC, and 15 control individuals. The diagnosis of IBD was confirmed by routine clinical, imaging, endoscopic, and histological parameters. All patients with IBD had active disease, according to the Harvey–Bradshaw Activity [26] and the Clinical Colitis Activity index [27]. Patients with malignant and non-malignant conditions, who had histologically normal tissue, constituted the control group.

Patients with CD included 7 women and 8 men, with a mean age of 38 years (range 19–62 years). The mean duration of CD was 7 years (range 1–15 years). In 8 patients, the CD diagnosis was established before 40 years of age, whereas, in 7 patients, the CD diagnosis was established at 40 years of age or older. Nine patients were treated with azathioprine (100–200 mg/day), 4 were treated with prednisone (10–40 mg/day) and azathioprine (100–200 mg/day), and 2 were treated with prednisone (10–40 mg/day) and mesalamine (2.4–4 g/day). The disease involved both the ileum and the colon in 9 patients, while the disease involved the colon exclusively in 6 patients. None of the patients had proximal small bowel involvement. Regarding the predominant form of CD, 5 patients had the non-stricturing and non-penetrating form, 6 had the penetrating form, and 4 had the stricturing form.

Patients with UC included 8 women and 7 men, with a mean age of 41 years (range 20–65 years). The mean duration of UC was 7 years (range 2–21 years). Clinically, 8 patients had moderate disease and 7 had moderate-to-severe disease at the time of the study. Six patients were treated with azathioprine (100–200 mg/day); 6 patients were treated with prednisone (20–40 mg/day), azathioprine (100–150 mg/day), and mesalamine (2.4–4 g/day); and 3 patients were treated with mesalamine (2.4–4 g/day) only. Nine patients had pancolitis, and 6 patients had limited left-sided colitis. IBD samples from patients with CD and UC were inflamed colon specimens with comparably active disease and similar degrees of histological inflammation. Samples from non-inflamed CD tissues were obtained whenever possible and served as additional controls in some experiments.

The control group consisted of 6 men and 9 women, with a mean age of 42 years (range 24–62 years), including 7 with diverticular disease and 8 with benign polyps. Patients underwent colonoscopy, and tissues were obtained from at least 10 cm from the diverticula or polyps in macroscopically non-inflamed areas. All mucosal specimens of the control patients were histologically normal. None of the control patients were taking any medication at the time of the study.

Antibodies, reagents, and treatments

The antibodies used in this study were: rabbit polyclonal anti–Shh-antibody, rabbit polyclonal anti–Ihh-antibody, rabbit polyclonal anti–Gli1-antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA), mouse monoclonal anti–CD3 fluorescein isothiocyanate (FITC) antibody, mouse monoclonal anti–CD68 FITC antibody, mouse monoclonal anti–α-SMA antibody, mouse monoclonal anti–Vimentin antibody, biotinylated mouse monoclonal anti-rabbit IgG antibody Dual Link System-HRP (all from Dako, Glostrup, Denmark), mouse monoclonal anti–β-tubulin antibody (Sigma-Aldrich, St. Louis, MO, USA), Alexa 488-conjugated anti-rabbit IgG, and Alexa 633-conjugated anti-mouse IgG (Molecular Probes, Eugene, OR, USA).

Recombinant human N-Shh was purchased from R&D Systems (Minneapolis, MN, USA) and was used at 500 ng/ml. Cyclopamine (Sigma-Aldrich, St. Louis, MO, USA) was used as a selective chemical hedgehog inhibitor and was diluted in dimethyl sulfoxide (DMSO) (control vehicle) (Merck, Darmstadt, Germany) at 2 µM. GANT61 (Tocris Bioscience, Bristol, UK), a small molecule that acts downstream of cyclopamine to inhibit GLI transcription, was diluted in DMSO at 5 µM. The chemical hedgehog agonist purmorphamine (Alexis Biochemicals, Plymouth, PA, USA) was diluted in DMSO at 2 µM. Sodium butyrate was added to the cultures at 2.5 µM, while interferon-gamma (IFN-γ) was used at 2 ng/ml (Sigma-Aldrich, St. Louis, MO, USA).

Mucosal specimens

Colon specimens were obtained from inflamed mucosa from all patients with IBD and from macroscopically normal mucosa of the same patients with CD whenever possible. In the control group, biopsies were obtained from macroscopically normal colonic mucosa. Biopsy specimens were aliquoted for immunohistochemistry, immunofluorescence, qRT-PCR, organ culture, and caspase-3 activity assays. For the experiments involving human intestinal fibroblasts, additional samples were obtained as mucosal explants from surgically resected colon specimens from three patients with benign and non-inflammatory bowel diseases (colonic polyps, and diverticular disease), and three with actively inflamed CD, as previously described [28]. Briefly, specimens were washed in calcium- and magnesium-free Hank’s balanced salt solution (Gibco/Invitrogen, Grand Island, NY, USA), and strips of normal mucosa were dissected, cut in small fragments, and laid on the bottom of tissue culture dishes. Medium consisting of Dulbecco’s modified Eagle medium containing 2.5 % l-glutamine and 25 mmol HEPES buffer (Gibco/Invitrogen, Grand Island, NY, USA), supplemented with 10 % heat-inactivated fetal calf serum (FCS), and a 1 % mixture of penicillin, streptomycin, and Amphotericin B solution (Sigma-Aldrich, St. Louis, MO, USA) was added to the dishes. After 3–5 days in culture, incubated at 37 °C in humidified 5 % CO2, cells morphologically compatible with fibroblasts were seen emerging from the fragments on the bottom of the dishes. After 3–4 weeks, subconfluent fibroblasts were detached with Triple Express (Thermo Fisher Scientific, Waltham, MA, USA), established in long-term cultures, being fed twice weekly and split at confluency every 4–7 days. Cell monolayers presenting a typical fibroblastic morphology used for experiments were all between passages 3 and 6.

Histological analysis

Specimens were fixed in 40 g/L formaldehyde saline, embedded in paraffin, cut into 5-mm sections, stained with hematoxylin and eosin (HE) stain, and examined microscopically by two independent observers who were unaware of the clinical or endoscopic diagnoses. Colon sections were studied according to routine histological parameters. The histological assessment of inflammation in patients with CD was based on the Seldenrijk et al. [29] grading system. The histological analysis of patients with UC was based on the Truelove and Richards [30] grading system. Histological studies also served to confirm the absence of inflammatory changes in the macroscopically normal mucosa of patients with CD. To further analyze histopathological changes and tissue remodeling in IBD, we utilized picrosirius red and periodic acid-Schiff techniques (Supplementary materials).

