Plasma endocannabinoid levels in lean, overweight, and obese humans: relationships to intestinal permeability markers, inflammation, and incretin secretion

Abstract

Intestinal production of endocannabinoid and oleoylethanolamide (OEA) is impaired in high-fat diet/obese rodents, leading to reduced satiety. Such diets also alter the intestinal microbiome in association with enhanced intestinal permeability and inflammation; however, little is known of these effects in humans. This study aimed to 1) evaluate effects of lipid on plasma anandamide (AEA), 2-arachidonyl-sn-glycerol (2-AG), and OEA in humans; and 2) examine relationships to intestinal permeability, inflammation markers, and incretin hormone secretion. Twenty lean, 18 overweight, and 19 obese participants underwent intraduodenal Intralipid infusion (2 kcal/min) with collection of endoscopic duodenal biopsies and blood. Plasma AEA, 2-AG, and OEA (HPLC/tandem mass spectrometry), tumor necrosis factor-α (TNFα), glucagon-like peptide-1 (GLP-1), and glucose-dependent insulinotropic peptide (GIP) (multiplex), and duodenal expression of occludin, zona-occludin-1 (ZO-1), intestinal-alkaline-phosphatase (IAP), and Toll-like receptor 4 (TLR4) (by RT-PCR) were assessed. Fasting plasma AEA was increased in obese compared with lean and overweight patients (P < 0.05), with no effect of BMI group or ID lipid infusion on plasma 2-AG or OEA. Duodenal expression of IAP and ZO-1 was reduced in obese compared with lean (P < 0.05), and these levels related negatively to plasma AEA (P < 0.05). The iAUC for AEA was positively related to iAUC GIP (r = 0.384, P = 0.005). Obese individuals have increased plasma AEA and decreased duodenal expression of ZO-1 and IAP compared with lean and overweight subjects. The relationships between plasma AEA with duodenal ZO-1, IAP, and GIP suggest that altered endocannabinoid signaling may contribute to changes in intestinal permeability, inflammation, and incretin release in human obesity.

INTRODUCTION

Obesity is associated with chronic low-grade inflammation and metabolic disturbances, including insulin resistance and type 2 diabetes. One of the major drivers of the obesity epidemic is the consumption of an energy-dense high-fat diet, which promotes hyperphagia secondary to attenuation of the gut-brain signaling mechanisms involved in the control of food intake (24). In rodents, high-fat diet-induced hyperphagia is driven by alterations in the composition of the gut microbiota with consequent development of low-grade inflammation in the gastrointestinal tract. This is evidenced by increased expression of toll-like receptor 4 (TLR-4, the receptor for the microbial toxin, lipopolysaccharide (LPS or endotoxin)), in the ileum (11), increased systemic levels of LPS and proinflammatory cytokines (3), and altered intestinal permeability characterized by decreased intestinal expression and localization of tight junction proteins, including zona-occludin 1 (ZO-1) and occludin (4). Furthermore, the intestinal expression of intestinal alkaline phosphatase (IAP), an enzyme that inactivates LPS, is decreased by a high-fat diet in obesity-prone, but not obesity-resistant, rodents (11). Recent studies have also highlighted an important role of interactions between the gut microbiota and the endocannabinoid system in the regulation of energy homeostasis (29).

