Endocannabinoid cerebrospinal fluid levels in migraine and its relation to symptoms of depression

Abstract

Background

Disruption of the endocannabinoid system has been implicated in both migraine and depression, which are frequently comorbid in patients. Nevertheless, supporting evidence for this hypothesis is limited. Hence, we investigated whether the endocannabinoid system is disrupted in interictal migraine patients and whether confounding factors were of influence.

Methods

To this end, we measured the levels of endocannabinoids N-arachidonoylethanolamine (AEA) and 2-arachidonoylglycerol (2-AG), and the endocannabinoid analogue docosahexaenoylethanolamine (DHEA) in the cerebrospinal fluid (CSF), obtained from 194 individuals with episodic migraine between attacks (97 migraine with aura and 97 migraine without aura individuals) and from 94 healthy volunteers, using micro-liquid chromatography – tandem mass spectrometry (micro-LC-MS/MS). Groups were age- and sex-matched. Timing and processing of sampling was protocolized. Multivariate linear regression analyses for log₁₀-transformed endocannabinoid concentrations were performed to compare groups. Covariates that were included were sex, age, BMI, substance use and depressive symptoms.

Results

No differences were detected between individuals with migraine with or without aura and healthy controls in the CSF for AEA (p = 0.46, p = 0.08), 2-AG (p = 0.46, p = 0.90) and DHEA (p = 0.58, p = 0.79), respectively. For 2-AG, levels were, however, positively correlated with depressive symptoms (β = 0.053; 95% CI: 0.014 to 0.093, p = 0.008).

Conclusion

CSF endocannabinoids (AEA, 2-AG, and DHEA) may not serve as effective biomarkers for distinguishing interictal episodic migraine patients from controls. Instead, CSF levels of 2-AG are linked to depressive symptoms, shedding further light on the relevance of endocannabinoids in patients with migraine and depression. This study highlights the intricate nature of the endocannabinoid system concerning episodic migraine and confounding factors.

Graphical abstract

This is a visual representation of the abstract.

Keywords

migraine depression cerebrospinal fluid endocannabinoids metabolomics

Introduction

Migraine is a highly prevalent, disabling, primary brain disorder characterized by attacks of severe headache and associated symptoms.¹ Depression and anxiety are common comorbid disorders.^2,3 Research into neurochemistry in individuals with migraine between attacks may help elucidate its pathophysiology and provide useful treatment targets. Given the repeated nature of migraine attacks there likely is a difference between individuals with migraine and healthy controls that is due to the disease itself or is a consequence of experiencing multiple attacks. A possible contributor that makes patients vulnerable to migraine attacks could be changes in the endocannabinoid system, which has a strong influence on neurotransmission, the neuroendocrine and neuroimmune systems,^4–6 which are all involved in migraine pathophysiology.^7,8 In addition, it has been shown that the endocannabinoid system is involved in pain perception, depending on the mode of action, activation of the endocannabinoid system can both lead to the increase and decrease of pain either directly or indirectly via anti-inflammatory properties.⁹ Previous studies have suggested that endocannabinoids consisting of N-acylethanolamines (NAEs), act in the trigeminovascular system, a key component in migraine pathophysiology.^10,11 One study of mRNA expression of cannabinoid receptors in rat trigeminal ganglia has revealed that 30% of the fibers contain mRNA of this receptor, however, expression was mainly in Aβ-fibers and there was little colocalization with putative nociceptor markers, illustrating that it is unclear how endocannabinoids are involved in peripherally mediated antinociceptive effects.¹² Thus, it remains poorly understood what the exact role of endocannabinoids in migraine is. Still, cannabinoids are currently considered as treatment for migraine, thus it is highly relevant to determine the role of endocannabinoids in migraine.¹³

The endocannabinoid system consists of two major cannabinoid receptors (CB₁ and CB₂) and their lipid signaling ligands; the endocannabinoids.¹⁴ Endocannabinoid signaling is a key regulator of synaptic communication through retrograde inhibition of the release of inhibitory and excitatory neurotransmitters.¹⁵ CB₁ receptors are expressed throughout the peripheral and central nervous system, whereas CB₂ receptors are mainly located on immune cells. The two most widely studied endocannabinoids are N-arachidonoylethanolamine, also called anandamide (AEA) and 2-arachidonoylglycerol (2-AG). AEA is a partial agonist for both cannabinoid receptors, with a high affinity for CB₁ and a low affinity for CB₂.¹⁴ Additionally, AEA is a full agonist at the transient receptor potential vanilloid 1 (TRPV1), where it modulates the release of calcitonin gene-related peptide (CGRP).^16,17 CGRP is a key neurotransmitter of the trigeminovascular system and a known contributor in causing migraine attacks. 2-AG is a full agonist at CB₁ and CB₂ receptors and an important endocannabinoid ligand in nervous tissue.^14,18 Docosahexaenoylethanolamine (DHEA) is a CB₂ receptor partial agonist,¹⁹ however, its exact role in the endocannabinoid system has not been as extensively studied as that of AEA and 2-AG.²⁰

In a relatively small study (n = 50), AEA was suggested to be lower and palmitoylethanolamide (PEA) to be higher in cerebrospinal fluid (CSF) of chronic migraine patients.²¹ Fatty acid amide hydrolase (FAAH), which degrades AEA, and endocannabinoid membrane transporter (EMT) activity in platelets was found increased in women but not in men with migraine suggesting an altered endocannabinoid system in female patients.²² Gene expression and peripheral protein levels of endocannabinoid system components (those involved in biosynthesis of NEAs, e.g., NAPE, MAGL and DAGL) were elevated in the blood of people with migraine.²³ Additionally, evidence has emerged for a role of the endocannabinoid system in depression.²⁴ Depression has a bidirectional relationship with migraine.^2,25 Women with depression had reduced endocannabinoid levels in blood serum,²⁶ and chronic treatment with antidepressants seems to alter the endocannabinoid system.²⁷

As brain endocannabinoid concentrations do not directly correlate with peripheral endocannabinoid levels,⁹ measuring endocannabinoids in CSF, the biofluid closest to the brain, likely is the best choice of body fluid to assess what activity has occurred in the brain. Hence, we compared CSF levels of endocannabinoids in a large cohort of people with migraine with or without aura and healthy controls matched for sex and age, whilst taking depressive symptomology into account. Our goal was to investigate whether levels of endocannabinoids are different between migraine with or without aura compared to controls. The primary outcome was to detect a difference in AEA, 2-AG and/or DHEA levels using micro-liquid chromatography – tandem mass spectrometry (micro-LC-MS/MS) between those with or without aura compared to controls, whilst controlling for possible confounders. As secondary outcome the effects of confounding factors, such as depressive symptoms, on endocannabinoid levels were investigated. Thus, we sought to detect differences in the endocannabinoid system that hopefully aid in the development of better treatment for migraine.