Immunohistochemistry

Paraffin sections were cut onto slides pretreated with poly-lysine and were used to characterize Hh pathway-expressing cells using the indirect immunoperoxidase technique. Briefly, deparaffinized sections were first incubated at 90 °C in 0.01 M sodium citrate buffer (pH 6.0) for 30 min for antigen retrieval. Then, slides were immersed in 3 % hydrogen peroxide in methanol for 10 min to block endogenous peroxidase activity. After being rinsed in phosphate buffered saline (PBS) containing 0.5 % Tween 20 for 10 min, tissue sections were incubated with non-immune horse serum for 30 min and, subsequently, with the respective monoclonal antibody.

Immunohistochemical staining was carried out using the following primary antibodies: rabbit polyclonal anti-Shh (1:50) antibody; rabbit polyclonal anti-Ihh (1:50) antibody; and mouse polyclonal anti-Gli-1 (1:50) antibody. Two sections from each sample were incubated with either PBS alone or with isotype monoclonal IgG (concentration-matched) and served as negative controls.

After incubation in a humidified chamber overnight at 4 °C, the tissue sections were rinsed with PBS and incubated with Dual Link System-HRP (Dako, Glostrup, Denmark) for 30 min at room temperature. Additional rinsing was followed by development with a solution containing hydrogen peroxide and diaminobenzidine (Dako, Glostrup, Denmark). Preparations were lightly counterstained in Harris’s hematoxylin, dehydrated, and mounted in Permount (Fisher Scientific, Pittsburgh, PA, USA).

Immunofluorescence and confocal laser-scanning microscopy

In a double direct or indirect immunofluorescence study, sections were incubated overnight at 4 °C with 2.5 % bovine serum albumin, 2.0 % skimmed milk, and 8.0 % fetal bovine serum blocking buffer while shaking. Slides were rinsed once with PBS and 0.05 % Tween-20 and then incubated with appropriately diluted primary antibodies in PBS. Tissue sections were incubated for 1 h at room temperature with: rabbit monoclonal anti-Gli-1 (1:50) antibody and mouse monoclonal anti-CD3 FITC (1:50) antibody, mouse monoclonal anti-CD68 FITC (1:50) antibody, mouse monoclonal anti-α-SMA FITC (1:50) antibody, mouse monoclonal anti-Vimentin FITC (1:50) antibody, or mouse monoclonal anti-β-Tubulin FITC (1:2000) antibody.

After incubation, the slides were rinsed three times and incubated with Alexa 488-conjugated anti-mouse IgG (1:1000) or Alexa 633-conjugated anti-rabbit IgG (1:1000) for 30 min at room temperature. Sections from each sample were incubated with either PBS alone or with secondary antibody, and these served as negative isotype controls. Slides were air-dried, fixed for 5 min in a 1 % paraformaldehyde solution, and mounted in an anti-fading medium containing 4′,6-diamidino-2-phenylindole (DAPI) (Vector Labs, Burlingame, CA, USA). The localization and fluorescence intensity of the proteins were observed with a Leica TCS-SP5 AOBS confocal laser-scanning microscope (Leica, Heidelberg, Germany), which was used for capturing representative images of each sample.

RNA isolation, cDNA synthesis, and quantitative real-time PCR

The expression levels of selected genes were validated by quantitative real-time PCR (qRT-PCR). First, total RNA isolation from colonic biopsy specimens was performed using SV Total RNA isolation systems (Promega, Madison, WI, USA), following the manufacturer’s protocol. A Nanodrop 2000 UV–Vis Spectrophotometer (Thermo Scientific, Wilmington, DE, USA) was used for quantifying and determining the purity of the RNA samples. Equal amounts of total RNA were reverse transcribed using a High-Capacity cDNA Archive kit.

To quantify the mRNA, real-time RT-PCR was performed on the ABI Prism 7500 (Applied Biosystems, Foster City, CA, USA) using RT2 Real Time ™ SYBR Green/Rox PCR Master Mix (SABiosciences, Frederick, MD, USA). For this purpose, we used a customized commercially available RT2 Profiler PCR Array for detecting the IHH, SHH, GLI-1, GLI-2, GLI-3, PTCH1, SMO, HHIP, WNT1, BMP4, and BMP7 genes (SABiosciences, Frederick, MD, USA). Levels of mRNA were normalized to the expression levels of the control genes glyceraldehyde phosphate dehydrogenase (GAPDH), beta-actin, and ribosomal protein L32 (RPL32). For data analysis, the ΔΔC t method was used to determine the fold change of all of the target genes in each sample with 95 % confidence. qRT-PCR for each gene was performed in duplicate, and each experiment was repeated at least three times. A positive value indicates up-regulation of a gene and a negative value indicates down-regulation of a gene. PCR cycles were performed according to the manufacturer’s instructions.

Organ culture and cytokine measurements

Colonic explants were cultured in RPMI 1640 medium supplemented with 10 % fetal calf serum (Gibco-Invitrogen, Carlsbad, CA, USA), 2 mM l-glutamine, 50 mM 2-mercaptoethanol, 10 mM HEPES, penicillin (100 killiunits/L), and streptomycin (100 mg/L) (all from Sigma Chemical Co., St. Louis, MO) for 24 h at 37 °C in a 5 % CO2 humidified incubator. Samples were centrifuged, and the supernatants were used to measure the concentration of the cytokines TNF α, IL-17 (both from Peprotech Inc., Rocky Hill, NJ, USA), TGF β (R&D Systems, Minneapolis, MN, USA), and IL-10 (Invitrogen, Camarillo, CA, USA) using commercially available sensitive enzyme-linked immunosorbent assays (ELISAs). The total protein content of the biopsy specimens was estimated using the Pierce BCA protein assay kit (Thermo Scientific, Rockford, IL, USA) and was used to normalize the results. The minimum detectable concentration of human TNF-α, IL-17, TGF β, and IL-10 was 5.0 pg/ml.

