Article Text


Dietary modulation of peroxisome proliferator-activated receptor gamma
  1. R Marion-Letellier1,
  2. P Déchelotte1,
  3. M Iacucci2,
  4. S Ghosh2
  1. 1
    Appareil Digestif Environnement Nutrition (ADEN EA 4311), Institute for Biomedical Research, European Institute for Peptide Research (IFR23), Rouen University and Rouen University Hospital, Rouen, France
  2. 2
    Gastroenterology Section, Imperial College Faculty of Medicine, Hammersmith Hospital Campus, London, UK
  1. Dr R Marion-Letellier, Appareil Digestif Environnement Nutrition (ADEN EA 4311), Faculté de Médecine-Pharmacie, Institut Fédératif de Recherches Multidisciplinaires sur les Peptides (IFR23), 22, boulevard Gambetta, 76183 Rouen cedex, France; marion_rachel{at}


Peroxisome proliferator-activated receptor gamma (PPARγ) is a nuclear receptor that regulates intestinal inflammation. PPARγ is highly expressed in the colon and can be activated by various dietary ligands. A number of fatty acids such as polyunsaturated fatty acids or eicosanoids are considered as endogenous PPARγ activators. Nevertheless, other nutrients such as glutamine, spicy food or flavonoids are also able to activate PPARγ. As PPARγ plays a key role in bacterial induced inflammation, anti-inflammatory properties of probiotics may be mediated through PPARγ. The aims of the present review are to discuss of the potential roles of dietary compounds in modulating intestinal inflammation through PPARγ.

Statistics from

Peroxisome proliferator-activated receptor gamma (PPARγ) is a member of a family of nuclear receptors that interact with nuclear proteins which act as co-activators and co-repressors.1 In the non-liganded state, PPARγ interacts with co-repressors that, in the deacetylated state of activity, inhibit gene expression (fig 1). After binding to a ligand, PPARγ forms a heterodimer with the retinoid X receptor (RXR), recruits co-activators containing histone acetylase activity and binds to the peroxisome proliferator response element gene promoter, leading to regulation of gene transcription. Acetylation of histone proteins seems to relieve the tightly packed structure of the chromatin, allowing the RNA polymerase II complex to bind and initiate transcription. PPARγ is activated by dietary ligands2 or by synthetic ligands such as thiazolidinediones, a class of anti-diabetes drug.1 PPARγ also transactivates genes in the metabolic and cell proliferative pathways, which are not the subject of this review.

Figure 1

Peroxisome proliferator-activated receptor gamma (PPARγ) interacts with nuclear proteins that act as co-repressors (cR) and co-activators (cA). In the unliganded state, interactions between PPARγ and co-repressors inhibit gene expression. After binding to a ligand (L), PPARγ forms a heterodimer with the RXR, recruits co-activators and binds to the peroxisome proliferator response element gene promoter, leading to regulation of gene transcription.

Although PPARγ is expressed in various tissues and cell types including pancreas, liver, kidney, immune cells such as lymphocytes, monocytes, macrophages and dendritic cells, high amounts are found in adipose tissue and colon.3 The colon is a major tissue which expresses PPARγ in epithelial cells and, to a lesser degree, in macrophages and lymphocytes.3 PPARγ plays a role in the regulation of intestinal inflammation. Indeed, both natural2 and synthetic4 PPARγ ligands have beneficial effects in different models of rodent colitis59 (table 1). Moreover, administration of synthetic PPARγ ligands reduces the production of inflammatory cytokines in dendritic cells10 and can inhibit several other pro-inflammatory pathways in Caco-2 cells.11 In addition, patients with ulcerative colitis have been shown to have a reduced expression of PPARγ, particularly in colonic epithelial cells, without any mutation in the PPARγ gene,12 while in human Crohn’s disease, the gene encoding PPARγ (pparγ) has recently been identified as a susceptibility gene.13 PPARγ is also a key receptor mediating the effect of 5-aminosalicylic acid (5-ASA) in inflammatory bowel disease (IBD).14 A synthetic PPARγ agonist, rosiglitazone, has been evaluated in a small pilot study15 and then in a randomised placebo-controlled trial with 105 patients and this drug seems to have anti-inflammatory effects in mild-to-moderately active ulcerative colitis.16

Table 1 Demonstrated anti-inflammatory effects of nutrients through peroxisome proliferator-activated receptor gamma (PPARγ) in intestine

The aims of the present review are to discuss of the potential roles of dietary compounds in modulating intestinal inflammation through PPARγ.


