Endocannabinoid System in Psychotic and Mood Disorders, a review of human studies

Abstract

Despite widespread evidence of endocannabinoid system involvement in the pathophysiology of psychiatric disorders, our understanding remains rudimentary. Here we review studies of the endocannabinoid system in humans with psychotic and mood disorders. Postmortem, peripheral, cerebrospinal fluid and in vivo imaging studies provide evidence for the role of the endocannabinoid system in both psychotic and mood disorders. Psychotic disorders and major depressive disorder exhibit alterations of brain cannabinoid CB1 receptors and peripheral blood endocannabinoids. Further, these changes may be sensitive to treatment status, disease state, and symptom severity. Evidence from psychotic disorder extend to endocannabinoid metabolizing enzymes in the brain and periphery, whereas these lines of evidence remain poorly developed in mood disorders. A lack of studies examining this system in bipolar disorder represents a notable gap in the literature. Despite a growing body of productive work in this field of research, there is a clear need for investigation beyond the CB1 receptor in order to more fully elucidate the role of the endocannabinoid system in psychotic and mood disorders.

1. Introduction

Disturbances of the endocannabinoid system can underlie various psychiatric disorders including psychotic and mood disorders. The objective of this review is to summarize human evidence for the role of endocannabinoids in patients with psychotic and mood disorders. Endocannabinoids could represent novel targets for diagnosing and treating psychiatric disorders.

The endocannabinoid system is composed of receptors, their endogenous lipid ligands (endocannabinoids) and their respective biosynthetic and catabolic enzymes. The two most well-known endocannabinoids are arachidonic acid derivates – N-arachidonoylethanolamide, (anandamide), and 2-arachidonoyglycerol (2-AG; Mackie, 2005; Lu and Mackie, 2016). Anandamide and the related N-acyl ethanolamines palmitoylethanolamide (PEA) and oleoylethanolamide (OEA) are synthesized by N-acyl-phosphatidylethanolamine phospholipase D (NAPE-PLD) and catabolized primarily by fatty acid amide hydrolase (FAAH; Hillard, 2015). PEA and OEA do not bind to cannabinoid receptors but do bind to non-cannabinoid receptors that are secondary targets of anandamide (Cristino et al., 2020). 2-AG is synthesized by diacylglycerol lipase (DAGL) and catabolized primarily by monoacylglycerol lipase (MAGL), and to a lesser extent by α/β-hydrolase domain-containing 6 (ABHD6; Cao et al., 2019; Ueda et al., 2013; Figure 1).

Figure 1:

Endocannabinoid system. Anandamide, N-palmitoylethanolamide (PEA) and N-oleoylethanolamide (OEA) are synthesized from N-acylphosphatidylethanolamine. 2-AG are synthesized from phospholipids by their enzymes diacylglycerol lipases α and β. Anandamide and 2-AG are synthesized on-demand. Activation of presynaptic CB1 receptors by anandamide and 2-AG leads to a decrease in neurotransmitter release by the presynaptic neuron. 2-AG is catabolized to glycerol and arachidonic acid primarily by MAGL in the presynaptic neuron, but also to a lesser extent by ABHD6 in the postsynaptic neuron (Cao et al., 2019). Anandamide is catabolized by FAAH to arachidonic acid and ethanolamine in the postsynaptic neuron (Ahn et al., 2008; Katona and Freund, 2012).

AEA: anandamide; 2-AG: 2-arachidonoylglycerol; NAPE: N-acylphosphatidylethanolamine; NAPE-PLD: N-acyl phosphatidylethanolamine-specific phospholipase D; DAG: diacylglycerol; DAGL: diacylglycerol lipase; PEA: N-palmitoylethanolamide; OEA: N-oleoylethanolamide; CB1: cannabinoid receptor type 1; MAGL: monoacylglycerol lipase; FAAH: fatty acid amide hydrolase; ABHD6: α/β-hydrolase domain-containing 6; AA: arachidonic acid.

This image was created using Servier Medical Art (https://smart.servier.com) and modified under a Creative Commons Attribution 3.0 Unported License.

Endocannabinoids activate cannabinoid (CB1 and CB2) receptors. CB1 receptors are densely expressed in the brain on glial cells and on neurons at most synapses where their activation leads to inhibition of neurotransmitter release (Pertwee and Ross, 2002; Pertwee et al, 2010). A major function of the endocannabinoid system is to provide negative feedback control of neurotransmission through retrograde neurotranmission: post-synaptic neurons release endocannabinoids onto pre-synaptic CB1 receptors, thereby reducing neurotransmitter release from the pre-synaptic neuron (figure 1; Pertwee, 2006; Katona and Freund, 2008, Kano 2014). CB1 receptors are also found in mitochondria and are expressed by non-neuronal brain cells including oligodendrocytes, microglia, and astrocytes (Mackie, 2005). CB2 receptors are expressed primarily by immune cells and glia however they are also expressed to a lesser degree, in neurons (Pertwee, 2005; Van Sickle et al., 2005).

In this review, we summarize studies of the endocannabinoid system studies in humans with psychosis-spectrum and affective disorders. All articles included are from peer reviewed journals indexed in PubMed or google scholar. We prioritized studies that tested differences between patients and non-psychiatric controls. Studies measured chemical species, mRNA, protein levels or activity of endocannabinoid system components (CB1/CB2, endocannabinoid-metabolizing enzymes, anandamide, OEA, PEA or 2-AG) in peripheral blood, cerebrospinal fluid (CSF), postmortem brain, and in vivo imaging. Although we have aimed to provide a comprehensive and in depth interpretation of the state of the literature, it is not a systematic review and the authors regret any notable omissions. Studies that examined mood and endocannabinoids in non-clinical populations or in response to treatments known to directly affect endocannabinoids (e.g., physical training), though relevant, are beyond the scope of this review. For an overview of endocannabinoid genetics and psychiatric disorders, readers can see (Hillard et al., 2012). Non-cannabinoid receptor targets that interact with endocannabinoid and related ligands are not well-studied in psychotic or affective disorders and are beyond the scope of this review (Cristino et al., 2020).

2. Endocannabinoids in Psychotic disorders

2.1. Peripheral blood endocannabinoids in psychotic disorders

In peripheral blood, elevated anandamide was observed in untreated and treated patients with schizophrenia relative to healthy controls (table 1; De Marchi et al., 2003; Koethe et al., 2019; Potvin et al., 2008; Wang et al., 2018). These changes in anandamide were observed when measured in serum, plasma and in whole blood (De Marchi et al., 2003; Koethe et al., 2019; Potvin et al., 2008; Wang et al., 2018). In contrast to patients with schizophrenia, elevated serum anandamide was not observed in antipsychotic-naïve patients with first-episode psychosis (Giuffrida et al., 2004; Reuter et al., 2017). Three of four studies that measured other N-acylethanolamines observed elevated levels of either PEA or OEA in patients with schizophrenia who were treated or clinically stable (Koethe et al., 2019; Wang et al., 2018), or of both PEA and OEA in patients with co-morbid substance use disorders (Potvin et al., 2008). Finally, plasma 2-AG did not differ significantly between healthy controls and clinically stable patients with schizophrenia, with or without a co-morbid substance use disorders (Koethe et al., 2019; Potvin et al., 2008).