Assessment of apoptosis in the colon

Apoptosis was assessed in tissue sections of the colon by terminal deoxynucleotidyl transferase (TdT)-mediated dUTP nick-end labeling (TUNEL), and apoptosis was assessed in epithelial cells isolated from colonic samples by caspase-3 activity assay.

For the TUNEL assay, samples from control, CD and UC patients were analyzed using the ApopTag Peroxidase In Situ Apoptosis Detection Kit (Millipore Corporation, Billerica, MA). First, paraffin sections were deparaffinized, hydrated, and incubated with proteinase K solution. After blocking endogenous peroxidase activity, slides were covered with the equilibration buffer and then incubated with a solution containing the TdT enzyme. For negative controls, we incubated a second section of each sample without the TdT enzyme. For positive controls, we pre-treated samples with DNAase I (Sigma, Deisenhofen, Germany). After the reaction was terminated, specimens were incubated with non-immune horse serum and then incubated with an anti-digoxigenin peroxidase conjugate. As described above, the sections were developed with diaminobenzidine, counterstained in Harris’s hematoxylin, dehydrated, and mounted in mounting medium. Morphologically preserved TUNEL-positive cells and apoptotic bodies were referred to as apoptotic cells.

For the assessment of caspase-3 activity, colonic biopsy specimens were incubated in a Ca2+ and Mg2+ -free Hanks’ balanced salt solution (HBSS) containing 0.5 mM EDTA at 37 °C with constant stirring for 20 min, and this incubation was repeated two additional times. The epithelial cells detached from the basal membrane were passed through a 40-μm-pore-size nylon mesh strainer (BD Bioscience, San Jose, CA, USA), counted and divided into four equal parts for the subsequent cell culture.

Epithelial cells were then incubated at 2 × 105 cells per well in a 24-well plate at 37 °C in a 5 % CO2 humidified incubator in Dulbecco’s modified Eagle medium (Gibco/Invitrogen, Grand Island, NY, USA) buffered with 3.7 g/L sodium bicarbonate and 5 g/L HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (Sigma-Aldrich, St. Louis, MO, USA) and supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin (Gibco/Invitrogen, Grand Island, NY, USA), 1 % l-glutamine (Sigma-Aldrich, St. Louis, MO, USA) and heat-inactivated 10 % fetal bovine serum (LGC Bio, São Paulo, Brazil).

Cells were treated with recombinant human N-Shh and GANT61, alone or in combination, and with anti-CD95/APO-1 monoclonal antibody (Southern Biotech, Birmingham, AL, USA) to induce apoptosis or with vehicle control (DMSO, 0.2 %). All treatments were performed in duplicate. After 24 h, cells were harvested, washed twice in cold PBS, and resuspended in chilled lysis buffer. The cytosolic extracts were then subjected to caspase-3 activity measurement according to the manufacturer’s instructions (Caspase-3 Colorimetric Activity Assay Kit, Chemicon Millipore, Billerica, MA, USA).

Quantitative analysis of tissue sections

Quantitative measurements of tissue sections (under light microscopy) were performed using a computer-assisted image analyzer (Leica QWin Plus version 3.5.1, Leica Microsystems Ltd, Switzerland). Cells exhibiting identifiable reactivity distinct from background were considered positive. In the immunoperoxidase studies, percentages of the different cell subsets were defined by the number of immunoreactive cells relative to the total number of cells (immunoreactive and non-immunoreactive cells; 400× magnification), and these were counted in at least ten different areas or in at least 500 epithelial cells (total, superficial and crypts) of longitudinally sectioned colonic crypts. The results of the quantitative analysis of Shh, Ihh, Gli-1, and TUNEL were expressed as the percentages of positive cells. Two independent observers who were unaware of the experimental data examined all of the tissue sections and captured images.

Transwell migration assay

Monocytes were obtained from the peripheral blood of four healthy human volunteers, and peripheral blood mononuclear cells (PBMCs) were layered over a Ficoll–Hypaque cushion and centrifuged at 400 g for 20 min. PBMCs were harvested at the interface, washed twice with HBSS and then resuspended in RPMI 1640 medium supplemented with 10 % heat-inactivated FBS (LGC Bio, São Paulo, Brazil), 100 U/ml penicillin, and 100 mg/ml streptomycin (Sigma Chemical Co., St. Louis, MO, USA). PBMCs were then plated on culture flasks, and monocytes were allowed to attach and spread into the surface for 24 h at 37 °C in a humidified atmosphere with 5 % CO2. The adherent monocytes were washed twice with ice-cold Hank’s balanced solution and then detached by trypsinization. The recovered cells were centrifuged at 720 g and resuspended in RPMI 1640 medium supplemented with 2.5 % FCS. Cell count and viability were determined by Trypan blue staining.

To determine the chemotactic activity, we utilized the QCM Chemotaxis Cell Migration Assay (Millipore Corporation, Billerica, MA, USA) fitted with a polycarbonate membrane of 5-mm pore size. Human monocytes (1×105 cells) were incubated with 2.5 % FBS complete medium in the upper chambers, while rhShh, GANT61, rhShh + GANT61, or vehicle (DMSO) were diluted in 600 µl of the culture medium and added to the lower chambers. Additional control wells received 20 ng/ml of recombinant monocyte chemotactic protein-1 (rMCP-1), anti-MCP-1 antibody (e-Bioscience, San Diego, CA, USA), or 10 % FCS in the bottom chambers of the plate. Cells were allowed to incubate at 37 °C for 4 h, after which the cell density in the bottom chamber was determined using a colorimetric assay read at 560 nm in a 96-well plate.

Culture, cell proliferation and migration of fibroblasts

In addition to human intestinal fibroblasts, primary human skin fibroblasts obtained from 4 donors were analyzed to determine the effects of the Hh pathway on cell proliferation and migration. Fibroblasts (passage 6–8) were maintained in DMEM medium supplemented with 10 % heat-inactivated FBS (Gibco-Invitrogen, Carlsbad, CA, USA), 100 U/ml penicillin, and 100 mg/ml streptomycin (Sigma Chemical Co., St. Louis, MO, USA) at 37 °C in a humidified atmosphere with 5 % CO2.

