Article Text

Download PDFPDF

Original article
Intestinal steroidogenesis controls PPARγ expression in the colon and is impaired during ulcerative colitis
  1. Guillaume Bouguen1,2,3,4,5,
  2. Audrey Langlois1,2,
  3. Madjid Djouina1,2,
  4. Julien Branche1,6,
  5. Dine Koriche1,7,
  6. Edmone Dewaeles1,2,
  7. Alice Mongy1,2,
  8. Johan Auwerx8,9,
  9. Jean-Frederic Colombel1,2,6,10,
  10. Pierre Desreumaux1,2,6,
  11. Laurent Dubuquoy1,2,
  12. Benjamin Bertin1,2,11
  1. 1Université Lille Nord de France, Lille, France
  2. 2Inserm U995, Lille, France
  3. 3Service des Maladies de l'Appareil digestif, University Hospital of Rennes, Pontchaillou, France
  4. 4Inserm, UMR991, Liver Metabolism and Cancer, Rennes, France
  5. 5Université de Rennes 1, Rennes, France
  6. 6CHU Lille, Service des Maladies de l'Appareil Digestif et de la Nutrition, Hôpital Claude Huriez, Lille, France
  7. 7CHU Lille, Service de Chirurgie Digestive et Transplantations, Hôpital Claude Huriez, Lille, France
  8. 8Institut Clinique de la souris, Illkirch, France
  9. 9Laboratory of Integrative and Systems Physiology, Ecole polytechnique fédérale de Lausanne, Lausanne, Switzerland
  10. 10Department of Gastroenterology, Icahn School of Medicine at Mount Sinai, New York, USA
  11. 11UDSL, Faculté des Sciences Pharmaceutiques et Biologiques, Lille, France
  1. Correspondence to Dr Benjamin Bertin, Inserm U995, Amphi J&K, Boulevard du Prof. Jules Leclercq, Lille cedex 59045, France; benjamin.bertin-2{at}


Background and aims Immune tolerance breakdown during UC involves the peroxisome proliferator-activated receptor-γ (PPARγ), a key factor in mucosal homoeostasis and the therapeutic target of 5-aminosalycilates, which expression is impaired during UC. Here we assess the impact of glucocorticoids (GCs) on PPARγ expression, focusing especially on extra-adrenal cortisol production by colonic epithelial cells (CECs).

Methods Activation of PPARγ in the colon was evaluated using transgenic mice for the luciferase gene under PPAR control (peroxisome proliferator response element-luciferase mice). Protein and mRNA expression of PPARγ were evaluated with colon fragments and purified CEC from mice. Cortisol production and steroidogenic factor expression were quantified in human CEC of patients with UC and those of controls. Gene expression knockdown by short hairpin RNA in Caco-2 cells was used for functional studies.

Results GCs were able to raise luciferase activity in peroxisome proliferator response element-luciferase mice. In the mice colons and Caco-2 cells, PPARγ expression was increased either with GCs or with an inducer of steroidogenesis and then decreased after treatment with a steroidogenesis inhibitor. Cortisol production and steroidogenic factor expression, such as liver receptor homologue-1 (LRH-1), were decreased in CEC isolated from patients with UC, directly correlating with PPARγ impairment. Experiments on Caco-2 cells lacking LRH-1 expression confirmed that LRH-1 controls PPARγ expression by regulating GC synthesis in CEC.

Conclusions These results demonstrate cortisol control of PPARγ expression in CEC, highlighting cortisol production deficiency in colonocytes as a key molecular event in the pathophysiology of UC.

View Full Text

Statistics from

Significance of this study

What is already known on this subject?

  • During UC, peroxisome proliferator-activated receptor-γ (PPARγ) expression involved in gut homoeostasis and targeted by 5-aminosalicylates is initially decreased.

  • Glucocorticoids (GCs) increase the expression of PPARγ in adipocytes.

  • Colonic epithelial cells (CECs) in mice are able to produce a significant amount of GCs under the control of liver receptor homologue-1 (LHR-1).

What are the new findings?

  • Synthetic and endogenous GCs control PPARγ expression in CEC.

  • Human CECs are able to produce cortisol that is defective during UC as well as LRH-1 expression.

  • Combination therapy of 5-aminosalycilates and GCs enhance PPARγ activity as compared with monotherapy with 5-aminosalycilates alone.

How might it impact on clinical practice in the foreseeable future?

  • These results establish a molecular basis for developing and testing therapeutic strategies in the field of UC by targeting cortisol production and PPARγ expression and strongly support the enhancement of therapeutic effects with the use of corticosteroids in association with PPARγ agonist.