Table 1.

Peripheral blood endocannabinoids in psychotic disorders.

Condition & treatment	n (cases/controls)	Type	AEA	PEA	OEA	2-AG	Reference
Untreated
First-episode psychosis	47/84	serum	n.s.	-	-	-	Giuffrida et al., 2004
First-episode psychosis	28/81	serum	n.s.	-	-	-	Reuter et al., 2017
Schizophrenia	115/88	serum	↑	-	↑	-	Wang et al., 2018
Schizophrenia	12/20	whole blood	↑	-	-	-	De Marchi et al., 2003
Treated
Schizophrenia	115/88	serum	n.s.	-	↑	-	Wang et al., 2018
Schizophrenia	25/16	plasma	↑	↑	n.s.	n.s.	Koethe et al., 2019
Schizophrenia	12/20	whole blood	n.s.	-	-	-	De Marchi et al., 2003
Schizophrenia	25/27	plasma	n.s.	-	n.s.	-	Desfossés et al., 2012
Schizophrenia with substance use disorder	29/17	plasma	↑	↑	↑	n.s.	Potvin et al., 2008
Psychosis Risk
Unaffected twin of Schizophrenia	25/16	plasma	↑	↑	n.s.	n.s.	Koethe et al., 2019
Clinical high risk	33/58	plasma	↑	n.s.	n.s.	↑	Appiah-Kusi et al., 2019
Psychosis risk	27/81	serum	n.s.	-	n.s.	-	Koethe et al., 2009

↑: higher; ↓: lower; n.s.: no significant difference from healthy controls; -: no available data;

AEA: anandamide; PEA: N-palmitoylethanolamide; OEA: N-olcoylethanolamide; 2-AG: 2-arachidonoyglycerol.

In untreated patients with elevated baseline anandamide, antipsychotic treatment and symptom remission was accompanied by normalization of peripheral blood anandamide levels (De Marchi et al., 2003; Wang et al., 2018). Normalization of anandamide following treatment was observed in the largest sample to date (n=115) of untreated patients with schizophrenia (Wang et al., 2018). Although OEA levels also declined significantly following treatment, they remained higher than observed in controls (Wang et al., 2018).

Anandamide was elevated in the blood of individuals at elevated risk for developing psychotic disorders (Appiah-Kusi et al., 2019; Koethe et al., 2009; Koethe et al., 2019). Unaffected twins of individuals with schizophrenia exhibited elevated anandamide and PEA relative to twins with no family history of psychotic disorders (Koethe et al., 2019). In individuals at clinical high risk for developing psychotic disorders, both anandamide and 2-AG were elevated relative to healthy controls (Appiah-Kusi et al., 2019); but see (Koethe et al., 2009; table 1). None of the studies reported showed elevated OEA in psychosis risk populations. Finally, unaffected twins in Koethe et al., (2019) who subsequently developed psychotic disorders had lower baseline plasma anandamide and 2-AG compared to unaffected twins who did not develop psychotic disorders within the follow-up period (Koethe et al., 2019).

Altogether, studies of peripheral blood endocannabinoids in psychotic disorders provide evidence that peripheral endocannabinoids are elevated in psychotic disorders and are sensitive to antipsychotic treatment and symptom remission. Further evidence suggests that peripheral blood endocannabinoids may be related to risk of developing psychosis, although this finding has yet to be replicated in an independent sample (Minichino et al., 2019).

2.2. Peripheral blood mononuclear cell endocannabinoids in psychotic disorders

In untreated and treated patients with schizophrenia CB1 and CB2 receptor mRNA were elevated in peripheral blood cells relative to controls (table 2; D’Addario et al., 2017; Moretti et al., 2018; Chase et al., 2016). Evidence suggests that alterations in peripheral cannabinoid receptor levels in schizophrenia exhibit cell-type specificity. Patients with schizophrenia showed an elevation of CB1 and CB2 receptor protein in neutrophils and monocytes, whereas only CB2 was elevated in natural killer cells. Further, elevations of CB1 were observed in CD4+ but not in CD8+ T-cells (de Campos-Carli et al., 2017). In contrast, B-cells from patients exhibited reduced levels of CB1 and CB2 protein compared to controls (de Campos-Carli et al., 2017). Patients also exhibited changes in the biosynthetic and catabolic enzymes for anandamide and 2-AG. Treated patients with first-episode psychosis exhibited reduced protein levels of NAPE-PLD and DAGL, and increased levels of FAAH and MAGL compared to controls (Bioque et al., 2013). In contrast, treated patients with schizophrenia did not differ from controls in peripheral blood mRNA levels of NAPE-PLD, FAAH, DAGL or MAGL (D’Addario et al., 2017).

Table 2.

Peripheral blood mononuclear cell endocannabinoids in psychotic disorders.

Condition & treatment	n (cases/controls)	Cell type	Receptors		Metabolic enzymes		Greater symptom severity	Reference
			1CB1	1CB2	1Anandamide	12-AG
Mixed (treated & untreated)
First-episode psychosis	95/90	PBMC	-	n.s.	↓NAPE -PLD	↓DAGL	↓FAAH	Bioque et al., 2013
					↑FAAH	↑MAGL
Cannabis+	46/71			↓CB2	↓NAPE-PLD	↓DAGL
Cannabis−	49/71			n.s.	n.s.	↓DAGL
Untreated
Schizophrenia	35/35	PBMC	↑mRNA	↑mRNA	-	-	↑CB1	Chase et al., 2016
							↑CB2
Treated
Schizophrenia	162/94	PBMC	↑mRNA	-	-	-	-	Moretti et al., 2018
Schizophrenia	25/34	PBMC	↑mRNA		n.s.NAPE-PLD	n.s. DAGL	-	D’Addario et al., 2017
				n.s.	n.s. FAAH	n.s. MAGL
				mRNA	mRNA	mRNA
Schizophrenia	55/48	Neutrophils	↑CB1	↑CB2	-	-	-	de Campos-Carli et al., 2017
		Monocytes	↑CB1	↑CB2	-	-
		NK-cells	n.s.	↑CB2	-	-
		B-cells	↓CB1	↓CB2	-	-
		T cells CD8+	n.s.	n.s.	-	-
		T cells CD4+	↑CB1	n.s.	-	-
Schizophrenia	53/22	Monocytes	n.s.	n.s.	-	-		Ferretjans et al., 2014
		NK-cells	n.s.	n.s.	-	-	n.s.
		B-cells	n.s.	n.s.	-	-
		T cells CD8+	n.s.	n.s.	-	-
		T cells CD4+	n.s.	n.s.	-	-

↑: higher; ↓: lower; n.s.: no significant difference from healthy controls; -: no available data;

Cannabis+: patients that met criteria for cannabis abuse or dependence;

Cannabis−: patients that did not meet criteria for cannabis abuse or dependence;

PBMC: Peripheral Blood Mononuclear Cells; 2-AG: 2-arachidonoyglycerol; NK-Cells: Natural Killer cells; NAPE-PLD: N-acyl Phosphatidylethanolamine Phospholipase D; FAAH: Fatty Acid Amide Hydrolase; DAGL: Diacylglycerol Lipase; MAGL: Monoacylglycerol Lipase;

¹Note: result refers to protein levels unless otherwise indicated.