For the assessment of cell proliferation, we utilized the BrdU Cell Proliferation Assay Kit (Chemicon International, Temecula, CA, USA). To modulate the Hh pathway and measure the incorporation of BrdU, primary human intestinal and skin fibroblasts were cultured in 96-well plates and, after overnight attachment, cells were treated for 24, 48, and 72 h in triplicate with different reagents, such as DMSO, Shh, butyrate, purmorphamine, cyclopamine, GANT61, and IFN-γ. BrdU was added to the wells 4 h prior to the plate reading at a wavelength of 450 nm.

The migration ability of primary human fibroblasts exposed to Hh pathway agonists or antagonists was assessed using a scratch wound assay, which measures the expansion of a cell population onto surfaces. Briefly, 2 × 105 cells/well were seeded into 6-well plates in complete culture medium overnight, as described above. Then, the confluent layer of cells was scratched using a sterile 20- to 200-μl pipette tip. After washing with PBS, cells were treated for 8 and 24 h in triplicate with different reagents, such as DMSO, Shh, GANT61, and IFN-γ. Images were captured at 0, 8, and 24 h after the monolayers were scratched. Scratched areas were measured using the computer-assisted image analyzer Leica QWin Plus version 3.5.1 (Leica Microsystems Ltd, Switzerland).

Statistical analysis

Statistical analyses were performed using the statistical software SPSS for Windows (Version 10.1, SPSS, USA). Significant differences among the experimental groups were evaluated using the Wilcoxon signed rank test for comparisons between the patients with CD and the Kruskal–Wallis ANOVA on ranks test, in which multiple comparisons were carried out using the Dunnett’s test, as appropriate. Values are expressed as medians with inter-quartile ranges. The correlation between caspase-3 activity and relative GLI-1 mRNA levels was assessed using the Spearman’s rank correlation coefficient. The level of significance was set at p < 0.05.

Results

Histopathological findings

In terms of histological grading of inflammation, of the specimens obtained from the macroscopically inflamed colons of patients with CD, 8 were classified as severe and 7 were classified as moderate. All specimens obtained from the macroscopically non-inflamed areas of the same patients with CD were histologically normal. Of the specimens obtained from patients with UC, 8 were classified as severe and 7 were classified as moderate. All mucosal biopsies from the control group were histologically normal. Complementary histopathological assessment is shown in the supplementary materials (Supplementary Figures S1 and S2).

The expression of Hh pathway components is decreased in the inflamed colonic mucosa from patients with IBD

We first characterized the expression pattern of Hh pathway components in the inflamed and non-inflamed colonic mucosa of CD patients, UC patients, and controls by immunohistochemistry. Overall, a marked decrease in Hh components was detected in the inflamed mucosal samples from patients with IBD. The expression levels of Shh, Ihh, and Gli-1 were significantly reduced in the colonic epithelium of inflamed CD and inflamed UC mucosa compared to the controls (p < 0.001). In particular, the inflamed areas of CD mucosa also exhibited significantly reduced levels of Shh, Ihh, and Gli-1 compared to the non-inflamed areas of CD (p = 0.001) and, to a lesser extent, compared to the inflamed UC mucosa (p < 0.04) ( Fig. 1a, b).

Fig. 1
figure 1

Localization and distribution of Hedgehog (Hh) components in the colonic mucosa of patients with IBD. The ligands Sonic hedgehog and Indian hedgehog (Shh and Ihh, respectively) and the transcription factor Gli-1 were characteristically expressed at the superficial epithelium, and a marked decrease in these Hh components was observed in the inflamed mucosa of CD and UC patients. Length bars represent 100 µm (a). The percentages of Hh-positive cells in the colonic epithelium (superficial and crypt) are individually represented. Horizontal bars represent the medians of 15 samples in each group. Significant differences are highlighted: *p = 0.001, **p < 0.001 (b)

Full size image

In addition to the quantitative differences, the normal colonic mucosa from control individuals and the non-inflamed CD mucosa displayed a distinctive distribution of Hh components that was characterized by the consistent and homogeneous staining of the superficial epithelium, with only few positive cells within the lamina propria. On the other hand, in inflamed CD and UC mucosa, the localization of Hh components was characteristically restricted to the crypt epithelium. Regarding the staining of the lamina propria for Shh, Ihh, or Gli-1, cells presented a scattered distribution throughout the mucosa, and the majority of them were morphologically compatible with non-inflammatory cells (Fig. 1a).

Gli-1-positive cells are predominantly mesenchymal cells in the normal colonic mucosa

To further investigate Hh pathway expression in the colonic lamina propria, we assessed the expression of Gli-1, a main activator of Hh target genes and an indicator of Hh pathway activation, by double immunofluorescence using confocal microscopy analysis.

Immunofluorescence images revealed that Gli-1 was predominantly expressed in the lamina propria in morphologically elongated, spindle-shaped, primarily non-immune, and non-inflammatory cells, and it was preferentially localized in the sub-epithelial layer of the normal control colonic mucosa. To determine which cells co-express Gli-1, sections were incubated with anti-Gli-1 in combination with anti-α-SMA, anti-vimentin, anti-β-tubulin, anti-CD3, or anti-CD68. Double-positive cells were counted in relation to the specific immune and non-immune cells of the intestinal lamina propria. Representative images and data from the co-localization studies are summarized in Fig. 2.

Fig. 2
figure 2

Co-localization of Gli-1 in lamina propria mononuclear cells. Co-localization of Gli-1 (red) with α-smooth muscle actin (α-SMA), vimentin, β-tubulin, T cells (CD3), and macrophages (CD68) (green) was determined by double immunofluorescence with confocal microscopy analysis. In the normal control lamina propria, Gli-1 co-localizes predominantly with α-smooth muscle actin (α-SMA), vimentin, and β-tubulin-positive cells and to a lesser extent with macrophages and T cells (magenta). Gli-1 co-localization with α-SMA, vimentin, and β-tubulin is significantly greater in controls than in inflamed CD mucosa. Nuclei are stained with 40,6-diamidino-2-phenylindole (blue). The micrograph panel is representative of 7 control, 7 CD, and 7 UC mucosal samples (original magnification, ×1000). Length bars represent 50 µm (a). The percentages of double-positive cells in the colonic lamina propria are individually represented. Horizontal bars represent the medians. Significant changes are highlighted: *p < 0.001, **p = 0.007 (b) (color figure online)

Full size image

In the normal control lamina propria, the overall expression of Gli-1 reflects its predominant co-localization with β-tubulin-positive cells (median 30.7 %), vimentin-positive cells (median 21.3 %), and α-SMA-positive cells (median 19.8 %). Only few Gli-1-positive cells co-localized with CD3-positive (7 %) or CD68-positive (7.5 %) cells. The inflamed mucosa from patients with CD expresses proportionally less Gli-1 in vimentin-positive cells compared with the controls and UC patients (p < 0.001) as well as fewer α-SMA-positive cells compared with controls (p < 0.001) and fewer β-tubulin-positive cells compared with controls (p = 0.007).

mRNA expression of Hh pathway components in the colonic mucosa

Next, several genes intrinsically involved in the Hh pathway and other genes that potentially interact with Hh signaling were quantitatively analyzed by real-time PCR. The levels of all genes investigated were expressed as the fold changes relative to the control group and arbitrarily normalized to 1 (Fig. 3).

