Pharmaceutical targeting of the cannabinoid type 1 receptor impacts the crosstalk between immune cells and islets to reduce insulitis in humans

Abstract

Aims/hypothesis: Insulitis, a hallmark of inflammation preceding autoimmune type 1 diabetes, leads to the eventual loss of functional beta cells. However, functional beta cells can persist even in the face of continuous insulitis. Despite advances in immunosuppressive treatments, maintaining functional beta cells to prevent insulitis progression and hyperglycaemia remains a challenge. The cannabinoid type 1 receptor (CB1R), present in immune cells and beta cells, regulates inflammation and beta cell function. Here, we pioneer an ex vivo model mirroring human insulitis to investigate the role of CB1R in this process.

Methods: CD4⁺ T lymphocytes were isolated from peripheral blood mononuclear cells (PBMCs) from male and female individuals at the onset of type 1 diabetes and from non-diabetic individuals, RNA was extracted and mRNA expression was analysed by real-time PCR. Single beta cell expression from donors with type 1 diabetes was obtained from data mining. Patient-derived human islets from male and female cadaveric donors were 3D-cultured in solubilised extracellular matrix gel in co-culture with the same donor PBMCs, and incubated with cytokines (IL-1β, TNF-α, IFN-γ) for 24-48 h in the presence of vehicle or increasing concentrations of the CB1R blocker JD-5037. Expression of CNR1 (encoding for CB1R) was ablated using CRISPR/Cas9 technology. Viability, intracellular stress and signalling were assayed by live-cell probing and real-time PCR. The islet function measured as glucose-stimulated insulin secretion was determined in a perifusion system. Infiltration of immune cells into the islets was monitored by microscopy. Non-obese diabetic mice aged 7 weeks were treated for 1 week with JD-5037, then euthanised. Profiling of immune cells infiltrated in the islets was performed by flow cytometry.

Results: CNR1 expression was upregulated in circulating CD4⁺ T cells from individuals at type 1 diabetes onset (6.9-fold higher vs healthy individuals) and in sorted islet beta cells from donors with type 1 diabetes (3.6-fold higher vs healthy counterparts). The peripherally restricted CB1R inverse agonist JD-5037 arrested the initiation of insulitis in humans and mice. Mechanistically, CB1R blockade prevented islet NO production and ameliorated the ATF6 arm of the unfolded protein response. Consequently, cyto/chemokine expression decreased in human islets, leading to sustained islet cell viability and function.

Conclusions/interpretation: These results suggest that CB1R could be an interesting target for type 1 diabetes while highlighting the regulatory mechanisms of insulitis. Moreover, these findings may apply to type 2 diabetes where islet inflammation is also a pathophysiological factor.

Data availability: Transcriptomic analysis of sorted human beta cells are from Gene Expression Omnibus database, accession no. GSE121863, available at https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSM3448161 .

Keywords: CB1R; Cannabinoid receptor; Endocannabinoid; Insulitis; Islet of Langerhans; Peripheral CB1R inverse agonist; Type 1 diabetes.

Fig. 1

Circulating CD4⁺ T cells have higher CNR1 expression levels at type 1 diabetes onset compared with levels in healthy donors. Expression of CNR1 (a, c) and CNR2 (b, d) mRNA in PBMCs (a, b) and circulating CD4⁺ T cells (c, d) from the blood of healthy donors (control; n=11) and donors with recent onset of type 1 diabetes (n=22). The expression of ACTB was used as a control. The box and whiskers graphs (Tukey) show data distribution (top and bottom quartiles in boxes and the minimum and maximum value with lines), median and the outliers as single data points outside the box. *p<0.05 (by Mann–Whitney U test). T1D, type 1 diabetes

Fig. 2

Pharmacological blockade of CB1R by JD-5037 arrests insulitis in an ex vivo human model. (a) Schematic representation of the experimental ex vivo human insulitis model (created using Servier Medical Art templates). Human islets and PBMCs were freshly isolated from the pancreas and blood, respectively, of a deceased organ donor. Islets were embedded in Matrigel to create a 3D culture. PBMCs were stained with 488 CellTrace CFSE before adding them to the media in co-culture with the islets. Co-cultures were treated with a mix of cytokines (IFN-γ, IL1-β and TNF-α) and monitored over 6 days. Within the first 18 h, there was chemoattraction of the immune cells towards the border of the 3D culture, followed by their infiltration through the Matrigel towards the islets. (b, c) Representative photomicrograph (b) and quantification (c) of the chemoattraction in 3D co-cultures treated with vehicle or increasing concentrations of JD-5037. (d) Illustrative photomicrograph after 6 days of insult of the PBMCs (green) and bright field. The border of the Matrigel is delineated with a dotted red line. The islets are delineated with a dotted white line. (e, f) Quantification of islet infiltration in a single donor (same donor as the photomicrographs) (e) and three donors (f). Data are mean ± SEM, n=3 donors. *p<0.05, **p<0.01 and ***p<0.001 (by one-way [c] or two-way ANOVA [f]; Tukey post hoc test). Scale bar, 200 μm. BF, bright field

