Longitudinal changes in endocannabinoidome and dietary lipid profile in severe obesity: impact of sleeve gastrectomy

Abstract

Background

The endocannabinoid system and its extension, the endocannabinoidome (eCBome), are lipid-based signalling systems involved in regulating energy balance and metabolic homoeostasis. The eCBome includes a variety of bioactive lipids derived from fatty acids, such as N-acylethanolamines (NAEs) and 2-monoacyl-glycerols (2‑MAGs), which have been linked to different patterns of adiposity. Dietary intake, particularly fatty acid intake, plays a key role in shaping the circulating eCBome profile. Bariatric surgery and subsequent diet modifications can significantly modulate the eCBome, potentially restoring metabolic balance in individuals with obesity.

Aim

The aim of the study is to identify longitudinal changes in the circulating profile of eCBome mediators in individuals living with severe obesity, before and after a sleeve gastrectomy, and to relate these changes to metabolic improvements and changes in dietary intake.

Methods

The cohort includes 33 adults with severe obesity (BMI ≥ 35 kg/m²), awaiting a sleeve gastrectomy. Blood samples, anthropometric measurements, metabolic profile and dietary intakes (assessed via 24 h dietary recalls) were collected before and 4 months after surgery. Circulating eCBome mediators were quantified by liquid-chromatography tandem mass spectrometry in fasting plasma samples.

Results

Four months after sleeve gastrectomy, reductions in adiposity (BMI, body fat mass, waist and neck circumferences) and improvements in metabolic parameters (triglycerides, high-density lipoprotein cholesterol levels and fasting insulin) were accompanied by significant changes in the levels of some circulating eCBome mediators. Post-surgery circulating levels of EPEA and DHEA reduced, whereas levels of 2-AG, 2-LG, 2-DPG and 2-DHG increased after adjusting for pre-surgery levels and sex (P < 0.05). Interestingly, the circulating levels of eCBome mediators within each family (i.e., NAEs and 2-MAGs) were more closely intercorrelated after surgery than before. Adiposity measurements and dietary fatty acid intakes, such as arachidonic acid and omega-3 fatty acids, were associated with the circulating eCBome profile only after surgery.

Conclusion

Weight loss and metabolic profile improvements induced by sleeve gastrectomy correlate with changes in the circulating eCBome profile. In severe obesity, neither adiposity nor dietary fatty acid intake appear to directly influence the circulating eCBome profile. The contribution of these factors, which had been previously observed in individuals with normal weight to moderate obesity, only becomes evident following weight loss.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12944-026-02988-1.

Background

The endocannabinoid (eCB) system and its extension, the endocannabinoidome (eCBome), are large lipid signalling systems [1]. The eCB system comprises two main ligands, 2-arachidonoyl-glycerol (2-AG) and anandamide (AEA), which derive from arachidonic acid. It also features the cannabinoid receptors 1 and 2 (CB₁ and CB₂), and the enzymes required to synthesize and degrade the eCBs [2]. The concept of the eCBome emerges with the identification of numerous congeners of AEA, the N-acylethanolamines (NAEs), and of 2-AG, the 2-monoacyl-glycerols (2-MAGs), having distinct fatty acyl chains linked to ethanolamine or glycerol moieties [3]. Interestingly, while these congeners partly share their anabolic and catabolic pathways with eCBs, they possess a distinct specificity to a large array of molecular targets, including several peroxisome proliferator-activated receptors (PPARs), transient receptor potential channels and G protein-coupled receptors [1, 4].

The eCBome plays multiple roles in maintaining homoeostasis [1]. As for endocannabinoids, eCBome mediators play a key role in the regulation of food intake and energy balance [1, 5]. Associations between obesity and the eCBome mediator profile, both in circulation and several tissues, are well established [6, 7]. For instance, circulating 2-MAG mediators tend to be positively associated with visceral adiposity, while NAE mediators show associations with total adiposity [8].

Moreover, the profile of circulating eCBome mediators has previously been associated with recent dietary intake, especially the relative intake of different fatty acid classes [8]. As these mediators are derived from fatty acids, the composition of the lipid pools from which they are produced plays a key role in determining their levels. This has been evidenced both in response to long-term diets rich in omega-3 fatty acids and along day-to-day variations in dietary fatty acid profile [9].