Fig. 3
figure 3

Expression of Hh components in the colonic mucosa of patients with IBD measured by quantitative PCR. Histograms of individual gene expression fold changes in colon explants: IHH and SHH (a); GLI1, GLI2 and GLI3 (b); SMO, PTCH1 and HHIP (c); and WNT1, BMP4 and BMP7 (d). Dot plot of gene expression showing fold changes for individual explants. Horizontal bars represent the medians of 10 samples in each group and were normalized to the RPL32 and glyceraldehyde phosphate dehydrogenase RNA genes. Significant changes are highlighted

Full size image

Levels of IHH and SHH mRNA were significantly lower in the inflamed CD mucosa compared with normal controls (p < 0.001 and p = 0.003, respectively) and compared with UC mucosal samples (p = 0.004 and p = 0.002, respectively) (Fig. 3a).

GLI-1 and GLI-2, which are the main activators of Hh target genes and indicators of Hh pathway activation [31], mRNA levels were significantly lower in CD samples compared with controls (p < 0.001 and p = 0.001, respectively) and compared with UC samples (p < 0.001). GLI-1 and GLI-2 mRNA levels were also lower in the UC samples compared with the control samples (p = 0.002). Levels of GLI-3 were significantly higher in the normal control samples compared with the CD or UC samples (p < 0.001) (Fig. 3b).

The expression levels of both SMO and PTCH1, transmembrane receptors of the Hh pathway, were significantly lower in the CD samples compared with the controls (p < 0.001) or the UC samples (p < 0.001 and p = 0.009, respectively). The expression levels of SMO and PTCH1 were also lower in the UC samples compared with the control samples (p = 0.047 and p < 0.001, respectively). Similarly, the expression of the inhibitor HHIP was significantly lower in the CD samples compared with the controls (p < 0.001) or the UC (p = 0.008) samples. HHIP expression was also lower in the UC samples compared with the controls (p = 0.001) (Fig. 3c).

In contrast to the lower levels of Hh components in the inflamed colonic mucosa, the expression of WNT1, a component of the Wnt signaling pathway, was significantly higher in the CD (p = 0.001) and UC (p = 0.015) samples compared to the control samples. The expression of BMP4, which is present in intestinal epithelial cells and is a target of both the Hh [22] and Wnt [23] pathways, was significantly lower in the CD and UC samples compared with the control samples (p < 0.001). BMP4 expression was also lower in the CD samples compared with the UC samples (p < 0.023). The expression of BMP7, which is also present in the colon [32], was significantly lower in the CD and UC (p < 0.001) samples compared with the control samples (Fig. 3d). To further investigate the Wnt signaling pathway in the context of IBD, we studied the expression of β-catenin, a target of the Wnt pathway, by immunohistochemistry (Supplementary Figure 3).

The Hh pathway is involved in the control of apoptosis in the colonic epithelium

Because the differential expression of Hh components was predominantly observed in the epithelial layer, we next investigated the potential role of the Hh pathway in the control of epithelial cell death by apoptosis.

Initially, we analyzed apoptosis in the inflamed and non-inflamed colonic mucosa of CD patients, UC patients, and controls using the TUNEL technique. A significant increase in TUNEL-positive cells was detected in the colonic epithelium of the inflamed CD mucosa compared with the non-inflamed areas (p = 0.005) and with controls (p < 0.001). UC samples also showed a significantly higher number of TUNEL-positive cells in the colonic epithelium compared with controls (p = 0.001) (Fig. 4a, b).

Fig. 4
figure 4

Hh signaling is inversely associated with apoptosis in the colonic mucosa. Apoptotic cells in the colonic mucosa of patients with IBD were detected using the TUNEL assay, as shown by representative photomicrographs. Negative controls were incubated without the Tdt enzyme. Length bars represent 50 µm (a). The percentages of TUNEL-positive cells in the colonic mucosa are individually represented. Horizontal bars represent the medians of 15 samples in each group (b). Isolated colonic epithelial cells were analyzed for caspase-3 activity under various conditions. Significant changes are highlighted for control (a and b, p = 0.018), CD (c, p = 0.018; d, p = 0.028), and UC (e and f, p = 0.018) mucosa (c). GLI1 mRNA levels quantified by real-time PCR were significantly inversely correlated with caspase-3 activity in the colonic epithelial cells (d)

Full size image

To confirm the results obtained from the studies of the tissue sections, we isolated epithelial cells from colon explants and measured the activity of caspase-3 under different conditions. The basal levels of caspase-3 activity were significantly higher in the CD colonic epithelial cells than in the control (p = 0.002) or UC (p = 0.01) colonic epithelial cells. In addition, the activity of caspase-3 significantly increased, compared to the non-stimulated, vehicle-treated cultures, following the incubation of control (p = 0.018), CD (p = 0.018 and p = 0.028), and UC (p = 0.018) epithelial cells with Gant61 or anti-CD95. In contrast, caspase-3 activity significantly decreased (p = 0.018) after treatment of control, CD and UC epithelial cells with Shh (Fig. 4c).

To further investigate the involvement of the Hh pathway in colonic epithelial apoptosis, we measured GLI-1 mRNA expression by real-time PCR. Levels of GLI-1 showed a significant negative correlation with the activity of caspase-3 in colonic epithelial cells (r −0.704; p = 0.001) (Fig. 4d).

