A review of the effects of acute and chronic cannabinoid exposure on the stress response

Abstract

While cannabis has been used for centuries for its stress-alleviating properties, the effects of acute and chronic cannabinoid exposure on responses to stress remain poorly understood. This review provides an overview of studies that measured stress-related endpoints following acute or chronic cannabinoid exposure in humans and animals. Acute cannabinoid exposure increases basal concentrations of stress hormones in rodents and humans and has dose-dependent effects on stress reactivity in humans and anxiety-like behavior in rodents. Chronic cannabis exposure is associated with dampened stress reactivity, a blunted cortisol awakening response (CAR), and flattened diurnal cortisol slope in humans. Sex differences in these effects remain underexamined, with limited evidence for sex differences in effects of cannabinoids on stress reactivity in rodents. Future research is needed to better understand sex differences in the effects of cannabis on the stress response, as well as downstream impacts on mental health and stress-related disorders.

1. Introduction

Stress is a prevalent issue in North America that places a heavy financial burden on taxpayers and healthcare providers. Over 75% of adults in the United States (US) report physical or emotional symptoms of stress, including headache, fatigue, or changes in sleeping habits (American Psychological Association [APA], 2019). Stress is estimated to cost the US $300 billion per year and can shorten the life expectancy of those under high levels of chronic stress by 10–15 years (APA, 2017). As the prevalence of stress in modern society continues to rise, so too does the prevalence of cannabis use. Accordingly, the most cited reason for habitual cannabis use is to cope with stress (Hyman & Sinha, 2009). Moreover, the shifting social and political landscape surrounding cannabis in North America has been met with reductions in perceived harm (Carliner et al., 2017; Okaneku et al., 2015; Schuermeyer et al., 2014; Stolzenberg et al., 2016) and increased availability of a variety of cannabis products. As a result, daily cannabis use as a stress coping strategy is becoming increasingly common (Mauro et al., 2018). Meanwhile, the impacts of cannabis use on the stress response are not well understood. Thus, the primary objective of this paper is to review the effects of acute and chronic cannabinoid exposure on stress-related endpoints in both humans and rodent models. In doing so, we aim to explore sex differences in these effects and discuss their potential implications for mental health.

1.1. The Endocannabinoid System Regulates the Stress Response

The endocannabinoid (ECB) system has recently emerged as a fundamental component of the neuroendocrine and behavioral response to stress. This system is generally comprised of two primary endogenous ligands, anandamide (AEA) and 2-arachidonoylglycerol (2-AG), which exert their effects via activation of type-1 (CB1) and type-2 (CB2) cannabinoid receptors (Katona & Freund, 2012). CB1 receptors are abundantly expressed in the central nervous system (CNS), as well as in the liver, intestines, vasculature, and muscle tissues (Devane et al., 1988), whereas CB2 receptors are expressed primarily in peripheral immune cells and at much lower concentrations in the brain, where they are typically expressed on microglia (Atwood & Mackie, 2010; Munro et al., 1993). Importantly, CB1 receptors are also expressed on major endocrine glands and at all levels of the hypothalamic-pituitary-adrenal (HPA) axis in both humans and rodents (see Hillard et al., 2016 for a review). Specifically, CB1 receptors are densely expressed in regions that regulate emotional and neuroendocrine stress responses, including the amygdala, hippocampus, and prefrontal cortex (PFC) (Herkenham et al., 1991; Glass et al., 1997). CB1 receptor density is relatively low in the hypothalamus compared to other brain regions (Herkenham et al., 1991). However, these receptors play a critical role in mediating rapid feedback inhibition of the HPA axis locally within the PVN (Evanson et al., 2010) by regulating excitability of glutamatergic parvocellular cells via a mechanism involving CORT-induced mobilization of 2-AG (Di et al., 2003; 2009; Malcher-Lopes et al., 2006). Δ⁹-tetrahydrocannabinol (THC), the primary psychoactive constituent of cannabis, exerts its effects largely via activation of CB1 receptors, where it binds as a partial agonist with a binding affinity comparable to AEA (Devane et al., 1988).

Accumulating evidence from rodent studies indicates that AEA signaling in the basolateral amygdala (BLA) critically constrains activation of stress-responsive neurons and inhibits the expression of emotional behaviors under unstressed conditions (Gray et al., 2015). Upon encountering a stressor, corticotropin-releasing hormone receptor-1(CRHR1) activation in the BLA triggers an increase of fatty acid amide hydrolase (FAAH) activity, which is the primary enzyme responsible for AEA hydrolysis (Deutsch et al., 2002). The CRHR1-mediated increase in FAAH activity reduces AEA signaling in the BLA, which is thought to be a “gatekeeper” of HPA axis activation (Gray et al., 2015). This decrease in AEA reduces inhibition onto glutamatergic BLA neurons that further drive the HPA axis response to stress and the downstream release of glucocorticoids (Gunduz-Cinar et al., 2013; Yasmin et al., 2020). Glucocorticoids are then essential for negatively regulating HPA axis activation via actions that involve the ECB system. Specifically, corticosterone (CORT) binds to membrane-bound glucocorticoid receptors in the PVN and medial PFC, which causes the synthesis and mobilization of 2-AG (Hill & Tasker, 2012). In the prelimbic region of the PFC, the increase in 2-AG activates CB1 receptors on GABAergic inputs to glutamatergic pyramidal neurons, thereby disinhibiting these projections that serve to promote recovery to baseline upon cessation of stress (Hill et al., 2011). Thus, the ECB system is not only the biological target for exogenous cannabinoids including THC, but it is also intimately linked to activation and termination of the HPA axis stress response (Hill & McEwen, 2010; Patel et al., 2004; Patel & Hillard, 2008).