Weight loss interventions have the potential to induce favourable changes in the circulating eCBome profile among individuals living with obesity [10, 11]. Reductions in adiposity, especially in the visceral fat compartment, as well as changes in dietary habits, notably variations in the intake of unsaturated and polyunsaturated fatty acids, may contribute to the improvements [8, 12–14]. Bariatric surgery, including sleeve gastrectomy (SG), is designed to reduce food intake and induce significant, sustained weight loss along with metabolic improvements [15]. Following SG, individuals typically exhibit changes in food preferences and eating behaviour, such as consuming smaller but more frequent meals [16]. We hypothesize that these dietary changes, combined with weight loss, may have beneficial effects on the circulating eCBome profile. Indeed, studies in rat models of bariatric surgery have evidenced that the intestinal levels of several eCBome mediators can be affected by this process [17]. Given the role of eCBome mediators in regulating food intake, any change in their circulating profile after bariatric surgery could significantly contribute to the improvements in adiposity and metabolic outcomes observed post-surgery. While the characterization of 2-AG and AEA post-bariatric surgery has been reported [10], there is limited information regarding the modulation of the circulating mediators of the eCBome. This study first aimed to characterize longitudinal changes in circulating eCBome mediator profiles in individuals living with severe obesity before and four months after SG, and to examine their associations with metabolic parameters and dietary intake.

Methodology

Participants and study design

This study presents secondary and exploratory analyses of longitudinal follow‑up data from a prospective cohort study investigating the impact of bariatric surgery on brain and cognitive health. Participants with severe obesity scheduled to undergo laparoscopic SG were recruited between January 2021 and February 2024 at the Institut universitaire de cardiologie et de pneumologie de Québec-Université Laval (IUCPQ-ULaval). Recruitment was conducted through referrals from IUCPQ-ULaval surgeons, followed by direct contact from a trained research team member. The SG procedure consisted of a 150 to 250 cm³ vertical gastrectomy starting 5–7 cm from the pylorus to the Hiss angle, using a 34–44 French Bougie for guidance to create a gastric tube [18]. The greater curvature and fundus of the stomach were removed. Study assessments were conducted at two timepoints: prior to surgery (1–2 months) and 4 months postoperatively.

Eligible participants were women or men aged between 18 and 65 years, non-diabetic, with a body mass index (BMI) ≥ 35 kg/m² and normal cognitive function, as indicated by a Montreal Cognitive Assessment (MOCA) score ≥ 26/30. Exclusion criteria were the following conditions: (1) history of neurological or gastrointestinal diseases, including irritable bowel syndrome; (2) previous brain or gastrointestinal surgery; (3) unexplained intermittent vomiting, severe abdominal pain, diarrhoea, or chronic constipation; (4) current pregnancy or intention to become pregnant during the study period; (5) substance or alcohol abuse; (6) uncontrolled hypertension; (7) psychiatric conditions that could impair the ability to understand the procedure or comply with medical and surgical recommendations; (8) history of severe renal, hepatic, cardiac, or pulmonary disease; (9) inability to provide informed consent or lack of fluency in French; (10) contraindications to MRI, including claustrophobia, the presence of metal implants, or implanted medical devices.

Of the 50 participants enrolled in the study, 5 had their surgeries cancelled, 2 voluntarily withdrew, 6 lacked sufficient plasma samples for secondary analysis, and 4 had incomplete pre- or post-surgery dietary data. Consequently, the final analysis was conducted on 33 participants (27 women and 6 men). The trial was approved by the Research Ethics Committee of the Research Centre of IUCPQ-UL (2021–3456). Each participant provided written informed consent.

As prescribed following SG, all participants consumed a soft diet for up to 6 weeks after surgery. After this period, participants gradually resumed consumption of solid foods. However, several high-fibre and low-digestible foods remained proscribed for up to 3 months post-surgery to minimize the risk of gastrointestinal complications. Participants were encouraged to consume small but frequent meals and to take adequate time to chew properly [19].

Body composition and biochemical measures

All participants underwent 12 h overnight fasting before each visit. Blood samples were collected in EDTA-coated tubes for analysis. All samples were immediately placed at 4° C, centrifuged, aliquoted and stored at − 80 °C. Plasma levels of total cholesterol, high-density lipoprotein cholesterol (HDL-C), triglycerides (TG) and fasting glucose were measured at each visit by the Biochemistry Department of IUCPQ-ULaval. Plasma fasting insulin levels were measured at each visit using the Mercodia Insulin ELISA kit (10-1113-01). Non-HDL cholesterol levels were calculated by subtracting the HDL-C from total cholesterol. Anthropometric measurements (weight, BMI, waist circumference, neck circumference, and waist-to-hip ratio) and body composition measurements (percent of body fat and body fat mass), assessed using a bioimpedance scale (Tanita DC-430U, Arlington Heights, IL, USA), were collected at both visits using standardized procedures.