The Hh pathway modulates inflammatory cytokines in colon explants

To investigate the potential role of the Hh pathway in intestinal inflammation, we cultured colon explants from control, CD and UC under different conditions, and the respective supernatants were harvested for measurement of the concentration of cytokines by ELISA. Colon explants were exposed to the agonist recombinant Shh peptide; to the antagonists anti-Shh, Gant61, or cyclopamine; and to combinations of Shh + Gant61 or Shh + cyclopamine.

The basal levels of TNF-α were significantly higher in CD supernatants compared with control (p < 0.001) and UC samples (p = 0.047), while, in the UC samples, the levels were also higher than in the controls (p < 0.045). In CD explants, TNF-α levels further increased following exposure to anti-Shh (p = 0.012), Gant61 or cyclopamine (p = 0.018), while levels decreased after treatment with Shh (p = 0.012) (Fig. 5a). Similarly, the basal levels of IL-17 were significantly higher in CD samples compared with the control (p = 0.002) and UC (p = 0.036) samples. IL-17 concentrations further increased in the CD samples following the Hh pathway blockade with anti-Shh (p = 0.009), Gant61 or cyclopamine (p = 0.005), and IL-17 concentrations decreased after Shh (p = 0.005) stimulation. In the UC samples, IL-17 levels increased after blockade with anti-Shh (p = 0.02) and Gant61 (p = 0.012), and IL-17 levels decreased after Shh (p = 0.01) stimulation (Fig. 5b). The basal levels of TGF-β were also significantly higher in CD explants compared with the controls (p = 0.001) and the UC (p = 0.037) samples. Following the exposure to the Hh inhibitors anti-Shh, Gant61, or cyclopamine, the concentrations of TGF-β increased in the CD samples (p = 0.022), and TGF-β levels decreased after Shh (p = 0.005) stimulation. In UC explants, TGF-β levels in the supernatants also increased after blockade with anti-Shh (0.007) or Gant61 (p = 0.025) (Fig. 5c). In contrast, the basal levels of IL-10 were significantly lower in the CD samples compared with the controls (p = 0.042), and the basal levels of IL-10 increased following treatment with Shh (p = 0.007) (Fig. 5d).

Fig. 5
figure 5

Effect of Hh modulation on cytokine production in colon explants. Colon explants from control, CD, and UC were cultured for 24 h at 37 °C. Supernatants were used to measure the concentration of cytokines by enzyme-linked immunosorbent assay. Dot plots show the levels of TNF-α (a), IL-17 (b), TGF-β (c), and IL-10 (d) under various conditions. Significant differences are highlighted: a, p = 0.012; b, p < 0.018; c, p = 0.005; d, p < 0.01; e, p = 0.01; f, p < 0.02; g, p = 0.005; h, p = 0.022; i, p < 0.03; and j,p = 0.007. Values are expressed as picograms of cytokine per milliliter of culture supernatant and are normalized by protein content. Horizontal bars represent the medians of 10 experiments in each group

Full size image

Hh pathway blockade induces monocyte chemotaxis

In another set of experiments, we investigated the ability of the Hh pathway to interfere with monocyte chemotaxis in a transwell system. While Shh failed to induce the migration of monocytes, a remarkable response was observed upon Hh blockade. Treatment with Gant61 significantly increased the attraction of monocytes (p = 0.044) to a level similar to that observed with the well-established monocyte chemoattractants FCS (p = 0.048) or MCP-1 (p = 0.039) (Fig. 6).

Fig. 6
figure 6

Monocyte chemotaxis induced upon exposure to Hh agonist or antagonist. Purified peripheral blood monocytes were incubated in the upper chambers of a 5-µm-pore-size polycarbonate filter in a transwell system, with cell culture medium treated with rShh or Gant61 in the lower chambers. Cell culture media without any treatments were used as negative controls, while fetal calf serum (FCS) and the chemokine MCP-1 were used as positive controls. Values represent the fold changes for individual experiments and were arbitrarily normalized to 1. Horizontal bars represent the medians. Significant changes are highlighted. Differences were analyzed using ANOVA on ranks with Dunnett’s test

Full size image

Hh pathway blockade induces proliferation but inhibits migration of human intestinal fibroblasts

To investigate additional functional effects of the Hh pathway on mesenchymal cells, we analyzed the potential role of Hh in cell proliferation, as measured by the cellular incorporation of BrdU at 24, 48, and 72 h, and in cell migration, as measured in modified Boyden chamber assays, both for HIF (Figs. 7, 8) and skin fibroblasts (Supplementary Figure S4).

Fig. 7
figure 7

The effects of different stimuli involving the Hh signaling pathway on the proliferation of human intestinal fibroblasts (HIF). For analyzing changes in the proliferative activity of HIF in response to different treatments, the cellular incorporation of BrdU was measured at 24, 48, and 72 h. BrdU incorporation was significantly greater in Gant61-treated control HIF compared with the Shh-treated cells at 48 h (p = 0.031), and with Shh- (p = 0.011) and vehicle-treated (p = 0.024) cells at 72 h (a). Significant increases in BrdU incorporation were also detected in the Gant61-treated CD HIF compared with Shh- (p = 0.015) or vehicle-treated cells (p = 0.042), at 72 h (b). Horizontal bars represent the medians, boxes represent the 25th and 75th percentiles, and vertical bars represent the ranges of 4 independent experiments. Significant values are highlighted

Full size image
Fig. 8
figure 8

The effects of different stimuli involving the Hh signaling pathway on the migration of human intestinal fibroblasts (HIF). The light microscopy panels show images of primary control (a) and Crohn’s disease HIF (c) migration rates into scratch sites following growth in medium containing vehicle (DMSO), rShh, Gant61, or Shh with Gant61, for 8 and 24 h (b, d). Length bars represent 20 µm. Horizontal bars represent the medians, boxes represent the 25th and 75th percentiles, and vertical bars represent the ranges of 4 independent experiments. Significant values are highlighted

Full size image

In regard to HIF proliferation, we analyzed the response of CD and control samples upon exposure to different treatments concerning the Hh pathway. No significant differences among the treatment groups were detected in the first 24 h. However, significant increases in BrdU incorporation were detected in the Gant61-treated control HIF compared with the Shh-treated cells at 48 h (p = 0.031), and compared with Shh- (p = 0.011) and vehicle-treated (p = 0.024) cells at 72 h (Fig. 7a). Significant increases in BrdU incorporation were also detected in the Gant61-treated CD HIF compared with Shh- (p = 0.015) or vehicle-treated cells (p = 0.042), at 72 h (Fig. 7b).