Methods

Study design and participants

The collection of the study population has been described in Onderwater et al.²⁸ In brief, we enrolled n = 198 individuals with migraine (n = 100 migraine with aura and n = 98 migraine without aura) and n = 96 healthy controls and group-matched them for age (using 5-year strata) and sex. Migraine was diagnosed according to the International Classification of Headache Disorders (ICHD3) by experienced physicians (supervised by G.M.T. and M.D.F.).¹ For the first few years, we instead used the ICHD2 version without essential differences between patients, at least for this study of episodic cases, so we feel that using different versions of ICHD2 did not influence the outcome of the study. The majority of participants were recruited from the LUMINA (Leiden University Medical Center Migraine Neuro Analysis) database, which encompasses both migraine patients and healthy volunteers from the Dutch population, all of whom volunteered to contribute to research focused on migraine. Some patients and healthy volunteers were recruited through a public advertisement or through the Leiden University Medical Center Headache Clinic. All participants were between 18 and 65 years (41 ± 14) and had a body mass index (BMI) between 18 and 28 (24 ± 3). Individuals suffering from migraine did not use acute migraine drugs on more than 8 days per month. None of the patients suffered from chronic migraine or medication overuse headache. Also, they did not suffer from frequent episodic tension-type headache. Healthy volunteers had no signs or symptoms of a disease and had no history of headache (except for infrequent episodic tension-type headaches) or other pain-related syndromes. They also did not have first-degree relatives with migraine or trigeminal autonomic cephalalgia. Other exclusion criteria were (1) severe psychiatric disorder (such as major depressive disorder), (2) history of oncological disease, or (3) contra-indication for lumbar puncture (LP) (signs and symptoms of increased intracranial pressure, local skin infection, or a coagulopathy including use of anti-coagulant drugs or platelet-inhibitors).

Participants were asked to fill out a questionnaire on the day of the LP that included questions related to their migraine characteristics and possible history of depression. To compute symptoms of depression (either current of past depressive symptoms), additional questionnaires were preferentially completed on the day of LP but otherwise extracted from electronic medical records. The presence of symptoms of depression was determined based on previously described criteria.²⁹ In short, symptoms of depression were determined using validated cut-off scores for the (CES-D). Depressive symptoms was defined as HADS-D ≥ 8 or CES-D ≥ 16 or (past) depression diagnosed by a physician or (past) use of antidepressants for depression.

Cerebrospinal fluid collection

Sampling and handling of CSF samples was described in Onderwater et al.²⁸ In brief, samples were collected between 2008 and 2016, the time of the LP was between 8:30 a.m. and 1:00 p.m. and in random order to minimize diurnal and seasonal variation. Participants were required to refrain from eating or drinking, apart from water and to abstain from smoking, use of hash and other recreational drugs, for at least 8 h prior to sampling. Sampling occurred interictally with migraine individuals being attack-free for a minimum of 3 days. Participants were urged to refrain from the use of prophylactic or any other chronic medication (other than oral contraceptives) in the four weeks preceding the LP. The LP and subsequent CSF collection were performed in a standardized matter to limit inter-individual variation.

For endocannabinoid measurements, 3 mL CSF was sampled directly in a 15-mL polypropylene Falcon tube (Cat. No. 188271; Greiner Bio-One B.V.) that already contained 6 mL of ice-cold ethanol (to instantly stop any metabolic activity). The tube was immediately put back on ice after CSF collection. After gently shaking the tube, the CSF was divided in aliquots of 1.5 mL in 1.8-mL cryotubes (Art. No. 368632; NUNC Brand, Sigma-Aldrich, Zwijndrecht, The Netherlands). The cryotubes were transferred to a −80°C freezer as soon as possible. All CSF samples remained at −80°C until sample preparation for analysis.

Analysis of endocannabinoids using micro-LC-MS/MS

The analysis of endocannabinoids in CSF was performed using a validated method.³⁰ Sample preparation, except the evaporation step, was carried out under ice-cold conditions. Analytes were extracted using liquid-liquid extraction. In brief, 750 µL of CSF samples (CSF:ethanol = 1:2, v/v) was evaporated to approximately 100 µL using a SpeedVac (Thermo Fisher Scientific, Waltham, MA, USA). Then, 10 µL of isotopically labelled internal standard solution and 50 µL of a buffer solution (0.1 M ammonium acetate solution, adjusted to pH 4 with formic acid) were added to each sample. One milliliter of methyl-tert-butylether (MTBE) was then added prior to agitation for 10 min using a bullet blender and subsequent centrifugation for 5 min at 16,000 g. The organic supernatant (900 µL) was collected and evaporated to dryness using the SpeedVac. The dry residues were then reconstituted in 20 µL of an ice-cold solution of water and acetonitrile (ACN) 50:50 (v/v), prior to agitation (20 min) and centrifugation at 16,000 g for 50 min. Finally, 15 µL of the supernatant was transferred into a glass vial and 3 µL was injected into a nanoAcquity micro-LC system (Waters Corporation, Mildford, MA, USA) hyphenated with a Shimadzu 8060 triple quadrupole mass spectrometer (Shimadzu Corporation, Kyoto, Japan). Separation was performed using a micro C18 column (0.3 × 150 mm, 2.6 μm) from Phenomenex (Torrance, CA, USA) maintained at 45°C. The mobile phase was composed of 2 mM ammonium formate and 10 mM formic acid in water (A), and acetonitrile (B). The flow rate was 4 μL/min. Ionization of the compounds was performed using a customized micro-electrospray ionization (micro-ESI) source in positive mode. Selected Reaction Mode (SRM) was used for MS/MS acquisition. SRM transitions were individually optimized for targeted analytes and respective isotopically labelled internal standards. Data acquisition and pre-processing was performed using LabSolutions LCMS Version 5.97 SP1 (Shimadzu Corporation). Data quality was monitored using regular injection of quality control (QC) samples prepared from pooled CSF. Samples were injected in a randomized order. QC samples were used to correct for inter-batch variations using the in-house developed mzQuality workflow (available at http://www.mzQuality.nl). Relative standard deviations (RSDs) of peak area ratios were calculated for each targeted analyte detected in the QC samples. The absolute concentration of the endocannabinoids in CSF samples was calculated using calibration samples. Quality control was set on RSDqc < 15% and blank effect < 15%. The data as relative response ratios (target area/ISTD area; unit free) of these metabolites after QC correction were calculated and converted to concentrations.