A single study provides direct evidence that peripheral blood endocannabinoid system proteins are sensitive to treatment and symptom remission, by testing pre- and post-treatment outcomes (De Marchi et al., 2003). From an initial 12 patients with untreated schizophrenia, 5 completed treatment, showed symptom remission and provided a second blood sample. Following treatment and symptom remission, mRNA transcripts for FAAH and CB2 declined significantly from pre-treatment levels (De Marchi et al., 2003). In contrast, no significant changes in CB1 mRNA were observed following treatment and symptom remission (De Marchi et al., 2003).

Peripheral blood endocannabinoid system proteins were associated with psychotic symptom severity in one study. Higher CB2 mRNA was associated with greater positive symptom severity (Chase et al., 2016). Similarly, higher CB1 and CB2 mRNA levels were associated with greater negative symptom severity (Chase et al., 2016).

Endocannabinoid receptor and enzymes were also associated with cognitive symptoms and performance. Higher CB1 receptor mRNA was associated with greater severity of cognitive symptoms, and greater CB1 and CB2 mRNA were associated with poorer attention performance (Chase et al., 2016). Higher expression of CB2 receptors on T-lymphocytes was associated with greater cognitive impairment (Ferretjans et al., 2014). Further, cognitive perfomance was associated with endocannabinoid metabolizing enzymes. Lower protein levels of NAPE-PLD, DAGL and FAAH were related to poorer short-term memory, and higher levels of MAGL were associated with greater deficits in attention (Bioque et al., 2016).

Past or current cannabis exposure may influence peripheral endocannabinoid system protein expression (Jacobson et al., 2019). Cannabis-using patients had lower levels of CB2, NAPE-PLD, DAGL whereas only DAGL was lower in non-using patients, relative to healthy controls (Bioque et al., 2013). To our knowledge, it is not known whether similar changes in CB2 and NAPE-PLD expression are observed in healthy cannabis users, or if this is observed only in cannabis-using patients with psychotic disorders.

Taken together, studies in peripheral blood cells from patients with psychotic disorders provide evidence of elevated expression of cannabinoid receptors and endocannabinoid-catabolizing enzymes and reduced levels of biosynthetic enzymes. Further, FAAH and CB2 (but not CB1) receptors were related to antipsychotic treatment status. Current cannabis use status was associated with altered expression of CB2 and NAPE-PLD. Finally, higher expression of CB1 and CB2 receptors was associated with greater severity of positive, negative or cognitive symptoms and with poor cognitive performance. Likewise, lower levels of NAPE-PLD, FAAH and DAGL, or higher levels of MAGL were related to poor cognitive performance. These studies therefore suggest that disturbances to the peripheral endocannabinoid system in psychotic disorders is widespread and may extend to all major components of this system.

2.3. Cerebrospinal fluid endocannabinoids in psychotic disorders

Cerebrospinal fluid (CSF) anandamide concentrations were elevated in patients with first-episode psychosis, chronic schizophrenia and individuals at high risk for developing psychosis (table 3; Giuffrida et al., 2004; Koethe et al., 2009; Koethe et al., 2007; Leweke et al., 2007; Leweke et al., 1999; Minichino et al., 2019; Reuter et al., 2017). Levels of PEA and OEA were reported in only two studies, one of which had an extremely small sample (Giuffrida et al., 2004; Leweke et al., 1999). In contrast to anandamide, PEA was significantly reduced in antipsychotic-naïve patients with first-episode psychosis, but not in treated patients with schizophrenia whereas OEA did not differ between patients and controls in either study (Giuffrida et al., 2004; Leweke et al., 1999). Finally, 2-AG was not reliably observed in CSF at levels above the detection limit in the initial report, therefore 2-AG was not quantified in subsequent studies (Leweke et al., 1999).

Table 3.

Cerebrospinal fluid endocannabinoids in psychotic disorders.

Condition & treatment	n (cases/controls)	AEA	PEA	OEA	Greater symptom severity	Reference
Untreated
First-episode psychosis	47/84	↑	↓	n.s.	↓ AEA	Giuffrida et al., 2004
First-episode psychosis	28/81	↑	-	-	-	Reuter et al., 2017
First-episode psychosis	47/81					Leweke et al., 2007
(cannabis >20 times)	19/26	n.s.			↓ AEA
(cannabis <5 times)	25/55	↑	-	-	↓ AEA
Mixed (treated & untreated)
First-episode psychosis	10/11	↑	↑	n.s.	-	Leweke et al., 1999
Treated
Schizophrenia (atypical AP)	35/84	↑	n.s.	-	-	Giuffrida et al., 2004
Schizophrenia (typical AP)	37/84	n.s.	n.s.	-	-	Giuffrida et al., 2004
Psychosis risk
Psychosis risk	27/81	↑	-	n.s.	↓ AEA	Koethe et al., 2009

↑: higher; ↓: lower; n.s.: no significant difference from healthy controls; -: no available data;

AP: antipsychotics; AEA: anandamide; PEA: N-palmitoylethanolamide; OEA: N-olcoylethanolamide.

In contrast to patients who were antipsychotic-naïve or treated with atypical antipsychotic medications, first-episode psychosis patients treated with typical antipsychotic medications did not exhibit elevated CSF anandamide (Giuffrida et al., 2004).

Two studies with antipsychotic-naïve first-episode schizophrenia patients reported that higher anandamide levels were associated with lower psychotic symptom severity (Giuffrida et al., 2004; Leweke et al., 2007). In general, higher anandamide was associated with lower severity of symptoms across positive, negative and general subscales, although only the negative subscale survived Bonferroni correction.

When divided by history of cannabis exposure, Leweke et al., (2007) observed elevated anandamide only in patients with low exposure (≤5 uses) but not in patients with greater past cannabis exposure (≥20 uses) (Leweke et al., 2007). Lifetime cannabis exposure in controls (20-50 times) and patients (>100 times in 16 of 19 patients) differed considerably in Leweke et al., (2007), thus it is not clear whether differences in anandamide might be observed if compared to a control group matched for cannabis exposure (Leweke et al., 2007).