Human studies provide further evidence that acute stress recruits the ECB system. For instance, Dlugos et al. (2012) found that acute social stress increased circulating AEA concentrations in healthy participants, but baseline concentrations of AEA were negatively correlated with baseline ratings of anxiety. In a study by Choukèr et al. (2010), AEA concentrations were similarly shown to increase in individuals that did not experience symptoms of motion sickness in response to parabolic flight maneuvers, but those that did experience acute motion sickness demonstrated lower AEA levels and higher perceived stress scores (Choukèr et al., 2010). Correlations have been documented in women subjected to electrical shock administration following acute aerobic exercise, such that those with greater exercise-induced increases in plasma 2-AG (but not AEA) had greater reductions in fear and anxiety ratings (Crombie et al., 2021). Similarly, exposure to a multidimensional stressor increased salivary levels of 2-AG (but not AEA) in healthy participants (Ney et al., 2021). These inconsistencies in the ECB recruited (AEA vs. 2-AG) are likely due to the time point at which ECBs were measured, the biological source for ECB quantification (plasma vs. saliva), and/or the nature of the stressor used (psychological, physiological, multidimensional). Nevertheless, these findings indicate that recruitment of ECBs under conditions of stress likely serves a protective role in healthy individuals.

Consistent with findings suggesting that ECBs are protective against stress, low levels of circulating ECBs have been associated with stress-related disorders in humans (Hillard, 2018). Concentrations of circulating 2-AG are decreased in women diagnosed with major depression and are negatively correlated with the duration of the depressive episode (Hill et al., 2008). While AEA was not associated with major depression in this sample of women, there was a strong negative correlation between serum AEA content and ratings for cognitive and somatic anxiety on the Hamilton Rating Scale for Depression (Hill et al., 2008). A subsequent study found decreased basal serum concentrations of 2-AG and AEA in women with major depression (Hill et al., 2009). Circulating concentrations of 2-AG were also found to be significantly reduced in individuals with PTSD following exposure to the 9/11 World Trade Center attacks, whereas AEA concentrations were negatively correlated with the degree of intrusive symptoms indicated by these individuals (Hill et al., 2013). Similarly, Wilker et al. (2016) found strong negative relationships between hair concentrations of AEA and other N-acyl-ethanolamides and PTSD symptom severity in a small sample of rebel war survivors from Uganda. However, it should be noted that some have failed to demonstrate relationships between ECB concentrations and symptoms of PTSD (Croissant et al., 2020), while others have even detected higher plasma concentrations of AEA and 2-AG in trauma-exposed patients with PTSD relative to healthy controls and trauma-exposed patients without evidence of PTSD, respectively Hauer et al., 2013). Nevertheless, the collective evidence indicates that an ECB deficiency can increase vulnerability to stress-related psychopathology.

In rodents, genetic and pharmacological disruption of CB1 receptor signaling also leads to increased depressive-like behaviors (see Gorzalka & Hill, 2011 for a review) and can enhance stress-related anxiety (Haller et al., 2002; Navarro et al., 1997). Conversely, higher levels of AEA due to a genetic deficiency in FAAH-mediated AEA degradation is associated with lower levels of anxiety (Habib et al., 2019; Hariri et al., 2009) and facilitates fear extinction (Mayo et al., 2018) in humans. Moreover, male mice that express a polymorphism in the FAAH 385A allele that results in higher levels of AEA exhibit decreased anxiety-like behavior (Dincheva et al., 2015) and are protected against swim stress-induced reductions in AEA content in the amygdala and PFC (Mayo et al., 2018). Interestingly, a recent double-blind placebo-controlled study has shown that 10-day treatment with a FAAH inhibitor significantly potentiated recall of fear extinction memories and attenuated autonomic stress reactivity and stress-induced negative affect in healthy adults (Mayo et al., 2020). Thus, the majority of studies indicate an important protective role for the ECB system in regulating stress and stress-related behaviors, making it an attractive target for the treatment of stress-related illness.