Statistical analysis

Principal component analysis (PCA) was performed on the concentration measurements of all measured endocannabinoids together. Based on the first two principal components outliers were identified by visually inspecting the plots. Log₁₀-transformed endocannabinoid concentrations (pM) were used for further statistical analyses. Multivariate linear regression was used to detect metabolite differences between individuals with migraine and healthy controls for each endocannabinoid separately. Based on literature, BMI, sex, age, weekly alcohol consumption, cigarettes per day, drug use and depressive symptoms were identified as potential covariates and therefore included as independent variables. Prior to data analysis, missing data for relevant covariates was imputed with Multiple Imputation by Chained Equations (MICE).³¹ For MICE, we used 10 imputations and 30 iterations as settings and included the following variables: Log₁₀ endocannabinoid concentrations, age, sex, BMI, migraine without aura/migraine with aura/control-stratus, depression score, drug use and the amount of alcohol and smoking. Regression analyses were performed on the imputed data and results were pooled. All data analyses were performed using R version 4.3.1. and RStudio version 2023.06.0 + 421 (RStudio: Integrated Development for R. RStudio, PBC).³² P-values of ≤ 0.05 were considered statistically significant.

Results

Clinical characteristics

Of all measured endocannabinoids the levels of AEA, 2-AG, and DHEA in CSF of 195 individuals with episodic migraine (n = 98 migraine with aura (MA); n = 97 migraine without aura (MO)) and 94 healthy controls were reliably measured; for 5 individuals there was not enough CSF sample available (< 1.5 mL). Next, one population outlier with migraine with aura was excluded based on the PCA plot, leaving 194 people with migraine and 94 controls for the statistical analysis. The clinical characteristics are shown in Table 1. Sex distribution (two third female) of this selected group was similar between cases and controls. There were no differences between the migraine groups, except for a higher monthly attack frequency in migraine without aura (2.0 IQR 1.0–3.5) versus migraine with aura (1.5 IQR 1.0–3.0) (p = 0.017). There was a difference between the three groups in, total HADS related score (p < 0.001), total CES-D score (p < 0.001) and weekly alcohol consumption (p = 0.009) and headache days (p < 0.001) (Table 1). Additional analyses showed a difference between those with migraine and controls, as those with migraine had more headache days (p < 2.2 × 10⁻¹⁶) a higher total HADS score (p = 9.147 × 10⁻⁹) and CES-D score (p = 7.217×10⁻⁷), but consumed less alcohol in units of alcohol per week (p = 0.006). This is in line with prior results on migraine comorbidities and lifestyle characteristics, as those with migraine more often suffer from headache, symptoms of depression and are more inclined to refrain from alcohol.^2,3,33 Data was assumed to be missing at random for weekly alcohol consumption (n = 6 (3 MA and 3 MO)), migraine days (cases n = 7 (5 MA and 2 MO), depressive symptoms (n = 45 (cases n = 13 (8 MA and 5 MO), controls n = 32)), HADS scores (n = 9 (1 MA and 1 MO), controls n = 7)), total CES-D (n = 48 (cases = 16 (9 MA and 7 MO), controls = 32)) and drug use (n = 1 (cases n = 1 (1 MA), controls n = 0)). The missing values used for the regression analysis were imputed. No group differences were observed for fasting time (in hours) and time of CSF collection (data not shown). In addition, no time effect of collection time or storage time was observed (data not shown).

Table 1.

Characteristics of control and migraine patient cohort

Participants characteristics	MA (n = 97)	IQR	MO (n = 97)	IQR	Control (n = 94)	IQR	P-value maMoCo
General characteristics
Sex, female (n, %)	63 (64.9)		59 (60.8)		54 (57.4)		0.567^a
BMI (median)	23.9	22.0–25.8	23.8	21.9–25.4	23.5	21.5–25.3	0.433^b
Age (median)	40	29–52	42	32–53	39	26–51	0.262^c
Alcohol use (n, %)	78 (80.4)		67 (69.1)		77 (81.9)		0.068^a
Units alcohol per week^# (median)	4.0	1.8–7.0	2.8	1.0–5.1	5.0	1.4–8.0	0.009^c
Smoking (n, %)	13 (13.4)		13 (13.4)		20 (21.3)		0.231^a
Cigarettes per day^# (median)	4.0	1.5–15.0	6.5	2.5–10.0	6.0	2.8–10.5	0.693^c
Migraine characteristics
Migraine days (days/month) (median)	1.5	1.0–3.0	2.0	1.0–3.5	NA	NA	0.017^d
Headache days (days/month) (median)	4	2.4–6.0	5	3.0–8.0	0	0.0–0.5	<0.001^c
Depression/Anxiety characteristics
Depressive symptoms (n, %)	32 (33.0)		26 (26.8)		17 (18.1)		0.424^a
Total HADS score (median)	8.0	5.0–13.0	8.0	4.0–11.3	4.0	2.0–7.0	<0.001^c
CES-D score (median)	10.5	5.0–17.3	8.0	4.0–12.0	3.5	1.0–7.8	<0.001^c
Drugs related characteristics
Drug use, (n, %)	4 (4.1)		4 (4.1)		9 (9.8)		0.188^a