Dietary assessments

Nutritional intakes were assessed using a validated, web-based, self-administered 24-hour dietary recall (R24W) developed by the Institut sur la nutrition et les aliments fonctionnels (INAF) at Université Laval [20, 21]. Participants received an email invitation to complete the dietary recall questionnaire on three randomly selected, unannounced days (two weekdays and one weekend day) within a 21-day period, both before and 4 months after surgery. Only participants who completed at least two dietary recalls were included in the analyses. The data included metrics such as absolute daily intake of fatty acids (in grams) and relative intake of specific fatty acid classes expressed as a proportion of total lipid intake. Food groups were categorized according to the Healthy Eating Food Index-2019 (HEFI) score, including vegetables and fruits, refined grains, whole grains, plant-based proteins, animal-based proteins, dairy products, and sugary foods and beverages [22].

Quantification of circulating endocannabinoidome mediators

Circulating levels of NAEs and 2-MAGs in plasma were determined using high-performance liquid chromatography (Kinetex C8, 150 × 2.1 mm, 2.6 μm, Phenomenex) coupled with a Shimadzu 8050 triple quadrupole tandem mass spectrometer (LC-MS/MS). Extraction and quantification methods for eCBome mediators were previously described, but modifications of the protocol happened to enhance quantification sensitivity [23]. Frozen (-80 °C) plasma samples (200 µl) were slowly thawed on ice then combined with 300 µl of Tris-HCl (50 mM, pH 7) and 5 µl of our deuterated internal standards. Lipid extraction was performed by adding 2 ml of toluene containing 11,5 µl/ml of acetic acid to each sample, followed by thorough vortexing (30 s). Samples then were centrifuged (4000 × g, 10 min, 4 °C and low brakes). Tubes were next placed in an ethanol-dry ice bath (-80 °C) to freeze the bottom aqueous phase, allowing the collection of the toluene (and lipid-containing) upper phase. Toluene fractions were evaporated under a stream of nitrogen steam, and samples were reconstituted in 60 µl of a mix (50%, v/v) of mobile phase A (H₂O containing 1 mM ammonium acetate and 0.05% acetic acid) and solvent B (MeCN: H₂O, 95:5, v/v, containing 1 mM ammonium acetate and 0.05% acetic acid).

This method allows quantification of NAEs, including AEA, N-palmitoyl-ethanolamine (PEA), N-oleoyl-ethanolamine (OEA), N-linoleoyl-ethanolamine (LEA), N-docosapentaenoyl-ethanolamine (DPEA), N-eicosapentaenoyl-ethanolamine (EPEA) and N-docosahexaenoyl-ethanolamine (DHEA), as well as 2-MAGs, including 2-AG, 2-palmitoyl-glycerol (2-PG), 2-oleoyl-glycerol (2-OG), 2-linoleoyl-glycerol (2-LG), 2-eicosapentaenoyl-glycerol (2-EPG), 2-docosapentaenoyl-glycerol (2-DPG) and 2-docosahexaenoyl-glycerol (2-DHG). While DPEA isomers can be differentiated by our method, only the omega-3 isomer was analysed and identified as DPEA. MAG isomers at the sn-1 and sn-2 positions can be differentiated. However, given their rapid interconversion and the preferential esterification of polyunsaturated fatty acids (PUFA) on the sn-2 position of phospholipids, monounsaturated fatty acids (MUFA) and PUFA-derived 2-MAGs were summed up and identified as 2-MAGs.

Statistical analyses

Paired t-tests were performed to assess significant changes in outcomes from pre- to post-surgery, after verifying the normality of delta values using Shapiro-Wilk test. The effect of sex on outcome measures was explored using a mixed linear model. However, this analysis was underpowered due to sex group imbalance, limiting our capacity to detect smaller but potentially clinically significant effects. Plasma eCBome mediator levels were corrected for batch effects using ratio-based scaling. To identify significant postoperative changes in circulating levels of eCBome mediators, generalized linear models encompassing potential covariates (i.e., baseline age, BMI, and baseline eCBome mediator levels) were computed. For exploratory analyses (i.e., correlation analyses), data were imputed under the assumption of missing not at random (MNAR, < 10% of samples below the limit of quantification). Given the limited set of a priori, biochemically related metabolites measured in a single targeted run, FDR correction was not applied. Instead, interpretation focused on effect sizes, directionality, and consistent patterns within metabolite families rather than isolated P values. The relationships between variables were analysed using Pearson’s or Spearman’s correlation coefficients, as indicated in the figure legends. A two-tailed P value below 0.05 was deemed statistically significant. All analyses and data representations were performed using R Studio (version 4.4.3).

Results

Sleeve gastrectomy reduces adiposity and improves metabolic status

Anthropometric and metabolic outcomes were assessed at baseline and 4 months post‑SG, as summarized in Table 1. As expected, SG resulted in significant decreases in anthropometric and adiposity measures. Fasting plasma glucose and triglyceride levels showed significant improvements 4 months post-surgery compared to baseline, while total cholesterol and HDL‑C levels remained unchanged. Sex-related differences were observed in longitudinal changes in body weight, body fat percentage and neck circumference. These differences may be attributed to women’s lower pre-surgery body weight combined with the greater post-surgery weight reduction in men. Taken together, these findings confirm that SG leads to significant metabolic improvements within 4 months after the procedure.

