A preliminary investigation of tobacco co-use on endocannabinoid activity in people with cannabis use

Abstract

Tobacco is commonly co-used with cannabis. This is unfortunate because tobacco co-use exacerbates select clinical consequences associated with cannabis use. Evidence demonstrates that low levels of anandamide, a prominent endocannabinoid, correlate with worse clinical outcomes. Fatty acid amide hydrolase (FAAH) degrades anandamide, and greater FAAH levels may underlie poorer clinical outcomes in people who co-use relative to those who use only cannabis. Therefore, we tested whether tobacco co-use increases FAAH levels beyond those associated with cannabis use alone. Cannabis-using participants (N = 13) were parsed into individuals with daily tobacco use (CT, n = 5) and no current tobacco use (CAN, n = 8). We evaluated group differences in FAAH, quantified using positron emission tomography and [¹¹C]CURB, while controlling for sex and FAAH genotype in the prefrontal cortex, hippocampus, thalamus, sensorimotor striatum, substantia nigra, and cerebellum. A significant group x ROI interaction for [¹¹C]CURB λk₃ [F(5, 45)= 3.15, p = 0.016] emerged. Bonferroni-corrected post-hoc tests indicated greater FAAH levels in CT compared to CAN in the substantia nigra (p = 0.023, d=1.54) and cerebellum (p = 0.003, d=1.76), while a trend emerged in the sensorimotor striatum (p = 0.054, d=1.33). Preliminary findings suggest that tobacco co-use is associated with elevated FAAH activity relative to cannabis-only use, which may underlie poorer clinical outcomes associated with co-use.

Highlights

• •The neurobiological underpinnings of cannabis-tobacco co-use are underexplored.
• •Tobacco co-use is associated with elevated FAAH levels relative to cannabis-only use.
• •Greater tobacco exposure is associated with higher cerebellar FAAH levels.
• •Higher FAAH levels may predict worse clinic outcomes in people with co-use.

1. Introduction

Cannabis use is prevalent, with approximately, 25 % of Americans and Canadians reporting past-year use (Government of Canada, 2024, Substance Abuse and Mental Health Services Administration, 2023). Daily cannabis consumption is rising (Caulkins, 2024, Goodman et al., 2024), which parallels relaxation in cannabis legislation (Caulkins, 2024). This shift has sparked concerns about the impact of frequent cannabis use on mental health outcomes, especially among adolescents and young adults (Lorenzetti et al., 2020, Volkow et al., 2014), when individuals may be at greater risk for the development of psychopathology triggered by cannabis use (e.g., depression, anxiety, psychosis, and suicide ideation) (Gobbi et al., 2019, Lowe et al., 2024).

Notably, up to 80 % of people who use cannabis co-use a tobacco product (Government of Canada, 2017, Gravely et al., 2020). This statistic may even underestimate rates given that co-use is often overlooked in population surveys (Weinberger et al., 2020), as many surveys fail to specifically assess or distinguish co-use behaviors. Cigarettes are the most common mode of tobacco administration in people who use cannabis (Cohn, 2022). However, other methods, such as adding tobacco to joints (mulling) and vaping both substances, are also prevalent (Baggio et al., 2014). Several mechanisms have been proposed to explain the high rates of tobacco co-use in people who use cannabis (Rabin and George, 2015). For example, tobacco use may enhance and prolong the euphoric effects of cannabis (Lee et al., 2010, Penetar et al., 2005, Ramo et al., 2013, Tullis et al., 2003). Other explanations for the high rates of co-use include shared route of administration (e.g., smoking) and common predisposing factors such as genetics, personality, and environment (Agrawal et al., 2012, Agrawal and Lynskey, 2009).

The endocannabinoid (eCB) system is an important modulatory network in the brain. The eCB system consists of naturally circulating eCBs, such as anandamide, which is degraded by the enzyme fatty acid amide hydrolase (FAAH) into ethanolamine and arachidonic acid (Ueda et al., 2000). FAAH is important because it keeps anandamide levels balanced (Cravatt et al., 2001, Di Marzo, 2011), and lower levels of anandamide have been linked to greater clinical symptoms and the development of psychiatric disorders (Dlugos et al., 2012, Gao et al., 2020, Garani et al., 2021, Hill et al., 2008).