For continuous variables the median value is depicted. IQR = inter quartile range; MA = migraine with aura; MO = migraine without aura; BMI = body mass index; HADS = Hospital Anxiety and Depression Scale; CES-D = Center for Epidemiologic Studies Depression Scale. Values are expressed as absolute values and percentage or median and IQR. P-values ≤ 0.05 are marked in bold. ^a Chi-squared Test, ^b One-way Anova, ^c Kruskal-Wallis test, ^d Mann-Whitney U test, ^#Only users (alcohol/smokers/drugs) are used for number per day/week/month. Missing values in: units alcohol per week (cases n = 6 (3 MA and 3 MO), controls n = 0), migraine days (cases n = 7 (5 MA and 2 MO), depressive symptoms (cases n = 13 (8 MA and 5 MO), controls n = 32), HADS scores (cases n = 2 (1 MA and 1 MO), controls n = 7) CES-D score (cases = 16 (9 MA and 7 MO), controls = 32), and Drug use (cases n = 1 (1 MA), controls n = 0).

CSF levels of AEA, 2-AG and DHEA in migraine patients and controls

Multivariate linear regression analyses showed that there were no differences in concentrations in migraine with aura or migraine without aura versus control for each of the three endocannabinoids (Figure 1A-C) p = 0.46 and p = 0.08 for AEA, p = 0.46 and p = 0.90 for 2-AG and p = 0.58 and p = 0.79 for DHEA, respectively (Supplementary Table 1). Some of the covariates, however, did have a significant effect on endocannabinoid concentration in the regression model. AEA concentration was associated with sex (β = -0.081; 95% CI: −0.12 to −0.04, p < 0.001) and number of cigarettes per day (β = 0.007; 95% CI: 0.00 to 0.01, p = 0.004). Female sex was associated with a lower AEA concentration (Figure 2A) and the number of cigarettes per day was positively correlated with the AEA concentration (Figure 2B). 2-AG concentration was also negatively correlated with female sex (β = -0.084; 95% CI: −0.12 to −0.05, p < 0.001) (Figure 3A) and positively correlated with depressive symptoms (β = 0.054; 95% CI: 0.014 to 0.09, p = 0.008) (Figure 3B and 3C). DHEA concentration was positively correlated with age (β = 0.003; 95% CI: 0.00 to 0.00, p < 0.001), BMI (β = 0.010; 95% CI: 0.00 to 0.02, p = 0.013), smoking per day (β = 0.006; 95% CI: 0.00 to 0.01, p = 0.03), alcohol units per week (β = 0.004; 95% CI: 0.00 to 0.01, p = 0.033), and also negatively correlated with female sex (β = -0.046; 95% CI: −0.08 to −0.00, p = 0.034), and drug use (β = -0.102; 95% CI: −0.20 to −0.00, p = 0.042) (Figure 4A-F). These results were comparable to not imputed data (data not shown).

Figure 1.

Endocannabinoid concentration per subgroup. A) Violin plot of log₁₀ transformed AEA concentrations in each subgroup. B) Violin plot of log₁₀ transformed 2-AG concentrations in each subgroup. C) Violin plot of log₁₀ transformed DHEA concentrations in each subgroup. Each dot represents an individual, the median is marked by a red diamond. The horizontal lines represent the 0.25, 0.5 and 0.75 quantiles. AEA = N-arachidonoylethanolamine, 2-AG = -arachidonoylglycerol, DHEA = Docosahexaenoylethanolamine. Plots are based on pre-imputed data and uncorrected

Figure 2.

Relation of AEA concentration with covariates. A) Violin plot of log₁₀ transformed AEA concentrations and sex. Each dot represents an individual, the median is marked by a red diamond. The horizontal lines represent the 0.25, 0.5 and 0.75 quantiles. B) Correlation plot of log₁₀ transformed AEA concentration and cigarettes per day. The black line indicates the linear regression line and grey area the 95% confidence interval. Cases are depicted in red triangles and controls in blue circles. r = Pearson's correlation coefficient. AEA = N-arachidonoylethanolamine. Plots are based on pre-imputed data and uncorrected

Figure 3.

Relation of 2-AG concentration with covariates. A) Violin plot of log₁₀ transformed 2-AG concentrations and sex. B) Violin plot of log₁₀ transformed 2-AG concentrations and depressive symptoms. C) Violin plot of log₁₀ transformed 2-AG concentrations and depressive symptoms subdivided by diagnosis with and without depressive symptoms (control = M-, case = M+, positive for depressive symptoms = D+ and negative for depressive symptoms = D-). Each dot represents an individual, the median is marked by a red diamond. The horizontal lines represent the 0.25, 0.5 and 0.75 quantiles. 2-AG = 2-arachidonoylglycerol. Plots are based on pre-imputed data and uncorrected

Figure 4.

Relation of DHEA concentration with covariates. A) Correlation plot of log₁₀ transformed DHEA concentration and age. B) Correlation plot of log₁₀ transformed DHEA concentration and BMI. C) Correlation plot of log₁₀ transformed DHEA concentration and cigarettes per day. The black line indicates the linear regression line and grey area the 95% confidence interval. Cases are depicted in red triangles and controls in blue circles. r = Pearson's correlation coefficient. D) Correlation plot of log₁₀ transformed DHEA concentration and units alcohol per week. E) Violin plot of log₁₀ transformed DHEA concentrations and sex. F) Violin plot of log₁₀ transformed DHEA concentrations and drug use. Each dot represents an individual, the median is marked by a red diamond. The horizontal lines represent the 0.25, 0.5 and 0.75 quantiles. DHEA = Docosahexaenoylethanolamine, BMI = Body mass index. Plots are based on pre-imputed data and uncorrected

Discussion

Here we measured and compared concentrations of the endocannabinoids AEA, 2-AG, and DHEA in CSF of 194 individuals with migraine with (n = 97) or without aura (n = 97) outside migraine attacks (interictal state) compared with 94 healthy matched controls to determine whether the endocannabinoid system is disrupted in interictal episodic migraine patients and whether depressive symptoms were of influence. The CSF sampling and processing was strictly protocoled, and the analysis was performed with our validated micro-LC-MS/MS method.³⁰ As for our primary outcome we found no differences in endocannabinoids related to migraine status, so we regard this a negative study. For our secondary outcome, however, covariates age, sex, daily number of cigarettes, units alcohol per week, drug use, depressive symptoms and BMI, all had an effect on the different endocannabinoid concentrations.

