Cannabinoid Receptor-2 Ameliorates Inflammation in Crohn’s Disease

This article originally appeared at https://academic.oup.com/ecco-jcc/article/11/11/1369/3977952?login=false

Abstract

Background and Aims

Cannabinoid receptor stimulation may have positive symptomatic effects on inflammatory bowel disease [IBD] patients through analgesic and anti-inflammatory effects. The cannabinoid 2 receptor [CB₂R] is expressed primarily on immune cells, including CD4⁺ T cells, and is induced by active inflammation in both humans and mice. We therefore investigated the effect of targeting CB₂R in a preclinical IBD model.

Methods

Employing a chronic ileitis model [TNF^ΔARE/+ mice], we assessed expression of the CB₂R receptor in ileal tissue and on CD4⁺ T cells and evaluated the effect of stimulation with CB₂R-selective ligand GP-1a both in vitro and in vivo. Additionally, we compared cannabinoid receptor expression in the ilea and colons of healthy human controls with that of Crohn’s disease patients.

Results

Ileal expression of CB₂R and the endocannabinoid anandamide [AEA] was increased in actively inflamed TNF^∆ARE/+ mice compared with controls. CB₂R mRNA was preferentially induced on regulatory T cells [Tregs] compared with T effector cells, approximately 2.4-fold in wild-type [WT] and 11-fold in TNF^∆ARE/+ mice. Furthermore, GP-1a enhanced Treg suppressive function with a concomitant increase in IL-10 secretion. GP-1a attenuated murine ileitis, as demonstrated by improved histological scoring and decreased inflammatory cytokine expression. Lastly, CB₂R is downregulated in both chronically inflamed TNF^∆ARE/+ mice and in IBD patients.

Conclusions

In summary, the endocannabinoid system is induced in murine ileitis but is downregulated in chronic murine and human intestinal inflammation, and CB₂R activation attenuates murine ileitis, establishing an anti-inflammatory role of the endocannabinoid system.

Inflammatory bowel disease, endocannabinoid, regulatory T cell

1. Introduction

Inflammatory bowel disease [IBD], including Crohn’s Disease [CD] and ulcerative colitis, is one of the most costly, prevalent gastrointestinal disorders and affects approximately 1.6 million people in the USA.¹ Chronic inflammation of the gastrointestinal tract in IBD places a significant burden on affected individuals, not only due to poor quality of life and the long-term risk of malignancy, but also due to recurrent symptoms of abdominal pain, bloody diarrhoea, and weight loss, as well as the growth failure, impaired bone mineralisation and pubertal delay commonly seen in paediatric-onset IBD.²

The complex factors underlying IBD pathogenesis are still incompletely understood, but it is generally recognised that IBD emerges due to an aberrant immune response. CD4⁺ T cells act as key mediators of intestinal inflammation and therefore represent an ideal target for therapeutic treatment of IBD. Of particular significance are CD4⁺FoxP3⁺ regulatory T cells [Tregs], which play a crucial role in maintaining gut homeostasis by suppressing inappropriate immune responses. Expression of the transcription factor FoxP3 is vital for the regulatory activity of Tregs, and thus mutated FoxP3 in both humans and mice results in severe autoimmunity, including immune dysregulation, polyendocrinopathy, enteropathy, and X-linked syndrome [IPEX] in humans.³ Novel modulatory approaches which promote the regulatory pathway in T cells, whereby excessive effector cell responses are curtailed, hold considerable potential as future therapeutic approaches for the treatment of IBD.

The limited benefits, rapid loss of efficacy, and multiple long-term risks, including malignancy, infection, and low bone density,⁴ of current immunosuppressive therapies have influenced affected individuals and physicians alike to urge the development of alternate therapeutic strategies for the treatment of IBD. The endogenous cannabinoid [endocannabinoid] system is one pathway that is rapidly gaining attention as a novel therapeutic target for its immunosuppressive effects⁵ and, specifically in IBD, for its possible analgesic and anti-inflammatory effects.⁶ As of March 2017, 28 states and Washington DC have legalised cannabis for medical purposes, and limited observational studies have demonstrated that medical marijuana, through interaction with the endogenous cannabinoid system [ECS], can potentially be used to alleviate a number of IBD-associated intestinal symptoms including reducing nausea, stool frequency, and abdominal pain, while improving appetite and weight gain.⁷ However, it is unclear if the symptomatic relief experienced by patients using cannabis in IBD is accompanied by improvement of intestinal inflammation or any real disease-modifying effect.⁸ Indeed, one such study showed that the patients with Crohn’s disease who used cannabis had worse disease prognosis and higher tendency toward surgical intervention.⁹ It is this uncertainty of the risks, benefits, and true impact of cannabis on IBD that has driven the intensification of basic science research investigating the biological effects of manipulation of the ECS in intestinal inflammation.

The ECS consists of endogenous ligands termed endocannabinoids, including anandamide [AEA] and 2-arachidonoylglycerol [2-AG], which mainly act via two classic cannabinoid heterotrimeric G-protein-coupled receptors: cannabinoid receptor 1 [CB₁R] and cannabinoid receptor 2 [CB₂R]. They also act through non-classical receptors: transient receptor potential vanilloid type 1 [TRPV1] and the orphan G-protein-coupled receptor GPR55.^2,10 The CB₁R modulates neurotransmitter release and is found on central and peripheral neurons, including within the enteric nervous system.¹⁰ The CB₂R has been shown to be mostly expressed on neutrophils, macrophages, subsets of T and B cells, and epithelial cells,¹¹ and thus is mainly associated with immune function.¹² The ECS also includes the proteins involved in endocannabinoid biosynthesis and degradation; fatty-acid amide hydrolase [FAAH] degrades AEA to arachidonic acid and ethanolamine,¹³ and monoacylglycerol lipase [MAGL] degrades 2-AG to arachidonic acid and glycerol.¹⁴