Similarly to other nuclear hormone receptors, PPARγ possesses a central DNA-binding domain, a C-terminal ligand-binding domain and two transcription-activation function motifs (AF-1 and AF-2).4 The secondary structure of the ligand-binding domain forms a very large pocket called ligand-binding pocket which allows the binding of different size ligands.17 Binding of ligands to PPARγ leads to a conformational change in the receptor which allows recruitment of co-activator proteins to then induce transcriptional activation. Ligands bind to PPARγ at different docking sites.18 The transcriptional activity of PPARγ is regulated by post-translational changes such as phosphorylation or ubiquitination.19

PPARγ interferes with inflammatory pathways by interactions with transcription factors such as nuclear factor kappa B (NF-κB),20 activating protein-1 (AP-1),21 signal transducer and activator of transcription (STAT)22 and nuclear factor activated T cell (NFAT).23 For example, PPARγ is able to form a complex with the NF-κB subunit p65 at a nuclear lever and this complex is exported from the nucleus leading to an altered expression of pro-inflammatory NF-κB-mediated gene expression.20

Dietary lipids include mono-, di- and triglycerides, sphingolipids, free fatty acids, cholesterol, plant sterols, various pigments and fat-soluble vitamins. A number of fatty acids are considered endogenous PPARγ ligands and activators. Butyrate is a short-chain fatty acid that is produced by commensal intestinal bacteria. In vitro, incubation with 2 mmol/l of butyrate for 3 and 6 days strongly increased PPARγ expression in enterocyte-like Caco-2 cells.24 This effect is specific of butyrate because it was not reproducible by incubation with two other short-chain fatty acids, propionate and valerate.24 Using dominant negative mutant HT-29 cells, Schwab et al25 have shown that butyrate at 4 mmol/l inhibits lipopolysaccharide (LPS)-induced NF-κB activation through PPARγ. Butyrate from 0.01 to 1 mmol/l also decreased permeability in HT-29 cells in a dose-dependent manner26 and this effect is reproducible by a synthetic PPARγ agonist troglitazone (1 μmol/l).26


Fatty acids and their metabolites can affect gene expression via binding to PPARγ.27 The effect of n-3 PUFAs on PPARγ is well documented. In human monocytic THP-1 cells in response to LPS, pre-treatment with 100 μmol/l of α-linolenic acid or docosahexaenoic acid (DHA) downregulate production of pro-inflammatory cytokines interleukin 1β (IL1β), IL6 and tumour necrosis factor α (TNFα) through activation of PPARγ28 but the effect of DHA is more marked. In vitro, we have recently tested a range of five PUFAS in enterocyte-like Caco-2 cells in response to a pro-inflammatory inducer IL1β, and PUFAs, in particular eicosapentaenoic acid (EPA) and DHA, are able to reduce production of IL6 and IL8 through PPARγ.29 Indeed, the anti-inflammatory effects were reproducible by troglitazone, a PPARγ agonist and abolished by GW9662, a PPARγ antagonist.29 Recently, it has been shown in reporter assays that EPA at 100 μmol/l is able to activate PPARγ in non-stimulated HT-29 cells and this effect is abolished by co-administration of GW9662, a PPARγ antagonist.30 The effects of n-3 PUFAs in IBD have also been evaluated in clinical trials.31 32

The beneficial effects of n-3 PUFAs on intestinal inflammation are also mediated independently of PPARγ (fig 2). Indeed, n-3 PUFAs could modulate fatty acid composition in cell membrane phospholipids leading to a decrease of eicosanoids derived from arachidonic acid, such as prostaglandin E2 (PGE2) or leukotriene B4 (LTB4). In addition, resolvins, compounds derived from EPA such as RvE1, exerts strong anti-inflammatory properties, in particular in 2,4,6-trinitrobenzene sulfonic acid (TNBS)-induced colitis.33 The beneficial effects of n-3 PUFAs in intestinal inflammation may also be mediated through inhibition of toll-like receptor 4 (TLR4).3436