The effect that the covariates have on the endocannabinoids concentrations in CSF emphasizes the complexity of the endocannabinoid system regarding confounding factors. Interestingly, 2-AG levels were positively correlated with depressive symptoms, which has implications for further research as comorbidity with depression in migraine patients makes them also more prone for chronification and medication overuse.³⁴

So far, only two studies have assessed the levels of endocannabinoids in body fluids of individuals with migraine, but the results are contradictory.^21,35 In one study, concentrations of AEA were lower and those of the endocannabinoid analogue PEA were higher in the CSF of 15 individuals with chronic migraine versus 20 age-matched controls.²¹ However, in that study, the analysis of endocannabinoids was based on a different extraction method (i.e., the Bligh and Dyer liquid-liquid extraction method), which may generate AEA during the analytical step (see Figure S1, supplemental material). Most importantly, there was no correction for confounding factors, as was done in our study. Given that chronic migraine often coincides with depression one should wonder whether the difference found by Sarchielli et al.²¹ is actually due to depression or other confounding factors rather than migraine status. Addmittedly, the current study encompasses a less severly affected migraine cohort that could in part also be explanatory for the difference in results. In the other study, 16 different endocannabinoids were measured by LC-MS/MS in blood plasma from 38 females with episodic migraine between attacks and 26 healthy controls collected on days 1 and 14 of the menstrual cycle.³⁵ In line with our study, there were no differences for AEA and N-acylethanolamines. In addition, in a subset of study participants endocannabinoids were measured in plasma collected during attacks and compared with the interictal measurement; again, no differences were found.³⁵ In both studies, the number of participants was limited, the body fluids in which the measurements were done were different, the sampling conditions were not protocoled, and no correction for possible confounding factors was done. Because no prior CSF endocannabinoid study in migraine with low-frequency attacks existed, we estimated 50 patients per group from comparable literature³⁶ and our preliminary data, but enrolled nearly twice as many to minimize the chance of missing clinically meaningful differences.

In addition to the well-established relation between depression and endocannabinoids, endocannabinoids have also been associated with BMI, sex, age, weekly alcohol consumption, smoking and drug use.^9,37–39 Energy homeostasis is one of the processes best known to be controlled by the endocannabinoid system.^9,40,41 Although the pathways underlying sex differences in the endocannabinoid system are not fully elucidated, it is generally accepted that differences exist between the male and female endocannabinoid system.^42,43 Female rats had different concentrations of AEA and 2-AG compared to male rats in different brain structures.⁴⁴ The endocannabinoid system is involved in the reward circuits of the brain, so an effect of drug use on the endocannabinoid system is not surprising.^39,45 Additionally, studies in mice have found that chronic exposure to ethanol leads to increased levels of AEA and 2-AG.⁴⁶ There is some evidence on a decrease in the expression of CB₁ molecules and an increase in the AEA level in smokers.^47,48 Given that these factors could influence both the exposure and the outcome, they were considered when studying endocannabinoid levels in the present study.

Contrary to most previous presented evidence on the influence of endocannabinoids in migraine pathophysiology,^10,21,49,50 in our CSF study, no association was detected between the levels of certain endocannabinoids and (episodic) migraine, at least not in the interictal phase. A dysregulation of the endocannabinoid system might have remained undetected in case it involves only a localized (e.g., near the synapse) or short-lasting migraine-attack related disturbance of the endocannabinoid system, and not a steady-state adaptation of the body to experienced migraine attacks. Future studies should focus on endocannabinoid levels during an attack to discern whether there is a difference in concentration that can be directly related to a migraine attack itself. Further limitations of our study are that due to the size of the study and the great care taken to match our study groups, including for age, sex and diurnal and seasonal time of sampling, that the samples were collected over a long period of time. This could have potentially affected the stability of the compounds studied, even though they were stored at the recommended −80°C.⁵¹ We acknowledge that the collection of samples was done over a fairly long time, which is hard to avoid for a difficult to obtain body fluid as CSF and the samples had been stored for variable time in the freezer (in all cases for at least 6 years) until measurement in 2022, as the method to analyze endocannabinoids first needed to be developed. However, the actual pre-analytical phase was protocoled and running of the analyses was done in consecutive batches in only a few days, so we feel that the variation in collection and storage did not have a major influence on the outcome, also because we took along quality control and blank samples for each run. Furthermore, only minor changes of oxylipins were observed after prolonged storage at −80°C of blood samples and the metabolome and lipidome of CSF were found to be stable for longer periods.^51,52 Additionally, we have focused on patients with episodic migraine to investigate whether there is a difference with controls outside of an attack. Strong exclusion criteria and guidelines were used in this study, such as medication use prior to the study; hence most participants are on the milder end of the episodic migraine spectrum. This study population was chosen to investigate the relation with endocannabinoids as “purely” as possible. However, this might limit the generalizability of our results.