IBDs, namely UC and Crohn's disease (CD), are chronic, relapsing immune-mediated disorders of the GI tract with no known aetiology. Despite sharing some clinical manifestations, UC and CD are distinct pathophysiological diseases. In contrast to CD, inflammation during UC is mainly restricted to colonic mucosa.1 Recent studies suggest that intestinal epithelial cell dysfunction may actively participate in the pathogenesis of UC.1 ,2 Such dysfunction may be attributed to abnormal expression of inflammatory regulators in the colons of patients with UC, such as peroxisome proliferator-activated receptor-γ (PPARγ).

The colon is a major tissue primarily expressing PPARγ in epithelial cells.3 ,4 PPARγ belongs to the nuclear receptor superfamily of ligand-inducible transcription factors that has been described as a main regulator of adipocyte differentiation as well as a factor of insulin sensitisation.5 In addition, PPARγ is a critical regulator of the inflammatory response,5 ,6 involving positive gene regulation by binding the PPARγ-retinoid X receptor (RXR) heterodimer to the peroxisome proliferator response element (PPRE) located in the promoter of target genes. PPARγ also interferes with transcription factors involved in the inflammatory response, such as NF-κB, activating protein-1 (AP-1) or retinoic acid receptor-related orphan receptor (ROR)γt, a mechanism often referred to as ‘transrepression’.7

Evidence suggests that PPARγ plays a key anti-inflammatory role during intestinal inflammation, where PPARγ ligands inhibit inflammation and reduce disease severity in various experimental models of colitis.8–10 Conversely, disruption of PPARγ expression in mouse colonic epithelial cells (CECs) increases susceptibility to dextran sodium sulfate induced colitis.11 In patients with mild to moderate UC, administration of PPARγ agonists such as thiazolidinediones and 5-aminosalicylic acids (5-ASA) has induced and maintained clinical remission.12–14 Of note, an impaired expression of PPARγ mRNA and protein in the CEC was observed in patients with UC but not CD.13 ,15 Given the role of PPARγ in colonic homoeostasis and its therapeutic implications, this defect observed in inflamed and healthy colons of patients with UC suggests a primary role of colonic PPARγ impairment in the pathophysiology of UC.15 Importantly, the lack of PPARγ expression may prevent the optimal efficacy of PPARγ agonist. Thus, restoring normal levels of this receptor's expression in the colon of patients with UC may improve thiazolidinediones or 5-ASA treatment. However, aetiology of impaired PPARγ expression in CECs is still unknown.

Endogenous local glucocorticoids (GCs) synthesised by epithelial cells of the intestinal crypt are a significant source of extra-adrenal endogenous GCs in mice and are of growing interest with an increased insight to their numerous properties, particularly by the ability to maintain immune homoeostasis in the intestinal mucosa.16 Given that GCs increase PPARγ expression in adipocytes,17 this study investigates whether GCs are able to control the expression of PPARγ in vitro and in vivo and the relationship between the expression of intestinal PPARγ and epithelial steroidogenesis, especially in the colon of patients with UC.

Material and methods

Animal procedure

Animal experiments were performed in the accredited Pasteur Institute animal facilities (Institut Pasteur de Lille, France—licence no B59-35009) according to governmental guidelines and those of the Nord-Pas de Calais Ethical Committee for animal use.

In order to study PPAR activation in vivo, transgenic mice were generated carrying a PPRE controlled luciferase (Luc) reporter gene cassette (PPRE-Luc mice). The construction of this model has been previously described.18 PPRE-Luc mice (male) were divided into four groups (3<n<6 per group): control (intraperitoneal injection of normal saline buffer+intrarectal administration of Roswell Park Memorial Institute medium (RPMI)), MetPred (intraperitoneal injection of methylprednisolone 50 mg/kg+intrarectal administration of RPMI), 5-ASA (intraperitoneal injection of normal saline buffer+intrarectal administration of 5-ASA 10 mg/kg) and MetPred+5-ASA (intraperitoneal injection of methylprednisolone 50 mg/kg+intrarectal administration of 5-ASA 10 mg/kg). Mice received treatments once daily according to study group during 4 days. Bioluminescence was measured at day 0 (baseline before treatment), day 2 and at day 4, so that each mouse was its own control. Bioluminescence activity obtained at day 0 was subtracted from the activity measured at day 2 and day 4. Bioluminescence reporter imaging was assessed by a charge-coupled device camera (IVIS Imaging System 100 Series, Caliper Life Sciences). Mice received an intraperitoneal injection of 30 mg/kg D-luciferin (Euromedex) 10 min before bioluminescence quantification under isofluran (2%) anaesthesia. Mice were placed in the light-tight chamber where a grayscale photo of the animals was first taken with dimmed light. Photon emission was then integrated over a 5 min period. Images were processed using Living Image 3.2 Software (Caliper Life Sciences, France). Luminescence measurements are expressed as the integration of the average brightness/pixel unit expressed as photon counts per second (p/s) on a surface representative of the rodent abdomen.