Figure 2

Mechanisms of action of n-3 polyunsaturated fatty acids (PUFAs) in intestinal inflammation. The n-3 PUFAs activate peroxisome proliferator activated receptor γ (PPARγ), which inhibits the nuclear factor kappa B (NF-κB) signalling pathway. The effects of n-3 PUFAs may also inhibit toll-like receptor 4 (TLR4). The n-3 PUFAs could also modulate fatty acid composition in cell membrane phospholipids leading to a decrease of inflammatory eicosanoids derived from arachidonic acid and to an increased production of anti-inflammatory compounds such as resolvins. These regulatory pathways lead to decreased generation of pro-inflammatory cytokines and decreased expression of adhesion molecules, resulting in an inhibition of intestinal inflammation.

Linoleic acid is the major PUFA in human diet and several derivatives from linoleic acids have shown properties of activation of PPARγ. For example, conjugated linoleic acid (CLA) is a group of isomers of linoleic acid with biological functions. Beef and bovine milk are two major sources of CLA (table 2). In vitro, in reporter assays conducted in the mouse macrophage cell line RAW264.7, various isomers of CLA at 200 μmol/l were able to activate PPARγ37 and inhibit interferon (IFN)γ-induced pro-inflammatory mediators such as inducible nitric oxide synthase (iNOS) and cyclo-oxygenase-2 (COX-2). The inhibitory effect of CLA in RAW264.7 was specific for PPARγ since the transfection of a construct with a mutation of the ligand-binding domain of PPARγ leading to a protein with a strong dominant-negative activity abolished the effect of CLA on the iNOS promoter.37

Table 2 Dietary sources of natural peroxisome proliferator-activated receptor gamma (PPARγ) modulators

In vitro incubation with free fatty acid isolated from bitter gourd seed rich in 9-cis,11-trans,13-trans-CLA at 25 μmol/l upregulated PPARγ mRNA levels in non-stimulated Caco-2 cells.38 We also tested the effect of the isomer (9Z,11E)-CLA on Caco-2 cells treated with IL1β but we failed to attenuate inflammation by incubation with CLA.29 CLA can also be produced by lactic acid bacterial strains.39 Interestingly, Ewaschuk et al40 have recently evaluated the ability of probiotics to produce CLA. They have shown that incubation with conditioned medium from CLA-producing probiotics induced apoptosis in two intestinal epithelial cell lines HT-29 and Caco-2 cells and increased PPARγ expression.40 Lactobacillus-produced CLA inhibits the activation of NF-κB in Helicobacter pylori-infected gastric epithelial cells leading to decreased production of IL8.41 In vivo, Bassaganya-Riera et al5 and Bassanganya-Riera and Hontecillas42 have evaluated the effects of CLA in several models of colitis. Using a model of recombinant mice with specific deletions of PPARγ in the colon, they demonstrated the specific involvement of PPARγ in the protective effect of CLA in mouse colitis.5

Nitrolinoleic acid (LNO2) is a nitrated derivative from linoleic acid and LNO2 has been shown to be a strong endogenous PPARγ ligand (Ki = 133 nmol/l) in reporter assays43 but its effects in intestine needs to be investigated. Other derivatives of linoleic acid, such as 13-hydroxyoctadecandienoic acid and 13-oxooctadecadienoic acid, have been also identified in reporter assays as endogenous ligands for PPARγ in intestinal cell lines (Caco-2 and HCT-116).44 Recently, 13-oxo-ODE has been identified as a strong inducer of PPARγ in human intestinal epithelial cell line HT-29.45 The effects on PPARγ and NF-κB binding in intestinal epithelial cells from pigs fed with oxidised frying oils or fresh fat have been investigated but no significant difference was found between these treatments.46

Gamma linolenic acid (GLA) is an n-6 PUFA and we tested its effect in IL1β-treated Caco-2 cells and found that GLA reduced IL8 production.29