Alterations in the eCB system are commonly reported in people with chronic cannabis use. Positron emission tomography (PET) studies demonstrate selective and reversible downregulation of cannabinoid 1 receptors (CB1R) in frontolimbic areas in the brain in individuals with daily cannabis use compared to controls (Ceccarini et al., 2015, Hirvonen et al., 2012). In clinical populations, cannabis use has been associated with low plasma-derived anandamide levels (Bassir Nia et al., 2023, Leweke et al., 2007). Further, frequent cannabis use has been associated with lower anandamide levels in the cerebrospinal fluid compared to infrequent cannabis use (Morgan et al., 2013). Chronic cannabis use has also been associated with changes in brain levels of FAAH and circulating eCBs relative to people who do not use cannabis (Boileau et al., 2016, Jacobson et al., 2021).

Like cannabis use, alterations in the eCB system have also been documented with tobacco use (Buczynski et al., 2013, Gonzalez et al., 2002). Hirvonen and et al. (2018) found a 20 % reduction in CB1R binding across multiple brain regions, including the prefrontal cortex, hippocampus, thalamus, midbrain, and cerebellum, in individuals with tobacco use compared to healthy controls. Further, a preclinical study demonstrated that chronic nicotine administration decreased anandamide levels in the cerebral cortex, hippocampus, and striatum (Gonzalez et al., 2002). It is plausible that low anandamide levels may reflect CB1R downregulation due to chronic tobacco exposure (Hill et al., 2005).

Taken together, the use of cannabis alone and the use of tobacco alone have been linked to alterations in the eCB system, including reduced anandamide levels, which may reflect higher FAAH concentrations. Given the rising rates of co-use (Rubenstein et al., 2024) and the accumulating evidence linking co-use to more severe mental health consequences compared to using either substance alone (Do et al., 2024, McClure et al., 2020, Rabin et al., 2023, Walsh et al., 2020), we sought to investigate whether co-use exacerbates eCB dysregulation beyond what is observed with cannabis use alone.

The radiotracer, [¹¹C-carbonyl]6-hydroxy-[1,1′-biphenyl]-3-yl cyclohexylcarbamate ([¹¹C]CURB) is the most advanced PET imaging tool for quantifying brain FAAH levels in humans in vivo (Wilson et al., 2011). [¹ ¹C]CURB binds irreversibly to FAAH, enabling measurement of λk₃, a kinetic parameter that reflects the rate of irreversible binding and serves as a reliable index of FAAH availability (Rusjan et al., 2013). Therefore, in this preliminary analysis, we compared FAAH in brain, in vivo, in individuals with co-use to individuals with cannabis-only use. We focused on regions of interest (ROIs) in the brain that have high densities of overlapping cannabinoid and nicotinic receptors, which are likely critical for facilitating interactions between these two signalling systems (Adermark, 2011, Narushima et al., 2007, Picciotto et al., 2000, Tsou et al., 1998).

2. Methods and materials

2.1. Ethics approval and consent to participate

This study was performed under a repository protocol that allowed analysis of previously acquired data, approved by the CAMH Research Ethics Board (REB) and now approved under Clinical and Translational Sciences (CaTS) BioBank (REB number: 2021–304, IUSMD-21–02)

by REB of the Centre intégré universitaire de santé et de services sociaux (CIUSSS) de l’Ouest-de-l′Île-de-Montréal – Mental Health and Neuroscience subcommittee. The study was performed in accordance with Good Clinical Practice guidelines, regulatory requirements, and the Code of Ethics of the World Medical Association (Declaration of Helsinki). Written informed consent was obtained from all participants at the beginning of screening after a full explanation of anticipated study procedures. Participants received compensation for their study participation.