In antipsychotic-naïve first-episode patients, CSF anandamide was not significantly associated with past cannabis exposure (Reuter et al., 2017), and within a sample of individuals at high risk for developing psychosis, CSF anandamide did not differ by cannabis use history (above or below 20 lifetime exposures) (Koethe et al., 2009). Giuffrida et al. similarly noted that across the whole sample, no overall effect of past cannabis use was observed, although they did not report a statistical test (Giuffrida et al., 2004).

Psychosis related alterations of anandamide may precede psychosis onset, as individuals at high risk for developing psychosis exhibited elevated anandamide, but not OEA, in CSF (Koethe et al., 2009). Remarkably, within the high risk group those with higher CSF anandamide tended to have a lower relative risk of subsequently developing psychotic disorders (Koethe et al., 2009).

Taken together, elevated anandamide levels in CSF appears to be present in acutely ill patients with first-episode psychosis, chronic schizophrenia patients treated with atypical, but not typical antipsychotic medications, and those at clinical high risk for developing psychotic disorders. Although one study reported that patients with greater lifetime cannabis exposure had lower CSF anandamide levels compared to those with negligible exposure, a similar effect was not observed by others. Thus there is a need for a priori study designs and larger sample sizes in order to clarify the impact of cannabis exposure on CSF anandamide in psychosis-spectrum populations. Given the reported associations of CSF anandamide with psychotic symptom severity, alterations in endocannabinoid metabolism and endocannabinoid system function may therefore be associated with disease state.

2.4. Postmortem brain endocannabinoids in psychotic disorders

Postmortem studies consistently report elevated CB1 receptor density in cortex in schizophrenia using autoradiography, while reporting reductions in the same cortical regions using antibody-labelling or in situ hybridization techniques. Postmortem studies investigating endocannabinoids and their metabolic enzymes provide preliminary evidence for disturbed metabolism of anandamide and 2-AG.

Receptor autoradiography studies using CB1 agonists and antagonists report higher CB1 binding density in patients compared to healthy controls in dorsolateral prefrontal cortex (table 4; Brodmann areas 9 and/or 46; Dean et al., 2001; Jenko et al., 2012; Volk et al., 2014; Dalton et al. 2011), anterior cingulate cortex (Zavitsanou et al., 2004), posterior cingulate cortex (Newell et al., 2006) but not in medial temporal lobe (Dean et al., 2001), superior temporal gyrus (Deng et al., 2007) or striatum (Dean et al., 2001). One study reported higher binding in patients with paranoid schizophrenia that was not observed in those with nonparanoid schizophrenia (Dalton et al., 2011). There was no difference in maximum efficacy or potency of CB1 receptor mediated [³⁵S]GTPɣS binding stimulation between patients and controls (Muguruza et al., 2019).

Table 4.

Postmortem studies of the CB1 receptors in psychotic disorders.

Condition & brain region	n (cases/controls)	Autoradiography	1Antibody	mRNA	Reference
Dorsolateral prefrontal cortex
Schizophrenia	20/20	-			Uriguen et al., 2009
(untreated)	5/5		n.s. IB	n.s. qPCR
(treated)	11/11		↓IB	n.s. qPCR
Schizophrenia	37/37		-		Dalton et al., 2011
(paranoid)	16/37	↑[3H]CP55,940		n.s. qPCR
(non-paranoid)	21/37	n.s. [3H]CP55,940		n.s. qPCR
Schizophrenia	14/14	↑[3H]CP55,940	-	-	Dean et al., 2001
Schizophrenia	47/43	↑[3H]MePPEP	-	-	Jenko et al., 2012
Schizophrenia2	12/12	↑[3H]OMAR	↓CB1	↓CB1 ISH	Eggan et al., 2008;
	21/21				Volk, D. W. et al., 2014
Schizophrenia3	14/14	-	↓CB1	-	Eggan et al., 2010
Anterior cingulate cortex
Schizophrenia	10/9	↑[3H]SR141716A	-	-	Zavitsanou et al., 2004
Schizophrenia	15/15	-	n.s. CB1 (neurons and glia)	-	Koethe et al., 2007
Posterior cingulate cortex
Schizophrenia	8/8	↑[3H]CP55,940	-	-	Newell et al., 2006
Medial temporal lobe
Schizophrenia	14/14	n.s. [3H]CP55,940	-	-	Dean et al., 2001
Superior temporal gyrus
Schizophrenia	8/8	n.s. [3H]CP55,940 n.s.[3H]SR141716A	-	-	Deng et al., 2007
Striatum
Schizophrenia	14/14	n.s. [3H]CP55,940	-	-	Dean et al., 2001

↑: higher; ↓: lower; n.s.: no significant difference from controls; -: no available data;

CB1 agonist: [³H]CP55,940; CB1 antagonist: [³H]MePPEP, [³H]OMAR, [³H]SR141716A;

IB: Immunoblotting; ICC: immunocytochemistry; IHC: immunohistochemistry; qPCR: Quantitative real-time polymerase chain reaction; ISH: in situ hybridization;

¹Measured using immunocytochemistry unless noted as immunoblot IB.

²Eggan et al. 2008 and Volk et al., 2014 tested tissue from the same patients.

³This sample consists of a subset from Eggan et al, 2008 study.

In contrast to the observed elevations of CB1 receptor binding, patients with schizophrenia exhibited reduced CB1 immunoreactivity in dorsolateral prefrontal cortex (Dalton et al., 2011; Dean et al., 2001; Jenko et al., 2012; Volk, D. W. et al., 2014; but see Uriguen et al., 2009; table 4), but not in anterior cingulate cortex (Koethe et al., 2007). Patients with schizophrenia similarly exhibited lower CB1 receptor mRNA relative to controls (Dalton et al., 2011; Eggan et al., 2008; Uriguen et al., 2009).

Changes to CB1 receptors are paralleled by alterations in brain endocannabinoids and endocannabinoid metabolism. Patients with schizophrenia exhibited lower anandamide levels in hippocampus and cerebellum but not in dorsolateral prefrontal cortex and superior frontal gyrus (table 5; Muguruza et al., 2013; Yu et al., 2018). Lower levels of PEA were observed in cerebellum but not in dorsolateral prefrontal cortex or hippocampus (Muguruza et al., 2013) whereas OEA levels did not differ between patients and controls in any of these regions. FAAH activity, but not mRNA, was elevated in dorsolateral prefrontal cortex of patients, relative to controls (Muguruza et al., 2019). Finally, the ratio of 2-AG to anandamide, PEA and OEA was elevated in prefrontal cortex, hippocampus and cerebellum of patients with schizophrenia relative to controls (Muguruza et al., 2019).

Table 5.

Postmortem studies of endocannabinoid metabolism in psychotic disorders.