Experimental procedures on C57BL/6 mice are described in online supplementary material and methods.

Cell culture and gene expression knockdown

Liver receptor homologue-1 (LRH-1) and PPARγ knockdown were obtained using the pSUPER.retro system (OligoEngine). Forward and reverse target sequences corresponding to nucleotides 105–123 of the human PPARγ1 mRNA (5′-GCCCTTCACTACTGTTGAC-3′) and to nucleotides 1446–1464 of the human LRH-1 variant1 mRNA (5′-AGGATCCATCTTCCTGGTT-3′) were cloned into the BglII/XhoI restriction sites of the pSUPER.retro vector giving the ShPPAR and ShLRH-1 construct, respectively. A negative control pSUPER.retro vector plasmid containing the sequence 5′-ACGCTGAGTACTTCGAAAT-3′ targeted against the luciferase gene was also generated (ShLuc construct). Each construction was transfected in Caco-2 cells (ATCC n°CRL-2102) using Nucleofector technology from Amaxa Biosystems according to the manufacturer's protocol. Stably transfected clones were selected 24 h post-transfection with complete culture medium supplemented with puromycin (5 µg/mL). The silencing of PPARγ expression was checked by quantitative reverse transcription PCR and western-blot analysis. The silencing of LRH-1 expression was checked by quantitative reverse transcription PCR. Once established, ShPPAR, ShLRH-1 and ShLuc cell lines were maintained in complete medium supplemented with 2.5 µg/mL puromycin.

Other experiments on Caco-2 cells are described in online supplementary material and methods.

Patient tissue samples

A local ethics committee approved the study and all subjects gave informed consent (No. DC-2008-642). Colonic samples were obtained from control patients (patients having undergone surgery for colorectal cancer and patients with diverticulitis of the sigmoid) and from patients with an established diagnosis of UC according to international criteria.

Intestinal epithelial cells were isolated from resected colonic specimens as previously described.15 Mucosal layers were dissected away from the muscular and serosal layers and incubated in RPMI medium 1640 containing 0.5 mmol/L dithiothreitol for 15 min. Mucosal fragments were then rinsed in calcium-free and magnesium-free Hank's balanced salt solution supplemented with heat-inactivated fetal calf serum (5%), L-glutamine (5%), and 500 mmol/L EDTA (5%), ampicillin (5%), gentamicin (1%), amphotericin B (1%) adjusted for pH 7. After manual extraction of residual mucosal layers, samples were centrifuged for 5 min and filtered through a nylon mesh. Cells were rinsed twice in RPMI medium. The epithelial cells were collected at the interface of a Percoll gradient (1 mL of Percoll 1.085 of density, 4 mL of Percoll 1.055 of density with cells and 3 mL of RPMI, centrifugation 600 g for 20 min) and rinsed. The viability of epithelial cells was evaluated by Trypan blue staining, which was greater than 85%. After several washes, cells were frozen at −80°C. Purity was evaluated by PCR using A33, cytokeratin-18 and CD3ε oligonucleotides.

Cortisol dosage

Intracellular cortisol was measured in intestinal epithelial cells from patients using the The DetectX Cortisol Immunoassay kit (ArborAssays, Michigan, USA). Epithelial cell pellets were suspended in 300 µL of dichloromethane and air-dried overnight under a chemical hood. The pellets were solubilised in 100 µL of the provided Assay Buffer and cortisol was dosed in duplicate according to manufacturer's instructions. Epithelial cortisol concentration was determined using a cortisol standard curve and normalised with cell number.

For measuring cortisol in culture media, cortisol was first concentrated by freeze-drying. One mL of culture media was freeze-dried, suspended in 100 µL of assay buffer and cortisol dosage was duplicated according to manufacturer's instructions.

Statistical analysis

All graphics were drawn and analysed with the GraphPad Prism 5 Software (GraphPad Software, San Diego, California, USA) and the StatXact V.7.0 (Cytel Studio) software using a non-parametric Wilcoxon-Mann-Whitney test. Values of p≤0.05 were considered statistically significant. Correlation tests were made using the non-parametric Spearman test. In each figure *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.


The association of steroids with 5-ASA increases the bioluminescent activity of transgenic PPRE-Luc mice

In order to approach temporal and spatial regulation of PPARγ expression and activation in vivo, PPRE-Luc mice were used to assess a potential relationship between steroids and PPARγ expression. In this genetic model, luciferase activity from PPRE-Luc mice is a marker of PPARγ activation. Based on the hypothesis that GCs induce PPARγ expression without triggering its biological activity, we used a combination of methylprednisolone and a PPARγ agonist (5-ASA) to determine whether GCs are able to induce PPARγ expression. As shown in figure 1, intra-abdominal luciferase activity was significantly enhanced in co-treated mice compared with mice treated with vehicles or 5-ASA alone. This result suggested induction of PPARγ expression in the abdominal cavity of mice by GCs and its activation by 5-ASA lead to an overall increase in bioluminescence activity in PPRE-Luc mice. Given the intrarectal administration of the PPARγ agonist and the low bioavailability of 5-ASA, our results suggested that GCs increased PPARγ expression in the colon.