15-Deoxy-Δ12,14-prostaglandin J2 (15d-PGJ2) is a fatty acid-derived metabolite and is used as a strong PPARγ inducer in many studies. In a reporter assay, 15d-PGJ2 at 1 μmol/l significantly activated PPARγ in non-stimulated HT-29 cells.47 In both intestinal epithelial cell lines Caco-2 and HT-29, 15d-PGJ2 downregulated IL1β-induced IL8 secretion6 by preventing the activation of NF-κB. In LPS-treated HT-29 cells, 15d-PGJ2 at 10 μmol/l decreased IL8 secretion and TLR4 expression whereas it increased PPARγ mRNA expression.48 In vivo, daily administration of 15d-PGJ2 improved colitis in a rat model by reducing colonic damage and by inhibiting pro-inflammatory mediators such as the production of TNFα, myeloperoxidase activity, iNOS expression and intracellular adhesion molecule-1 (ICAM-1).49 Administration of 15d-PGJ2 and a synthetic PPARγ agonist, rosiglitazone, were also tested in a rat model of stress-induced colonic inflammation.50 In this model, rats were exposed to various immobilisation stress periods resulting in upregulated production of pro-inflammatory mediators such as iNOS, PGE2, myeloperoxidase and an enhanced colonic permeability and bacterial translocation. Pre-treatment with 15d-PGJ2 (120 μg/kg) as well as rosiglitazone counteracted the stress-induced gut dysfunction.50

9-cis-Retinoic acid (9cRA) is a lipid molecule derived from vitamin A (retinol) and 9cRA is able to modulate gene expression by binding to two classes of nuclear receptors: retinoic acid receptors and RXR, the receptor with which PPARγ forms a heterodimer. Nuclear magnetic resonance analysis has recently demonstrated that 9cRA-bound RXRα can modulate PPARγ activation51 serving as the basis of immunomodulator properties of 9cRA. Indeed, activation of naive T cells by dendritic cells in presence of transforming growth factor β (TGFβ) leads to the differentiation into proinflammatory T helper17 (Th17) cells and this differentiation into Th17 is suppressed in presence of retinoic acid and 9cRa.52 Moreover, 9cRA modified human dendritic cell maturation and function53 and this effect is mediated through PPARγ because these changes are abolished by GW9662, a specific inhibitor of PPARγ.


Glutamine is the preferential substrate of enterocytes and is considered as a conditionally essential amino acid in stress situations. To our knowledge, only a single study has investigated the effect of glutamine on PPARγ expression. Using a rodent model of gut ischaemia–reperfusion, the authors studied the effects of luminal glutamine and arginine on iNOS and haem oxygenase-1 (HO-1) expression and PPARγ activity by DNA binding activity.54 Whereas luminal glutamine is able to increase PPARγ, arginine decreases its activity. Moreover, in the same study, pre-treatment with GW9662, a PPARγ inhibitor, abrogates the effect of glutamine. These results suggest that glutamine may also be a PPARγ agonist. A recent study has investigated the effect of arginine, an amino acid considered to be an immunonutrient. This study evaluated the effect of arginine supplementation in a weaned pig model of LPS-induced gut injury.55 It was found that arginine treatment demonstrated a beneficial effect on gut injury such as a decrease of jejunal TNFα mRNA level and arginine also enhanced PPARγ mRNA level although this failed to reach statistical significance (p<0.094).


Curcumin (or diferuloylmethane) is a natural phenolic compound derived from Curcuma longa and is widely used as a food additive. Curcumin is the yellow pigment in turmeric (curry powder). In vitro, curcumin is well known to have anti-inflammatory properties as it is an inhibitor of NF-κB, a key transcription factor in intestinal inflammation.5658 The beneficial effect of curcumin in rodent models of colitis is also well documented5962 but very few studies have investigated whether its effect is mediated through PPARγ. In a rat model of TNBS-induced colitis, Zhang et al63 have shown that curcumin administration (30 mg/kg/day) decreased Th1 cytokine mRNA expression whereas it restored PPARγ mRNA level and protein expression compared with TNBS alone. The authors did not perform experiments using co-administration of a PPARγ antagonist to determine whether it blocks the curcumin effect. In a rat model of sepsis, pre-treatment with curcumin improved sepsis via reduction of TNFα expression in liver and this effect is mediated through PPARγ because the beneficial effect of curcumin was abolished by GW9662.64 Curcumin has also been evaluated in a small pilot trial that included five patients with ulcerative colitis and five with Crohn’s disease.65 This open-label study has shown some encouraging effect of curcumin on IBD such as a reduction of medication requirement. Nevertheless, a double-blind clinical trial is now required to confirm these preliminary results. One of the mechanisms of beneficial effects of curcumin in inflammation may be via PPARγ agonism (table 3).