				1Metabolic enzymes
Condition & brain region	n (cases/controls)	AEA/PEA/OEA	2-AG	AEA	2-AG	Reference
Dorsolateral prefrontal cortex
Schizophrenia	42/42	-	-	-	n.s. DAGL	Volk et al., 2010
					n.s. MAGL
Schizophrenia	42/42	-	-	-	n.s. ABHD6	Volk et al., 2013
>15 years duration2	30/42				n.s.ABHD6
<15 years duration2	12/42				↑ABHD6
Schizophrenia	19/19	n.s.AEA	↑	n.s. FAAH, ↑FAAH activity	n.s. MAGL	Muguruza et al., 2013; Muguruza et al., 2019
		n.s.PEA
		n.s.OEA
Superior frontal gyrus (anterior)
Schizophrenia	27/396	n.s.AEA	↑	-	-	(Yu et al., 2018)
Hippocampus
Schizophrenia	19/19	↓AEA	↑	-	-	(Muguruza et al., 2013)
		n.s.PEA
		n.s.OEA
Cerebellum
Schizophrenia	19/19	↓AEA	n.s.	-	-	(Muguruza et al., 2013)
		↓PEA
		n.s.OEA

↑: higher; ↓: lower; n.s: no significant difference from controls; -: no available data;

AEA: anandamide; PEA: N-palmitoylethanolamide; OEA: N-oleoylethanolamide; 2-AG: 2-arachidonoyglycerol; NAPE-PLD: N-acyl phosphatidylethanolamine phospholipase D; FAAH: fatty acid amide hydrolase; DAGL: diacylglycerol lipase; MAGL: monoacylglycerol lipase; ABHD6: α/β-hydrolase domain-containing 6;

¹Results reflect mRNA levels unless noted as enzyme activity.

²This is a subgroup within Volk et al., 2013 study.

Alterations in levels of 2-AG and its respective metabolic enzymes followed a different pattern than observed for anandamide and FAAH. Patients with schizophrenia exhibited elevated 2-AG in dorsolateral prefrontal cortex, superior frontal gyrus and hippocampus but not in cerebellum, relative to controls (Muguruza et al., 2013; Yu et al., 2018). Despite the elevation of 2-AG observed in dorsolateral prefrontal cortex, patients did not exhibit significant differences in MAGL activity or of MAGL or ABHD6 mRNA levels compared to age-matched controls (Muguruza et al., 2019; Volk et al., 2010).

Postmortem studies provide some evidence that antipsychotics affect the endocannabinoid system. In dorsolateral prefrontal cortex, reductions of CB1 immunoreactivity were observed only in treated patients, not in antipsychotic-free patients (Uriguen et al., 2009; table 4). Most studies, however, report no significant relationship between lifetime or recent antipsychotic exposure and CB1 receptor binding, immunoreactivity or mRNA (Dean et al., 2001; Deng et al., 2007; Eggan et al., 2008; Jenko et al., 2012; Muguruza et al., 2019; Volk, D. W. et al., 2014; Zavitsanou et al., 2004).

Antipsychotics may also affect brain endocannabinoid levels in a region-specific manner (Muguruza et al., 2013). In prefrontal cortex, anandamide was lower in treated than antipsychotic-free patients. However, 2-AG was elevated in antipsychotic-free but not treated patients in prefrontal cortex and hippocampus (Muguruza et al., 2013). Lastly, in cerebellum, anandamide and PEA were lower in antipsychotic-free patients but not treated patients.

Alterations in 2-AG metabolism may be related to disease stage. When subdivided by duration of illness, ABHD6 mRNA levels were elevated in patients within 15 years of disease onset, whereas patients beyond 15 years since onset did not differ from controls (Volk et al., 2013). To our knowledge, similar data have not been reported for other endocannabinoid system targets in postmortem studies.

Taken together, regional alterations in CB1 receptors, anandamide, 2-AG and their enzymes were observed in brains of patients with schizophrenia. In general, patients exhibited increased CB1 receptor binding but lower immunoreactivity and mRNA in prefrontal cortex, with similar changes in binding observed in cingulate. Altered endocannabinoid levels were observed in several brain regions, with hippocampus exhibiting changes in anandamide and 2-AG levels. Endocannabinoid-metabolizing enzymes have been less well-studied in schizophrenia, although available evidence is suggestive of elevated metabolism of anandamide and 2-AG. Finally, the observed changes to CB1 receptors, endocannabinoids and enzymes appear to vary according to brain region, treatment status and disease stage.

2.5. In vivo brain Positron Emission Tomography (PET) endocannabinoids in psychotic disorders

In vivo PET imaging studies in patients with schizophrenia report increased (table 6; Ceccarini et al., 2013; Wong et al., 2010) and decreased (Borgan et al., 2019; Ranganathan et al., 2016) CB1 receptor availability in comparison to healthy controls. The sole PET study to measure FAAH reported no significant differences in brain FAAH between untreated primarily first-episode patients as compared to healthy controls (Watts et al., 2020).

Table 6.

In vivo brain PET imaging endocannabinoids in psychotic disorders.

Condition & treatment	n (cases/controls)	1Radioligand	Result	Primary regions of interest	Greater symptom severity	Reference
Untreated
First-episode psychosis	20/20	[11C]MePPEP	↓CB1	Hipp, striatum, ACC, thalamus	↓CB1	Borgan et al., 2019
Schizophrenia	16/12	[18F]-MK-9470	↑CB1	SF, IF, TC, MTL, PC, OC, insula, CC, putamen, caudate nucleus, NAc,, thalamus, Cb	↓CB1	Ceccarini et al., 2013
Schizophrenia	27/36	[11C]CURB	n.s. FAAH	Hipp, amygdala, LST, AST, SMST, ACC, mPFC, DLPFC	↓FAAH	Watts et al., 2020
Mixed(treated & untreated)
Schizophrenia	25/18	[11C]-OMAR	↓CB1	amygdala, caudate, Cb, cingulum anterior, cingulum posterior, FL, hipp, hypothalamus, insula, OC, pallidum, PC, putamen, TC, thalamus	↑CB1	Ranganathan et al., 2016
Treated
First-episode psychosis	7/11	[18F]FMPEP-d2	↓CB1	Hipp, striatum, ACC, thalamus	-	Borgan et al., 2019
Schizophrenia	51/12	[18F]-MK-9470	↑CB1	SF, IF, TC, MTL, PC, OC, insula, CC, putamen, caudate nucleus, NAc,, thalamus, Cb	n.s.	Ceccarini et al., 2013
Schizophrenia	9/10	[11C]-OMAR	↑CB1 (pons only)	FL, cingulate, hipp, putamen, globus pallidus, Cb, pons	↑CB1*	Wong et al., 2010

↑: higher, ↓: lower; n.s.: no significant difference from controls; -: no available data

CB1: cannabinoid CB1 receptor; FAAH: fatty acid amide hydrolase; Hipp: hippocampus; ACC: anterior cingulate cortex; SF: superior frontal; IF: inferior frontal; TC: temporal cortex, MTL: mesotemporal lobe; PC: parietal cortex; OC: occipital cortex; CC: cingulate cortex; NAc: nucleus accumbens; Cb: cerebellum; LST: limbic striatum; AST: associative striatum; SMST: sensory motor striatum; mPFC: medial prefrontal cortex; DLPFC: dorsolateral prefrontal cortex; FL: Frontal lobe.