1.2. Sex Differences in Cannabis Use, the Stress Response, and the ECB System

There are marked sex differences in rates of cannabis use, metabolism of THC, and the prevalence of stress-related disorders. For instance, in animal studies, female rodents are more sensitive to the reinforcing properties of cannabinoids (Fattore, 2012; Fattore et al., 2009, 2010) and will self-administer higher quantities of the drug relative to males (Glodosky et al., 2020). Further, female rats produce more of the major active metabolite 11-OH-THC (Britch et al., 2017; Narimatsu et al., 1991; Ruiz et al., 2020; Tseng et al., 2004; Wiley & Burston, 2014). There have been similar findings in humans; after oral administration of THC, women had significantly higher 11-OH-THC levels compared to men, although it was present for a shorter duration (Nadulski et al., 2005). Increased production of the 11-OH-THC metabolite is important because like THC, 11-OH-THC is psychoactive, but it is 2–7x more potent in producing cannabinergic effects and diffuses more readily into the brain than THC (Grotenhermen, 2003; Lemberger et al., 1972; Perez-Reyes et al., 1972). Therefore, greater production of 11-OH-THC in females could result in sex-dependent effects that have not been adequately studied to date. Indeed, although women use cannabis less frequently and in lower quantities than men (Cuttler et al., 2016; Reed et al., 2019) they are at greater risk for experiencing the consequences of chronic cannabis use on negative affect. For instance, women who use cannabis are at an increased risk for depression and anxiety compared to men (Patton et al., 2002), and women who are heavy cannabis users report more negative and less positive moods than light users (Lex et al., 1989).

There are also well-known sex differences in all components of the HPA axis (see Bale & Epperson, 2015 for a review). For example, women have higher circulating cortisol and increased CRH receptor activation in response to stress (Gunn et al., 2016). In rodents, female rats exhibit a more robust adrenocorticotropic hormone (ACTH) and CORT response to stress (Young, 1996), and stress-related hormones remain elevated for longer in females than in males (Figueiredo et al., 2002). Additionally, CRHR1-expressing neurons have been shown to be more sensitive to low levels of CRH and less adaptable to high levels of CRH in female rats (Bangasser et al., 2013), which might be a mechanism underlying increased prevalence of stress-related disorders in women (American Psychiatric Association, 2013; Bekker & van Mens-Verhulst, 2007). At the behavioral level, sex differences in stress coping strategies may also exist (Bale & Epperson, 2015), such that females are more likely to engage in passive coping strategies (Hänninen & Aro, 1996; Mueller & Bale, 2008). Many of these sex differences are due to the organizational and activational effects of gonadal hormones (see Oyola & Handa, 2017 for a review). In males, testosterone suppresses HPA axis activation after puberty (Bale & Epperson, 2015), while higher circulating estradiol concentrations in females elevate stress hormone concentrations (Oyola & Handa, 2017). Moreover, changes in hormone concentrations across the menstrual and estrous cycles also affect the stress response. Female rats in the proestrus phase have higher peak ACTH and CORT concentrations following restraint stress compared to those in estrus and diestrus phases (Viau & Meaney, 1991), while in humans, blunted cortisol responses to stress are observed during the luteal phase compared to the follicular phase (Bale & Epperson, 2015).

Recently, sex differences in the function of the ECB system have emerged that could influence the degree to which cannabis use affects the stress response of male and female users (see Ney et al., 2018 for review). For instance, CB1 receptor expression differs considerably by sex (Liu et al., 2020) and ovarian hormone status (Castelli et al., 2014), such that CB1 receptor density is significantly lower in the PFC and amygdala of regularly cycling female rats compared to male and ovariectomized female rats (Castelli et al., 2014). Additionally, sex differences in CB1 receptor signaling in the locus coeruleus may also contribute to differential effects on the stress response, which has recently been discussed by Wyrofsky et al. (2019). Moreover, estradiol administration increases both AEA content (Scorticati et al., 2004) and AEA signaling (Huang & Woolley, 2012). Notably, Woolley and colleagues have recently demonstrated that estradiol mobilizes AEA in the hippocampus in a sex-specific manner, such that only females exhibit estrogen-receptor-alpha dependent AEA biosynthesis that ultimately acts to suppress GABA release from inhibitory terminals that express CB1 receptors (Huang & Woolley, 2012; Tabatadze et al., 2015). Moreover, an estrogen response element on the FAAH gene has been shown to negatively regulate FAAH transcription, which could result in more pronounced differences in AEA content between sexes (Waleh et al., 2002). Therefore, circulating estradiol levels could influence CB1 receptor expression and activity of the ECB system, which could then result in differential neuroendocrine and behavioral responses to stress.

Altogether, these differences highlight the need to account for sex and hormonal status in stress research. Research on the acute and chronic effects of cannabinoid exposure on the stress response in both sexes is imperative to better understanding how cannabis use could differentially impact susceptibility to stress-related disorders in men and women.

1.3. Objectives of Current Review

There are numerous reviews concerning the actions of ECBs and their relationship to stress (Hill & Tasker, 2012; Hillard et al., 2016; Micale & Drago, 2018; Morena et al., 2016; Murphy et al., 1998) and anxiety (Bedse et al., 2020; Lisboa et al., 2017; Patel et al., 2017; Petrie et al., 2021) as well as many reviews of the literature on exogenous cannabinoids and anxiety (Bahji et al., 2020; Crippa et al., 2009; Petrie et al., 2021; Raymundi et al., 2020; Viveros et al., 2005). Further there has been one review of the acute and chronic effects of cannabis on HPA-axis functioning in humans (Cservenka et al., 2018). However, few have focused on sex differences and there have been no comprehensive reviews of the acute and chronic effects of exogenous cannabinoids on the stress response in both humans and animal models. Since THC and other cannabinoids bind to CB1 receptors located at each arm of the HPA axis, as well as in brain regions that respond to stressors, there are many sites where exogenous cannabinoid exposure could contribute to alterations in the neuroendocrine stress response. Thus, the remainder of this review will focus on the effects of acute and chronic cannabinoid exposure on the stress-related endpoints (e.g., stress-related hormones, subjective stress/distress, amygdala activity) in humans and animal models, making note of sex differences when possible.