Exposure of either control or CD HIF to Shh did not change cell migration significantly. However, after treatment with Gant61, both control and CD HIF migrated significantly less at 24 h compared to Shh-treated cells (p = 0.049 and p = 0.039, respectively), or vehicle-treated cells (p = 0.033 and p = 0.018, respectively) (Fig. 8).

Discussion

In this study, we investigated whether Hh signaling could be involved in the intestinal inflammation underlying IBD pathogenesis. We demonstrated that the expression levels of Hh pathway components are markedly reduced in inflamed IBD mucosa, where apoptotic rates are greater than in the controls and in the non-inflamed mucosa of patients with CD. In the normal and non-inflamed mucosa, Hh components were predominantly present in the superficial colonic epithelium, while, in the lamina propria, co-localization of the Hh transcription factor Gli-1 was predominantly detected among non-immune cells, mostly mesenchymal cells. In addition, we showed, for the first time, that the higher concentrations of TNF-α, IL-17 and TGF-β produced by IBD mucosal explants are increased even more after exposure to Hh inhibitors but are decreased after treatment with recombinant Hh. To further investigate the potential mechanistic role of the Hh pathway in intestinal inflammation, we demonstrated in vitro that Hh signaling is a critical modulator of monocyte chemotaxis and of fibroblast migration and proliferation.

In the past few years, Hh signaling has been reported to be a regulatory element in the gastrointestinal tract, with important roles in maintaining intestinal stem cells [15], patterning the intestinal crypt-villus axis and mediating epithelial-mesenchymal interactions [13, 33, 34]. In addition, the down-regulation of the Hh pathway has been associated with intestinal inflammation in experimental models [14, 21, 35]. On the other hand, the results from a recent study using the murine dextran sodium sulfate model of colitis proposed the existence of a crosstalk, mediated by miR-146a, between NOD2 and Hh signaling in the context of gut inflammation [36]. Regarding human IBD, the few data available on Hh expression have been controversial. For instance, similar to a previous study [20], we also found a remarkable reduction in Hh pathway expression in the intestinal mucosa of patients with IBD. In contrast, the results from another study suggested that the Hh signaling pathway could be up-regulated in the chronic intestinal inflammation of IBD [37]. These discrepancies are likely to result from the intrinsic details of the experiments, including the use of distinct antibodies, differences in patient selection, the use of different therapeutic regimens, and differences in mucosal sampling. These discrepancies may also result from the time-dependent outcome of cell–cell interactions within the gut in the context of a dynamically fine-controlled system.

In terms of the complex interplay between the epithelial and mesenchymal compartments, it is believed that Hh ligands secreted by the intestinal epithelium deliver a paracrine signal to the underlying mesenchyme [24, 33], which expresses signaling targets such as Ptch-1 and Gli-1 [14, 21]. Indeed, Hh signaling from the epithelium toward mesenchymal cells, including intestinal subepithelial myofibroblasts, was suggested to modulate epithelial proliferation [13] and inflammatory signaling in the intestinal lamina propria [21]. In agreement with this, our findings showing the concomitant reduction of Hh expression in the lamina propria, especially among mesenchymal cells, further support the notion of an overall lower Hh signaling within the intestinal mucosa of patients with IBD, particularly in CD. Moreover, these findings underscore the importance of the Hh pathway as a key mediator of the crosstalk between epithelium and mesenchyme, with a major role in modulating the inflammatory response within the intestinal mucosa.

To further investigate the expression of Hh components in human IBD, we assessed genes directly and indirectly related to the Hh pathway by quantitative real-time PCR. Overall, the inflamed areas of the colonic mucosa of patients with IBD expressed lower mRNA levels of the ligands IHH and SHH; the transcription factors GLI1, GLI2, and GLI3; the receptors PTHC1 and SMO; and the pathway regulator HHIP. These results confirm that there is a general down-regulation of the HH pathway in IBD, which is more pronounced in CD. Although the concomitant down-regulation of HHIP, negative regulator of the HH pathway, may appear paradoxical in this study, it is in accordance with findings from a recent investigation on HHIP effects in human bronchial epithelial cells. Unexpectedly, the authors did not observe a significant enrichment of canonical HH pathway members other than HHIP among genes differentially expressed by microarray analysis after HHIP knockdown. Similar to our data, these results appear to support the idea of a crosstalk between HHIP and genes from distinct pathways, probably determining reciprocal influences among each other [38]. On the other hand, the expression of WNT1, an antagonist of the Hh pathway in epithelial cells [13], was markedly increased in IBD samples. Nevertheless, regarding bone morphogenetic proteins (Bmps), molecules with a well-established role in epithelial homeostasis [39] known to antagonize the Wnt pathway in the intestine [40, 41], this study revealed a marked down-regulation of both BMP4 and BMP7 in IBD samples. In addition to representing direct targets of the Hh pathway, at least during development [42], Bmps have also been suggested to be mediators of the epithelial–mesenchymal interaction in response to Hh signaling in both the mouse [22] and the amphibian intestine [43]. Although gene expression in the chronically inflamed intestine may reflect the response to inflammatory signals and epigenetic modifications, these findings point to the existence of a complex network comprised of multiple genes and interacting pathways, of which Hh appears to play a central role.

In this study, the lower expression of Hh pathway components in the intestinal mucosa, characteristically located at the superficial epithelium coincided with the higher apoptotic rates of the inflamed colonic mucosa of patients with IBD. These findings prompted us to investigate a possible role of the Hh pathway in the control of epithelial cell death by apoptosis. Previous studies have established that epithelial cell death constitutes a ubiquitous phenomenon in IBD [44, 45], and it has been associated with chronic intestinal inflammation [46–48], implications to barrier function and increased permeability to luminal contents [49, 50]. Similar to a previous study from our group, in which the Hh pathway was shown to play a critical role in the fate of colon cancer cell lines [25], here, we demonstrated that Hh signaling protects epithelial cells from apoptotic cell death. In fact, in isolated colonic epithelial cells, Hh stimulation resulted in a reduction in apoptotic rates, in contrast to a drastic increase in apoptotic rates observed following Hh pathway blockade. The involvement of the Hh pathway in colonic apoptosis was further corroborated by the demonstration of a negative correlation between apoptosis and GLI-1 mRNA expression in isolated epithelial cells. In addition to effects on apoptosis, inhibition of Hh signaling has also been previously shown to increase crypt cell proliferation and epithelial cell migration from crypt to villous tip [14]. In this regard, Wnt signaling, an important regulator of epithelial cell differentiation that antagonizes Indian Hh in the colon [23] was also overexpressed in IBD inflamed samples in this study. Moreover, we showed that β-catenin, a target of the Wnt pathway is strongly expressed in the superficial epithelium of IBD inflamed samples, as opposed to the predominant localization in the bottom of the crypts in the normal, non-inflamed intestinal mucosa. This finding is in agreement with previous data suggesting that β-catenin transactivation can be induced by inflammation [51], and this finding supports the concept of an imbalance between Hh and Wnt signaling in IBD.