Table 3 Peroxisome proliferator-activated receptor gamma (PPARγ) modulators which have shown some efficacy in inflammatory diseases of the human intestine

In various spicy foods, two other compounds derived from pepper have been shown to activate PPARγ: capsaicin (from cayenne pepper) in human intestinal epithelial cell line HT-2966 and Kochujang extract (Korean fermented red pepper paste) in adipocytes67 but their respective effect on intestinal inflammation have not yet been investigated. Ginsenosides are compounds derived from ginseng but their effects on PPARγ seem to be dependent on the type of ginsenoside. Indeed, both studies conducted in 3T3–L1 adipocytes have shown opposite effects of ginsenoides: ginsenoside-(20S)-protopanaxatriol has been described as a strong PPARγ inducer68 whereas ginsenoside Rh2 has been described as an inhibitor of PPARγ.69 These compounds are bioactive and require further investigation as anti-inflammatory agents.


A recent study has highlighted, by docking simulations, the potential of two flavonoids as new strong PPARγ agonists: ψ-baptigenin (EC50 = 2.9 μmol/l) and hesperidin (EC50 = 6.6 μmol/l).70 At 30 μmol/l, both compounds induced 2.9-fold and 2.5-fold PPARγ expression in THP-1 cells, respectively, and this effect was abolished by GW9662 at 5 μmol/l.70 Other compounds derived from botanicals have been identified as PPARγ agonists but their effects have been mainly identified in lipid metabolism. For example, epigallocatechin gallate, a compound from green tea, has also been shown as a PPARγ inducer in adipocytes.71 Resveratrol is a polyphenol derived from grapes and its effect has been evaluated in intestinal cell lines.71 By using reporter assays, Ulrich et al72 have shown that resveratrol at a concentration of 100 μmol/l inhibits cell growth through PPARγ72 in two intestinal epithelial cell lines, Caco-2 and HCT-116.

Micronutrients such as vitamin E (α-tocopherol and γ-tocopherol) or selenium also upregulates PPARγ expression in the SW480 human colon cancer cell line73 and macrophages,74 respectively.


Bacterial products are sensed by the TLRs and there is preliminary evidence of cross-talk between TLR and PPARγ pathways. Dubuquoy et al12 have illustrated the cross-talk between bacteria and PPARγ in intestinal inflammation. First, they have shown in reporter assays that LPS treatment activates PPARγ in enterocyte-like Caco-2 cells.12 They also validated in vivo the existence of the cross-talk between TLR4 and PPARγ by analysing the expression of PPARγ in colon of mice with different flora: germ-free or colonised with commensal or human intestinal flora.12 Bacteroides thetaiotaomicron, a commensal bacterium, is able to attenuate inflammation in intestinal epithelial cells by activating the nuclear export of the NF-κB subunit.20 Infection with Helicobacter pylori also upregulated the expression and the mRNA level of PPARγ in the human gastric cancer cells Kato III.75 In nutritional intervention, the effects of probiotics such as Saccharomyces boulardii and Lactobacillus casei have been evaluated in vitro in the intestinal epithelial cell line HT-29 in response to pro-inflammatory inducers.76 77 S boulardii upregulated PPARγ mRNA and protein levels whereas it decreases cytokine-induced IL8 production.76 Pre-treatment with L casei also exerts anti-inflammatory effects such as a decreased mRNA expression of IL8, TLR4 and COX-2 and increased PPARγ mRNA in a dose-dependent manner.77 L casei at 107 cfu/ml induced a 2-fold activation of PPARγ in a reporter assay as well as ciglitazone and 15d-PGJ2, both PPARγ agonists at 10 μmol/l.77 As described below, Ewaschuk et al40 have studied in vitro the effect of cocktail of probiotics (VSL#3) on PPARγ expression. They also determine the effects of probiotics (VSL#3) in a murine model of sepsis and pre-treatment with oral probiotics improves gut barrier function such as a decrease of pro-inflammatory cytokine and an altered bacterial translocation.78 They evaluated whether the effects of probiotics were mediated through PPARγ using GW9662, a PPARγ inhibitor, and found that administration of GW9662 abolished the protective effect of probiotics.78 A very recent study has also investigated the effect of probiotics (Lactobacillus crispatus M247) on murine intestinal epithelial cells.79 This probiotic strain is able to produce H2O2 and uses it as a molecular signal transducer to induce PPARγ in intestinal epithelial cells.79