Our study also has several strengths. First, we studied and compared CSF from uniquely large groups of people with migraine and controls. Moreover, unlike other CSF studies, control samples were collected from people who can be considered healthy and had not undergone a lumbar punction because of neurological symptoms and almost all had no first-degree relatives with migraine.⁵³

Second, sample preparation for assays and the demographic characteristics of patients can contribute to variability of metabolite concentrations. Thus, when investigating a putative role of endocannabinoids in disease, it is crucial to use standardized protocols as the collection and processing of endocannabinoid samples can affect the concentration.^51,54

Furthermore, it is imperative to correct the statistical model for possible confounders, in our case most importantly, depression. We want to stress that multivariate linear regression analysis was used in our study to model the relationship between endocannabinoids and explanatory variables simultaneously. Although people with major depression were excluded in our study, we found a positive correlation between 2-AG and depressive symptoms, as screened with the validated tools (HADS-D and CESD) for depressive symptomatology. Previous research on the role of AEA and 2-AG in depression has been conducted primarily in individuals with major depressive disorder (MDD).^26,55 The hypothalamic–pituitary–adrenal (HPA)–axis, neurogenesis and neuroinflammation have all been implicated in the etiology of MDD,⁵⁶ and all are dependent on proper functioning of the endocannabinoid system,²⁴ suggesting a crucial role for the endocannabinoid system in MDD. Given the increased risk of depression in migraine, it is important to realize that depressive symptoms may have a significant effect on endocannabinoid levels, even when MDD has been ruled out. The association between migraine and endocannabinoids may therefore depend on depressive symptomatology rather than on migraine itself. Our study shows that depressive symptoms functions as confounder on the relationship between migraine and endocannabinoid levels. Again, it is important to note that our study was not designed to specifically investigate the role of depression, but given our results and the current literature,²⁶ it is likely that depression has a confounding effect on the relationship between migraine and endocannabinoids. One should realize that in studies that are similar to ours, finding an association between migraine and endocannabinoid levels, or not, might possibly be due to whether or not depressive symptoms were taken along as confounding factor in the statistical model. Afterall, given that migraine and depression are comorbid conditions, it is very likely that in such studies a considerable number of migraine cases may have suffered from depressive symptoms. Again, future studies should therefore take depressive symptomology into account when studying the role of the endocannabinoid system in migraine. Contrary to many studies to minimize possible confounding effects, we have added drug-use of participants as a variable to our statistical model. Thus, we believe the effect of drug-use on our outcome to be negligible. However, a complete abolishment of any confounding effect can never be claimed.

In conclusion, CSF endocannabinoids may not serve as effective biomarkers for distinguishing interictal migraine individuals from controls. However, 2-AG levels are likely linked to depressive symptoms, but also sex, BMI, smoking and alcohol use were of influence on AEA, 2-AG, and DHEA levels. This study shows the relevance of endocannabinoids in patients with migraine and highlights the intricate and complex role of the endocannabinoid system in relation to both migraine and confounding factors, such as depressive symptoms.

Highlights

The study reveals a significant link between CSF levels of the endocannabinoid 2-AG and depressive symptoms, offering new insights into the neurochemical basis of comorbid migraine and depression.

These findings underscore the broader importance of the endocannabinoid system in understanding the complex interplay between episodic migraine and depression.

By highlighting the role of endocannabinoids beyond diagnostic biomarkers, our research opens new avenues for exploring therapeutic targets in migraine-related mood disorders.

Supplemental Material

sj-xlsx-1-cep-10.1177_03331024261451250 - Supplemental material for Endocannabinoid cerebrospinal fluid levels in migraine and its relation to symptoms of depression

Supplemental material, sj-xlsx-1-cep-10.1177_03331024261451250 for Endocannabinoid cerebrospinal fluid levels in migraine and its relation to symptoms of depression by Aster V.E. Harder, Xinyu Di, Wendy M. Winter, Bingshu He, Robin M. Van Dongen, Gerrit L.J. Onderwater, Erik W. Van Zwet, Elke H.J. Krekels, Isabelle Kohler, Michel D. Ferrari, Amy C. Harms, Arn M.J.M. Van Den Maagdenberg, Thomas Hankemeier and Gisela M. Terwindt in Cephalalgia

Supplemental Material

sj-png-2-cep-10.1177_03331024261451250 - Supplemental material for Endocannabinoid cerebrospinal fluid levels in migraine and its relation to symptoms of depression

Supplemental material, sj-png-2-cep-10.1177_03331024261451250 for Endocannabinoid cerebrospinal fluid levels in migraine and its relation to symptoms of depression by Aster V.E. Harder, Xinyu Di, Wendy M. Winter, Bingshu He, Robin M. Van Dongen, Gerrit L.J. Onderwater, Erik W. Van Zwet, Elke H.J. Krekels, Isabelle Kohler, Michel D. Ferrari, Amy C. Harms, Arn M.J.M. Van Den Maagdenberg, Thomas Hankemeier and Gisela M. Terwindt in Cephalalgia

Footnotes

ORCID iDs

Aster V.E. Harder

Robin M. VanDongen

Michel D. Ferrari

Arn M.J.M. Van Den Maagdenberg

Gisela M. Terwindt

Consent to participate

All participants provided written informed consent prior to participation in the study.

Consent for publication

All authors have agreed to publish this work in Cephalalgia.

Ethical considerations

The study was conducted according to the criteria of the Declaration of Helsinki and was approved by the Leiden University Medical Center institutional ethics committee.

Author contributions

A.V.E. Harder: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Visualization; Writing - original draft. X. Di: Conceptualization; Data curation; Formal analysis; Methodology; Resources; Validation; Writing - review & editing. W.M. Winter: Formal analysis; Investigation; Methodology; Project administration; Visualization; Writing - review & editing. B. He: Data curation; Methodology; Resources; Validation; Writing - review & editing. R.M. van Dongen: Conceptualization; Investigation; Project administration; Resources; Writing - review & editing. G.L.J. Onderwater: Conceptualization; Investigation; Project administration; Resources; Writing - review & editing. E.W. van Zwet: Investigation; Methodology; Supervision; Writing - review & editing. E.H.J. Krekels: Investigation; Resources; Writing - review & editing. I. Kohler: Investigation; Resources; Writing - review & editing. M.D. Ferrari: Conceptualization; Funding acquisition; Supervision; Writing - review & editing. A.C. Harms: Investigation; Supervision, Writing - review & editing. A.M.J.M. van den Maagdenberg: Conceptualization; Funding acquisition; Supervision; Writing - review & editing. T. Hankemeier: Conceptualization; Funding acquisition; Supervision; Writing - review & editing. G.M. Terwindt: Conceptualization; Funding acquisition; Supervision; Writing - review & editing.