2. Methods

A literature search was conducted using the EBSCOhost, PubMed, and Google Scholar databases to review articles published through March 2021 using combinations of the following search terms: “cannabis,” “THC,” “cannabinoid,” “stress,” “stress reactivity,” “HPA axis,” “sex differences.” After screening these terms for articles based on titles, keywords, and abstracts, 28 articles were identified, and the references of these studies were used to find additional articles. Studies were retained if they met the following criteria: (1) administered cannabis or cannabinoids to human or animal subjects or conducted research on humans or animals chronically exposed to cannabis or cannabinoids, and (2) measured endpoints related to stress including endocrine measurements, functional magnetic resonance imaging (fMRI) of brain regions implicated in stress, self-reported stress or distress. This approach yielded a total of 44 studies that were included in this literature review. An additional 10 articles on acute cannabinoid-induced alterations of anxiety-like behavior were added to supplement the section on stress reactivity in rodents due to a paucity of research on neuroendocrine reactivity in rodent models.

3. Acute Effects of Cannabinoids on Stress

3.1. Cannabis is Used for Stress Relief

Given that ECBs suppress HPA axis activation and promote stress recovery, it is not surprising that cannabis is used to cope with stress more than any other drug (Green et al., 2003). In fact, stress relief and coping with stress are frequently reported by cannabis users as their primary motivations for use (Buckner et al., 2016; Copeland et al., 2001; Crippa et al., 2009; Green et al., 2003; Hyman & Sinha, 2009; Lee et al., 2007; Segal et al., 1982), with daily cannabis users being the most likely to report using cannabis to relax or relieve tension (Copeland et al., 2001; Hyman & Sinha, 2009; Johnston & O’Malley, 1986). Moreover, higher rates of cannabis use are seen during times of distress and in those who experience greater life stress (Copeland et al., 2001; Hyman & Sinha, 2009; Segal et al., 1982), and trying cannabis at a time of psychological distress predicts escalated use (Kaplan et al., 1986).

Surveys of medical cannabis users further support the notion that cannabis is used to manage stress and anxiety. In a large survey of medical cannabis users, the top two reasons for medical cannabis use were to manage pain followed by managing symptoms of anxiety (58%) (Sexton et al., 2016), with women being more likely than men to report using cannabis to cope with anxiety (Cuttler et al., 2016). Similarly, in a survey assessing the efficacy of medical cannabis to treat pain, 50% of those surveyed wrote in to indicate that they also experienced relief from anxiety and stress (Webb & Webb, 2014). Individuals with social anxiety disorder also frequently report using cannabis for coping motives (e.g., Buckner et al., 2007), although research has shown that they also are at greater risk for developing cannabis use disorder (Buckner et al., 2008). Cannabidiol (CBD), the primary non-psychoactive constituent of cannabis, is also frequently used for anxiety and stress relief (Corroon & Phillips, 2018; Wheeler et al., 2020).

Anecdotal reports (Iversen, 2003) and a recent naturalistic study (Cuttler et al., 2018) also provide evidence for cannabis’ acute stress-relieving properties. Cuttler et al. (2018) analyzed data gathered from the Strainprint® app which was designed for medical cannabis users to track symptom changes for a variety of conditions in relation to the quantity and type of cannabis they used. Medical cannabis users reported significant reductions in stress (58%) and anxiety (58%) from before to after inhaling cannabis, with higher doses associated with the largest reductions in stress. While men and women reported comparable decreases in stress following cannabis use, women reported significantly larger decreases in anxiety than did men. While this method is high in ecological validity, as participants could purchase the product desired and use it in their preferred setting, there is also a high probability of sampling bias. Those who continue to use cannabis and report their symptoms in the app are also likely to be those who experience the most symptom relief (Cuttler et al., 2018). Indeed, it has been suggested that those who have positive experiences with cannabis could be more likely to continue to use it for stress relief, while those with aversive experiences discontinue cannabis use (Becker, 1953).