The endocannabinoids, 2-arachidonyl-sn-glycerol (2-AG) and anandamide (AEA), are lipid-derived mediators that are involved in the control of food intake (1, 13). 2-AG and AEA stimulate food intake by activating cannabinoid-1 (CB₁) receptors (16). In the small intestine of rodents, 2-AG and AEA concentrations increase with fasting (14, 21) and upon oral sensing of dietary fats (13), and both 2-AG and AEA decrease with refeeding (21). Intestinal levels of endocannabinoids are altered in obesity, for example, obese Zucker rats exhibit higher duodenal levels of 2-AG and AEA than wild-type rats following both food deprivation and refeeding, while in high-fat diet induced obese rats, duodenal 2-AG is higher, but AEA lower, than in control rats fed a standard laboratory diet (21). In Western diet (high-fat, high-sucrose)-induced obese mice, both 2-AG and AEA are increased (1). While studies are inconsistent, human obesity is generally associated with increased concentrations of endocannabinoids in plasma and adipose tissues, and most commonly elevated 2-AG, but not AEA, correlates with BMI and waist circumference (2, 7, 12). While now withdrawn from the market due to adverse psychological effects, the CB₁ receptor antagonist Rimonabant promoted weight loss in obese humans (23). Interest in the endocannabinoid system as a target for obesity therapy has re-emerged since rodent studies demonstrated that obesity-associated changes in the gut microbiota activated the intestinal endocannabinoid system (via LPS) to increase gut permeability, which further enhanced plasma LPS levels and peripheral inflammation (5, 29). Studies examining the relationships between intestinal endocannabinoids, intestinal permeability and inflammation are lacking in humans, although a recent study using human colonic explants from patients with inflammatory bowel disease has reported that cannabidiol and palmitoylethanolamine exert anti-inflammatory effects (8).

Endocannabinoids may also determine the secretion of the incretin hormones, glucose-dependent insulinotropic peptide (GIP) and glucagon-like peptide-1 (GLP-1). CB₁ receptors are expressed on enteroendocrine K- and L-cells in rodents (19, 28), and activation of the CB₁ receptor by AEA before an oral glucose tolerance test inhibits GIP secretion in rodents (28). In contrast, in humans, acute administration of the CB₁ receptor agonist, nabilone, enhances fasting plasma GIP by ~80%, has no effect on glucose-mediated GIP release, but decreases the GLP-1 response to oral glucose (6). Therefore, while it has not been directly evaluated, changes in incretin levels observed in obesity (35) may arise due to altered endocannabinoid levels.

Endocannabinoid-related molecules, such as the fatty acid-derived n-acylethanolamines, and in particular, OEA, have also been shown to play a role in energy homeostasis, where they reduce energy intake (by prolonging the inter-meal interval and reducing meal frequency) via activation of the nuclear receptor, peroxisome-proliferator-activated receptor-α (PPAR-α) (17, 25, 31). OEA is also an endogenous ligand for the fatty acid sensing receptor, GPR119, which stimulates GLP-1 secretion (22, 30). Whether OEA levels are altered in human obesity is currently unclear.

We recently characterized the effects of intraduodenal lipid infusion on the duodenal expression of fat sensing molecules in lean, overweight and obese humans. We reported that human obesity was associated with altered capacity for the sensing of dietary fat, including decreased duodenal expression of free fatty acid receptors (FFAR) 1 and 4, and increased expression of cluster of differentiation 36 (CD36) (9, 10). Using plasma samples and duodenal biopsies collected in these studies, the aim of the current investigation was to (i) evaluate effects of intraduodenal lipid infusion on plasma levels of AEA, 2-AG and OEA in humans, and (ii) examine relationships between plasma concentrations of endocannabinoid related molecules with BMI, intestinal permeability and inflammation markers, and incretin hormone secretion.

MATERIALS AND METHODS

Participants

Twenty lean [10 males, body mass index (BMI) 22 ± 0.5 kg/m², age: 28 ± 2 yr] 18 overweight (12 males, BMI 27 ± 0.3 kg/m², age: 32 ± 3 yr), and 19 obese (12 males, BMI 35 ± 1 kg/m², age: 30 ± 2 yr) participants were enrolled in the study. Participant demographics have been published previously (9, 10). In brief, all participants were weight stable for at least 3 mo before inclusion in the study and reported no gastrointestinal (GI) symptoms, had no prior GI surgery, did not take medications or supplements known to affect GI motility or appetite (including fish oil), consumed <20 g of alcohol per week, and were nonsmokers. The study protocol was approved by the Royal Adelaide Hospital Research Ethics Committee and carried out in accordance with the Declaration of Helsinki. Each participant provided written, informed consent before inclusion in the study. The trial was registered with the Australia and New Zealand Clinical Trial Registry (http://www.anzctr.org.au Trial no. ACTRN12612000376842).

Study Design and Protocol

Endoscopy and collection of duodenal mucosal biopsies.