Fig. 3

JD-5037 prevents cytokine-induced cell death and nitric oxide (NO) production in human islets ex vivo. (a, b) Apoptosis and cell death were measured as caspase 3 activity (a) and dead-cell protease activity (b) in islets stimulated with or without cytokines and treated with vehicle or JD-5037. (c, d) Representative photomicrographs showing mitochondrial superoxide (O₂⁻) (c) and cellular NO production (d) as stained with MitoSox (red) and DAF-FM (green), respectively, in islets stimulated with or without cytokines and treated with vehicle or JD-5037 alone or in combination with ACEA (25 nmol/l). Bright-field photomicrographs and fluorescence imaging of the same field are shown; the graph shows quantification of the relative fluorescence units for MitoSox or DAF-FM staining. Scale bar, 200 μm. Data are mean ± SEM and individual data points, n=3 donors; n=20–25 islets per group. *p<0.05, **p<0.01 and ***p<0.001 (by one-way ANOVA; Tukey post hoc test). BF, bright field; RFU, relative fluorescence units; RLU, relative light units

Fig. 4

CNR1 ablation in human islets prevents cytokine-induced NO production and insulitis ex vivo. CNR1 was knocked down in human islets or PBMCs using CRISPR/Cas9, and cytokine-mediated damage was investigated in the human insulitis ex vivo model. (a) mRNA expression of CNR1 full length (CB1FL) and b isoform (CB1b) in CNR1KD and control islets. (b) GSIS in a perifusion system. (c) CNR1 expression after cytokine treatment. (d) NO production in wild-type and CNR1KD islets. Representative photomicrographs showing NO production as stained with DAF-FM (green), and bright-field photomicrographs of the same field; the graph shows quantification of the relative fluorescence units for DAF-FM staining. Scale bar, 200 μm. (e) mRNA expression of CNR1 in CNR1KD and wild-type PBMCs. (f, g) Infiltration of wild-type immune cells in co-culture with wild-type or CNR1KD islets (f) and infiltration of wild-type or CNR1KD immune cells in co-culture with wild-type islets (g). (h) Early infiltration (24 h) of islets in the ex vivo model using CNR1KD islets, and CNR1KD PBMCs with or without JD-5037 (n=3 independent experiments). *p<0.05, **p<0.01 and ***p<0.001 (by t test or two-way ANOVA; Tukey post hoc test). BF, bright field; WT, wild-type

Fig. 5

Targeting CB1R modulates ER stress response to cytokines in human islets ex vivo. Human islets were treated with JD-5037 (10 nmol/l), JD-5037 in combination with ACEA (25 nmol/l), or vehicle before an insult with cytokines, and samples were collected at 1 and 4 h for RNA extraction. (a) Schematic representation of ER stress signalling (created using Servier Medical Art templates). (b–f) Expression of GRP78 (also known as HSPA5 and BIP) (b), DDIT3 (also known as CHOP) (c), ATF6 (d), ATF4 (e) and PDL1 (also known as CD274) (f) after 1h (GRP78) and 4 h (the rest) of insult, respectively. The expression of RPLP0 was used as a control. Data are mean ± SEM, n=4 donors. *p<0.05, **p<0.01 and ***p<0.001 (by one-way ANOVA; Tukey post hoc test)

Fig. 6

CB1R modulates chemokine expression upon cytokine insult in human islets. Human islets were treated with JD-5037 (10 nmol/l), JD-5037 in combination with ACEA (25 nmol/l), or vehicle before an insult with cytokines, and samples were collected after 24 h for RNA extraction. Expression of CXCL10 (a), CCL2 (b), IL1B (c), TNF (d), ICAM1 (e), HLA-ABC (f), PDX1 (g), SLC2A1 (h) and SLC2A2 mRNA (i) was measured. The expression of RPLP0 was used as a control. Data are mean ± SEM, n=4 donors. *p<0.05, **p<0.01 and ***p<0.001 (by one-way ANOVA; Tukey post hoc test)

Fig. 7

JD-5037 protects beta cell function in the face of inflammation. (a–c) Dynamic GSIS after 2 h of insult with a mix of cytokines (a), stimulation index (b) and intra-islet insulin content (c) of islets from n=4 donors. (d–f) Dynamic GSIS after 24 h of insult with a mix of cytokines (d), stimulation index (e) and intra-islet insulin content (f) in islets from n=6 donors (except for ACEA [n=3 donors]). Data are mean ± SEM (shadowed in a, d or error bars in b, c, e, f). *p<0.05 and **p<0.01 for cytokine vs control groups; †p<0.05 and †††p<0.001 for JD-5037+cytokine vs vehicle+cytokine group; ‡p<0.05 for ACEA-JD-5037+cytokine vs JD-5037+cytokine group (by one- or two-way ANOVA; Tukey post hoc test)

Fig. 8

Schematic representation of the regulation of crosstalk between beta cells and immune cells by the ECS. Activated proinflammatory immune cells express CB1R in addition to CB2R. Cytokines secreted by immune cells (or another source) activate their receptors in beta cells, leading to the production of NO and the synthesis and secretion of endocannabinoids (AEA and 2-AG). The endocannabinoids activate the cannabinoid receptors in both beta cells (CB1R) and immune cells (CB1R and CB2R). In beta cells, activation of CB1R further enhances NO–ER stress signalling pathways, leading to chemokine secretion and beta cell death. Chemokines further attract immune cells to the site. Activation of CB1R also leads to a reduction in GSIS. The effect of cytokines is illustrated with red lines/arrows and the effect of JD-5037 with green lines/arrows. CCL2, chemokine ligand 2; EC, endocannabinoid. This figure was created using Servier Medical Art templates