Human CB1 Receptor Isoforms, present in Hepatocytes and β-cells, are Involved in Regulating Metabolism

Abstract

Therapeutics aimed at blocking the cannabinoid 1 (CB1) receptor for treatment of obesity resulted in significant improvements in liver function, glucose uptake and pancreatic β-cell function independent of weight loss or CB1 receptor blockade in the brain, suggesting that peripherally-acting only CB1 receptor blockers may be useful therapeutic agents. Neuropsychiatric side effects and lack of tissue specificity precluded clinical use of first-generation, centrally acting CB1 receptor blockers. In this study we specifically analyzed the potential relevance to diabetes of human CB1 receptor isoforms in extraneural tissues involved in glucose metabolism. We identified an isoform of the human CB1 receptor (CB1b) that is highly expressed in β-cells and hepatocytes but not in the brain. Importantly, CB1b shows stronger affinity for the inverse agonist JD-5037 than for rimonabant compared to CB1 full length. Most relevant to the field, CB1b is a potent regulator of adenylyl cyclase activity in peripheral metabolic tissues. CB1b blockade by JD-5037 results in stronger adenylyl cyclase activation compared to rimonabant and it is a better enhancer of insulin secretion in β-cells. We propose this isoform as a principal pharmacological target for the treatment of metabolic disorders involving glucose metabolism.

Figure 1. Human CNR1 Gene Splice Variants.

Human CNR1 (6q15) gene structure (A) and its alternatively spliced transcript variants (B): the open bar represents the exons with exon numbers inside and horizontal lines represent the introns. Human CNR1 has splicing sites in two of its exons (A). Within exon 1 there are two splicing sites (1A and 1B) and within coding exon 4 there are four splicing sites (4A-D) indicated by arrows above the exons. Alternatively-spliced transcript variants are shown (B) and GenBank accession numbers are included in parenthesis; the corresponding protein isoforms are indicated with an arrow. CB1 receptor full length protein and N-terminal isoforms are represented by CB1, CB1a and CB1b, respectively. CB1 full length amplicons align the common deleted region in between exon 4C-4D, recognizing all full length splice variants. Each TaqMan probe recognizes the adjacent and inter-exonal spliced exons and is represented by short horizontal bars in black, grey and white for CB1, CB1a and CB1b receptors, respectively. The specific full length CB1 receptor amplicon (black bar) aligns the intra-exonal sequence between splicing sites 4B and 4C, and does not recognize the spliced variants CB1a and CB1b (C). Alignment of the protein N-terminal amino acid sequences of CB1a and CB1b with full length CB1 receptor (D). Gaps are represented by dashed lines, identical amino acids by the asterisks. The beginning of the first transmembrane domain is underlined. Increasing concentrations of ACEA did not influence on Ex4-mediated cAMP accumulation in CHO cells when the cells were not transduced with any CB1 receptor (E). Cells were pre-treated with ACEA for 20 min before addition of Ex4 for a further 15 min. **p < 0.01, n.s. = not significant compared to vehicle. Treatment with the synthetic CB1 receptor agonist ACEA attenuated Ex4-mediated cAMP accumulation in CHO-GLP-1R cells when transduced with CB1 and CB1a, and to a greater extent with CB1b (C). All values were normalized to protein concentration. Data represent mean percentage over maximum response ± SEM from at least three independent experiments. ***p ≤ 0.005 compared to CB1 full length. For cell line validation, see Supplementary Fig. S1.

Figure 2. Relative abundance of CB1 receptor isoforms mRNA in human tissues.