The ECS regulates gastrointestinal function through its effects on gastric secretions, gut-brain signalling, gut motility, and the gut microbiome, and significantly influences murine intestinal inflammation.¹⁵ A recent review on cannabinoids in murine colitis models demonstrated that manipulation of the endocannabinoid system had immunomodulatory effects dependent on complex interactions of specific receptor pathways.² CB₁R activation^16–21 and CB₂R activation^{19,20,22–26} both ameliorated acute colitis. Furthermore, inflamed colonic epithelium and lamina propria [LP] immune cells have increased expression of the CB₂R,²⁷ and activation of epithelial CB₂R inhibits tumour necrosis factor alpha [TNF-α]-induced CXCL8 release in vitro.²⁸

Our studies made use of the TNF^∆ARE/+ mouse model of Crohn’s-like ileitis.²⁹ These mice are generated by the deletion of 69 bp within the AU-rich element of the TNF gene, causing systemic overproduction of TNF and consequently chronic ileal inflammation representative of human CD. The compelling similarities with human CD in the pathogenesis and histological presentation of disease in this model reinforce its value and relevance in our investigation of the role of cannabinoids in mediating T cell homeostasis.

In summary, development of novel therapeutic pathways is imperative in the treatment of IBD. Limited data are supportive of symptomatic improvement with cannabis use in IBD, but the role of cannabinoids in immune response and immunostimulation is incompletely understood. Because CB₂R is expressed by CD4⁺ T cells³⁰ and is upregulated in inflamed LP immune cells, and based on the amalgamation of the murine colitis studies described, a central anti-inflammatory role of the CB₂R is suggested. We hypothesised that stimulation of CB₂R would positively affect disease outcomes in murine ileitis. In our study, we investigated inflammation-induced changes in cannabinoid tone in both murine and human intestinal tissue, and assessed the role of cannabinoids as endogenous modulators of intestinal inflammation by defining the effect of pharmacological CB₂R targeting on inflammation, T cell function, and phenotype.

2. Materials and Methods

2.1. Mice

The B6.129S-Tnftm2GKl/Jarn strain [TNF^∆ARE/+; MGI: 3720980] was generated by continuous backcrosses between TNF^∆ARE on a mixed background²⁹ to C57BL/6J mice. Homozygous wild-type [WT] C57BL/6J littermates served as non-inflamed controls. An alternate model of Salmonella enterica serovar Typhimurium infectious colitis was induced in adult C57BL/6J mice as previously described.³¹ CB₂R^-/- knockout mice [B6.129P2-Cnr2^tm1Dgen/J] were obtained from Jackson Laboratories. Mice were kept under specific pathogen-free conditions, and faecal samples were routinely negative for Helicobacter species, protozoa, and helminthes. Animal procedures were approved by the Institutional Animal Care and Use Committee of the University of Colorado Denver.

2.2. Analysis of endogenous cannabinoid tone

Tissue samples from WT and TNF^∆ARE/+ terminal ilea were assessed for presence of endogenous cannabinoids using-high performance liquid chromatography [HPLC]. RNA was isolated from ileal tissue or T cell subsets using the RNeasy Mini Kit [QIAGEN] and converted to complementary DNA [cDNA] using a high-capacity cDNA Reverse Transcription Kit [Applied Biosystems]. Relative quantitation of mRNA expression was performed using Taqman Gene Expression Assays and the ABI PRISM 7900HT Sequence Detection System [Applied Biosystems]. Polymerase chain reaction [PCR] assays for Cnr1 [Mm01212171_s1] and Cnr2 [Mm02620087_s1] were carried out, with ribosomal 18s as an endogenous control.

2.3. Western blot analysis

Ilea from inflamed TNF^∆ARE/+ and non-inflamed WT littermates were homogenised in RIPA lysis and extraction buffer [Thermo Scientific] containing Halt protease inhibitor cocktail [Thermo Scientific]. Protein concentration was determined using the BCA Protein Assay [Pierce]. Protein extracts were separated by SDS-PAGE using Bio-Rad Mini-PROTEAN® TGX Stain-Free™ Gels. Rabbit anti-CB₂R was raised by immunising rabbits with the C-terminal fragment of recombinant CB₂R [Ac-CKYLQGLGPEGKEEGPR-amide; New England Peptide]. Blots were probed overnight at 4°C with rabbit anti-CB₂R [1:2500] or rabbit anti-Na+/K+-ATPase alpha [sc28800; 1:200; Santa Cruz], followed by a 1-h incubation at room temperature with a goat anti-rabbit peroxidase-conjugated antibody [sc2004; 1:5000; Santa Cruz]. Blots were then exposed to enhanced chemiluminescence substrate [Clarity Western ECL Substrate; Bio-Rad] and imaged using the Bio-Rad ChemiDoc™ MP system.

2.4. In vitro suppression assays

Treg suppression assays were performed as previously described.³² Briefly CD4⁺CD25⁺ Tregs and CD4⁺CD25^Neg effector T cells [Teff] were magnetically isolated from splenocytes using the EasySep Mouse CD25 Regulatory T Cell Positive Selection Kit II [Stem Cell Technologies]; 5 × 10⁴ CellTrace Violet-labeled [Invitrogen] CD4⁺CD25^Neg Teff per well were stimulated with anti-CD3 mAb [1 μg/mL] in the presence of 7.5 × 10⁴ irradiated syngeneic antigen-presenting cells [APC] and varying ratios of Tregs. Where indicated, cells were cultured with 10 μM GP-1a [Tocris Bioscience]. The percentage of Teff proliferation was determined by the CellTrace profile of live CD4⁺ cells at 72 h.