2.2. Study participants

Participants for this secondary analysis were extracted from The CaTS BioBank repository. Participants were recruited from the Toronto-area community between February 2015 and March 2016.

The repository is composed of retrospective [¹¹C]CURB data from healthy volunteers with and without cannabis use, individuals with psychotic disorder, and individuals at clinical high risk for psychosis between the ages of 18 and 35. In our analysis we included participants with current regular cannabis use (defined as using cannabis at least four days per week for one year) and a positive urine toxicology for cannabis.

Exclusion criteria were as follows: (i) DSM-IV Axis I disorder, as determined by the Structured Clinical Interview for DSM-IV (SCID) (American Psychiatric Association, 2000) with the exception of nicotine abuse/dependence, and cannabis abuse/dependence; (ii) individuals at clinical high risk for psychosis; (iii) significant current or past medical conditions; (iv) neurological illnesses or head trauma; (v) use of medications that might affect the central nervous system, (vi) the presence of metal implants precluding a magnetic resonance imaging (MRI) scan, and/or pregnancy or breastfeeding.

Following a phone screen, potentially eligible participants completed an in-person screening visit. Demographic information was collected, and current/history of drug use was assessed by an in-house semi-structured interview. The Fagerstrom Test for Nicotine Dependence was used to assess nicotine dependence (Heatherton et al., 1991). Participants were then asked to abstain from cannabis overnight (~12 h) prior to the scheduled PET scan, as previously done by our group (Jacobson et al., 2021). No instructions were given to participants regarding tobacco use; therefore, they used ad libitum.

Eligible participants completing all measures were then parsed according to their daily cigarette use: those who consumed ≥ 1 cigarettes per day (CT); and individuals with no current tobacco use (CAN).

2.3. PET and MRI data acquisition and analysis

[¹¹C]CURB PET data were acquired according to the validated method reported elsewhere (Rusjan et al., 2013). [¹¹C]CURB was synthesized as previously described (Wilson et al., 2013). Briefly, participants underwent a transmission scan followed by an intravenous bolus injection of [¹¹C]CURB (9.25 ± 1.1 mCi in CT, 9.69 ± 0.8 mCi in CAN) and 60 min PET scan using a 3D HRRT brain tomograph (CPS/Siemens, Knoxville, TN, USA) (Rusjan et al., 2013). A 2D filtered-back projection algorithm, with a HANN filter at Nyquist cutoff frequency, was applied to the 2D sinograms to reconstruct the images. Arterial blood samples were collected automatically using an automated blood sampling system (ABSS; Model PBS-101, Veenstra Instruments, The Netherlands) at a rate 350 mL/h for the first 7.25 min after [¹¹C]CURB-injection and 150 mL/h for the next 15 min and samples were collected manually at 3, 7, 12, 20, 30, 45, and 60 min post-injection, to measure radioactivity in blood and plasma and determine the relative proportion of radiolabeled metabolites. A metabolite-corrected plasma input function was generated as previously described (Rusjan et al., 2013). To permit delineation of regions of interest, a standard proton density (PD) weighted brain MRI was acquired for each participant, using a Discovery MR750 3 T MRI (General Electric, Milwaukee, WI, USA).

Time-activity curves for each ROI were extracted using an in-house imaging pipeline (Rusjan et al., 2006). [¹¹C]CURB binding was quantified using the composite rate constant λk3 (λ=K1/k2), as derived from an irreversible two-tissue compartment model, which is the validated method for quantifying [¹¹C]CURB binding in vivo (Boileau et al., 2015a).

2.4. rs324420 FAAH genotyping

The FAAH gene is polymorphic (rs324420, C385A), resulting in lower FAAH protein levels and associated [¹¹C]CURB binding in carriers of one of more copy of the A allele (Boileau et al., 2015b). Thus, all participants were genotyped using a commercially available (Life Technologies, Burlington, Ontario, Canada) taqman assay as performed as previously described (Boileau et al., 2015b).