Figure 1

The combination of methylprednisolone and 5-aminosalicylate (5-ASA) increased the abdominal bioluminescent activity of the peroxisome proliferator response element-luciferase (PPRE-Luc) mice. Bioluminescent activity from transgenic PPRE-Luc mice was assessed by a charge-coupled device camera after intraperitoneal injection of D-luciferin. (A) Pictures show the luminescence distribution of representative mice treated with vehicles (control), methylprednisolone alone (MetPred) and methylprednisolone+5-ASA (MetPred+5-ASA). (B) Photon emission was quantified using an electronic grid drawn in the abdomen area below the liver area in mice treated with a normal saline buffer (intraperitoneal vehicle) or methylprednisolone (intraperitoneal MetPred) in association or not with intrarectal delivery of 5-ASA. Results represent the mean±SEM of the fold average radiance (photons/sec/cm2/sr) of bioluminescent activity measured at day 4. The bioluminescent activity measured for control animals was used as a reference (two independent experiments, 3<n<6 animals per group for each experiment; *p<0.05).

GCs control PPARγ expression in CECs

To further explore regulation of PPARγ expression by GCs in the colon, variations of PPARγ expression in the colon of C57BL/6 wild type mice were assessed. Mice treated with methylprednisolone displayed a significant rise of PPARγ mRNA expression restricted to the colon, without any variation in PPARγ mRNA levels in the small intestines of mice treated either with methylprednisolone or with vehicle (figure 2A). Western blot analysis of colon samples showed that steroid treatment enhanced PPARγ protein (figure 2B). These results indicate that GCs increased PPARγ expression in the colon. The relationship between GCs and PPARγ was confirmed using metyrapone, a well-known potent inhibitor of GC synthesis. As shown in figure 2C, PPARγ mRNA expression was significantly decreased in colon samples of mice treated with metyrapone compared with controls (figure 2C), confirming the control of PPARγ colonic expression by GCs in vivo.

Figure 2

Upregulation of colonic peroxisome proliferator-activated receptor-γ (PPARγ) expression by glucocorticoids. (A) mRNA expression of PPARγ in the colon and ileum of mice treated with methylprednisolone (n=10, MetPred) and compared with control animals (n=9, Vehicle). (B) Representative immunoblots of the colon of mice treated with methylprednisolone or vehicle showing that PPARγ expression is increased in the colon of mice treated with methylprednisolone. Vehicle and MetPred samples were run under the same conditions in two separate gels. (C) mRNA expression level of PPARγ in the colon following in vivo treatment of mice with intraperitoneal metyrapone (100 mg/kg; n=10). (D and E) Colonic epithelial cells were purified from mice treated with methylprednisolone (n=10, MetPred) or vehicle (n=10, Vehicle). The expression of PPARγ was analysed at the mRNA (D) and protein levels (E). In each experiment, the mRNA expression of PPARγ was assessed by quantitative RT-PCR (reverse transcription PCR); β-actin was used to normalise expression level. In the dot plots, horizontal bar indicates the median value. In panel D, results represent the mean±SD.*p<0.05; **p<0.01.

In order to determine whether GCs are able to regulate PPARγ in CECs, CECs were isolated from mice treated with or without methylprednisolone and assessed for PPARγ expression. The purity of isolated cells was assessed by PCR amplification of CEC markers (cytokeratin 18 and A33) and lymphocyte markers (CD3ɛ) (see online supplementary figure S1). Figure 2D,E show that PPARγ mRNA (p<0.05) and PPARγ protein were increased in purified CEC from methylprednisolone treated mice compared with untreated mice.

Taken together, the results demonstrate that GCs (endogenous and synthetic) control in vivo PPARγ expression in the colon and this regulation occurs, at least in part, in CECs.