Funding

This study was supported by the Dutch Research Council (NWO) (VICI 918.56.601 and Spinoza 2009 to M.D. Ferrari, Vidi 917.11.31 to G.M. Terwindt), the Netherlands Organization for Health Research and Development (ZonMw) (Clinical Fellowship 90700217 to G.M. Terwindt), and European Community (EC) FP7-EUROHEADPAIN – no. 602633 (to M.D. Ferrari and A.M.J.M. van den Maagdenberg).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

Data not published within the article is available from the corresponding author upon reasonable request.

Open practices

Not applicable.

Supplemental material

Supplemental material for this article is available online.

References

Headache Classification Committee of the International Headache Society (IHS). The International Classification of Headache Disorders, 3rd edition. Cephalalgia 2018; 38: 1–211.

Louter

Pijpers

Wardenaar

, et al. Symptom dimensions of affective disorders in migraine patients. J Psychosom Res 2015; 79: 458–463.

Breslau

Lipton

Stewart

, et al. Comorbidity of migraine and depression: investigating potential etiology and prognosis. Neurology 2003; 60: 1308–1312.

Hillard

Beatka

Sarvaideo

. Endocannabinoid signaling and the hypothalamic-pituitary-adrenal axis. Compr Physiol 2016; 7: 1–15.

Pandey

Mousawy

Nagarkatti

, et al. Endocannabinoids and immune regulation. Pharmacol Res 2009; 60: 85–92.

Szabo

Schlicker

. Effects of cannabinoids on neurotransmission. Handb Exp Pharmacol 2005; 168: 327–365. DOI: 10.1007/3-540-26573-2_11

Goadsby

Holland

Martins-Oliveira

, et al. Pathophysiology of migraine: a disorder of sensory processing. Physiol Rev 2017; 97: 553–622.

Bruno

Carpino

, et al. An overview on immune system and migraine. Eur Rev Med Pharmacol Sci 2007; 11: 245–248.

Hillard

. Circulating endocannabinoids: from whence do they Come and where are they going? Neuropsychopharmacology 2018; 43: 155–172.

10.

Akerman

Kaube

Goadsby

. Anandamide is able to inhibit trigeminal neurons using an in vivo model of trigeminovascular-mediated nociception. J Pharmacol Exp Ther 2004; 309: 56–63.

11.

Akerman

Holland

Goadsby

. Cannabinoid (CB1) receptor activation inhibits trigeminovascular neurons. J Pharmacol Exp Ther 2007; 320: 64–71.

12.

Price

Helesic

Parghi

, et al. The neuronal distribution of cannabinoid receptor type 1 in the trigeminal ganglion of the rat. Neuroscience 2003; 120: 155–162.

13.

Kuruvilla

. Cannabinoids in headache: helpful or harmful? Curr Opin Neurol 2025; 38: 277–280.

14.

Zou

Kumar

. Cannabinoid receptors and the endocannabinoid system: signaling and function in the central nervous system. Int J Mol Sci 2018; 19: 33.

15.

Ohno-Shosaku

Maejima

Kano

. Endogenous cannabinoids mediate retrograde signals from depolarized postsynaptic neurons to presynaptic terminals. Neuron 2001; 29: 729–738.

16.

Engel

Izydorczyk

Mueller-Tribbensee

, et al. Inhibitory CB1 and activating/desensitizing TRPV1-mediated cannabinoid actions on CGRP release in rodent skin. Neuropeptides 2011; 45: 229–237.

17.

Ross

. Anandamide and vanilloid TRPV1 receptors. Br J Pharmacol 2003; 140: 790–801.

18.

Sugiura

Kondo

Kishimoto

, et al. Evidence that 2-arachidonoylglycerol but not N-palmitoylethanolamine or anandamide is the physiological ligand for the cannabinoid CB2 receptor. Comparison of the agonistic activities of various cannabinoid receptor ligands in HL-60 cells. J Biol Chem 2000; 275: 605–612.

19.

Brown

Cascio

Wahle

, et al. Cannabinoid receptor-dependent and -independent anti-proliferative effects of omega-3 ethanolamides in androgen receptor-positive and -negative prostate cancer cell lines. Carcinogenesis 2010; 31: 1584–1591.

20.

Mock

Gagestein

van der Stelt

. Anandamide and other N-acylethanolamines: a class of signaling lipids with therapeutic opportunities. Prog Lipid Res 2023; 89: 101194.

21.

Sarchielli

Pini

Coppola

, et al. Endocannabinoids in chronic migraine: CSF findings suggest a system failure. Neuropsychopharmacology 2007; 32: 1384–1390.

22.

Cupini

Bari

Battista

, et al. Biochemical changes in endocannabinoid system are expressed in platelets of female but not male migraineurs. Cephalalgia 2006; 26: 277–281.

23.

Greco

Demartini

Zanaboni

, et al. Peripheral changes of endocannabinoid system components in episodic and chronic migraine patients: a pilot study. Cephalalgia 2021; 41: 185–196.

24.

Gallego-Landin

García-Baos

Castro-Zavala

, et al. Reviewing the role of the endocannabinoid system in the pathophysiology of depression. Front Pharmacol 2021; 12: 762738.

25.

Peres

MFP

Mercante

JPP

Tobo

, et al. Anxiety and depression symptoms and migraine: a symptom-based approach research. J Headache Pain 2017; 18: 37.

26.

Hill

Miller

, et al. Serum endocannabinoid content is altered in females with depressive disorders: a preliminary report. Pharmacopsychiatry 2008; 41: 48–53.

27.

Hill

Hillard

, et al. Differential effects of the antidepressants tranylcypromine and fluoxetine on limbic cannabinoid receptor binding and endocannabinoid contents. J Neural Transm (Vienna) 2008; 115: 1673–1679.