2.5. Apoptosis assay

WT splenocytes were stimulated at a concentration of 2 × 10⁶/mL with plate-bound anti-CD3 mAb [1 µg/mL] and soluble anti-CD28 [4 µg/mL] for 24 h in complete RPMI-1640 [with L-glutamine, Corning Inc., supplemented with 10% FBS, 100 IU penicillin and 100 µg/ml streptomycin; Invitrogen]. Where indicated, GP-1a was added to cultures at 10 µM. Determination of apoptosis was performed using FITC Annexin V Apoptosis Detection Kit I [BD Biosciences] according to the manufacturer’s instructions.

2.6. In vivo GP-1a treatment

Twice per week for 2 weeks, 8-week-old actively inflamed TNF^∆ARE/+ mice were treated with selective CB₂R ligand GP-1a [5 mg/kg per mouse³³] or vehicle [10% DMSO in PBS] by retro-orbital injection. At the time of sacrifice, spleens, mesenteric lymph nodes [MLN], and ilea were excised for flow cytometric analysis, histological assessment of ileitis, and cytokine analyses. Ileal sections were fixed in 10% formalin, and stained with haematoxylin and eosin. Histological assessment of ileitis was performed using a standardised semi-quantitative scoring system, as described previously.³⁴ Representative micrographs were obtained using an Olympus IX83 inverted microscope with cellSens Dimension v1.16 software [Olympus], made available through the Gastrointestinal Liver Innate Immunology Program at the University of Colorado.

Splenocytes, MLN, and LP mononuclear cells were isolated as previously described.^35,36 Tissues were filtered to remove any remaining undigested material before antibody labelling and flow cytometric evaluation.

2.7. Flow cytometry and intracellular cytokine staining

Cells from indicated compartments were incubated with fluorescent rat anti-mouse antibodies for T cell subset evaluation. Intracellular cytokine staining was performed by stimulating cells for 5 h with 20 ng/mL phorbol 12-myristate 13-acetate [PMA, Sigma Aldrich #P8139] and 1 µg/mL calcium ionomycin [Sigma Aldrich #I0634] in the presence of 3 µg/mL brefeldin-A [Sigma Aldrich #B7651], followed by permeabilisation and staining with antibodies against CD4 [GK1.5], IFNγ [XMG1.2], IL-10 [JES5-16E3], and FoxP3 [FJK16S]. Live cells were identified using Live/Dead Fixable Aqua dye [Invitrogen]. FoxP3 staining was performed according to manufacturer’s instructions [FoxP3/Transcription Factor Staining Buffer Kit, eBiosciences]. Cells were washed and fixed with 2% paraformaldehyde and then analysed using the FACS^® Canto II system [Becton-Dickinson]. Post-analyses were performed using FlowJo software [Tree Star Inc.].

2.8. Quantification of cytokine production

Terminal ileal explants [1 cm²] were cultured for 24 h in RPMI-1640 [with L-glutamine; Corning Inc., supplemented with 10% FBS, 100 IU penicillin and 100 µg/ml streptomycin; Invitrogen]. Culture supernatants were then analysed for IL-10 concentration using enzyme-linked immunosorbent assay [ELISA], as per the manufacturer’s instructions [Mouse IL-10 ELISA Ready-Set-Go! eBioscience].

2.9. Evaluation of cannabinoid receptor expression in Crohn’s disease patients

The study protocol was approved by the Ethical Committee of the University Hospital, Tuebingen, Germany. All patients and controls gave their written, informed consent to participation. Biopsies from the terminal ileum and colon were sampled during routine colonoscopy at the Robert Bosch Hospital, Stuttgart, Germany, and were immediately stored in liquid nitrogen for subsequent processing and analysis. Diagnosis was performed according to standard criteria using clinical, radiological, endoscopic, and histopathological methods. Patients with CD were grouped according to standard Vienna classifications, designating L1 as CD with exclusive ileal involvement, L2 as CD with solely colonic involvement, and L3 as CD with ileocolonic involvement.

2.10. RNA preparation from human tissue and quantification of mRNA expression

RNA from human tissue was isolated using TRIzol Reagent [Life Technologies] according to manufacturer’s protocol. cDNA synthesis was performed using a reverse transcription kit [Promega] and oligo[dT]15-primer. An amount of cDNA corresponding to 10 ng of RNA was applied in all assays. For quantification, we used SYBR Green I [Roche] and the Roche LightCycler 480 system. Absolute copy numbers were calculated from the crossing point [Cp] values according to respective standard plasmid curves, generated for all targets. Plasmid standards were designed using a TOPO TA Cloning Kit [Invitrogen] and confirmed by sequencing. Results were normalised to β-actin [see Table 1 for primers].

Table 1.

Primer sequences.

Gene	Sense	Antisense
β-actin	5’-GCCAACCGCGAGAAGATGA-3’	5’-CATCACGATGCCAGTGGT A-3’
CXCL8	5’-ATGACTTCCAAGCTGGCCGTGGC-3’	5’-TCTCAGCCCTCTTCAAAAACTTC-3’
CNR1	5’-ATGCGAAGGGATTGCCCC-3’	5’-ATGGTGCGGAAGGTGGTATC-3’
CNR2	5’-TCCTGGGAGAGGACAGAAAAC-3’	5’-TTGTCTAGAAGGCTTGTTG-3’

Open in new tab

2.11. Statistics

Statistical analyses were performed using one- or two-tailed Student’s t test [as appropriate], Mann–Whitney U-test [for non-normally distributed groups], and Spearman rank analysis to test for correlation. In the absence of evidence regarding expression of CB receptors in the inflamed ileum, we chose to employ a more conservative two-tailed t test. In contrast, interventional studies were assessed using a one-tailed t test consistent with our directional research hypothesis to be tested. Graphs are presented as mean ± standard error of the mean [SEM], generated using GraphPad Prism software. Values of p < 0.05 were considered statistically significant.