CEC production of GCs regulates CEC PPARγ expression

Intestinal epithelial cells have recently been shown to be a significant source of extra-adrenal GC synthesis.16 ,19 Ex vivo fragments of mice colon were first stimulated with metyrapone and evaluated for PPARγ expression to assess whether intestinal epithelial GCs could be involved in the regulation of expression of PPARγ in the colon. As shown in figure 3A, the inhibition of steroidogenesis by metyrapone in colon fragments resulted in a decrease in PPARγ protein expression as compared with DMSO treated colons, suggesting that intestinal steroidogenesis is involved in the regulation of PPARγ expression. Next, PPARγ expression was modulated by controlling CEC steroidogenesis using lipopolysaccharide (LPS), which induces intestinal GC synthesis.20 Colonic tissues were harvested and stimulated ex vivo with or without metyrapone for 2 h or 6 h to block GC synthesis. Results showed that LPS treatment led to a strong and significant rise in PPARγ mRNA level in the mouse colon (figure 3B grey bars). When in situ GC synthesis was blocked with metyrapone, PPARγ mRNA returned to its basal level within 2 h, demonstrating that colonic production of GCs is able to regulate PPARγ expression (figure 3B black bar). After 6 h, PPARγ mRNA expression returned to its basal level in DMSO-treated tissue and remained lower in metyrapone-stimulated ex vivo colonic fragments. PPARγ protein expression in colonic tissue was decreased in the entire colon at 6 h following metyrapone treatment (figure 3C). Importantly, there was a robust decrease in the PPARγ protein level in CEC purified from ex vivo-stimulated colonic tissues (figure 3C), highlighting the control of PPARγ expression specifically by CEC steroidogenesis.

Figure 3

Glucocorticoids controlled peroxisome proliferator-activated receptor-γ (PPARγ) expression in epithelial colonic cells (A) Detection of PPARγ protein in mice colon stimulated with metyrapone ex vivo. Proteins were extracted from fragments of colon stimulated with metyrapone (200 µg/ml) or DMSO for 6 h. (B and C) Detection of PPARγ in the colon of mice that received intraperitoneal injection of LPS prior ex vivo treatment with metyrapone (200 µg/mL) or DMSO (B) Modulation of expression of PPARγ mRNA assessed by quantitative reverse transcription PCR in the whole colon. Results represent the mean±SD of the relative PPARγ expression normalised to β-actin level (*p<0.05). (C) Detection of PPARγ protein in the colon of mice that received intraperitoneal injection of LPS prior ex vivo treatment with metyrapone (200 µg/mL) or DMSO. Protein extracts were obtained from whole colons (left panel) or purified intestinal epithelial cells (right panel). Protein extracts from purified colonic epithelial cells were obtained from two to three pooled colons. (D and E) Modulation of expression of PPARγ assessed by western blot in Caco-2 colorectal cancer cell line; (D) Analysis of protein extracts from Caco-2 cells treated with metyrapone (200 µg/mL) or DMSO during 24 h in steroid-free medium and further cultured 24 h with DMSO or dexamethasone (a synthetic glucocorticoid) (E) Analysis of protein extracts from Caco-2 cells treated with metyrapone alone or concomitant to dexamethasone stimulation.

Besides being a model of CEC, the Caco-2 colorectal cancer cell line expresses steroidogenic enzymes and produces significant amounts of cortisol which could be blocked with metyrapone.19 Stimulation of Caco-2 cells with metyrapone strongly impaired PPARγ protein expression (figure 3D). This reduction was reversible and normal levels of PPARγ protein were recovered after changing the medium containing metyrapone with fresh medium lacking steroidogenic blocker (figure 3D). This recovery was even stronger when dexamethasone (a synthetic GC) was added to the fresh medium (figure 3D). In addition, co-treatment of Caco-2 cells with dexamethasone and metyrapone was able to reverse the PPARγ protein deficiency induced by metyrapone alone (figure 3E).

Combined, these data suggest that intestinal GCs (cortisol in human or corticosterone in mice) produced by CECs are involved in the regulation of expression of PPARγ within CECs.

Patients with UC display a cortisol deficiency in intraepithelial cells

Based on our previous results, a defect of endogenous cortisol synthesis in human CEC could impact PPARγ expression in patients with UC. To test this hypothesis, epithelial cells were isolated from surgical specimens of patients with and without UC (figure 4A). A significant decrease of expression of PPARγ in CECs from UC specimens compared with controls was confirmed. Considering CEC cortisol production, intraepithelial cortisol concentrations from isolated CECs from UC specimens were significantly lower than those from cells purified from control specimens (figure 4B). Based on previous results demonstrating the ability of GCs to regulate PPARγ expression, this lack of cortisol strongly supports the link between colonic epithelial cortisol production deficiency and impaired PPARγ expression during UC.

Figure 4

Defective epithelial expression of peroxisome proliferator-activated receptor-γ (PPARγ) and production of cortisol by colonic epithelial cells (CECs) in patients with UC. (A) CECs were purified from resected colon specimens of control subjects (n=8) or untreated patients with UC (n=12). Quantitative expression of PPARγ mRNA was assessed by real-time PCR and normalised to β-actin level (**p<0.01). (B) CECs were purified from resected colon specimens of control subjects (n=6) or untreated patients with UC (n=6). Intracellular cortisol quantity was measured by ELISA as described in the Methods section and normalised to the cell count. Horizontal bar indicates the median value. *p<0.05; **p<0.01.