Table 1.

Characteristics of study participants. Values are expressed as mean ± SD. Significant differences between pre- and post-surgery measurements were obtained using paired Student’s t-tests. Interactions between parameters and sex were evaluated using mixed linear models

	Pre-surgery N = 33 (W = 27, M = 6)	Post-surgery N = 33 (W = 27, M = 6)	P value
			Visit	Sex	Visit: Sex
Body weight (kg)	120.8 ± 14.7	98.2 ± 12.8	< 0.001	< 0.001	< 0.001
BMI (kg/m2)	43.1 ± 3.1	35.1 ± 3.3	< 0.001	NS	< 0.05
Body fat (%)	47.9 ± 5.0	44.6 ± 5.4	< 0.001	< 0.001	NS
Waist circumference (cm)	129.5 ± 11.3	110.8 ± 12.5	< 0.001	NS	NS
Neck circumference (cm)	42.0 ± 4.7	38.1 ± 3.3	< 0.001	< 0.001	NS
Triglycerides (mmol/L)	2.0 ± 0.8	1.5 ± 0.5	< 0.01	NS	NS
Total cholesterol (mmol/L)	6.2 ± 1.3	5.8 ± 1.0	NS	NS	NS
HDL-C (mmol/L)	1.6 ± 0.5	1.6 ± 0.6	NS	NS	NS
Fasting glucose (mmol/L)	6.7 ± 0.9	5.8 ± 0.8	< 0.001	NS	NS
Fasting insulin (pmol/L)	103.4 ± 38.8	59.8 ± 19.0	< 0.001	NS	NS

W women, M men, BMI body mass index, HDL-C high-density lipoprotein cholesterol

Sleeve gastrectomy modulates circulating levels of eCBome mediators

We next examined whether SG alters circulating eCBome mediator levels by calculating post-SG log2-fold changes in NAEs and 2‑MAGs before and after surgery (Fig. 1A). Four months after SG, circulating levels of EPEA and DHEA were significantly reduced, while levels of 2-AG and 2-DPG were significantly increased (Fig. 1A and B). These changes remained significant after accounting for baseline mediator levels and baseline BMI, sex, and age. Surgery also significantly influenced circulating levels of 2-PG, 2-LG and 2-DHG but only when models were adjusted for baseline mediator levels and sex (Fig. 1A and B). Interestingly, only a few changes in adiposity measures correlated with changes in circulating levels of NAEs and 2-MAGs between the pre-surgery and the 4-month post-surgery visits (Fig. 1C). Among NAEs, only the longitudinal decrease in PEA levels were significantly correlated with reductions in waist circumference. Regarding biochemical measures, longitudinal changes in 2-DPG were significantly correlated with improvements in HDL-C levels. However, the eCBome response following SG was significantly associated with reductions in TG levels, especially for all 2-MAG mediators as well as PEA and DPEA (Fig. 1C).

Fig. 1.

A Log2-fold changes in circulating levels of NAEs and 2-MAGs post-surgery relative to pre-surgery. ^a Unadjusted and adjusted P < 0.05 (baseline mediator levels, BMI, sex, and age), ^b Adjusted P < 0.05 only (baseline mediator levels, BMI, sex, and age), Error bars are expressed as mean ± SEM. B Boxplots showing changes in circulating levels of NAEs (EPEA and DHEA) and 2-MAGs (2-AG, 2-LG, 2-DPG and 2-DHG) before and after surgery. C Heatmap of Pearson’s correlation coefficients (r) between longitudinal changes in anthropometric and biochemical parameters and changes in circulating eCBome mediators. The intensity of the square colour represents the magnitude of Pearson’s correlation coefficient (r). * P < 0.05

We then explored the intra- and inter-family (i.e., NAEs and 2-MAGs) correlations of circulating levels of eCBome mediators before and 4 months after SG (Fig. 2). Prior to surgery, strong positive correlations were observed within NAEs, and similar associations were noted within 2-MAGs (Fig. 2A). Intra-family correlations within NAEs and 2-MAGs were slightly stronger after surgery (Fig. 2A). Trends for inverse correlations emerged between some NAEs and 2-MAGs after surgery compared to the baseline (Fig. 2A). The boxplot of Pearson correlation coefficients by eCBome family highlights the strengthening and convergence within the 2-MAG and the NAE mediator families at 4-month post-surgery compared to baseline (Fig. 2B). These observations further support the impact of bariatric surgery on the overall profile of circulating eCBome mediators, as it also involves correlations with mediators like DPEA and 2-OG that did not directly exhibit longitudinal changes in their levels following the surgery.