3. Results

3.1. Murine ileitis induces the endocannabinoid system

The ECS is induced in acute colitis models using dextran sulphate sodium [DSS] or 2,4,6-trinitrobenzenesulphonic acid [TNBS].^2,15 To establish the validity of this finding in the context of our preclinical CD model, we isolated terminal ileal tissue from acutely inflamed 8–12-week-old male TNF^∆ARE/+ mice [n = 7] and age-matched male WT mice [n = 5] and assessed for the presence of endogenous cannabinoids AEA, 2-AG, oleoylethanolamide [OEA], and palmitoylethanolamide [PEA] by HPLC. We found that the endocannabinoid AEA was significantly induced in inflamed TNF^∆ARE/+ ilea p[26.9 ± 1.6 pmol/g] as compared with WT ilea [18.8 ± 3.1 pmol/g; P < 0.05] [Figure 1A]. Furthermore, endocannabinoid OEA showed an overall trend of increased expression in inflamed TNF^∆ARE/+ ilea [182.0 ± 32.1 pmol/g] as compared with WT [147.5 ± 26.1 pmol/g; p = 0.45], whereas 2-AG showed a decreased although non-significant trend in expression [49.6 ± 11.7 pmol/g versus 26.3 ± 9.5 pmol/g; p = 0.15] and PEA demonstrated no significant difference in expression among the two groups [Figure 1A].

Figure 1.

Open in new tabDownload slide

Murine ileitis induces the endocannabinoid system. [A] HPLC analysis of endogenous cannabinoids anandamide, 2-AG, OEA, and PEA in terminal ilea of TNF^∆ARE/+ mice compared with WT [mean ± SEM of n ≥ 5 individual mice; two-tailed t test *p < 0.05]. [B] mRNA expression of Cnr1 and Cnr2 in whole ileal tissue of 4-week [pre-disease], 8–12 week [active inflammation] and ≥ 20-week [chronic inflammation] TNF^∆ARE/+ mice compared with WT [mean ± SEM, n ≥ 5 individual mice; two-tailed t test *p < 0.05, **p < 0.01]. [C] Representative Western blot of CB₂R expression in whole ileal tissue of 8–12 week TNF^∆ARE/+ mice and non-inflamed littermates. Na+/K+-ATPase is used as a loading control. HPLC, high-performance liquid chromatography; WT, wild type; SEM, standard error of the mean; TNF, tumour necrosis factor.

As previously described, TNF^∆ARE/+ mice are grouped into subsets of pre-disease [age 4 weeks], peak active inflammatory state [age 8–12 weeks], and chronic inflammatory state [age ≥ 20 weeks].²⁹ To further evaluate ECS expression in our inflamed TNF^∆ARE/+ mouse model, we isolated terminal ilea from each age subset [n ≥ 4 per age group] and measured expression of Cnr1 and Cnr2 [genes coding for CB₁R and CB₂R, respectively]. Neither Cnr1 nor Cnr2 expression was significantly altered in the ileum of pre-disease 4-week-old TNF^∆ARE/+ mice as compared with WT mice. In actively inflamed 8–12-week-old mice, however, Cnr1 was found to be significantly increased approximately 1.5-fold compared with WT, and Cnr2 increased 2.6-fold [p < 0.05 and p < 0.01]. In chronically inflamed mice [≥ 20 weeks], we saw a significant downregulation of both Cnr1 and Cnr2 [p < 0.01]. Thus, both CB₁R and CB₂R are induced by active ileitis and downregulated by chronic inflammation [Figure 1B]. We also investigated cannabinoid receptor expression in an infectious S. Typhimurium colitis model [n ≥ 3 per group], and found that there was no such induction in cannabinoid receptor expression, indicating that this phenomenon was not simply a general response to intestinal insult [Supplementary Figure 1, available as Supplementary data at ECCO-JCC online].

Due to the expression profile of the CB₂R in the ileum, and because CB₂R is preferentially expressed on lymphocytes and other immune cells, we identified the CB₂R as our receptor of interest. We found that in line with the results of real-time [RT]-PCR, CB₂R protein was prominently increased in the actively inflamed ilea of TNF^∆ARE/+ mice compared with WT littermates [Figure 1C].

In summary, the compilation of data in Figure 1 highlights the induction of the ECS at the receptor level, and to a much lesser extent at the ligand level, in the setting of active murine ileitis. Furthermore, CB₂R mRNA expression is more significantly elevated than CB₁R expression in the ilea of actively inflamed TNF^∆ARE/+ mice and is preferentially expressed on immune cells, distinguishing CB₂R as the key receptor of interest in our investigation.