Participants attended the Department of Gastroenterology and Hepatology, Royal Adelaide Hospital, at 0830 following a standardized evening meal [400 g beef lasagna (2,479 kJ, 20 g fat, 20 g protein, 80 g carbohydrate); McCain Foods, Wendouree, Victoria, Australia] and an overnight fast of 12 h from solids and 10 h from liquids. The protocol for endoscopic collection of mucosal biopsies has been described previously (9). Briefly, a small-diameter video endoscope was passed through the nose into the second part of the duodenum, and, once positioned, two duodenal biopsies were collected using standard endoscopic biopsy forceps and placed immediately in Allprotect Tissue Reagent (Qiagen, Australia). An intravenous cannula was inserted into a forearm vein and a baseline blood sample (10 ml) was collected (t = 0 min). Following this, an intraduodenal infusion of 10% Intralipid (Fresenius Kabi, China; 2 kcal/min, 109 ml/h) commenced via endoscope infusion channel for 30 min. Blood samples, for analysis of plasma TNFα concentrations, were collected at t = 10, 20, and 30 min and placed in ice-chilled EDTA-treated and serum tubes, as described previously (9). After the infusion (t = 30 min), two additional duodenal biopsies were collected.

Effect of intraduodenal lipid infusion on plasma concentrations of endocannabinoids and incretins.

Participants attended the Discipline of Medicine at 0830 following a standardized evening meal and overnight fast, as described above. Anesthetic spray and gel were administered into the nasal cavity (as above) before insertion of a small-diameter (3.5 mm) catheter (Dentsleeve International, Mui Scientific), which was allowed to pass via peristalsis through the pylorus into the second part of the duodenum. Accurate positioning of the catheter across the pylorus was achieved by monitoring the transmucosal potential difference using a monitoring electrode (Red Dot, 3M Healthcare) placed on the forearm as a reference (20). Once positioned, an intravenous cannula was inserted into a forearm vein, and a baseline blood sample (10 ml) was collected (t = 0 min). Intraduodenal infusion of 10% Intralipid then commenced at a rate of 2 kcal/min for 120 min (t = 0–120 min), during which blood samples were collected every 15 min and placed in ice-chilled EDTA-treated and serum tubes, as described previously (9).

Measurements

Analysis of AEA, OEA, and 2-AG levels in plasma.

Plasma samples collected at t = 0, 30, 60, and 120 min were analyzed. Lipid extraction and analysis were performed as previously described (1). Plasma (0.5 ml) was added to 1.0 ml of methanol solution containing the internal standards, [²H₅]2-AG, [²H₄]AEA, and [²H₄]OEA (Cayman Chemical, Ann Arbor, MI). Lipids were extracted with chloroform (2 ml) and washed with 0.9% saline (0.5 ml). Organic phases were collected and separated by open-bed silica gel column chromatography. The eluate was gently dried under a N₂ stream (99.998% pure) and resuspended in 0.1 ml of methanol-chloroform (9:1), with 1 μl injected for ultraperformance liquid chromatography-tandem mass spectrometry (UPLC/MS/MS) analysis.