Fig. 2.

A Heatmaps of Pearson’s correlation coefficients (r) between circulating NAEs and 2-MAGs pre- (left) and 4 months post-surgery (right). The intensity of the square colour represents the magnitude of Pearson’s correlation coefficient (r). B Boxplots of absolute values of Pearson’s correlation coefficients between NAEs and 2-MAGs, at pre- and post-surgery timepoints. * < 0.05

Circulating eCBome profile correlates with adiposity measures after sleeve gastrectomy

Given the concomitant changes observed in anthropometric, biochemical and eCBome measures following surgery, we performed correlation analyses both before and after SG to better capture their underlying dynamics. Figure 3 shows the results of Spearman correlation analyses between adiposity measures or metabolic parameters and circulating levels of eCBome mediators, both before and 4 months after surgery. Unlike most previous reports in individuals with less severe obesity phenotype [8], total and abdominal adiposity measures were not correlated with eCBome mediators before surgery. Only 2-AG and 2-LG showed significant positive associations with neck circumference and waist circumference, respectively. However, levels of TG and non-HDL cholesterol were positively correlated with circulating levels of several 2-MAGs, whereas HDL-C was negatively associated only with circulating levels of 2-PG. Following SG, positive correlations emerged between total adiposity measures (BMI and body fat mass) and a cluster of NAEs, including PEA, OEA, LEA, and EPEA. Moreover, 4 months after surgery, body fat distribution measures were positively associated with several 2-MAG mediators, such as 2-OG, 2-LG and 2-DHG. Waist circumference is a well-established marker of visceral adiposity, whereas neck circumference has been proposed as a surrogate marker of central and upper-body fat distribution, and has been linked to overall adiposity. Together, these anthropometric measures provide complementary indicators of ectopic fat distribution [24]. Similar to pre-surgery findings, most 2-MAGs remained correlated with higher TG levels and lower HDL-C levels. Therefore, these results suggest that the circulating eCBome profile becomes more closely associated with adiposity measures and metabolic parameters following weight loss induced by bariatric surgery. This finding aligns with a previous study involving individuals ranging from healthy weight to moderate obesity [8].

Fig. 3.

Heatmap of Spearman’s correlation coefficients (r) between anthropometric measures, biochemical parameters and circulating eCBome mediators before surgery (top) and 4 months post-surgery (bottom). The intensity of the colour of the square represents the magnitude of Spearman’s correlation coefficient (r). * P < 0.05

Dietary fatty acids correlate with eCBome mediators after sleeve gastrectomy

Given the strong impact of SG on dietary intake and prior evidence linking dietary habits to eCBome mediators, we investigated whether surgery‑induced dietary changes could contribute to post‑surgical alterations in circulating eCBome mediators. As expected, overall energy intake was reduced following SG (Fig. 4A). While the relative contribution of carbohydrates decreased after surgery, the relative contribution of proteins increased, whereas the relative contribution of lipids remained unchanged (Fig. 4A). Absolute dietary intakes of saturated fatty acids (SFA), MUFA and PUFA decreased significantly 4 months after surgery compared to baseline (Fig. 4B). However, no significant differences were observed between pre- and post-surgery time points for the relative intake of these fatty acid classes to total energy intake (Fig. 4C). Deeper insight into post‑surgical dietary patterns also revealed postoperative reductions in the consumption of vegetables and fruits, refined and whole grains, animal protein foods, and sugary foods and beverages (Fig. 4D).

Fig. 4.

Boxplots of dietary intake in terms of A energy intake (kcal) and proportion of macronutrient intake,B absolute fatty acid intake (in grams),C fatty acid intake relative to total lipid intake, and D relative intake classified by food groups, before and 4 months post-surgery. * P < 0.05

We next examined whether intraindividual changes in dietary intake were correlated with longitudinal changes in the circulating eCBome profile (Fig. 5). Reductions in total energy intake were positively associated with most NAEs, except for PEA and DHEA, which showed only a trend toward significant association. No associations were observed between energy intake and circulating 2‑MAGs. Interestingly, changes in specific circulating 2‑MAGs, including 2‑PG and 2‑LG, were significantly and inversely associated with changes in the relative intake of SFA. Conversely, increases in the relative intake of PUFA were positively associated with several 2‑MAG mediators. Aside from these fatty acid-specific associations, no significant longitudinal correlations were detected between changes in macronutrient distribution or food‑group intake and circulating eCBome mediators. Despite marked surgery‑induced dietary changes, their relationships with changes in circulating eCBome mediators were limited in scope and suggest that dietary intake is not the primary driver of the circulating eCBome response to SG.

Fig. 5.