3.2. Acute Effects of Cannabinoids on Stress Hormone Concentrations

3.2.1. Cortisol Concentrations in Humans

Although cannabis users commonly use cannabis to alleviate stress and report substantial stress reduction following use, intrapulmonary and intravenous (IV) cannabinoid administration have consistently been found to increase stress hormone concentrations in humans. For example, vaporized THC inhalation has been found to increase cortisol concentrations compared to placebo. One study found a significant increase in cortisol concentrations after participants inhaled successive doses of 200 and 100 μg/kg THC in vaporized cannabis (De Sousa Fernandes Perna et al., 2016), and another found significant increases in serum cortisol in men after three successive doses of 2, 4, and 6 mg vaporized THC compared to placebo (Kleinloog et al., 2012). Klumpers et al. (2012) also found increased plasma cortisol concentrations after three successive doses (2, 6, and 6 mg) of vaporized THC at 90 min intervals. However, it should be noted that two of these studies (Kleinloog et al., 2012; Klumpers et al., 2012) took several blood samples throughout the testing period (9–10 blood samples), and one was an fMRI study (Klumpers et al., 2012), both of which are factors that could impact cortisol concentrations if they were perceived as stressful by the participants.

Nevertheless, increased cortisol concentrations have also been found after inhaling smoked cannabis. Smoked cannabis raised serum cortisol concentrations compared to unintoxicated baseline levels for male and female frequent users asked to smoke a “socially relevant dose” of their own cannabis (Androvicova et al., 2017), and raised plasma cortisol after smoking a joint containing 2.8% THC compared to placebo in a sample of four men (Cone et al., 1986). However, one other study found no effect of smoking a joint containing 2% THC (18 mg) on plasma cortisol in abstinent cannabis users, although the sample contained only six men (Dax et al., 1989). The inconsistencies between these studies might be explained by small sample size, low THC concentration, and/or lack of female participants.

There have been similar findings in studies using THC delivered IV. Ranganathan et al. (2009) compiled data from two previous studies (D’Souza et al., 2004, 2008) in which healthy control participants and frequent cannabis users were administered THC IV (0.0286, 0.0357, 0.0714 mg/kg THC). Consistent with inhaled cannabis, IV THC led to dose-dependent increases in plasma cortisol concentrations compared to baseline; however, this increase was diminished in frequent cannabis users relative to healthy control participants, which the authors suggest may reflect tolerance to the acute effects of cannabis on cortisol concentrations (Ranganathan et al., 2009).

Two studies that used an oral route of administration had results that are inconsistent with findings from studies that used intrapulmonary or IV routes of administration. Specifically, one study found no differences in plasma cortisol after 12 abstinent cannabis using men were given placebo or 10 mg THC orally three times per day for three days and once on the morning of the fourth day (Dax et al., 1989). Another found no effect of administering 7.5 or 12.5 mg oral THC on salivary cortisol concentrations (Childs et al., 2017).

3.2.2. Corticosterone Concentrations in Rodents

Studies using rodent models have corroborated human studies, revealing significant elevations in ACTH and CORT concentrations following acute cannabinoid administration. THC (50 and 150 μg) administration via intracerebroventricular injection increased serum ACTH and CORT concentrations dose-dependently in male rats (Weidenfeld et al., 1994; Manzanares et al., 1999), 75 μg/kg CP 55,940 IP significantly increased serum CORT in male rats (Marín et al., 2003), and 5 mg/kg THC IP increased serum ACTH and CORT in male rats, although there were greater ACTH elevations in adult rats compared to adolescents (Schramm-Sapyta et al., 2007). THC (0.5 and 1 mg/kg, IV) also dose-dependently increased ACTH concentrations in ovariectomized female rats, and this effect may be mediated by CRH activation because a CRH receptor antagonist abolished this increase (Jackson & Murphy, 1997).

There is evidence that this cannabinoid induced CORT recruitment may occur via a recruitment of monoaminergic neurotransmission. McLaughlin et al. (2009) showed that a 100 μg/kg injection of the CB1 receptor agonist HU-210 reliably increased CORT concentrations in male rats, but pre-treatment with antagonists to 5-HT_1A (WAY100635; 0.5 mg/kg) or 5-HT_2A/C (ketanserin; 1 mg/kg) receptors significantly attenuated the HU-210-induced increase in CORT secretion. Similarly, pre-treatment with antagonists to the alpha-1-adrenoceptor (prazosin; 1mg/kg) and beta-adrenoceptor (propanolol; 2.5 mg/kg) also attenuated this effect, but pre-treatment with antagonists to the NMDA (MK-801; 0.1 mg/kg) and AMPA/kainate (DNQX; 10 mg/kg) receptors had no effect (McLaughlin et al., 2009). Thus, acute administration of cannabinoids likely results in indirect activation the HPA axis via a CRHR1-dependent mechanism involving recruitment of midbrain serotonergic and noradrenergic systems.

3.2.3. Summary of Acute Effects of Cannabinoids on Stress Hormone Concentrations

Collectively, these results which are further summarized in Table 1, clearly demonstrate that THC administration increases cortisol concentrations in humans and CORT concentrations in rodents. Pharmacological studies have revealed a mechanism involving activation of CRHR1 receptors (Jackson & Murphy, 1997) and recruitment of serotonergic and noradrenergic systems that serve to indirectly activate the HPA axis (McLaughlin et al., 2009). However, pharmacokinetics associated with the route of administration appears to play an important role in the ability of THC to increase circulating cortisol concentrations and/or the timing of these effects, as they have not been observed following oral administration. Finally, it is important to note that many of these studies relied predominantly (Childs et al., 2017; De Sousa Fernandes Perna et al., 2016; Klumpers et al., 2012; Ranganathan et al., 2009) or exclusively (Cone et al., 1986; Dax et al., 1989; Kleinloog et al., 2012; Manzanares et al., 1999; Marín et al., 2003; McLaughlin et al., 2009; Schramm-Sapyta et al., 2007; Weidenfeld et al., 1994) on male subjects and none attempted to assess potential sex differences in the acute effects of cannabinoids on stress hormone concentrations.