3.2. CB₂R is preferentially upregulated on regulatory T cells in murine ileitis, and targeting with GP-1a enhances suppressive function

In view of the fact that cannabinoid receptors are expressed by CD4⁺ T cells, key regulators of intestinal inflammation, and based on their upregulation in acute murine ileitis, we focused our investigational efforts on elucidation of the role of the cannabinoid receptors on CD4⁺ T cell phenotype and function during inflammation. First, we compared cannabinoid receptor expression on CD4⁺ Tregs with that on CD4⁺ non-Tregs isolated from actively inflamed 8–12-week-old male TNF^∆ARE/+ mice and age-matched male WT controls [n = 3 per group]. Tregs were magnetically enriched from splenocytes based on expression of CD25, and although CD25 can be induced on activated non-Treg cells especially in vitro, our experience with TNF^ΔARE/+ mice indicates that CD25 is not expressed on non-Tregs, even in the inflamed lamina propria.³² It was confirmed by flow cytometry that approximately 85% of enriched CD4⁺CD25⁺ expressed FoxP3. Our study demonstrated that relative mRNA expression of Cnr1 was preferentially upregulated 2-fold in TNF^∆ARE/+ Tregs as compared with Teff to a near-significant extent [p = 0.06], whereas there was no difference in expression of Cnr1 between WT Tregs and Teff [Figure 2A]. Moreover, relative mRNA expression of Cnr2 was increased 11-fold in Tregs as compared with Teff cells of actively inflamed TNF^∆ARE/+ mice [p < 0.001] and increased 2.4-fold on Tregs compared with Teff cells in WT mice, though this did not reach statistical significance [p = 0.09] [Figure 2A].

Figure 2.

Open in new tabDownload slide

CB₂R is preferentially upregulated on regulatory T cells in murine ileitis and targeting with GP-1a enhances suppressive function. [A] Preferential upregulation of CB₂R on regulatory T cells in murine ileitis. Cnr1 and Cnr2 mRNA expression in isolated CD4⁺CD25⁺ Tregs and CD4⁺CD25^Neg Teff from 8–12 week WT and TNF^∆ARE/+ mice [mean ± SEM of pooled cells from n ≥ 3 per group; two-tailed t test ***p < 0.001] [B] Treatment of co-culture with GP-1a restores capacity of TNF^∆ARE/+ Treg to suppress Teff. Data shown are percentages of proliferating cells at various ratios of Treg:Teff; representative of three experiments [mean ± SEM; two-tailed t test *p < 0.05, **p < 0.01, ***p < 0.001]. [C] Induction of Cnr1 and Cnr2 mRNA in TNF^∆ARE/+ CD4⁺CD25^Neg Teff with anti-CD3 and anti-CD28 stimulation and GP-1a treatment [mean ± SEM of n ≥ 4 per group; two-tailed t test **p < 0.01, ***p < 0.001]. WT, wild type; SEM, standard error of the mean; TNF, tumour necrosis factor.

In consideration of the preferential upregulation of the CB₂R on Tregs in acute murine ileitis, we performed a Treg suppression assay to assess the in vitro suppressive capacity of Tregs upon stimulation with selective CB₂R ligand GP-1a. CD4⁺CD25⁺ Tregs were isolated from actively inflamed 8–12-week-old male TNF^∆ARE/+ mice and co-cultured with CellTrace Violet-labelled CD4⁺CD25^Neg Teff at varying ratios in the presence of irradiated APC and anti-CD3 stimulation. Tregs treated with GP-1a showed enhanced suppressive capacity compared with control, indicated by significantly decreased Teff proliferation [Figure 2B] [p < 0.01 at 1:1 ratio, and p < 0.001 at 1:2, 1:4, and 1:8 ratios; assay repeated ≥ 3 times]. GP-1a did not affect proliferation of Teff in the absence of Treg. Furthermore, an ELISA was performed to assay IL-10 quantities in the culture supernatants of the suppression assay with and without GP-1a, revealing a concomitant increase in anti-inflammatory IL-10 secretion in the setting of CB₂R stimulation [vehicle = 125.4 ± 12.4 pg/ml, GP-1a = 224.7 ± 31.7 pg/ml; p < 0.05]. We then assessed the suppressive capacity of Tregs lacking the CB₂R, isolated from CB₂R^-/- knockout mice. Compared with WT, we observed no significant differences in suppression by Tregs lacking CB₂R [Supplementary Figure 2, available as Supplementary data at ECCO-JCC online].

In order to investigate a potential direct effect of GP-1a on Teff in the absence of Treg, we evaluated cannabinoid receptor expression on TNF^∆ARE/+ Teff stimulated with anti-CD3 and anti-CD28 for 24 h [n = 4 per group]. We found that there was a significant increase in mRNA expression of Cnr1 and Cnr2 upon stimulation and treatment with Gp-1a. Culture with GP-1a alone in the absence of stimulation resulted in a modest but significant induction of Cnr1 and Cnr2 approximately 1.1-fold [p < 0.01], and stimulation with anti-CD3 and anti-CD28 resulted in a 1.2-fold induction of both Cnr1 and Cnr2 [p < 0.01]. When stimulated and exposed to GP-1a simultaneously, Cnr1 mRNA increased 1.3-fold and Cnr2 increased 1.4-fold [p < 0.001] [Figure 2C].

In summary, Figure 2A identified a preferential expression of CB₁R and CB₂R on Tregs, and this effect was even more pronounced for CB₂R expression in the setting of active murine ileitis, implying involvement of the CB₂R on Tregs in the active inflammatory pathway. Figure 2B indicates the ability of CB₂R stimulation with GP-1a to enhance Treg suppression, with a concomitant increase in IL-10 production suggesting an anti-inflammatory role for CB₂R stimulation in CD4⁺ T cells. Figure 2C indicates a direct effect of GP-1a on the expression of CB₁R and CB₂R on TNF^∆ARE/+ Teff cells.