Heatmaps of Spearman’s correlation coefficients (r) between longitudinal changes in proportion of macronutrient intake, fatty acid intake relative to total lipid intake and relative intake classified by food groups and changes in circulating eCBome mediators. The intensity of the square colour represents the magnitude of Spearman’s correlation coefficient (r). * P < 0.05

Cross‑sectional correlation analyses revealed no significant associations between macronutrient intake (data not shown) or total fatty acid classes and circulating eCBome mediators at either the pre‑ or post‑surgery time point (Supplementary Fig. 1). Similarly, only limited associations were observed between food‑group consumption and circulating eCBome mediators at both time points. Notably, dairy intake at 4 months post‑surgery was positively correlated with circulating AEA and OEA but negatively correlated with several 2‑MAG species (Supplementary Fig. 2). Associations between dietary precursor fatty acids and circulating eCBome mediators were not observed before surgery. In contrast, several significant relationships emerged postoperatively, particularly for PUFA‑derived eCBome mediators (Fig. 6). Dietary arachidonic acid intake was positively correlated with several NAEs and 2‑MAGs, whereas omega‑3 PUFA intake was positively associated with circulating levels of DHEA and 2‑DHG. Collectively, these findings suggest that following bariatric surgery–induced weight loss, dietary PUFA intake becomes more closely linked to the circulating eCBome mediator profile. This pattern aligns with previous reports describing associations between fatty acid intakes and circulating eCBome mediators in individuals without obesity or with moderate obesity [25].

Fig. 6.

Heatmaps of Pearson’s correlation coefficients (r) between dietary fatty acid intake and circulating eCBome mediators before (top) and 4 months post-surgery (bottom). The intensity of the square colour represents the magnitude of Pearson’s correlation coefficient (r). * P < 0.05

Discussion

Previous studies have linked circulating levels of NAEs and 2-MAGs to food intake, adiposity and metabolic status. Several of these eCBome mediators are known to regulate energy balance and metabolic homoeostasis through various signalling pathways. Weight loss interventions, whether via lifestyle intervention or bariatric surgery, have the potential to significantly modulate the eCBome profile, which may represent a pivotal mechanism underlying the metabolic improvement observed in individuals. The study aimed to characterize the longitudinal changes in circulating eCBome mediator profile in individuals living with severe obesity undergoing SG. Circulating levels of eCBome mediators were assessed before and 4 months after surgery and analysed for concomitant changes in adiposity, metabolic outcomes and dietary intake. Briefly, weight loss and metabolic improvements following SG were accompanied by significant changes in the circulating eCBome mediator profile. Interestingly, circulating levels of eCBome mediators were more tightly intercorrelated and showed stronger, more physiologically expected associations with adiposity measurements and dietary fatty acid intake four months after the surgery than prior to the surgery.

Obesity and its metabolic complications have been associated with elevated circulating levels of the eCBs, AEA and 2-AG, as compared to individuals without obesity [26, 27]. We and others also reported higher levels of AEA and 2-AG congeners (i.e., NAEs and 2-MAGs) in the context of obesity [8, 28]. The impact of bariatric surgery on eCBs and eCB-like molecules has been examined in several studies. Azar and colleagues reported significant reductions in serum 2-AG and AEA levels 12 months following SG [29]. These findings suggest that decreases in these mediators may occur later in the postoperative trajectory, as we did not observe such reductions after 4 months. Similar to Azar et al., we detected positive correlations between changes in TG and changes in AEA levels, but we additionally identified correlations with several other eCBome mediators [29], highlighting a broader interplay between the circulating eCBome and the lipid profile. A shorter-term study also reported reductions in circulating AEA and PEA 6 months after SG or biliopancreatic diversion, with these decreases correlating with weight loss and reductions in waist circumference [30]. Consistent with these results, we found a significant correlation between changes in PEA levels and reductions in waist circumference at 4 months. Evidence from animal models of Roux‑en‑Y gastric bypass further supports these findings, showing decreased hepatic 2‑AG levels following surgery [31]. The liver being one of the most important regulators of lipid storage and release [32], this suggests that rapid weight loss induced by bariatric surgery may influence eCBome regulation in key metabolic organs. However, results on 2-AG levels after bariatric surgery remain incomplete or inconclusive across studies [29, 30, 33], and data on eCB congeners are even more limited. It is also important to nuance that none of the aforementioned studies incorporated dietary assessments into their experimental designs, unlike the present study, which allows us to better understand eCBome changes after metabolic and nutritional changes.