Table 1.

Effects of acute cannabinoid exposure on stress-related endpoints

Study	Species	Sample Size	Sample Characteristics	Drug Use/Administration	Study Design	Measurements	Findings
Stress Hormone Concentrations
Androvicova et al., 2017	Human	21 (9 F)	Casual cannabis users (at least once/month but not more than 2x/week)	Participants asked to inhale a “socially relevant dose” of their own cannabis All but one participant mixed tobacco with cannabis	Quasi-Experimental	After inhaling cannabis, fMRI was used to measure brain activation in response to erotic pictures and blood samples were collected to measure cortisol	Inhaled cannabis increased serum cortisol
Childs et al., 2017	Human	42 (13 F)	Participants with history of cannabis use (use in past year but not more than 1/week)	0 (n = 13), 7.5 (n = 14), or 12.5 (n = 15) mg oral THC	Quasi-Experimental	Assessed the effects of THC on TSST responses, measured salivary cortisol and self-reported distress	No cortisol diff following oral THC
Cone et al., 1986	Human	4 (0 F)	No female participants Frequent cannabis users	Smoked two cigarettes containing cannabis (2.8% THC) or placebo	Quasi-Experimental	Measured plasma cortisol before and after each session	Increased plasma cortisol following
Dax et al., 1989	Human	23 (0 F)	No female participants Occasional to heavy cannabis users (3 cannabis cigarettes/day for 6 months) following 2 weeks abstinence	10 mg oral THC or placebo (n = 12) 18 mg THC in smoked cannabis cigarette (n = 6)	Quasi-Experimental	Administered THC to participants 3x/day for 3 days and once on the fourth day; measured plasma cortisol	No ACTH or cortisol diff following oral or smoked THC
De Sousa Fernandes Perna et al., 2016	Human	61 (26 F)	Ages 18–28 Heavy alcohol (n = 20), regular cannabis users (n = 21; used 3–9x/week in past year), and controls (n = 20)	Participants received a total of 300 μg inhaled THC/kg bodyweight in two successive doses of 200 and 100 μg THC/kg	Quasi-Experimental	Received single doses of alcohol, THC, or placebo before aggression exposure; measured serum cortisol at baseline, before, and after exposure	Inhaled THC increased serum cortisol
Jackson & Murphy, 1997	Sprague Dawley Rat	10 F/group	Ovariectomized female rats No male subjects	0, 0.5, or 1.0 mg/kg IV THC	Experimental	THC was administered following ovariectomy and blood collected for plasma ACTH	0.5 and 1.0 mg/kg THC dose-dependently increased plasma ACTH compared to vehicle in ovariectomized female rats
Kleinloog et al., 2012	Human	49 (0 F)	Ages 18–45 Mild cannabis users (1/week or less in past year) No female participants	Inhaled vaporized THC in three doses of 2, 4, and 6 mg at 90 min intervals	Quasi-Experimental	Administered 10 mg olanzapine and/or THC to participants and collected serum cortisol	Inhaled THC increased serum cortisol
Klumpers et al., 2012	Human	12 (3 F)	Ages 18–45 Used cannabis no more than 1/week in the past year	Inhaled 2, 6, and 6 mg vaporized THC or placebo at 90 min intervals	Quasi-Experimental	Administered THC before fMRI scans on two occasions; collected blood before and after each dose for plasma cortisol	Inhaled THC increased plasma cortisol
Manzanares et al., 1999	Sprague Dawley Rat	5–7/group (0 F)	No female subjects	25, 50, or 100 μg/kg ICV THC	Experimental	Plasma ACTH and CORT at 30 min post-injection for each dose, or plasma ACTH and CORT at 0, 30, 60, 120, 240, and 480 min post-injection following 50 μg/kg ICV THC	THC administration dose-dependently increased plasma concentrations of ACTH and CORT, with peak concentrations occurring at 60 min post-infusion following 50 μg/kg ICV THC
McLaughlin et al., 2009	Sprague Dawley Rat	4–6/group (0 female)	No female subjects	0.1 mg/kg IP HU210	Experimental	Plasma CORT at 45 min post-injection	HU-210 injection increased plasma CORT in a serotonin and noradrenaline receptor-dependent manner
Marín et al., 2003	Wistar Rat	98 (0 F)	No female subjects	10 or 75 μg/kg IP CP 55,940	Experimental	Tested interactions between opioid-receptor antagonists and 10 or 75 μg/kg CP 55,940 on adrenocortical activity	75 mg/kg IP CP 55,940 increased serum CORT
Ranganathan et al., 2009	Human	76 (19 F)	Frequent cannabis users (n = 40) and healthy controls (n = 36) Data pooled from D’Souza et al., 2004, 2008	0, 0.0357 and 0.0714 mg/kg IV THC in study 1 0, 0.0286 mg/kg IV THC in study 2	Experimental	Measured plasma cortisol before and 70 min after IV THC delivery	IV THC increased plasma cortisol Frequent cannabis users displayed a blunted cortisol increase
Schramm-Sapyta et al., 2007	CD Rat	10–12/group (0 F)	Adolescent (PND 28) and adult (PND 64–66) rats No female subjects	5 mg/kg IP THC	Experimental	Tested the effects of THC in adolescent and adult rats and measured serum ACTH and CORT	THC increased serum ACTH and CORT; adults had higher ACTH levels after THC than adolescents
Weidenfeld et al., 1994	Hebrew University Rat	8/group (0 F)	No female subjects	0, 25, 50, or 150 μg ICV THC	Experimental	Measured serum ACTH and cortisol 90 min after ICV THC administration	50 and 150 μg ICV THC increased serum ACTH and CORT
Stress Reactivity