Another relevant finding of this study is the demonstration of an Hh-dependent modulation of cytokines in mucosal colonic explants. For example, the higher concentration of TNF-α, IL-17, and TGF-β from CD explants increased even more after the Hh pathway blockade, but decreased after stimulation with the Hh ligand. These results are in agreement with previous reports suggesting that Hh signaling blockade would lead to intestinal inflammation in humans [20] and mice [14, 20, 21, 35]. It is noticeable that the inflammatory process underlying IBD is complex and involves the interplay of several inflammatory mediators, including key molecules such as TNF-α and IL-17, especially in CD pathogenesis [52, 53]. However, the few studies previously addressing the relationship between Hh signaling and the production of pro-inflammatory cytokines revealed a reduction of pro-inflammatory cytokines, including TNF-α, in an experimental model of arthritis [54] and in a model of Helicobacter-induced gastric metaplasia [55]. In contrast, key molecules in IBD such as IL-12, IL-17, and IL-23 [56, 57] were markedly up-regulated in the dextran sodium sulfate-induced colitis model using Gli1 +/lacZ transgenic mice [20].

Likely as important as the numerous pro-inflammatory mediators, anti-inflammatory cytokines such as IL-10 and TGF-β play a crucial role in IBD pathogenesis. The role of IL-10 has been emphasized in CD, based on animal knockout models developing ileocolitis [58], and in studies showing increased disease susceptibility in association with IL-10 gene polymorphisms [59, 60]. TGF-β is a pleiotropic cytokine produced by both immune and non-immune cells and is well established as a key molecule for down-regulating inflammation in the intestine [61]. In patients with IBD, TGF-β expression is increased in the inflamed mucosa [62], consistent with the results from this study. In addition, TGF-β has been proposed to be a regulatory molecule capable of promoting the differentiation of naïve T cells into Th17 and T regulatory cells [63–65]. In fact, working in conjunction with IL-6, TGF-β was shown to drive the differentiation of naïve T cells to the Th17 phenotype [66]. Interestingly, pro-inflammatory mediators, including TNF-α, also appear to control SMAD7 posttranscriptionally, therefore indirectly blocking TGF-β signaling in IBD intestinal mucosa [67]. Of note, another study suggested that the prolonged loss of Ihh would lead to a progressive leukocyte infiltration in the lamina propria as well as infiltration of fibroblasts in parallel with TGF-β overexpression, resulting in the development of intestinal fibrosis [35]. Together with our results, this information highlights the complex interactions of inflammatory mediators within the intestinal mucosa and supports the direct or indirect role of Hh signaling in the interaction between non-immune and immune mechanisms as a negative feedback loop controlling intestinal inflammation and healing.

Because, in this study, we found reduced expression of Hh components in the inflamed mucosa, especially in CD samples, paralleled by an expected increased density of collagen fibers, we hypothesized that the loss of Hh signaling might disrupt the interaction between the epithelium and mesenchyme. Regarding this, the results obtained in vitro with isolated HIF showed that Hh signaling blockade induces cell proliferation, similar to the findings of a study on human idiopathic pulmonary fibrosis [68] and similar to the results obtained using an experimental model of kidney fibrosis [69]. On the other hand, the inhibition of HIF migration by Hh blockade that was demonstrated in this study appears to be in contrast with the findings obtained using lung fibroblasts [70]. Therefore, it is likely that distinct milieus, comprising different interacting cellular and molecular elements, may lead to divergent responses mediated by Hh signaling. In this sense, the abnormal regulation of activated mesenchymal cells in the intestinal mucosa, namely fibroblasts and myofibroblasts, is responsible for the overproduction of extracellular matrix in IBD in response to paracrine signals from multiple potential sources [71–73]. In fact, fibrogenesis represents the endpoint of a multifactorial process involving the activation of mesenchymal cells as well as several growth factors and mediators produced by immune and non-immune cells in the intestinal mucosa, frequently leading to stricture formation, a potentially serious complication, particularly in CD [74].

To date, the only candidate gene study investigating the possible link between germline GLI1 variation and IBD susceptibility supported the idea that Hh signaling is essential in the intestinal response to inflammatory stimuli and that reduced Hh signaling pathway and GLI1 function may be associated with IBD. Of note, a specific GLI1 single nucleotide polymorphism was highly associated with IBD, mostly with UC, in a large cohort of three independent Northern European populations [20]. In agreement with that study, our results strongly indicate that an abnormal down-regulation of the Hh signaling pathway is implicated in IBD; however, it is more pronounced in CD. This apparent discrepancy might be related to different genetic backgrounds of the populations but it would also be interesting to investigate the possible crosstalk between the different cell signaling pathways such as Wnt/β-catenin and Notch [75]. Furthermore, it is likely that posttranscriptional modifications of Hh genes, such as GLI1 [76], and epigenetic changes, including subsets of microRNAs targeting the Hh signaling pathway [77, 78], can also influence the outcome of Hh signaling.

In conclusion, the Hh signaling pathway plays a crucial functional role in intestinal homeostasis, as it is critically involved in epithelial integrity and cytokine production within the colonic mucosa. In vitro studies suggest that epithelial Hh signaling provides negative feedback to the lamina propria, inhibiting leukocyte migration and fibroblast proliferation, but favoring fibroblast migration. Taken together, these data support the notion of an abnormal intestinal down-regulation of Hh signaling in IBD, which might be implicated in disease pathogenesis and may represent a potential novel therapeutic target in chronic intestinal inflammation.