It has been reported that circulating levels of most NAEs appear primarily associated with total fat mass, while circulating levels of 2-MAGs are more strongly related to visceral fat accumulation [8]. In this cohort of individuals living with severe obesity undergoing SG, correlations with abdominal adiposity, likely reflecting fat mass, emerged 4 months after surgery but were not observed prior to surgery. It is interesting to note that the emergence of correlations between adiposity measures and circulating eCBome mediators is not exclusively attributable to changes in the eCBome profile along with weight loss. Indeed, only longitudinal fluctuations in PEA levels were significantly associated with changes in waist circumference. Regarding metabolic profile measures, prior to surgery, 2-MAGs were positively correlated with non-HDL-C and TG levels. After surgery, circulating levels of 2-MAGs were still positively associated with non-HDL-C and TG and negatively associated with HDL-C. The lack of associations between circulating eCBome mediators and adiposity measures in individuals with severe obesity suggests that very high adiposity or adipose dysfunction may impact the physiological regulation of circulating eCBome mediators. The mechanisms underlying this observation remain to be elucidated.

Adipose tissue acts as a key metabolic and endocrine organ, influencing multiple metabolic processes in the body [34, 35]. Excessive fat accumulation, especially in visceral adipose tissue, involves adipocyte hypertrophy, which promotes adipocyte dysfunction and contributes to obesity-related metabolic complications [36]. Adipose tissue possesses the enzymatic machinery needed to produce and degrade eCBome mediators from lipids [37]. It also expresses numerous eCBome receptors, including CB₁, which activation enhances lipid biosynthesis and fat storage [38]. The synthesis of most AEA and 2-AG congeners has been highlighted in primary adipocyte cultures and human adipose tissue biopsies [39–41]. Previous studies have shown that bariatric surgery not only reduces fat accumulation in adipose tissue, but also significantly reduces adipocyte size and improves its function [36, 42]. Reduced circulating levels of NAEs after bariatric surgery may be linked to improvements in adipocyte function, possibly through enhanced lipid metabolism and storage. Moreover, fatty acid mobilization by adipocytes for eCBome mediator synthesis tends to increase with the degree of unsaturation and decrease with chain length, with PEA being one of the most abundantly produced NAE in human adipocytes [39]. Our study found a significant correlation between weight loss and decreased PEA levels, highlighting a potential contribution of the adipose tissue in restoring PEA levels. Further analyses of circulating eCBome mediators over longer periods after SG will be necessary to clarify these dynamics.

Considering the importance of dietary lipids in shaping both the circulating fatty acid and eCBome mediator profile, it is reasonable to infer that changes in circulating eCBome levels may arise from dietary changes following bariatric surgery [43]. According to Canadian nutritional guidelines, the post-surgery diet is initially low in fat, moderate in carbohydrates, and high in protein [44]. Although solid foods are generally reintroduced within 4 to 6 weeks after SG, participants in the present study resumed solid food intake at 6 weeks. Thus, by the 4 months postoperative assessment, they had been consumed their usual diet for several weeks [45]. Apart from the drastically reduced total energy intake after SG, the most notable dietary change was a higher relative protein intake compared to baseline. Surprisingly, dietary intake was not a significant determinant of the circulating eCBome profile, especially before surgery in individuals with severe obesity. Only relative fatty acid dietary intake 4 months after surgery was significantly associated with circulating eCBome mediators. Indeed, dietary PUFA intake correlated with circulating levels of PUFA-derived eCBome mediators only after weight loss, suggesting that severe adiposity may conceal the previously reported association between dietary fatty acids and the circulating eCBome profile [43].

To achieve successful and long-term weight loss with minimal weight regain, bariatric surgery patients should adhere to healthy lifestyle habits, such as maintaining a balanced diet and practising regular physical activity [46, 47]. Previous studies have associated suboptimal weight loss or higher weight regain with lower protein intake, lower physical activity, disordered eating behaviours, older age and smoking habits [48, 49]. Healthy eating and exercise also impact circulating eCBome. MUFA and PUFA-rich diets, such as the Mediterranean diet, can increase circulating eCBome mediator levels by providing their fatty acid precursors [50]. Physical activity in bariatric surgery patients further increases long‑term HDL‑C levels, complementing the beneficial effects of surgery and contributing to an overall improvement in the lipid profile [51]. Physical activity can also acutely elevate circulating eCBome mediators [52], which may contribute to exercise-induced reward, mood enhancement and overall well-being [53, 54]. Interestingly, eCBome mediators produced in the brain during exercise can influence circulating eCBome levels [55, 56]. Maintaining a healthy lifestyle is therefore critical for achieving meaningful and sustained weight loss, but also for modulating circulating eCBome activity, partly through reward pathways linked to diet and physical activity. Longer postoperative follow‑up will help elucidate how these lifestyle factors influence the eCBome over time.