Enteral nutrition is likely to have a multitude of effects such as microflora modification or epithelial nutrition, which are not the subject of this review. In addition, anti-inflammatory effects of enteral nutrition are already available.32 8082 Table 3 gives a brief description of natural PPARγ modulators that have shown some efficacy in human intestinal inflammatory diseases. The mechanism of action of defined formula diets as anti–inflammatory agents is unclear.83 Since both elemental and polymeric diets appear equally effective, exclusion of a dietary antigen is unlikely to be the explanation for efficacy. Defined formula diets are ineffective in ulcerative colitis and hence alteration of intestinal bacterial flora as a result of a defined formula diet is unlikely to be the sole explanation of the mechanism of action of such dietary intervention.84 Increasing evidence implicates the fat composition of defined formula diet as being critically important in its anti-inflammatory effect.

An intriguing study was carried out by Gassul et al.85 Sixty-two consecutive patients with Crohn’s disease were randomised to receive two polymeric enteral diets with different fat composition to induce remission in active Crohn’s disease and compared with steroid treatment. One diet was rich in oleic acid (a mono-unsaturated fatty acid (MUFA)) and the other rich in linoleic acid (a PUFA). The hypothesis was that linoleic acid, being an arachidonic acid precursor, would have little anti-inflammatory efficacy. Surprisingly the MUFA diet was significantly less effective than the PUFA diet, with overall remission rates being respectively 20% for MUFA, 52% for PUFA and 79% for steroids. PPARγ agonism may be speculated to provide an explanation for this surprising finding.

Bamba et al86 demonstrated that a high fat diet (long-chain fatty acids) added to an elemental diet reduced the therapeutic in active Crohn’s disease in a small randomised study. The remission rate after 4 weeks was 80% in 10 patients who received low fat, 40% in 10 patients who received medium fat and 25% in eight patients who received high fat, respectively. A high-fat diet, in all likelihood, provides pro-inflammatory drive, but either a low fat diet or PUFAs, which act as PPARγ ligands, may attenuate inflammation.

The effects of omega-3 fatty acids in Crohn’s disease reduce the production of pro-inflammatory leukotrienes and cytokines such as IL1β and TNFα.87 Although a relatively small Italian study (n = 78) found fish oil supplementation to reduce relapses compared to placebo,88 this was not confirmed by a larger study,89 which found remission rates to be similar between placebo and ethyl ester fish oil concentrate (n = 204).89 European Prospective Investigation into Cancer and Nutrition-1 (EPIC-1) and EPIC 2 studies, recent large randomised double blind controlled trials, demonstrated no beneficial effects of omega-3 fatty acids for the prevention of relapse in patients with CD patients.31 A Cochrane systematic review prior to the EPIC studies concluded that there was no definite evidence of efficacy of fish oil in Crohn’s disease.90 However, both small intestinal and colonic Crohn’s disease were recruited into such studies, whereas a PPARγ agonist can be speculated to be more effective in colonic inflammation. 5-Aminosalicylic acid, a known PPARγ agonist, is also known to be largely ineffective in Crohn’s disease, but efficacious in ulcerative colitis.

Epidemiological data91 suggest that in ulcerative colitis a marginally significant positive association with an increasing percentage intake of energy from total PUFAs with an odds ratio for trend across quartiles of 1.19 (95% confidence interval, 0.99 to 1.43; p = 0.07). In the future further epidemiological studies in other populations, and randomised controlled trials in ulcerative colitis are fundamental in order to define better the role natural PPARγ ligands in ulcerative colitis. Given that both 5-aminosalicylic acid and rosiglitazone as PPARγ ligands are effective in ulcerative colitis, and PPARγ is expressed in colonic mucosa, the effect of such dietary manipulation using natural PPARγ ligands is worthy of further investigation in ulcerative colitis and colonic Crohn’s disease. The exact magnitude of exposure may be critical as the dose–response relationship of natural PPARγ ligands is not linear.


View Abstract


  • Competing interests: None.

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.