3.3. CB₂R stimulation attenuates murine ileitis

To investigate the potential anti-inflammatory role of CB₂R stimulation in active murine ileitis, we performed a treatment study using selective CB₂R ligand GP-1a. Eight-week-old male TNF^∆ARE/+ mice received pulsatile 2-week treatment with either GP-1a or vehicle [n = 9 mice per treatment group], and ilea were then isolated at 10 weeks of age for blinded histology review by a pathologist using the previously described Burns method.³⁴

We found that CB₂R stimulation with GP-1a led to histological improvement of ileitis of TNF^∆ARE/+ mice [Figure 3]. Notably, both active inflammation characterised by granulocytic infiltration, and chronic inflammation characterised by lymphocytic infiltration, significantly improved [p < 0.05] in response to CB₂R ligand treatment [Figure 3]. There was a trend of decreased villus architecture distortion in mice treated with GP-1a, although this improvement did not reach statistical significance [p = 0.07]. These data support the findings in Figure 2, suggesting an anti-inflammatory role of CB₂R stimulation in murine ileitis.

Figure 3.

Open in new tabDownload slide

Significant attenuation of ileitis by GP-1a treatment. Representative H&E ilea 10x micrographs of vehicle and GP-1a treated 10-week-old TNF^∆ARE/+ mice. Histological assessment of inflammatory indices from ilea of 10-week-old TNF^∆ARE/+ mice treated with GP-1a or vehicle for 2 weeks [mean ± SEM of n = 7 per group; one-tailed t test *p < 0.05]. H&E, haematoxylin and eosin; SEM, standard error of the mean; TNF, tumour necrosis factor.

3.4. CB₂R receptor stimulation reduces inflammatory infiltrate

To further elucidate the anti-inflammatory effect of CB₂R stimulation in murine ileitis, we assessed cells from spleen, MLN, and ileal LP from 10-week-old TNF^∆ARE/+mice following GP-1a treatment by flow cytometry. We found a significant decrease [2.2 ± 0.3% to 1.4 ± 0.2; p < 0.05] in IFNγ-producing Th1 cells in the TNF^∆ARE/+ LP when treated with GP-1a as compared with vehicle control, as well as a decrease, albeit non-significant, in the expression of IFNγ in both spleen and MLN [p = 0.1 and 0.3, respectively] [Figure 4A, B]. Furthermore, we found that whereas the frequency of CD4⁺FoxP3⁺ Tregs was unchanged in the LP of GP-1a treated mice, the proportion of Tregs producing IL-10 was significantly increased [3.9 ± 1% to 7.1 ± 0.9 %; p < 0.05]. Consistent with this finding, an ELISA revealed a significant increase in IL-10 in the supernatants of ileal explant culture of GP-1a treated mice [252.8 ± 43.7 pg/mg; p < 0.05] as compared with vehicle control [123.1 ± 26.3 pg/mg; p < 0.05] [Figure 4C].

Figure 4.

Open in new tabDownload slide

Reduction of inflammatory infiltrate with CB₂R stimulation. [A] Frequency of pro-inflammatory CD4⁺ IFNγ⁺ Th1 cells and CD4⁺FoxP3⁺IL-10⁺ Tregs in indicated organs isolated from vehicle or GP-1a treated 10-week-old TNF^∆ARE/+ mice [SP, spleen; MLN, mesenteric lymph node; LP, lamina propria] [mean ± SEM, n = 5 per group; one-tailed t test *p < 0.001]. [B] Representative dot plots display the CD4⁺ IFNγ⁺ and CD4⁺FoxP3⁺IL-10⁺ expression from vehicle and GP-1a treated mice. [C] IL-10 quantities from 24-h terminal ilea explant cultures measured by ELISA [mean ± SEM, n ≥ 3 per group; one-tailed t test *p < 0.05, **p < 0.01]. SEM, standard error of the mean; TNF, tumour necrosis factor.

Collectively, the data presented in Figures 3 and 4 indicate that CB₂R stimulation ameliorates inflammation and reduces inflammatory infiltrate in murine ileitis.

3.5. Cannabinoid receptor expression in Crohn’s disease patients

Classification of CD patient tissue as currently inflamed or non-inflamed was based on expert diagnosis [clinical, radiological, endoscopic, and histopathological methods]. Patients with CD were grouped according to standard Vienna classifications, designating L1 as CD with exclusive ileal involvement, L2 as CD with solely colonic involvement, and L3 as CD with ileocolonic involvement.

We also assessed mucosal CXCL8 expression to confirm the inflammatory status of the included samples. In the ileum, controls displayed a mean level of 546.3 copies/10 ng total RNA [417.4 SEM], non-inflamed ileal CD patients showed 380.2 ± 120.8, and inflamed samples 7873 ± 3036 copies of CXCL8/10ng total RNA. In the colon, controls displayed an average of 115.1 CXCL8 copies/10 ng total RNA, non-inflamed colonic CD patients showed on average 339.9 ± 138, and inflamed samples 4599 ± 2097 copies of CXCL8/10 ng total RNA. Patient characteristics are listed in Table 2. The age distribution of the studied patients was typical for this disease, as was the balanced gender ratio. For our ileal studies, we compared ilea of CD patients, who all belonged to Vienna classification L1 or L3 [n = 58], with those of healthy controls [n = 24], and for colonic tissue, we combined L2 and L3 subgroups as colonic CD [n = 49], which we again compared with healthy controls [n = 25].

Table 2.

Patient characteristics.

Ileum	N [total]	Gender		Age in years [mean ± SD]	Localisation		Inflammation
		F	M		L1	L3	Yes	No
Patients	57	52%	48%	51.4 ± 14.8	47%	53%	74%	25%
Controls	24	54%	46%	63.6 ± 18.6
Colon	N [total]	Gender		Age in years [mean ± SEM]	Localisation		Inflammation
		F	M		L2	L3	Yes	No
Patients	49	54%	46%	46.9 ± 11.16	41%	59%	67%	33%
Controls	25	60%	40%	59.8 ± 20.6

F, female; M, male; SD, standard deviation; SEM, standard error of the mean.