Bambico et al., 2012	Sprague-Dawley Rat	n = 7–9/group	No female subjects	1 mg/kg IP THC 3 mg/kg SR141716A	Experimental	Administered THC and/or SR141716A before testing with EPM, FST, light-dark box	Chronic (but not acute) treatment with THC increased active coping behaviors in the FST; this effect was reversed by SR141716A
Berrendero & Maldonado, 2002	CD1 Mouse	N = 38	No female subjects	0.3 mg/kg THC 0.5 mg/kg SR141716A	Experimental	Administered THC and/or SR141716A before the light-dark box test	THC reduced anxiety-like behaviors in the light-dark box; pretreatment with SR141716A blocked this effect
Bossong et al., 2013	Human	11 (0 F)	Used cannabis incidentally (using < 1x/week) No female participants	Inhaled 6 mg followed by three doses of 1 mg vaporized THC at 30 min intervals	Experimental	Administered THC to participants before an fMRI emotional processing task	THC led to poorer negative face-matching and reduced overall brain activity to fearful faces
Braida et al., 2007	Sprague-Dawley Rat	n = 10/group	No female subjects	0.075–0.75 mg/kg IP THC 0.75–1.25 mg/kg IP AM404	Experimental	Administered THC or AM404 before EPM	Both THC and AM404 reduced anxiety-like behaviors in the EPM
Childs et al., 2017	Human	42 (13 F)	Participants with history of cannabis use (use in past year but not more than 1/week)	0 (n = 13), 7.5 (n = 14), or 12.5 (n = 15) mg oral THC	Experimental	Assessed the effects of THC on TSST responses, measured salivary cortisol and self-reported distress	7.5 mg THC dampened distress ratings, 12.5 mg increased distress ratings
Marín et al., 2003	Wistar Rat	N = 81	No female subjects	10 or 75 μg/kg IP CP 55,940	Experimental	Used holeboard and EPM to test the effects of CP 55,940 on anxiety-like behaviors	75 μg/kg CP 55,940 increased anxiety-like behaviors in the holeboard and EPM
Marco et al., 2004	Wistar Rat	N = 60	No female subjects	1 or 50 μg/kg IP CP 55,940	Experimental	Administered CP 55,940 and tested anxiety-like responses in the EPM and holeboard	1 μg/kg reduced anxiety-like behaviors, 50 μg/kg increased anxiety-like behaviors
Onaivi et al., 1990	Sprague-Dawley Rat, ICR Mouse	n = 12/group	No female subjects	0.3–10 mg/kg IP THC for rats 1–20 mg/kg IP THC for mice	Experimental	Tested the effects of THC in rats and mice with the EPM	1–10 mg/kg increased anxiety-like behaviors in rats 10 and 20 mg/kg increased anxiety-like behaviors in mice
Patel et al., 2005	ICR Mouse	12–18/group (0 F)	No female subjects	2.5 mg/kg IP THC 0.3 mg/kg IP CP 55,940 10 mg/kg IP AM404 5 or 10 mg/kg IP SR141716A	Experimental	Administered THC, CP 55,940, AM404, or SR141716A before restraint stress and measured c-Fos expression in the BLA and CeA	THC and CP 55,940 alone increased c-Fos expression in the CeA (not BLA), and this effect was synergistic with the effect of restraint stress AM404 led to an additive interaction with restraint stress in the CeA SR141716 dose-dependently increased c-Fos in the BLA and CEA
Patel & Hillard, 2006	ICR Mouse	N = approx. 300	No female subjects	0.001–0.3 mg/kg IP CP 55,940 0.3–10 mg/kg IP WIN 55,212–2 0.25–10 mg/kg IP THC 1–10 mg/kg IP SR141716 1–10 mg/kg IP AM251	Experimental	Measured the effects of CP 55,940, WIN 55,212–2, THC, SR141716, and AM251 on anxiety-like behaviors with the EPM	Low doses of CP 55,940 and WIN 55,212–2 decreased anxiety-like behaviors THC dose-dependently increased anxiety-like behaviors SR141716 and AM251 dose-dependently increased anxiety-like behaviors
Phan et al., 2008	Human	16 (8 F)	Recreational cannabis users (used at least 10x in their life but did not use daily)	0 or 7.5 mg oral THC	Quasi-Experimental	Administered THC before completing an emotional face processing task while undergoing fMRI scans	THC dampened emotional arousal, amygdala reactivity, and impaired threatening face recognition THC increased amygdala activity in control condition
Rabinak et al., 2020	Human	71 (36 F)	Trauma-exposed adults without PTSD (n = 27), with PTSD (n = 19), and healthy controls (n = 25)	0 or 7.5 mg oral THC	Quasi-Experimental	Received THC or placebo before completing a threat processing paradigm during fMRI imaging	THC attenuated threat-related amygdala activation, increased PFC/adjacent rostral cingulate cortex activation, and increased corticolimbic functional connectivity to threat
Rey et al., 2012	C57BL/6 N Mouse	N = 88 (43 F)	Glu-CB1-KO GABA-CB1-KO	1 or 50 μg/kg IP CP 55,940	Experimental	Assessed the effects of CP 55,940 on anxiety in mice with CB1-KO mice in GABAergic or glutamatergic neurons	CB1 receptors on cortical glutamatergic terminals mediated anxiolytic-like effects of 1 μg/kg CB1 receptors on GABAergic terminals mediated anxiogenic-like effects of 50 μg/kg
Rubino et al., 2007	Sprague-Dawley Rat	n = 10/group	No female subjects	0.015–3 mg/kg IP THC 3 mg/kg IP AM251	Experimental	Measured anxiety-like responses to THC and AM251 with the EPM and cFos expression	0.015–0.75 mg/kg THC reduced anxiety-like behaviors THC reduced cFos in PFC and amygdala These effects were reversed by pretreatment with AM251
Valjent et al., 2002	CD1 Mouse	n = 10/group	No female subjects	0.03, 0.1, 0.3, 1, 2.5, and 5 mg/kg IP THC	Experimental	Injected THC before testing for anxiety-like behavior in the light-dark box	0.3 mg/kg decreased anxiety-like behaviors 5 mg/kg increased anxiety-like behaviors