Data were collected using an Acquity I Class UPLC system coupled to a Xevo TQ-S Mass Spectrometer (Waters, Milford, MA) with accompanying electrospray ionization (ESI) interface. Lipids were separated on an Acquity UPLC BEH C₁₈ column (2.1 × 50 mm ID, 1.7 µm, Waters) with inline Acquity guard column (UPLC BEH C₁₈ VanGuard Pre-column; 2.1 × 5 mm ID, 1.7 µm, Waters) and eluted by a gradient of methanol in water (0.25% acetic acid, 5 mM ammonium acetate) according to the following gradient at a flow rate of 0.4 ml/min: 80% methanol 0.5 min, 80–100% methanol 0.5–2.5 min, 100% methanol 2.5–3 min, 100–80% methanol 3–3.1 min). Column temperature was maintained at 40°C, and samples were maintained in the sample manager at 10°C. Argon (99.998%) was used as collision gas. MS detection was in positive ion mode and capillary voltage set at 0.1 kV. Cone voltage and collision energy were as follows, respectively: 2-AG = 30 V, 12 V; [²H₅] 2-AG = 25 V, 44 V; AEA = 30 V, 14 V; [²H₄]AEA = 26 V, 16 V; OEA = 28 V, 16 V; [²H₄]OEA = 48 V, 14 V. Lipids were quantified using a stable-isotope dilution method detecting protonated adducts of the molecular ions [M+H]⁺ in the multiple reaction monitoring (MRM) mode. Acyl migration from 2-AG to 1-AG is known to occur (32); thus, all reported values for 2-AG represent the sum of 2-AG and 1-AG. Tissue processing and LC/MS analysis from an individual experiment occurred independently of other experiments. Extracted ion chromatograms were used to quantify 2-AG (m/z = 379.3>287.3), AEA (m/z = 348.3>62.0), and OEA (m/z = 326.4 > 62.1), and [²H₅]2-AG (m/z = 384.3 > 93.4), [²H₄]AEA (m/z = 352.3>66.1), and [²H₄]OEA (m/z = 330.4 > 66.0), which were used as internal standards.

RNA extraction.

Frozen duodenal biopsies were disrupted using a bead-based tissue homogenizer (TissueLyser LT, Qiagen) and homogenized through Qiashredder columns (Qiagen). Total cellular RNA was isolated using the PureLink MicroKit (Invitrogen, Thermo Fisher Scientific), which included an on-column DNase digestion, per manufacturer’s instructions. RNA quantity was determined using a Nanodrop Lite Spectrophotometer (Thermo Fisher Scientific) and purity assessed using an A₂₆₀/A₂₈₀ ratio.

Quantification of duodenal gene expression by relative RT-PCR.

Real-time RT-PCR was performed using the 7500 Fast Real-Time PCR system (Applied Biosystems, Thermo Fisher Scientific). Taqman primers (Life Technologies, Thermo Fisher Scientific) were used to determine the expressions of TLR4 (Hs00152939_m1), alkaline phosphatase-1 (ALP1, Hs00357579_g1), occludin (OCLN, Hs00170162_m1), and ZO-1 (tight junction protein-1) (TJP1, Hs01551861_m1) relative to expression of the housekeeper β2-microglobulin (β2M) (Hs00984230_m1). All targets were assessed in triplicate, according to the manufacturer’s instructions.

Plasma TNFα and incretin hormone concentrations.

Multiplex assays were used to determine plasma TNFα) concentrations (pg/ml, t = 0, 10, 20, 30 min; Milliplex MAP Human Cytokine/Chemokine Magnetic Bead Panel, HCYTOMAG-60K), and total GLP-1 and GIP concentrations (both pg/ml, t = 0, 15, 30, 60, 90, 120 min; Milliplex MAP Human Metabolic Hormone Magnetic Bead Panel, HMHEMAG-34K; Millipore, Temecula, CA), using the Bio-plex MAGPIX Multiplex Reader (Luminex, Millipore) and xPONENT software (Luminex version 4.2, Millipore) according to the manufacturer’s instructions. There was negligible antibody cross-reactivity. Intraassay coefficients of variation (CVs) were ≤10%, and interassay CVs were ≤15% for all analytes. The minimum detection limits were for TNFα, 0.7 pg/ml; GLP-1, 2.5 pg/ml; and GIP, 0.6 pg/ml.

Data and Statistical Analyses

Statistical analysis was performed using SPSS software (SPSS, IBM, version 24), and all graphs were generated using GraphPad Prism 7 (GraphPad Software). One-way ANOVA with post hoc Bonferroni testing was used to compare values for all variables across BMI group (i.e., lean, overweight, obese) and time. Post hoc paired comparisons, corrected for multiple comparisons using Bonferroni’s correction, were performed if ANOVAs revealed significant effects. Relationships between variables were determined by correlation, with Pearson’s r values presented. Incretin hormone, AEA, 2-AG, and OEA data were expressed as incremental area under the curve [iAUC; calculated using the trapezoidal rule from t = 0 to 120 min (or t = 0–30 min for TNFα)]. This iAUC value was divided by the time of last measurement to obtain a final weighted average (AUC) to account for the rare times when samples could not be collected (e.g., bathroom breaks) or data could not be obtained from the assays (10). Data are expressed as means ±  SE, with statistical significance accepted at P < 0.05.