Decreased PPARγ expression in patients with UC is correlated with liver receptor homologue-1 (LRH-1, NR5a2), and enzymes involved in steroidogenesis

To strengthen the relationship between intestinal epithelium steroidogenesis and PPARγ expression, the potential correlation between PPARγ and other factors involved in the cortisol synthesis was assessed (see online supplementary figure S2 for a description of the cortisol synthesis pathway). Colonic UC specimens displayed an overall decrease of nearly all steroidogenesis enzymes compared with control specimens, with a significant decrease in mRNA expression of 3-β-hydroxysteroid dehydrogenase 1 and 2 (3βHSD1, 3βHSD2) and 11-β-hydroxysteroid dehydrogenase 1 (11βHSD1) (figure 5A and see online supplementary figure S3). Of note, CYP11A1 gene and CYP1B1 gene expression were not significantly detectable in isolated CECs from controls and UC specimens. A strong correlation between the expression level of PPARγ and 3βHSD1 was observed (rs=0.8372; p<0.0001) (figure 5A). LRH-1 is a nuclear receptor characterised as being a critical regulator of steroidogenic enzyme.21 In this study, LRH-1 expression was significantly decreased specifically in purified CECs from patients with UC (p<0.001), and correlated with PPARγ mRNA impaired expression (rs=0.8191; p<0.0001) (figure 5B). Transcription levels of impaired steroidogenic enzymes in patients with UC correlated with those from LRH-1 (rs=0.7729, p<0.0001 between 3βHSD-1 and LRH-1; rs=0.6321, p<0.05 between LRH-1 and 3βHSD-2; rs=0.6397, p<0.01 between LRH-1 and 11βHSD1) (see supplementary figure S3).

Figure 5

Defective epithelial expression of steroidogenic factors in patients with UC. In addition to the lack of peroxisome proliferator-activated receptor-γ (PPARγ) expression in colonocytes of patients with UC, a decrease of (A) 3-β-hydroxysteroid dehydrogenase 1 (3βHSD1) and (B) Liver receptor homologue-1 (LRH-1) was observed. Colonic epithelial cells (CECs) were purified from resected colon specimens of control subjects (n=8) or untreated patients with UC (n=12). Quantitative expression of mRNA was assessed by real-time PCR and normalised to β-actin level. Horizontal bar indicates the median value (*p<0.05; ***p<0.001). Transcript levels of (A) 3-βHSD1 and (B) LRH-1 mRNA levels in epithelial cells from controls and patients with UC were correlated with PPARγ mRNA level. A non-parametric Spearman test determined the p value.

These data demonstrate that the steroidogenic pathway is disrupted in CECs of patients with UC with a direct impact on PPARγ expression. They also suggest that LRH-1 could be indirectly involved in the control of PPARγ expression by regulating cortisol production.

LRH-1 regulates PPARγ expression via GC synthesis in CEC

In order to provide evidence of LRH-1 involvement in the control of PPARγ expression through GC production, LRH-1 expression in Caco-2 cells was downregulated using RNA interference. A short hairpin RNA strategy was used to create a stable cell line in which the expression of LRH-1 was reduced by 80% compared with the control cell line (named ShLuc), which expressed a control shRNA directed against the luciferase gene (figure 6A). Downregulation of LRH-1 resulted in a reduction of cortisol synthesis and release (figure 6B). The ShLRH-1 cell line expressed low level of 3βHSD1 (reduction of 75%) and 3βHSD2 (reduction of 80%) enzymes as compared with ShLuc control cells (figure 6C), confirming the ability of LRH-1 to control the expression of 3βHSD enzymes as suggested in CECs from patients with UC. In addition, downregulation of LRH-1 led to a nearly 40% reduction in mRNA expression encoding PPARγ, resulting in a strong decrease of PPARγ protein production (figure 6D). In order to assess whether decreased expression of PPARγ in LRH-1 deficient cells was due to the lack of cortisol synthesis rather than a direct control of PPARγ gene expression by LRH-1, ShLRH-1 cells were treated with exogenous GCs that aimed to restore a normal expression of PPARγ. Treatment with dexamethasone on ShLRH-1 cells re-established mRNA and protein expression level of PPARγ to a normal state (figure 6E).