Real time PCR quantification of relative expression of CB1 receptor isoforms in brain and peripheral tissues (A). CB1 receptor isoforms expression in human brain regions (B): nucleus accumbens (NAc), hippocampus and cortex. Expression is represented as fold change over CB1 expression in muscle. Data show mean ± SD (n = 3); p ≤ 0.001. Real time PCR quantification of CB1 receptor isoforms in purified human hepatocytes (C) (n = 5 separate livers) and isolated islets of Langerhans (D) (n = 15 donors). Chromatographic separations of the digested protein samples for immunoprecipitated CB1 protein isoforms from human islets (43.1 BMI; confirmed T2DM donor) (E) and primary human hepatocytes (F). Isotopically labeled synthetic versions of the corresponding proteotypic peptides were used to determine retention times (Supplementary Fig. S3). Relative expression of CB1 isoforms in hepatocytes (G) or in isolated islets (H) from donors with a BMI lower than 30 (n = 17 for hepatocytes; n = 3 for islets) or ≥ 30 (n = 5 for hepatocytes; n = 5 for islets). Data represented as fold change over CB1, with horizontal bars representing the median; **p ≤ 0.01, ***p ≤ 0.001. β-actin was used as housekeeping gene.

Figure 3. mRNA expression of CB1 receptor isoforms in single pancreatic β-cells.

Islets of Langerhans are a complex cluster of different cell types: endocrine cells (α-, β-, δ-, ε- and PP-cells), nerve fibers and blood cells (A). Disaggregated human islets were stained with two antibodies for cell surface markers and sorted into three endocrine fractions (P1, P2 and P3) and a non-endocrine fraction (NE) (B). RNA from P1, P2 and P3 was extracted and retro-transcribed into cDNA. RT-PCR of the five endocrine hormones was performed to confirm the enrichment in each endocrine cell type (C–E). Expression was normalized to insulin in P3. Data represent mean ± SD, p < 0.001. The endocrine fractions (P1-P3) were then sorted into individual single cells. From those, RNA was extracted and reverse transcribed into cDNA, pre-amplified for the hormone transcripts of interest and diluted prior to real time PCR amplification. Expression was normalized to glucagon (F), somatostatin (G) and insulin (H). p ≤ 0.001. CB1 isoforms were quantified and represented as fold difference relative to corresponding hormone levels (I). ***p ≤ 0.005. Data are represented in box and whisker plot (n = 10). Relative amount of copies of each CB1 receptor isoform was quantified using specific Taqman primers and probes.

Figure 4. CB1 receptor human isoforms differ in activity for cannabinoid receptor inverse agonists in metabolic systems.

Quantification of glucagon-stimulated cAMP accumulation in primary human hepatocytes (A). Hepatocytes from a human donor were pretreated with rimonabant (100 nM) or JD-5037 (100 nM) prior to glucagon stimulation. Data are the mean ± SEM from two independent experiments; *p ≤ 0.05 compared to vehicle. Quantification of cAMP accumulation in isolated human islets of Langerhans (B). Islets from one donor (20.4 BMI) were pre-cultured in 2 mM glucose for 2 hours followed by stimulation with 7.5 mM glucose in the presence or absence of increasing concentrations of rimonabant or JD-5037. Data are the mean ± SEM of 3 separated incubations; *p ≤ 0.05 compared to 7.5 mM glucose alone.

Figure 5. CB1 receptor blockade stimulates insulin secretion in isolated islets of Langerhans.

Experimental timeline for perfusion of human islets (A). Insulin secretion from human islets perfused with glucose alone (7.5 mM) (blue line) and in combination with 0.33 nM Ex4 (red line) (half maximum insulin secretion), 0.1 nM JD-5037 (green line) or 0.1 nM rimonabant (purple line) (B). Treatment time point is indicated with a black arrow. Islets were obtained from a single obese donor (34.7 BMI). Area under the curve (AUC) of insulin secretion in perfused islets from 3 independent donors (BMIs of 53.2, 32.8 and 34.7) with glucose alone or in combination with increasing concentrations of JD-5037 (C). Ex4 and rimonabant AUC are indicated with red and purple dotted lines respectively. Glucagon secretion from human islets perfused with glucose alone, glucose and Ex4 or glucose and JD-5037 (D). Data are the mean ± SEM of 3; *p ≤ 0.05 compared to vehicle.