¹CB1 receptor radioligand: [¹¹C]MePPEP, [¹⁸F]-MK-9470, [¹¹C]-OMAR and [¹⁸F]FMPEP-d2; FAAH radioligand: [¹¹C]CURB.

^*a ratio of positive and negative symptoms, not a direct measure from scale

Brain CB1 receptors and FAAH may be linked to disease stage and treatment status. Untreated patients had greater regional differences in CB1 receptor availability than controls (Borgan et al., 2019; Ceccarini et al., 2013; Ranganathan et al., 2016; table 6). Treatment-free patients showed lower CB1 receptor availability in amygdala, caudate, posterior cingulate cortex and pallidum whereas patients that were treated showed lower CB1 receptor availability only in posterior cingulate cortex (Ranganathan et al., 2016). Untreated and first-episode psychosis patients exhibited greater differences in CB1 receptor availability relative to healthy controls (Ceccarini et al., 2013; Borgan et al., 2019). Similarly, patients with longer duration of illness or duration of untreated psychosis had higher levels of brain FAAH (Watts et al., 2020).

Psychotic symptom severity was associated with the availability of CB1 receptors and FAAH. In untreated patients, lower CB1 receptor availability was related to greater severity of positive symptoms (Borgan et al., 2019) and of negative symptoms (Ceccarini et al., 2013; Borgan et al., 2019). In contrast, higher CB1 receptor availability was also related to higher ratio of positive to negative symptoms (Wong et al., 2010) and higher general or total symptom score; (Ranganathan et al., 2016, table 6). In some patient sub-samples, associations between CB1 receptor availability and symptoms were not significant (Ceccarini et al., 2013). Finally, lower FAAH levels were associated with greater severity of positive dimension psychotic symptoms (Watts et al., 2020).

The reasons for contradictory directions of CB1 receptor disturbances observed in PET studies are not immediately clear. The studies used different ligands and methods of quantification. Differences in clinical or demographic characteristics of the studies have been suggested as contributing to differences in study outcomes, although it is not clear that such differences are sufficient to explain differences in direction (for commentary, see Mihov 2016). Reduced CB1 availability was observed in male-only samples (Borgan et al. 2019, Ranganathan et al., 2016), however most control groups were matched for age and sex (Borgan et al, Ranganathan et al., Ceccarini et al., 2013).

While studies were not designed to test the effect of exposure to tobacco or cannabis, Ranganathan et al. reported that the greatest reduction in CB1 receptors availability was observed in patients who were non-smokers relative to smokers (Ranganathan et al., 2016), although this was not observed in other cohorts (Borgan et al., 2019; Ceccarini et al., 2013). Tobacco smoking was related to reduced CB1 receptor availability in otherwise healthy males, but not in males with alcohol or cannabis use disorders, suggesting the impact of tobacco use on CB1 receptor availability may vary across neuropsychiatric populations (Hirvonen et al., 2018). Past cannabis exposure did not have an effect on brain CB1 receptors although the lack of a control group matched for cannabis use history limits the interpretation of this analysis in some studies (Borgan et al., 2019; Ranganathan et al., 2016; Ceccarini et al., 2013).

Overall, PET studies provide strong evidence of disturbances in the brain endocannabinoid system in psychotic disorders. Brain CB1 receptors and FAAH may be associated with disease stage and severity of symptoms.

2.6. Summary of endocannabinoid system alterations in psychotic disorders

In psychotic disorders, peripheral studies provide evidence of elevated expression of CB1 and CB2 receptors, reduced expression of biosynthetic enzymes for anandamide and 2-AG, and elevated expression of catabolic enzymes for these endocannabinoids. Postmortem studies revealed elevated brain CB1 receptor ligand binding but reduced CB1 receptor mRNA and immunoreactivity. PET studies report widespread reductions of CB1 receptor availability in three independent cohorts of patients and controls, at odds with earlier reports. Elevated anandamide and N-acylethanolamines were observed in CSF and peripheral blood, whereas postmortem studies report region-dependent reductions of anandamide. In peripheral and CSF studies, higher anandamide levels were associated with lower symptom severity. Although higher FAAH activity was observed in peripheral blood and postmortem brain in schizophrenia, when measured in vivo, greater psychotic symptom severity was associated with lower FAAH. Disturbances in brain 2-AG metabolism were evidenced by increases in 2-AG and ABHD6 in the absence of changes in MAGL. Finally, elevations of anandamide and 2-AG observed in patients with high risk for psychosis suggest a relationship to disease stage. Taken together, studies using peripheral blood, CSF, postmortem tissue and in vivo PET imaging and provide evidence of widespread disturbances to the endocannabinoid system in psychotic disorders. These observations contribute toward evidence linking the peripheral and central endocannabinoid system with antipsychotic treatment, disease stage and symptom severity in psychotic disorders.

3. Endocannabinoids in mood disorders

3.1. Peripheral blood endocannabinoids in mood disorders

In medication-free patients with major depressive episode, serum anandamide and 2-AG was lower in major depressive episode and major depressive disorder relative to healthy controls (table 7; Hill et al., 2009; Hill et al., 2008). Medication-free patients with minor depression however, exhibited elevations in serum anandamide but not 2-AG (Hill et al., 2008). In a mixed cohort of treated and untreated patients with major depressive disorder, anandamide, OEA and 2-AG were elevated in plasma (Romero-Sanchiz et al., 2019). In this mixed cohort, elevations of OEA and 2-AG were related to antidepressant treatment.

Table 7.

Peripheral blood and cerebrospinal fluid endocannabinoids in mood disorders.

Condition & treatment	n (cases/controls)	Sample	AEA	PEA	OEA	2-AG	Reference
Untreated
Major depressive episode	15/15	serum	↓	n.s.	n.s.	↓	Hill et al., 2009
Major depressive episode	16/16	serum	n.s.	-	-	↓	Hill et al., 2008
Minor depressive episode	12/11	serum	↑	-	-	n.s.	Hill et al., 2008
Mixed (treated & untreated)
Major depressive disorder	69/47	plasma	↑	-	↑	↑	Romero-Sanchiz et al., 2019
Treated
Major depressive disorder & bipolar disorder	22/84	CSF	n.s.	n.s.	n.s.	-	Giuffrida et al., 2004
Treatment status not known
Bipolar affective disorder	7/16	plasma	↑	↑	n.s.	n.s.	Koethe et al., 2019
At risk
Unaffected twin of bipolar affective disorder	7/16	plasma	↑	↑	n.s.	n.s.	Koethe et al., 2019

↑: higher; ↓: lower; n.s.: no significant difference from controls; -: no available data;

CSF: cerebrospinal fluid; AEA: anandamide; PEA: N-palmitoylethanolamide; OEA: N-oleoylethanolamide; 2-AG: 2-arachidonoyglycerol.