Figure 6

Liver receptor homologue-1 (LRH-1) knockdown in Caco-2 cells led to peroxisome proliferator-activated receptor-γ (PPARγ) decrease through impaired glucocorticoid synthesis. (A) Caco-2 colorectal cell line knock-down for LRH-1 (ShLRH-1) expressed significantly fewer transcripts of LRH-1 than control cell line luciferase (ShLuc), (*p<0.05) and (B) produced significantly less cortisol. (C) ShLRH-1 cell line expressed significantly fewer transcripts of 3-β-hydroxysteroid dehydrogenase 1 (3βHSD1) and 3βHSD2 (***p<0.001; ****p<0.0001). (D) ShLRH-1 cell line expressed significantly less PPARγ than ShLuc cell line (*p<0.05). (E) PPARγ expression level was restored by stimulation of ShLRH-1 cells with dexamethasone (1 µM during 24 h) (**p<0.01). Quantitative expression of mRNA was assessed by real-time PCR. Results represent the mean values of ±SD of the fold change in the corresponding gene expression normalised to β-actin level (n=6). The expression level measured in ShLuc cells was used as a reference and set to one, except in panel E where relative expression was used. PPARγ protein level was assessed by a western blot test.

It might be argued that PPARγ could be upstream of LRH-1, regulating LRH-1 expression which would further impact steroidogenic enzyme expression. The same ShRNA strategy was therefore used to create a stable Caco-2 cell line in which PPARγ expression was strongly reduced (figure 7A). As shown in figure 7B,C, neither the expression of LRH-1 nor the expression of 3βHSD1 and 3βHSD2 were changed in the ShPPARγ cell line as compared with ShLuc control cells, supporting that PPARγ does not control the expression of LRH-1 and steroidogenic enzymes. These results confirm the regulation pathway identified in CECs in patients with UC, in which LRH-1 is upstream to PPARγ and controls the cortisol synthesis that ultimately regulates PPARγ expression.

Figure 7

Peroxisome proliferator-activated receptor-γ (PPARγ) controlled neither steroidogenic enzyme expression nor the expression of liver receptor homologue-1 (LRH-1). (A) Caco-2 colorectal cell line knockdown for PPARγ (ShPPARg) expressed significantly less transcripts of PPARγ (**p<0.01). (B and C) Compared with ShLuc cells, expression levels of LRH-1, 3-β-hydroxysteroid dehydrogenase 1(3βHSD1) and 3βHSD2 were not modulated in ShPPARγ cells. Quantitative expression of mRNA was assessed by real-time PCR. Results represent the mean values ±SD of the fold change of PPARγ expression normalised to β-actin level (n=6). The expression level measured in ShLuc cells was used as a reference and set to one. Protein level was assessed by a western blot test.


This study demonstrated that PPARγ expression in CECs is directly regulated by GCs in vitro and in vivo and that epithelial steroidogenesis plays a key role as an extra-adrenal source of cortisol in human CECs. It also highlighted the defect of cortisol production by colonocytes as a potential key molecular event involved in the pathophysiology of UC.

Since the 1950s, immunosuppressive and anti-inflammatory properties of GCs have been well recognised and used to treat inflammatory and immune-mediated disorders, although the mechanism of action remains to be understood.22 GCs may have rapid non-genomic effects, such as non-specific interactions of GCs with cytoplasmic and mitochondrial membranes, or mediated by the cytosolic GC receptor and by a membrane-bound GC receptor.23 However, the main effect of GCs is to control the expression of several genes through activation of the GC receptor that bind specific glucocorticoid response elements (GREs) in the promoter region of steroid-sensitive genes.23 In silico analysis of the PPARγ promoter region reveals several putative GREs with high probability of affinity, suggesting that GR might regulate PPARγ expression by direct binding on the promoter region. However, further investigations are required to fully understand the link between PPARγ and GCs. Indeed, it seems that only few genes regulated by GCs contain GRE in their promoter region, making it difficult to identify the bona fide GC target genes.22

Upstream of cortisol, the specific decrease of LRH-1 in CEC is of particular interest. First, lack of LRH-1 expression has been associated with an increased susceptibility of induced colitis in mice and LRH-1 level is reduced in mucosa of patients with UC.24 Second, a recent meta-analysis of genome-wide association scans identified 163 loci associated with IBD. A single nucleotide polymorphism of LRH-1 was found to be associated with UC.25 The current study strengthens the role of LRH-1 during UC pathophysiology and supports an indirect relationship between two nuclear receptors in the CECs. LRH-1 is a transcription factor critically involved in GC synthesis.21 Initially, LRH-1 was found to regulate CYP11A1 in mouse CEC, an enzyme converting cholesterol to pregnenolone, and CYP11B1, an enzyme catalysing inactive 11-deoxycorticosterone to active corticosterone.21 Additionally, our results show the control of two other enzymes involved in the steroidogenesis in Caco-2 cells lacking LRH-1, namely 3βHSD1 and 3βHSD2 which convert (17-OH-) pregnenolone to (17-OH-) progesterone, consistent with findings in ovarian cells.26 These results extend the control of LRH-1 as the main factor in steroidogenic enzyme expression and ultimately cortisol production in CEC. Importantly we found in patients with UC decreased 11βHSD1 specifically in the epithelial cell in addition to the impaired expression of 3βHSDs meaning a blockade of the two pathways of the intraepithelial steroidogenesis that led to a decreased cortisol production and ultimately PPARγ expression in patients with UC.