2.5. Statistical analysis

Group differences in demographic measures were determined using independent sample t-tests for continuous variable and Fisher’s exact tests for categorical variables. Group differences in [¹¹C]CURB λk3 were analyzed using a linear mixed models analysis, with group as the between subject factor (CT/CAN), [¹¹C]CURB λk3 as the dependent variable, ROI as a factor, and sex and FAAH genotype as covariates. The model tested for main effects of group, region, and genotype, and a group x ROI interaction. Six ROIs were included in the model: the prefrontal cortex (PFC), hippocampus, thalamus, sensorimotor striatum, substantia nigra, and cerebellum. Effect sizes (Cohen’s d) were calculated for brain regions that showed significant between group differences.

Brain regions that showed significant between group differences in [¹¹C]CURB λk3 were also tested for associations with daily tobacco exposure (using cigarettes per day) and daily cannabis exposure (using cannabis grams per day). We ran partial correlations to explore these associations including sex and FAAH genotype as covariates.

Statistical analyses were performed using SPSS (version 24.0; IBM, Armonk, NY, USA), with p < 0.05 considered to be significant.

3. Results

We extracted data from the repository for healthy volunteer participants (N = 102) who were subsequently screened based on inclusion criteria. Of these, 22 participants reported current regular cannabis use and had a positive urine toxicology for cannabis. After excluding early discharges (n = 8), and one participant with less than daily use of tobacco, data from 13 participants (CT, n = 5; CAN, n = 8) were included in the current preliminary analyses. See Figure S1 for Consort diagram.

The CT and CAN groups were well matched on demographic and cannabis use characteristics as described in Table 1. The groups did not differ with respect to age (M=22.84, SD=4.3), sex, or BMI, (ps>0.05). In addition, there were no significant differences between groups for any of the PET radiotracer parameters, including injected radioactivity, mass injected, and molar activity. There was no significant difference in the frequencies of C385A FAAH genetic polymorphism (rs324420) between groups. With respect to cannabis use, groups had on average, similar cumulative cannabis exposure, lifetime years of cannabis use, and past week use of cannabis.

Table 1.

Demographic and substance use characteristics.

	CT(n = 5)	CAN(n = 8)
Sexn (M/F)	3/2	6/2
Age	22.40 (1.8)	23.13 (5.4)
BMI	23.48 (5.0)	24.23 (5.3)
Genotypen (CC/AC)	4/1	6/2
Amount injected (mCi)	9.25 (1.1)	9.69 (0.8)
Mass injected (µg)	1.51 (0.4)	1.92 (1.1)
Molar Activity (mCi/µmol)	2034 (647)	1910 (797)
Cumulative cannabis uses	3441.18 (2335.9)
range: 260.00 – 5585.00	1973.18 (1293.2)
range: 390.50 – 3684.00
Lifetime cannabis exposure years	4.80 (2.8)	4.44 (2.8)
Past week cannabis use (g)	9.10 (3.1)	10.56 (6.9)
Last cannabis use (h)	16.28 (2.2)	18.54 (9.9)a
Cigarettes per day*	6.00 (4.3)	0.00 (0)
FTND*	2.20 (1.8)	0.00 (0)

a. one participant had missing data

BMI, body mass index; CAN, participants with cannabis-only use; CT, participants with cannabis-tobacco co-use; F, female; FTND, fagerstrom test for nicotine dependence; g, gram; h, hour; M, male.

^⁎p < 0.05

The average time of self-reported cannabis abstinence prior to the PET scan was 17.60 (SD=7.5) hours, which did not differ between groups. Among the CT group, participants consumed an average of 6.0 (SD=4.3) cigarettes per day and had an average FTND score of 2.20 (SD=1.8), indicative of low nicotine dependence (Heatherton et al., 1991).