RESULTS

Plasma 2-AG, AEA, and OEA Concentrations

There was no significant difference between BMI groups, or effect of lipid infusion, on plasma 2-AG concentrations (Fig. 1A). There was a significant effect of BMI group (P = 0.003), but not lipid infusion, on plasma AEA concentrations (Fig. 1B). Plasma AEA was higher in obese compared with both lean (P = 0.005) and overweight (P = 0.017) groups, with no difference between lean and overweight groups. There was no significant difference between BMI groups or effect of lipid infusion on plasma OEA concentrations (Fig. 1C).

Fig. 1.

Plasma 2-arachidonylglycerol (2-AG; A), anandamide (AEA; B), and oleoylethanolamide (OEA; C) concentrations at fasting (0 min) and during intraduodenal infusion of Intralipid (2 kcal/min for 120 min) in lean (n = 20), overweight (n = 18), and obese (n = 19) individuals. There was no significant difference between BMI groups, or effect of lipid infusion, on plasma 2-AG (A) or OEA (C) concentrations. Plasma AEA was higher in obese vs. lean (P = 0.005) and overweight (P = 0.017) subjects, with no difference between lean and overweight groups. *P < 0.05 vs. lean and overweight.

Duodenal Expression of TLR4, IAP, Occludin, and ZO-1

Baseline duodenal expression of IAP (P = 0.018) and ZO-1 (P = 0.002) was significantly different across BMI groups, with a trend for difference evident for TLR-4 (P = 0.08) and occludin (P = 0.08) (Fig. 2). Baseline IAP expression was significantly lower in obese compared with lean (P < 0.05) groups, with no difference between lean and overweight or obese and overweight groups. Baseline ZO-1 expression was significantly lower in obese compared with lean groups (P = 0.002), and in overweight compared with lean groups (P = 0.029), with no difference between overweight and obese groups. There was no effect of intraduodenal lipid on duodenal expression of TLR-4, IAP, occludin, or ZO-1 (data not shown).

Fig. 2.

Baseline duodenal expression of Toll-like receptor-4 (TLR4; A), intestinal alkaline phosphatase (IAP; B), occludin (C), and zona-occludin 1 (ZO-1; D) in lean (n = 20), overweight (n = 18), and obese (n = 19) individuals. Baseline IAP expression (B) was significantly lower in obese vs. lean (P < 0.05), with no difference between lean and overweight or obese and overweight groups. Baseline ZO-1 expression (D) was significantly lower in obese vs. lean (P = 0.002), and in overweight compared with lean (P = 0.029), with no difference between overweight and obese groups. *P < 0.05 obese vs. lean; #P < 0.05 overweight vs. lean.

Plasma TNFα Concentrations

There was no significant difference between BMI groups (P = 0.7) and no effect of lipid infusion (P = 0.2) on plasma concentrations of TNFα (Fig. 3).

Fig. 3.

Plasma tumor necrosis factor-α (TNFα) concentrations at fasting (0 min) and during intraduodenal infusion of Intralipid (2 kcal/min for 120 min) in lean (n = 19), overweight (n = 16), and obese individuals (n = 17). There was no significant difference between BMI groups (P = 0.7) and no effect of lipid infusion (P = 0.2) on plasma concentrations of TNFα. Sufficient plasma was not available for 1 lean, 2 overweight, and 2 obese individuals.

Plasma Incretin Hormone Concentrations

Data are available in the publication from Cvijanovic et al. (10).