The only study in bipolar disorder measured plasma endocannabinoids levels in twins discordant for bipolar disorder relative to matched healthy twin-pairs (Koethe et al., 2019). Both bipolar patients and their non-affected twins exhibited elevated plasma anandamide and PEA, but not OEA or 2-AG relative to healthy controls (Koethe et al., 2019). There was no difference in endocannabinoid levels between affected and non-affected twins (Koethe et al., 2019). Further, there was also no difference in endocannabinoid levels between non-affected twins that later converted to bipolar disorder (n=3) and twins who remained non-affected (n=4), in follow-up (Koethe et al., 2019).

Antidepressant treatment may influence blood endocannabinoids in major depression. For example, patients with major depressive disorder treated with serotonin reuptake inhibitors (n=22) had higher plasma OEA and 2-AG levels in comparison to untreated patients (n=47; Romero-Sanchiz et al., 2019). In contrast to observations in unipolar depression, medication status was not associated with plasma endocannabinoids in a small sample of bipolar patients (n=7; Koethe et al., 2019).

Low levels of 2-AG – but not AEA, OEA or PEA – were associated with longer duration of current major depressive episode in an all-female sample (Hill et al., 2008). Patients with moderate depression had higher anandamide in comparison to patients with mild depression as measured using Beck’s Depression Inventory-II (Romero-Sanchiz et al., 2019).

Romero-Sanchiz and colleagues reported that higher OEA was associated with greater severity of depression and higher AEA and OEA were associated with greater severity of somatic symptoms including fatigue, appetite change, sleep changes and decreased libido (Romero-Sanchiz et al., 2019). In contrast, neither anandamide nor 2-AG were associated with state depression scores in females with major or minor depression or in a heterogenous mixed-sex sample of patients with Axis I and/or personality disorders and non-psychiatric controls (Coccaro et al., 2018; Hill et al., 2009; Hill et al., 2008). In untreated patients with major depressive episode, higher levels of anandamide was associated with lower anxiety levels as measured on the Hamilton depression rating scale (Hill et al., 2008). Together, these studies provide preliminary evidence linking peripheral endocannabinoids to symptom severity in depression.

In comparison to studies in psychotic disorders, there are fewer studies of endocannabinoids in peripheral blood mononuclear cells in mood disorders. One study with major depressive disorder (n=24), bipolar disorder I (n=24), bipolar disorder II (n=24) did not find any alterations in peripheral blood mononuclear cells in CB1, CB2, NAPE-PLD, FAAH, DAGL or MAGL, relative to matched controls (D’Addario et al., 2017).

One study with a small combined mood disorder sample (unipolar depression or bipolar disorder, n=12) reported that CSF levels of anandamide, OEA and PEA did not differ between healthy controls and of treated patients with mood disorders (Giuffrida et al., 2004). There was lower CSF levels of ethanolamine in patients with major depressive disorder in comparison to controls. Further, low ethanolamine levels were associated with greater depression severity (Ogawa et al., 2015). A treatment study reported CSF measurements in patients with major depressive disorder (n=11) before and after treatment (Kranaster et al., 2017). Greater CSF anandamide levels were associated with greater severity of depressive symptoms, but no such association was observed with 2-AG. Following electroconvulsive therapy, CSF levels of anandamide levels were significantly elevated but these changes were not associated with post-treatment changes in depressive symptoms (Kranaster et al., 2017). Taken together, peripheral studies in patients with mood disorders report dysregulation in endocannabinoids which may be affected by medication, disease stage, and symptom severity. In major depressive disorder, studies reported reduced blood levels of anandamide, PEA and 2-AG, whereas anandamide and PEA were elevated in bipolar disorder. One study reported that antidepressant therapy was associated with elevation of OEA and 2-AG, but not anandamide. When significant associations with mood symptoms were observed, higher levels of anandamide and OEA in blood, or higher anandamide in CSF were associated with greater depressive symptom severity. All three studies of peripheral endocannabinoids in unipolar depression included samples composed primarily of female patients, indicating a need for studies with larger samples with sex balanced cohorts (Hill et al., 2009; Hill et al., 2008; Romero-Sanchiz et al., 2019). Future studies with demographically balanced samples of sufficient size are needed to better understand whether alterations in peripheral and CSF endocannabinoids are similar across males and females.

3.2. Postmortem brain endocannabinoids in mood disorders

Postmortem studies report elevated CB1 binding, immunoreactivity, mRNA and function in the dorsolateral prefrontal cortex of patients with major depressive disorder (table 8; Choi et al., 2012; Hungund et al., 2004; Mato et al., 2018; but see Eggan et al., 2010;). Lower CB1 receptor immunoreactivity was also observed in the anterior cingulate cortex of patients with depression (Koethe et al., 2007). CB1 receptor immunoreactivity did not differ significantly in the dorsolateral prefrontal cortex of patients with or without the presence of psychotic features (Eggan et al., 2010). Dorsolateral prefrontal cortex of depression patients exhibited elevated CB1 agonist-stimulated [³⁵S]GTPγS binding and activation of Gα_i/0 protein subunits (Hungund et al., 2004; Mato et al., 2018). Changes in functional activity were present even in the absence of changes to receptor density (Mato et al., 2018).

Table 8:

Postmortem studies of the CB1/CB2 receptors in mood disorders.

Condition	n (cases/controls)	Autoradiography	1Antibody	mRNA	Reference
Dorsolateral prefrontal cortex
Major depressive disorder	26/46	-	-	↑CB1 n.s. CB2	Choi et al., 2012
Major depressive disorder	14/14	-	n.s. CB1	-	Eggan et al., 2010
Major depressive disorder	10/10	↑[3H]CP55,940	↑CB1 (IB)	-	Hungund et al., 2004
Major depressive disorder	4/4	n.s. [3H]CP55,940	-	-	Mato et al., 2018
Anterior cingulate cortex
Major depressive disorder	15/15	-	↓CB1 (glia) n.s. CB1 (neurons)	-	Koethe et al., 2007
Dorsolateral prefrontal cortex
Bipolar disorder	31/46	-	-	n.s. CB1 n.s. CB2	Choi et al., 2012
Anterior cingulate cortex
Bipolar disorder	15/15	-	n.s. CB1 (glia and neurons)	-	Koethe et al., 2007

↑: higher; ↓: lower; n.s.: no significant difference from controls; -: no available data;

DLPFC: dorsolateral prefrontal cortex; ACC: anterior cingulate cortex;

[³H]CP55,940: CB1/CB2 receptor agonist; [³H]WIN55212-2: CB1/CB2 receptor agonist.

¹Measured using immunocytochemistry unless noted as immunoblot IB.

In contrast to CB1, CB2 receptor mRNA levels did not differ in the dorsolateral prefrontal cortex of patients with mood disorders relative to controls (Choi et al., 2012). In anterior cingulate cortex, depression patients exhibited reduced CB1 immunoreactivity in glia but not in neurons (Koethe et al., 2007). Bipolar patients and controls did not differ in CB1 or CB2 receptor mRNA levels or CB1 immunoreactivity in neurons or glial cells in dorsolateral prefrontal cortex or anterior cingulate cortex (Choi et al., 2012; Koethe et al., 2007).