28.

Onderwater

GLJ

van Dongen

Harms

, et al. Cerebrospinal fluid and plasma amine profiles in interictal migraine. Ann Neurol 2023; 93: 715–728.

29.

Louter

Pelzer

de Boer

, et al. Prevalence of lifetime depression in a large hemiplegic migraine cohort. Neurology 2016; 87: 2370–2374.

30.

Guled

, et al. Quantification of endocannabinoids in human cerebrospinal fluid using a novel micro-flow liquid chromatography-mass spectrometry method. Anal Chim Acta 2022; 1210: 339888.

31.

van Buuren

Groothuis-Oudshoorn

. mice: multivariate imputation by chained equations in R. J Stat Soft 2011; 45: 1–67, https://www.jstatsoft.org/index.php/jss/article/view/v045i03 .

32.

RStudio Team. RStudio: integrated development for R. Boston, MA: RStudio, Inc., 2015.

33.

Onderwater

GLJ

van Oosterhout

WPJ

Schoonman

, et al. Alcoholic beverages as trigger factor and the effect on alcohol consumption behavior in patients with migraine. Eur J Neurol 2019; 26: 588–595.

34.

Ashina

Terwindt

Steiner

, et al. Medication overuse headache. Nat Rev Dis Primers 2023; 9: 5.

35.

Gouveia-Figueira

Goldin

Hashemian

, et al. Plasma levels of the endocannabinoid anandamide, related N-acylethanolamines and linoleic acid-derived oxylipins in patients with migraine. Prostaglandins Leukot Essent Fatty Acids 2017; 120: 15–24.

36.

Wallin

Blennow

Andreasen

, et al. CSF Biomarkers for Alzheimer's Disease: levels of beta-amyloid, tau, phosphorylated tau relate to clinical symptoms and survival. Dement Geriatr Cogn Disord 2006; 21: 131–138.

37.

Engeli

Böhnke

Feldpausch

, et al. Activation of the peripheral endocannabinoid system in human obesity. Diabetes 2005; 54: 2838–2843.

38.

Piyanova

Lomazzo

Bindila

, et al. Age-related changes in the endocannabinoid system in the mouse hippocampus. Mech Ageing Dev 2015; 150: 55–64.

39.

Spanagel

. Cannabinoids and the endocannabinoid system in reward processing and addiction: from mechanisms to interventions. Dialogues Clin Neurosci 2020; 22: 241–250.

40.

Piazza

Cota

Marsicano

. The CB1 receptor as the cornerstone of exostasis. Neuron 2017; 93: 1252–1274.

41.

Blüher

Engeli

Klöting

, et al. Dysregulation of the peripheral and adipose tissue endocannabinoid system in human abdominal obesity. Diabetes 2006; 55: 3053–3060.

42.

Wagner

. Sex differences in cannabinoid-regulated biology: a focus on energy homeostasis. Front Neuroendocrinol 2016; 40: 101–109.

43.

Craft

Marusich

Wiley

. Sex differences in cannabinoid pharmacology: a reflection of differences in the endocannabinoid system? Life Sci 2013; 92: 476–481.

44.

Levine

Liktor-Busa

Lipinski

, et al. Sex differences in the expression of the endocannabinoid system within V1M cortex and PAG of Sprague Dawley rats. Biol Sex Differ 2021; 12: 60.

45.

Bystrowska

Frankowska

Smaga

, et al. Effects of cocaine self-administration and its extinction on the rat brain cannabinoid CB1 and CB2 receptors. Neurotox Res 2018; 34: 547–558.

46.

Vinod

Yalamanchili

Xie

, et al. Effect of chronic ethanol exposure and its withdrawal on the endocannabinoid system. Neurochem Int 2006; 49: 619–625.

47.

Scherma

Fadda

Le Foll

, et al. The endocannabinoid system: a new molecular target for the treatment of tobacco addiction. CNS Neurol Disord Drug Targets 2008; 7: 468–481.

48.

Hirvonen

Zanotti-Fregonara

Gorelick

, et al. Decreased cannabinoid CB1 receptors in male tobacco smokers examined with positron emission tomography. Biol Psychiatry 2018; 84: 715–721.

49.

Greco

Gasperi

Sandrini

, et al. Alterations of the endocannabinoid system in an animal model of migraine: evaluation in cerebral areas of rat. Cephalalgia 2010; 30: 296–302.

50.

Greco

Mangione

Sandrini

, et al. Effects of anandamide in migraine: data from an animal model. J Headache Pain 2011; 12: 177–183.

51.

Lehmann

. From bedside to bench-practical considerations to avoid pre-analytical pitfalls and assess sample quality for high-resolution metabolomics and lipidomics analyses of body fluids. Anal Bioanal Chem 2021; 413: 5567–5585.

52.

Jonasdottir

Brouwers

Toes

REM

, et al. Effects of anticoagulants and storage conditions on clinical oxylipid levels in human plasma. Biochim Biophys Acta Molecul Cell Biol Lipids 2018; 1863: 1511–1522.

53.

van Dongen

Zielman

Noga

, et al. Migraine biomarkers in cerebrospinal fluid: a systematic review and meta-analysis. Cephalalgia 2017; 37: 49–63.

54.

Zoerner

Gutzki

F-M

Batkai

, et al. Quantification of endocannabinoids in biological systems by chromatography and mass spectrometry: a comprehensive review from an analytical and biological perspective. Biochim Biophys Acta 2011 Nov; 1811: 706–723.

55.

Kranaster

Hoyer

Aksay

, et al. Electroconvulsive therapy enhances endocannabinoids in the cerebrospinal fluid of patients with major depression: a preliminary prospective study. Eur Arch Psychiatry Clin Neurosci 2017; 267: 781–786.

56.

Jesulola

Micalos

Baguley

. Understanding the pathophysiology of depression: from monoamines to the neurogenesis hypothesis model - are we there yet? Behav Brain Res 2018; 341: 79–90.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.01 MB

0.06 MB