Relationships Between Variables

Plasma AEA concentrations at baseline were positively related to BMI (r = 0.34, P = 0.014). Baseline expressions of IAP (r = −0.329, P = 0.014), occludin (r = −0.28, P = 0.038), and ZO-1 (r = −0.395, P = 0.003) were negatively related to BMI. Intralipid-evoked AEA concentrations (plasma iAUC) were negatively related to baseline TLR4 (r = −0.283, P = 0.04), IAP (r = −0.305, P = 0.032), occludin (r = −0.398, P = 0.001), and ZO-1 expression (r = −0.445, P = 0.001). Plasma TNFα concentrations at baseline related positively to expression of IAP (r = 0.401, P = 0.002) and ZO-1 at baseline (r = 0.256, P = 0.05). Finally, Intralipid-evoked AEA was positively related to evoked GIP (iAUC, r = 0.384, P = 0.005) but not to evoked GLP-1.

DISCUSSION

We have demonstrated that fasting plasma levels of the endocannabinoid AEA, but not 2-AG or the endocannabinoid-related OEA, were higher in obese than in lean and overweight individuals. Obese individuals also had lower duodenal expression of ZO-1 and IAP, and a trend toward lower duodenal occludin and TLR4 expression at baseline. These changes were negatively related to plasma AEA concentrations, suggesting that altered endocannabinoid signaling may contribute to changes in intestinal permeability and inflammation in human obesity. Plasma AEA levels were also positively related to plasma GIP, adding support that the endocannabinoid system plays a significant role in mediating GIP secretion in humans.

Our finding of increased plasma AEA in obesity is consistent with reports in animals and humans that documented increased peripheral endocannabinoid levels in obesity (1, 12, 21, 26). Whereas many studies report strong correlations between plasma 2-AG and body fat (2), we were not able to replicate these findings. Although mean levels of 2-AG were higher in obese compared with both lean and overweight individuals, the values were far more variable than those for AEA, and for this reason we believe that our study may have been underpowered to detect such a correlation. Rodent studies indicate that changes in peripheral endocannabinoid concentrations are induced by a “Western” style high-fat diet (1) and that elevated endocannabinoid levels in plasma and jejunal mucosa act via peripheral cannabinoid receptor 1 (CB1) receptors to promote hyperphagia (1). Mammals have an innate preference for dietary fats, and in sham-fed rats, where food is tasted but does not enter the small intestine, corn oil or emulsions containing oleic or linoleic acid rapidly and markedly increase jejunal levels of 2-AG and AEA (13, 15). Moreover, CB1 receptor blockade abolishes effects of oral fat on sham feeding and preference for fat (15). Accordingly, changes in peripheral endocannabinoid signaling in obesity may be important mediators of both fat preference and gut-to-brain signaling mechanisms involved in the regulation of energy intake. Acute intraduodenal infusion of lipid had no effect on plasma endocannabinoid levels in the present study. It is possible that the duration and/or load of lipid exposure was insufficient or that oral stimulation of fatty acid sensing receptors (13), such as CD36 and GPR40 and 120, is required to elicit changes in peripheral endocannabinoid levels, and further studies are needed to address this question in humans.

Duodenal expression of IAP and the tight-junction protein ZO-1 was reduced in obese compared with lean individuals, and there was a trend for reduced expression of TLR4 and occludin. These molecules are important markers of increased intestinal permeability and inflammation that have been shown in animal studies to be altered secondarily to high-fat diet- and obesity-induced changes in the intestinal microbiome (11), associated with increased systemic levels of LPS, termed “metabolic endotoxemia” (2, 3). These data suggest that the observed changes in duodenal expression of IAP and ZO-1 in obese compared with lean subjects are surrogate markers of an altered intestinal microbiome and enhanced intestinal permeability; however, this requires further direct investigation. Furthermore, the potential role of dietary fat intake remains to be elucidated. In humans, acute ingestion of fat has been associated with an increase in circulating LPS (18) and increased intestinal permeability (assessed by lactulose/mannitol tests) in obese individuals (33). Not all studies show changes in intestinal permeability despite evidence of increased local and systemic inflammation in obese individuals (34), which may reflect the relative insensitivity of the intestinal permeability tests, based on detection of urinary markers in humans. In the present study, we identified a relationship between the duodenal expression of IAP and ZO-1 with plasma levels of the proinflammatory cytokine TNFα, suggesting an important role for these molecules in mediating the chronic low-grade inflammation associated with obesity.