There is some evidence that antidepressant medications affected CB1 receptor function or protein levels in unipolar depression. In antidepressant-free patients with unipolar depression, CB1 receptor agonist-induced [³⁵S]GTPγS binding and activation of Gα_i/o protein subunits were elevated in dorsolateral prefrontal cortex, but this was not observed in patients taking antidepressants at the time of death (Mato et al., 2018). Specifically, elevated CB1 receptor coupling to the Gα_o was observed in patients who were antidepressant-free, and not in those who were taking antidepressants at the time of death (Mato et al., 2018). In contrast, antidepressant treatment at time of death was not associated with changes in CB1 immunoreactivity in patients with schizophrenia (Eggan et al., 2010). Finally, in the anterior cingulate cortex, treatment with antidepressant medications at the time of death was associated with lower density of CB1-immunoreactive neurons but not glial cells (Koethe et al., 2007).

In contrast to neuron-centric changes related to antidepressants, intake of first-generation antipsychotics was associated with lower density of CB1 receptor immunoreactive glial cells in patients with bipolar disorder (Koethe et al., 2007).

Overall, postmortem studies in affective disorders provide evidence of elevated CB1 receptor binding, protein, mRNA and function in unipolar depression, whereas no such differences have been reported in bipolar disorder. Changes to CB1 receptor function may be related to antidepressant treatment status. Finally, there were no in vivo endocannabinoid PET imaging studies in mood disorders.

3.3. Summary of endocannabinoid system alterations in mood disorders

Fewer investigations of endocannabinoids in affective disorders have been reported in comparison to psychotic disorders. Studies of peripheral blood in unipolar depression report lower levels of 2-AG, whereas anandamide may be increased or decreased according to disease stage. In contrast to the complex alterations in CB1 receptors reported in psychotic disorders, effects in postmortem studies in unipolar depression uniformly suggest elevated CB1 receptor expression and function in dorsolateral prefrontal cortex. Evidence includes increased levels of CB1 receptor mRNA along with higher CB1 receptor binding and functional activity in the brains of suicide victims with unipolar depression.

Taken together, studies using peripheral blood, CSF, and postmortem tissue provide evidence of alterations of anandamide, 2-AG and brain cannabinoid CB1 receptors in unipolar depression. These reports contribute toward emerging evidence linking the peripheral and central endocannabinoid system with antidepressant treatment, disease stage and symptom severity in unipolar depression. In comparison to unipolar depression, the literature describing endocannabinoid system in bipolar disorder remains in its infancy.

3. Discussion

Converging evidence from studies in peripheral blood, postmortem brain and in vivo molecular brain imaging provide strong evidence of altered peripheral blood endocannabinoids and altered brain cannabinoid receptors in psychotic and mood disorders.

In peripheral blood, patients with psychosis-spectrum and mood disorders exhibit changes in circulating anandamide, most commonly showing increased anandamide in psychotic disorders, psychosis risk as well as unipolar and bipolar depression (tables 1 and 7). The status of 2-AG is less well understood in both disorders. Disturbances of peripheral cannabinoid receptors and endocannabinoid-metabolizing enzymes for anandamide and 2-AG have been reported in psychotic disorders but have not been studied in mood disorders. Studies in CSF largely echo observations of disturbed anandamide and related N-acylethanolamines in peripheral blood in psychotic and mood disorders. Elevated anandamide in CSF was observed in patients with first-episode psychosis and chronic schizophrenia, as well as individuals with psychosis risk. In psychotic and mood disorders, preliminary evidence suggests a link between CSF anandamide and symptom severity.

Beyond CB1 and FAAH, other endocannabinoids have not been targeted with in vivo imaging. Moreover, at the time of review, no in vivo brain imaging studies of the endocannabinoid system have been reported in mood disorders. In psychosis-spectrum and mood disorders, peripheral and central endocannabinoid metabolism of anandamide and 2-AG remains an understudied area of inquiry. Finally, bipolar disorder has been largely neglected in peripheral and central, in vivo and postmortem studies of the endocannabinoid system, representing a marked gap in the literature.

Several studies have explored peripheral endocannabinoids in relation to disease stage, symptoms, and drug or medication exposure, however these outcomes were typically secondary or exploratory. Though informative, because of multiple testing, statistical power, and study design, these findings may be at elevated risk of type I and/or II error.

Taken together, the available literature provides strong evidence for disturbances of the peripheral and central endocannabinoid system in psychosis-spectrum and mood disorders. Both mood and psychotic disorders exhibit elevated anandamide and whereas psychotic disorders also exhibit decreased brain CB1 receptor availability when measured in vivo, this has not been measured in depression. Emerging evidence suggests that treatment with antipsychotics or antidepressants may be associated with normalization of peripheral anandamide levels. However, psychotic disorders and unipolar depression exhibit opposite associations of N-acylethanolamine levels with symptoms and opposite changes in CB1 receptor immunoreactivity and mRNA levels in postmortem dorsolateral prefrontal cortex and these observations may reflect differences in the nature of endocannabinoid system disturbances in these different disorders.

4. Conclusion

Overall, alterations in the peripheral and central endocannabinoid system are present in psychotic and mood disorders. The available evidence indicates that these changes are sensitive to treatment status, disease stage, and symptom severity. Evidence from psychotic disorder extend to endocannabinoid metabolizing enzymes in the brain and periphery, whereas these lines of evidence remain poorly developed in affective disorders. A lack of studies examining this system in bipolar disorder represents a notable gap in the literature.

Highlights.

• The endocannabinoid is altered in psychotic and mood disorders

• Changes in receptors, anandamide, 2-AG and their metabolism were disease-specific

• Endocannabinoids were related to treatment status, disease stage and symptom severity

• Endocannabinoid targets may be promising for development of novel therapeutics

Abbreviations

AEA

anandamide

2-AG

2-arachidonoylglycerol

NAPE

N-acylphosphatidylethanolamine

NAPE-PLD

N-acyl phosphatidylethanolamine-specific phospholipase D

DAG

diacylglycerol

DAGL

diacylglycerol lipase

PEA

N-palmitoylethanolamide

OEA

N-oleoylethanolamide

CB1

cannabinoid receptor type 1

CB2

cannabinioid receptor type 2

MAGL

monoacylglycerol lipase

FAAH

fatty acid amide hydrolase

ABHD6

α/β-hydrolase domain-containing 6

AA

arachidonic acid

Ca⁺²

calcium

PBMC

peripheral blood mononuclear cells

CSF

cerebrospinal fluid

IB

immunoblotting

ICC

immunocytochemistry

IHC

immunohistochemistry

qPCR

quantitative real-time polymerase chain reaction

ISH

in situ hybridization

PET

Positron Emission Tomography