3.1. Group differences in [¹¹C]CURB binding

We detected a significant effect of group [F(1, 9)= 5.41, p = 0.045], genotype [F(1, 9)= 6.71, p = 0.029], ROI [F(5,45)= 3.02; p = 0.020], and a significant group x ROI interaction [F(5, 45)= 3.15, p = 0.016]. [¹¹C]CURB λk3 was consistently higher in the CT group versus the CAN group in all brain regions examined. Bonferroni-corrected pairwise comparisons revealed significant group differences for [¹¹C]CURB λk3 in the substantia nigra (p = 0.023, d=1.54), and the cerebellum (p = 0.003, d=1.76), while a trend emerged in the sensorimotor striatum (p = 0.054, d=1.33). See Fig. 1.

Fig. 1.

Group Differences in FAAH Levels. [¹¹C]CURB λk3 was higher in CT relative to CAN in all brain regions examined. Significant group differences for [¹¹C]CURB λk3 emerged in the substantia nigra (p = 0.023, d=1.54), and cerebellum (p = 0.003, d=1.76), while a trend emerged in the sensorimotor striatum (p = 0.054, d=1.33). CAN, participants with cannabis-only use; Cer, cerebellum; CT, participants with cannabis-tobacco co-use Hipp, hippocampus; PFC, prefrontal cortex SMS, sensorimotor striatum; SN, substantia nigra; Thal, Thalamus. Means are adjusted for sex and FAAH genotype. Error bars represent standard error.

3.2. Associations between [¹¹C]CURB λk₃ and tobacco and cannabis use

Using partial correlations, controlling for sex and FAAH genotype, we tested associations between tobacco exposure and [¹¹C]CURB λk₃ in ROIs that showed significant between groups differences (the substantia nigra and the cerebellum) and cigarette consumption per day. A significant correlation emerged between [¹¹C]CURB λk3 and the cerebellum (r = 0.65, p = 0.03) (see Fig. 2), while a trend was observed between [¹¹C]CURB λk3 and the substantia nigra (r = 0.57, p = 0.07). No significant associations emerged between [¹¹C]CURB λk₃ in the cerebellum or in the substantia nigra and cannabis exposure.

Fig. 2.

Association Between Cerebellar FAAH Levels and Tobacco Consumption Cerebellar [¹¹C]CURB λk₃ significantly correlated with cigarettes per day across all participants, (r = 0.65, p = 0.030). Values have been corrected for sex and genotype.

4. Discussion

This is the first study to investigate differences in FAAH levels in brain, as indexed by [¹ ¹C]CURB PET imaging, between individuals who co-use cannabis and tobacco relative to individuals who use cannabis only. Our preliminary findings support our hypotheses and demonstrate significantly higher [¹¹C]CURB λk₃ in individuals who co-use compared to individuals with cannabis-only use in the substantia nigra, and cerebellum, with a trend emerging in the sensorimotor striatum. Notably, these effect sizes were in the large range. Further, we observed that [¹¹C]CURB λk₃ in the cerebellum, and at trend level in the substantia nigra, positively correlated with tobacco exposure. This suggests that greater tobacco consumption is associated with higher FAAH levels in select regions of the brain. Interestingly, there were no significant associations between [¹¹C]CURB λk₃ and cannabis use.

Of note, previous research by our group and others have observed lower FAAH levels in people with cannabis use relative to healthy controls (Boileau et al., 2016, Jacobson et al., 2021). In contrast to our null finding between [¹¹C]CURB and cannabis exposure, results from these studies demonstrated that cannabis exposure negatively correlated with [¹¹C]CURB λk₃ in the amygdala, ventral striatum, and medial PFC (Boileau et al., 2016, Jacobson et al., 2021); albeit these regions were not explored in the current study. Considering this evidence, our results of higher FAAH levels in people with co-use may solely reflect the effects of tobacco use on FAAH and not cannabis use. Currently, it remains unknown if tobacco use itself is associated with higher FAAH levels or if greater FAAH levels observed in people with co-use reflect interactions between cannabis use and tobacco use. Indeed, the latter would align with evidence of crosstalk between the eCB and nicotinic systems, which has been extensively documented (Scherma et al., 2016, Valjent et al., 2002; Viveros, 2007). Accordingly, eCB receptors and nicotine acetylcholine receptors may anatomically and/or functionally interact to produce greater FAAH levels than those seen with the use of each individual substance. For example, tobacco use may sensitize the eCB system to cannabis’ effects (Ponzoni et al., 2019, Sved et al., 2023), increasing demand for anandamide clearance via FAAH.