We also found that lipid-evoked plasma AEA levels were negatively related to baseline duodenal expression of TLR4, IAP, occludin, and ZO-1. It is possible that these sensors in the mucosa may play an increased role in lipid-evoked AEA release when permeability (and access) is increased. Rodent studies have provided evidence that the obesity-associated changes in the gut microbiota leads to activation of the endocannabinoid system (via LPS). Administration of the CB1 receptor agonist HU-210 to wild-type mice increased intestinal permeability and decreased ZO-1 and occludin expression in vitro, in Caco-2 cells (29). Conversely, in obese mice, pharmacological blockade of the CB1 receptor restored the distribution and localization of the tight-junction proteins ZO-1 and occludin and decreased plasma LPS levels (5, 29). Accordingly, in human obesity, altered plasma AEA may be an important mediator of the effects of obesity on intestinal permeability and inflammation. In rodents, alterations in the composition of the gut microbiota, with consequent development of low-grade inflammation in the gastrointestinal tract, have been shown to be responsible for the development of hyperphagia and obesity (11). Further studies in humans are now required to understand the relationships between these changes in endocannabinoid system activity, intestinal permeability, and food intake.

Plasma AEA concentrations related positively to plasma GIP, but not GLP-1, concentrations. Recent studies have suggested a role of the endocannabinoid system in mediating the secretion of incretin hormones (6). As discussed, CB1 receptors are expressed on enteroendocrine K and L cells in rodents (19, 28), and AEA inhibits the GIP response to oral glucose in rodents (28). In contrast, our data, and a recent study using acute administration of the CB1 receptor agonist nabilone (6), provide evidence that AEA enhances GIP secretion in humans. The reasons for these discrepancies are unclear; however, the endocannabinoid system is likely to play an influential role in mediating the GIP response to fat in humans.

Plasma OEA concentrations did not differ among lean, overweight, and obese humans during fasting and did not change in response to duodenal fat infusion. In rodents, the presence of fat in the duodenum stimulates the small intestinal conversion of oleic acid, found commonly in the diet, to OEA (31). When administered exogenously to rodents, OEA reduces overall energy intake by prolonging the time interval between meals (31). Our infused lipid emulsion comprised ~19–30% oleic acid, which may have been insufficient to stimulate OEA mobilization in the present study. Previous studies in humans have demonstrated a significant increase in plasma OEA concentrations in proportion to the oleic content of the ingested oil (30 g), which was associated with reduced subsequent energy intake (27).

In conclusion, we have demonstrated increased plasma levels of AEA, but not 2-AG or OEA, in obese compared with lean and overweight individuals. Obese individuals also displayed decreased fasting duodenal expression of ZO-1 and IAP. Plasma AEA concentrations were negatively related to duodenal expression of ZO-1 and IAP and positively related to plasma GIP, suggesting that altered endocannabinoid signaling may contribute to changes in intestinal permeability, inflammation, and GIP secretion in human obesity.

GRANTS

T. J. Little was supported by a National Health and Medical Research Council (NHMRC) Career Development Fellowship (Grant 1022706) and C. F. B by an NHMRC Senior Research Fellowship (Grant 1103020, 2016-21). The study was also supported by a Royal Adelaide Hospital Research Committee Clinical Project Grant (to T. J. Little, 2015-16), and National Institutes of Health Grant DA-034009 and Grant DK114978 to N. V. DiPatrizio.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

T.J.L., C.F.-B., and R.L.Y. conceived and designed research; T.J.L., N.C., N.V.D., D.A.A., and C.K.R. performed experiments; T.J.L. analyzed data; T.J.L., N.C., N.V.D., C.K.R., C.F.-B., and R.L.Y. interpreted results of experiments; T.J.L. prepared figures; T.J.L. drafted manuscript; T.J.L., N.C., N.V.D., D.A.A., C.K.R., C.F.-B., and R.L.Y. edited and revised manuscript; T.J.L., N.C., N.V.D., D.A.A., C.K.R., C.F.-B., and R.L.Y. approved final version of manuscript.