Endogenous local GCs from CECs are of growing interest with an increased insight into their numerous properties, particularly by the ability to maintain immune homoeostasis in the intestinal mucosa. Absence of endogenous GCs leads to increased disease severity in experimental models of colitis.24 Immunoregulatory properties have been assigned to this extra-adrenal source of GCs, such as modulation of intestinal T lymphocyte activation upon viral infection,16 or suppression of T cell activation and apoptosis promotion.19 Recently, endogenous GCs have been shown to alleviate endoplasmic reticulum stress that triggers and maintains inflammation in patients with active IBD.27 ,28 However, none of these studies described cortisol production in humans and the impact of its deficiency in epithelial cells in UC specimens. The link between endogenous GC synthesis and PPARγ expression in CECs opens a new outlook on understanding gut homoeostasis. Because of anti-inflammatory properties and involvement in UC onset, maintaining expression of PPARγ in CEC is critical.15 As a ligand-inducible transcription factor, activity of PPARγ is related to the level of expression and the presence of a ligand. An important aspect of the biology of PPARγ is the ability to be activated by the usual presence of several natural ligands in the human gut such as fatty acids along with their metabolites and their eicosanoid derivatives.29 ,30 Restoring PPARγ expression is a potential target to overcome treatment failure and maintain remission for patients with UC.13

In these patients, PPARγ ligands, such as 5-ASA or glitazone, and GCs are effective in induction of clinical remission as well as mucosal healing. Moreover, a growing body of evidence suggests that combination therapy with oral or rectally administered 5-ASA agents and corticosteroids may result in a higher rate of remission compared with monotherapy in patients with UC.31 To achieve clinical, endoscopic and histological remission among patients with active UC, the combination of beclomethasone dipropionate, a synthetic corticosteroid and oral or enema of mesalamine (a 5-ASA), has proven to be more effective when compared with monotherapy with either corticosteroid or 5-ASA.32 Although the higher efficacy of combination therapy may be due in part to the additive effects of the two drugs, our results also support that the efficacy of this therapy may be due to its synergistic effects. Highlighted on PPRE-Luc mice, synthetic GC potentiates the effect of 5-ASA by upregulating the expression of PPARγ; this leads to enhance its activity and 5-ASA efficacy. However a limitation from PPRE-Luc experiments is the modest increase in luciferase activity following 5-ASA treatment that could be due in part to the interference of natural ligand of the colon in these mice. Furthermore, the addition of GCs on ex vivo samples of mucosal fragment from patients would be of interest to clearly address this point but the limited viability of such samples precluded this experiment. Adverse events associated with systemic corticosteroid therapy preclude their use except for budesonide, a synthetic corticoid that targets the intestinal mucosa and acts as a topical steroid due to first-pass metabolic processes leading to minimal systemic effects.33 New tablet coatings such as the multimatrix structure allow the drug’s release throughout the colon, targeting specifically CEC.34 ,35 While budesonide multi-matrix system is effective in inducing and maintaining remission in UC,34 ,35 this study suggests better outcomes with combination therapy and new therapeutic approaches to avoid systemic steroid side effects by targeting CEC cortisol production.

Impaired epithelial steroidogenesis in the human colon leads to the impaired expression of PPARγ in patients with UC (see online supplementary figure S4). A new anti-inflammatory property of GC and a new action mechanism for PPARγ expression regulation in CEC were identified. An alternative strategy for patients with UC could be restoring PPARγ expression in epithelial cells by acting on intraepithelial steroidogenesis alone or in addition to treatment with PPARγ ligands.


Rachel Tipton for English revision.


View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • GB and AL contributed equally.

  • Contributors GB, AL, PD, LD and BB designed the study. GB, AL, MD, JB, ED, AM, LD and BB performed experiments and analysis of the data. GB, J-FC, PD, LD, BB wrote the paper and participated in the critical reading of the manuscript. DK and J-FC managed and provided human surgical samples. JA managed and provided transgenic mice.

  • Funding This work was supported by research grants from Ferring, the Association François Aupetit and BREMICI (Shering-Plough).

  • Competing interests GB received lecture fees from Ferring, MSD Pharma, Abbvie. PD has the following conflicts of interest for consulting fees, lecture fees or grant supports in the field of PPAR with Giuliani SpA, Milano, Italy, Procter and Gamble, London, UK, Shire Pharmaceuticals, USA and Lesaffre, Marcq en Baroeul, France. J-FC received consulting fees from Giuliani SpA.

  • Ethics approval Ethics committee of Lille University Hospital (No. DC-2008-642).

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.