Our findings hold clinical relevance as we posit that elevated FAAH in people who co-use may underlie their exacerbated symptomatology, such as greater depressive and anxiety symptoms, worse cannabis withdrawal symptoms, and more severe cannabis dependence, relative to people who use cannabis only (de Dios et al., 2009, Do et al., 2024, Gray et al., 2017, McClure et al., 2020, Moore and Budney, 2001, Nguyen et al., 2023, Peters et al., 2014, Rabin et al., 2023, Walsh et al., 2020, Yeap et al., 2023, Yeap et al., 2024). Studies investigating associations between FAAH levels and clinical outcomes support this, as do those linking anandamide levels to clinical outcomes. For example, recent reviews on the topic concluded that overexpression of FAAH might drive depressive and anxiogenic behavior (Petrie et al., 2021, Rafiei and Kolla, 2021). In line with this, studies demonstrate that individuals with decreased FAAH activity are at lower risk for developing cannabis dependence (Tyndale et al., 2007) and exhibit reduced cannabis withdrawal severity following cannabis abstinence (Schacht et al., 2009). Further, negative relationships have been reported between peripheral anandamide levels and depressive and anxiety symptoms (Dlugos et al., 2012, Gao et al., 2020, Hill et al., 2008), implying that lower anandamide levels correspond to greater affective symptomatology. Taken together, high FAAH levels and their corresponding low anandamide levels provide a neurobiological mechanism underlying the poor clinical phenotype associated with cannabis-tobacco co-use.

Several limitations are noteworthy in the current study. First, the sample size was small and thus results should be interpreted as preliminary. While three of the six regions reached significance or close to significance, it is unclear if the same pattern would be observed in other ROIs (PFC, thalamus, hippocampus) with larger samples. Second, participants used cigarettes ad libitum, and we did not control for recency of use in our analyses. Thus, the acute effects of tobacco use may have affected outcomes. Additionally, data on past-year tobacco use and pack-years were not collected, limiting our ability to examine potential associations between these variables and FAAH concentrations. Third, nicotine/tobacco use was assessed only in terms of cigarette smoking, as participants were not asked about other sources of nicotine or tobacco such as electronic cigarettes. As a result, additional nicotine/tobacco exposure from alternative products may have been overlooked. Fourth, we lacked a tobacco-only group and a healthy control group which would be essential to confirm putative interactive effects between cannabis and tobacco use on eCB dysregulation. Fifth, we did not assess clinical outcomes, such as depression and anxiety. Scales evaluating these symptoms would help determine whether elevated FAAH levels predict worse clinical outcomes in individuals with co-use.

Our preliminary findings indicate that people with co-use have greater FAAH levels relative to people with cannabis-only use, and that greater tobacco exposure correlates with elevated FAAH. These results support interactions between tobacco use and the eCB system, in vivo, in brain. Future studies should compare [¹¹C]CURB binding in the brain in well-powered samples of individuals with cannabis-tobacco co-use, cannabis-only use, tobacco-only use, and no substance use. This would help clarify whether elevated FAAH levels result from tobacco use alone or the combined use of cannabis and tobacco. Lastly, these studies should also investigate whether FAAH and anandamide levels predict clinical outcomes associated with co-use. This research might shed light on potential mechanisms underlying the heightened negative clinical outcomes associated with co-use. Consequently, regulating FAAH may hold great promise as a novel therapeutic target for the treatment of cannabis-tobacco co-use.

CRediT authorship contribution statement

Joseph Farrugia: Writing – review & editing. Rabin Rachel: Writing – review & editing, Writing – original draft, Formal analysis, Conceptualization. Romina Mizrahi: Writing – review & editing, Funding acquisition. Ranjini Garani: Data curation. Pablo Rusjan: Writing – review & editing, Formal analysis.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

Supplementary material