Our findings support the impact of SG on the circulating eCBome profile, which is paralleled with significant weight loss and marked improvements in metabolic health. The longitudinal design with complete data available at both pre-surgery and 4-month post-surgery time points represents a strength of this study. Nevertheless, the study design does not allow us to establish whether these effects are specific to bariatric surgery or simply reflect the impact of weight loss per se. Moreover, the imbalance between sexes in the study cohort limits our ability to explore whether changes in the eCBome mediator profile following the surgery differ between women or men, and if covariates (i.e., adiposity, metabolic status and food intake) are differentially associated with circulating eCBome mediators. While the 4-month follow-up is relatively late to assess the immediate impact of surgery on the eCBome, it may also be too short to evaluate long-term associations between weight loss (or even weight regain), metabolic improvements and the circulating eCBome profile. Despite evidence linking eCBome mediators to physical activity, the current study does not permit such analyses, as data on physical activity was not assessed. Nevertheless, weight loss after four months post-surgery was consistent across participants, with no evidence of weight regain, which typically occurs 24–36 months after surgery [57].

Conclusions

Our key findings reveal that severe obesity appears to mask the physiological associations between adiposity, dietary lipid intake and the circulating eCBome mediator profile. Weight loss following SG was accompanied by significant changes in the circulating eCBome profile and by the emergence of associations with adiposity and dietary lipid intake that were absent prior to surgery. Extrapolating from studies without severe obesity, it seems that eCBome signalling is disrupted in severe obesity and may be partially restored following SG, which may either contribute to or result from both the metabolic disturbances seen in severe obesity and the beneficial effects of bariatric surgery.

Abbreviations

eCB

Endocannabinoid

eCBome

Endocannabinoidome

2-AG

2-arachidonoyl-glycerol

AEA

Anandamide

CB₁

Cannabinoid receptor 1

CB₂

Cannabinoid receptor 2

NAEs

N-acylethanolamines

2-MAGs

2-monoacyl-glycerols

PPAR

Peroxisome proliferator-activated receptor

SG

Sleeve gastrectomy

BMI

Body mass index

R24W

Self-administered 24h dietary recall

SFA

Saturated fatty acid

MUFA

Monounsaturated fatty acid

PUFA

Polyunsaturated fatty acid

HEFI

Healthy Eating Food Index-2019

HDL-C

High-density lipoproteins cholesterol

TG

Triglycerides

LC-MS/MS

Liquid chromatography coupled to tandem mass spectrometry

PEA

N-palmitoyl-ethanolamine

OEA

N-oleoyl-ethanolamine

LEA

N-linoleoyl-ethanolamine

DPEA

N-docosapentaenoyl-ethanolamine

EPEA

N-eicosapentaenoyl-ethanolamine

DHEA

N-docosahexaenoyl-ethanolamine

2-PG

2-palmitoyl-glycerol

2-OG

2-oleoyl-glycerol

2-LG

2-linoleoyl-glycerol

2-EPG

2-eicosapentaenoyl-glycerol

2-DPG

2‑docosapentaenoyl-glycerol

2-DHG

2-docosahexaenoyl-glycerol

Authors' contributions

GSA contributed to literature review and data visualization, performed statistical analyses, and drafted the manuscript; JD managed patient recruitment, and contributed to data collection and curation; MP assisted with data curation, participated in the study investigation, and provided materials and resources for analyses; CG contributed to data curation and supplied methodological resources; LB was involved in the study’s conceptualization and provided surgical expertise; LBB acted as additional surgical expertise; ED processed the lipidomic data; NF supervised and validated lipidomic laboratory methods; VD secured funding and provided expertise on eCBome mediator role in obesity; AM managed the project administration, conceived the study and acquired funding; AV contributed to project conceptualization, participated in data visualization and assisted in writing the original draft. All authors read and approved the final manuscript.

Funding

This study was supported by the Foundation of Institut universitaire de cardiologie et de pneumologie de Québec (A.M. 2019–2023), the Cardiometabolic Health, Diabetes and Obesity Research Network, the Centre Nutrition, santé et société (NUTRISS) affiliated to Université Laval, the Canada Research Excellence Chair on the Microbiome-Endocannabinoidome Axis in Metabolic Health (CERC-MEND) (V.D., 2017–2024). AM is the recipient of a Research Scholars - Junior 1 award from the Fonds de recherche du Québec - Santé. JD received scholarships from the Canadian Institute of Health Research (CIHR, FBD-181468) and the Fonds de recherche du Québec - Santé (BF4-326155, 10.69777/326155). GSA received scholarships from Centre NUTRISS – Nutrition, santé et société (NUTRISS), the Chair in Nutrition, and the Fonds de recherche du Québec – Santé (BF2-360605, 10.69777/360605).

Data availability

The datasets generated and/or analysed during the current study are not publicly available due to restrictions imposed by the Research Ethics Committee of the Research Center of IUCPQ-UL. Data are, however, available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The trial was approved by the Research Ethics Committee of the Research Centre of IUCPQ-UL (2021-3456). Each participant provided written informed consent.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.