Open in new tab

Gene expression on the mRNA level for both CNR1 and CNR2 was relatively low in all studied groups. However, expression was consistently detected in all analysed samples. Interestingly, we found a profound decrease in both CNR1 [p < 0.001] [Figure 5A] and CNR2 [p < 0.05] [Figure 5C] in the ilea of non-inflamed CD patients. In active disease, the reduction was less pronounced but still present. In the colon, however, both receptors were markedly higher in non-inflamed colonic CD patients [p < 0.01 for CNR1 and p < 0.05 for CNR2] [Figure 5B, D]. Since the pattern of expression of the two receptors showed notable similarity, we checked for correlation. In both tissues, we found a moderate but statistically highly significant degree of correlation [p < 0.001] [Figure 5E, F].

Figure 5.

Open in new tabDownload slide

Cannabinoid receptor expression in Crohn’s disease patients. [A, C] mRNA expression of CNR1 and CNR2, respectively, in the ilea of inflamed and non-inflamed CD patients compared with healthy controls. [B, D] mRNA expression of CNR1 and CNR2, respectively, in the colon of inflamed and non-inflamed CD patients compared with healthy controls [mean ± SEM; two-tailed Mann-Whitney test *p < 0.05, **p < 0.01, ***p < 0.001]. [E] Correlation of CNR1 and CNR2 expression in the ilea of CD patients. [F] Correlation of CNR1 and CNR2 expression in the colon of CD patients [Spearman rank analysis p < 0.001]. CD, Crohn’s disease; SEM, standard error of the mean.

In summary, Figure 5 reveals an overall marked reduction in cannabinoid receptor expression in the ilea of CD patients with chronic inflammation compared with healthy controls, though expression of both CB₁R and CB₂R was increased in active disease compared with non-inflamed patients nonetheless. In the colon, we saw an increase in cannabinoid receptor expression with disease, to a significant extent in non-inflamed CD patients.

4. Discussion

Our experiments demonstrate that the ECS is induced by active intestinal inflammation at receptor mRNA and protein expression levels, and to a much lesser extent at the ligand level. Similar results of receptor expression have been reported in chemical models of intestinal inflammation.^{17,19,21,37–39} These findings are also supported from a broader perspective, where CB₂R expression is reported to increase in early inflammatory events in both the central nervous system [CNS] and peripheral tissue.^40,41 Thus, the effects of the ECS in models of gut inflammation appear to be primarily due to increased receptor expression and not ligand expression. In addition, these effects may be specific to the TNF^∆ARE/+ model of CD, as there was no induction of CB₁R and CB₂R seen in an acute infectious Salmonella colitis model. The compilation of past evidence and our current results validate the ECS as a promising therapeutic target in inflammatory conditions such as IBD.

Intriguingly, we found that both cannabinoid receptors were preferentially upregulated on Tregs when compared with CD4⁺CD25^Neg Teff cells. In light of the importance of Treg in maintaining normal gut homeostasis, we proceeded to examine the role of ECS activity in governing Treg function. We observed that treatment in vitro with CB₂R ligand GP-1a led to improved suppression of Teff proliferation in the setting of inflammation. Increased IL-10 in the supernatants of treated co-cultures suggests that the effect of GP-1a is a result of improved Treg suppressive function, although increased susceptibility of Teff to suppression is also a possibility. Our efforts in investigating the mechanism of decreased Teff proliferation with Gp-1a revealed that Gp-1a appears to have some direct effect on Teff; however, this effect was lost in the absence of Tregs, suggesting a possible effect on Teff resistance to suppression. In addition, GP-1a does not directly induce apoptosis [Supplementary Figure 3, available as Supplementary data at ECCO-JCC online].

We then demonstrated that CB₂R stimulation with GP-1a in vivo significantly attenuated active murine ileitis, as determined histologically and by assessment of inflammatory infiltrate. We found that histological features of ileitis, such as leukocyte infiltration, goblet cell hyperplasia, and muscularis hypertrophy, were markedly decreased with GP-1a treatment. Furthermore, we found a decrease in the frequency of infiltrating IFNγ-producing Th1 cells and an increase in the proportion of IL-10-producing Tregs, consistent with an overall increased secretion of IL-10 from the ileal tissue of GP-1a-treated mice. These in vivo results are supported by previous studies which showed GP-1a-induced inhibition of Th1 and Th17 differentiation in vitro, under both polarising and non-polarising conditions.³³ Overall, manipulation of the CB₂R has both pro-regulatory and anti-inflammatory effects, as evidenced by enhanced suppression of Teff, increased IL-10 production, and decreased Th1 cell infiltration.

Interestingly, the previously accepted classification of GP-1a as a selective CB₂R agonist^42,43 has more recently been debated. Soethoudt et al. profiled various CB₂R ligands and confirmed that GP-1a is one of the most selective compounds for the CB₂R and is an active AEA reuptake inhibitor but, in contrast to previous reports, it acts in vitro as an inverse agonist at the CB₂R while it is inactive at the CB₁R.⁴⁴ However, cannabinoid signalling is complex, and evidence shows that cannabinoid receptor ligands may have varying pharmacological roles in native versus recombinant systems.⁴⁵ Marini et al. characterised several cannabinoid receptor ligands in tissue that natively express CB₂R, and found that certain cannabinoid receptor ligands can act as an agonist in human spleen but act in vitro as an inverse agonist in cells over-expressing the human CB₂R.⁴⁵ Although GP-1a was not specifically investigated, this work demonstrated that some cannabinoid ligands that are classified as CB₂R inverse agonists can have anti-nociceptive actions in vivo, and therefore demonstrate CB₂R agonist properties.⁴⁵ As a G-protein-coupled receptor, CB₂R-specific properties such as tethering of G-protein subunits, G-protein signalling regulators, receptor phosphorylation, and compartmentalisation of signalling elements within the membrane influence the efficacy of CB₂R ligands,⁴⁶ and these factors may explain the differences in the pharmacological profile of ligands between species and between the native and recombinant cell systems.⁴⁵ Our experiments did not define the exact mechanism of GP-1a-attenuated inflammation, but we hypothesise that there is a critical importance in selecting a CB₂R inverse agonist in the therapeutic targeting of inflammation, as well as in the function of G-protein-coupled receptors such as the CB₂R.

It is well established that G-protein-coupled receptors such as the CB₂R are constitutively active, and will recruit known elements of cellular desensitisation leading to constitutive desensitisation.⁴⁷ Likewise, G-protein-coupled receptors are phosphorylated by G-protein-coupled receptor kinase and lead to desensitisation at a similar extent and rate in both the absence and presence of an agonist.⁴⁷ A study by Bouaboula et al. demonstrated that inverse agonism with SR144528 blocked the constitutive activation of CB₂R and did not allow desensitisation of the receptor to occur.⁴⁸ Therefore, G-protein-coupled receptors such as the CB₂R require treatment with an inverse agonist in order to reverse desensitisation, and thus allow activation of the receptor and its downstream pathways. We hypothesise that, in a similar mechanism, GP-1a acts upon the CB₂R with inverse agonism in order to reverse desensitisation and restore the innate anti-inflammatory signalling system of the ECS. One critical aspect of this assumption is that our experiments were performed using fetal bovine sera containing endocannabinoids,⁴⁹ and in contrast, our preliminary in vitro experiments using serum-free media have demonstrated that the effect of GP-1a treatment is abrogated in the absence of these endogenous cannabinoids [data not shown]. This incidental but significant observation reinforces our hypothesis that resensitisation of the CB₂R allows restoration of endogenous cannabinoid signalling and thereby resolution of inflammation.

Whereas we propose that GP-1a acts upon the CB₂R with inverse agonism in order to allow receptor resensitisation, promoting the endocannabinoid system to act at the CB₂R to ameliorate murine ileitis, our data reveal a novel and overarching role of Tregs in this process. It is well established that CD4⁺ T cells are key mediators of intestinal inflammation, and more specifically CD4⁺FoxP3⁺ cells [Tregs] suppress inappropriate immune responses and maintain gut homeostasis. Our experiments parallel this concept, and we have demonstrated that CB₁R and CB₂R are upregulated on Tregs in the setting of active murine ileitis. More importantly, usng a Treg suppression assay to assess the suppressive function of Tregs, we found that treatment in vitro with CB₂R ligand GP-1a led to improved suppressive function of Tregs in the setting of inflammation. When we investigated the suppressive capacity of Tregs lacking the CB₂R, our data showed no significant difference in their suppressive function. It should be noted, however, that this assay was conducted in the absence of inflammation-induced upregulation of the CB₂R. This experimental finding is therefore unsurprising, as CB₂R shows low expression overall in the absence of inflammation, and thus the CB₂R would not be expected to play a role in Treg suppressive function in a WT system. Efforts to cross CB₂R-deficient mice with the TNF^ΔARE/+ ileitis model are currently under way and may shed additional light on these findings.

One very important aspect of this study is the striking similarities in the decreased expression of the CB₁R and CB₂R in both murine and human chronic inflammation as compared with healthy controls, and in the increased expression of CB₁R and CB₂R in active murine ileitis and active, non-remission CD patients. This similarity lends credence to the translational applicability of our study in the search for novel therapeutic targets for IBD. However, we recognise that we see a decrease in CB₂R expression in 20-week TNF^∆ARE/+ mice as well as in human IBD, despite previous evidence that there is continued increase in Tregs in intestinal inflammation in 20-week TNF^∆ARE/+ mice³² and in IBD.^50,51 The difference in expression of Tregs and CB₂R is present because the CB₂R is not exclusively expressed on Tregs, but rather is found on many immune cell types.^2,11 In addition, whereas the differential expression of CB receptors in the colon has previously been classified,²⁷ the data collected in Figures 5A and C are novel in that, to the best of our knowledge, this is the first time the expression of the CB₂R has been classified in the ileum. Understanding the role of the ECS in human IBD is paramount; genetic studies have associated a common CB₂R functional variant, CB2-R63, with increased risk, worsening disease severity, and increased incidence of clinical relapse in paediatric IBD.⁵² Our human data demonstrate a marked reduction in expression of the CB₂R in diseased patients as compared with healthy controls, and this finding further justifies the importance of targeting the CB₂R for resensitisation to promote restoration of gut homeostasis.

In summary, our study has demonstrated the significance of the endocannabinoid system in IBD. We propose that GP-1a acts upon the CB₂R with inverse agonism in order sto allow receptor resensitization, promoting the endocannabinoid system to act at the CB₂R on Tregs, thus attenuating inflammation in a mouse model of Crohn’s disease. Given that CB₂R stimulation attenuates murine ileitis, the anti-inflammatory influences of the endocannabinoid system hold promise as a new potential treatment modality for IBD. Our work elucidates a possible mechanism in the anti-inflammatory pathways of the ECS, which will contribute to the interpretation of data and greatly benefit the ongoing clinical trials in cannabinoids and IBD.

Funding

This work was supported by the National Institutes of Health T32 Institutional Training Grant in Pediatric Gastroenterology [2T32DK067009-11]. CCs is supported by funding from NIH/NCATS Colorado CTSA Grant Number UL1 TR001082. This work was supported in part by the IRP of the National Institute on Aging/NIH [RM].