3.3. Acute Effects of Cannabinoids on Stress Reactivity

3.3.1. Stress Reactivity in Humans

Acute THC administration has dose-dependent effects on responses to acute stressors (i.e., stress reactivity) in humans. In a randomized, double-blind, placebo-controlled experiment, Childs et al. (2017) gave placebo, 7.5, or 12.5 mg of oral THC to participants of both sexes before administering the Trier Social Stress Test (TSST; Kirschbaum et al., 1993) or a non-stressful control task. Participants who received the low dose of THC reported lower subjective distress in response to the TSST compared to placebo, while those who were given the high dose reported higher levels of subjective distress compared to the low dose and placebo during both the control (no stress) and TSST sessions. Participants receiving 7.5 mg dose of THC also reported that the TSST was less stressful and challenging compared to those receiving placebo, but 12.5 mg THC had no effect on this measure. Nevertheless, there were no differences in salivary cortisol stress reactivity after the task (Childs et al., 2017).

A larger number of studies have examined the acute effects of low doses of THC on activity in brain regions associated with stress, with the amygdala emerging as a primary contributor to these effects. In studies with recreational cannabis users, 7.5 mg of oral THC dampened emotional arousal and amygdala reactivity and reduced recognition of threatening faces expressing anger or fear (Phan et al., 2008). Viewing threatening faces is known to enhance amygdala reactivity and also elicits exaggerated amygdala reactivity for individuals with anxiety (Rauch et al., 2000; Whalen et al., 1998). In another fMRI study, 11 male participants received 6 mg of vaporized THC followed by three doses of 1 mg at 30-minute intervals before completing an emotional processing task that involved matching happy or fearful faces (Bossong et al., 2013). THC administration resulted in poorer accuracy for matching negative (fearful), but not positive (happy) emotional content compared to placebo. Moreover, the task led to activation of the amygdala and other brain areas associated with stress (hippocampus, PFC), but after inhaling THC, overall brain activity was reduced when viewing fearful faces. Rabinak et al. (2020) also conducted an fMRI study in participants with and without a diagnosis of PTSD after being administered THC or a placebo. Using a similar emotional face processing task, participants were randomly assigned to receive either 7.5 mg oral THC or placebo. In those that received THC, fMRI imaging revealed attenuated threat-related amygdala activation and increased medial PFC/adjacent rostral cingulate cortex activation (Rabinak et al., 2020). These areas are important to activation and inhibition of the stress response, respectively, indicating that THC may dampen stress reactivity by both inhibiting areas that promote stress response activation and activating areas that inhibit it (Rabinak et al., 2020).

3.3.2. Stress Reactivity in Rodents

Surprisingly, no studies to date have directly assessed differences in neuroendocrine reactivity to a stressor following acute cannabinoid exposure in rodents. Instead, researchers have measured stress-induced neuronal activation or behavioral coping strategies as proxies of stress reactivity. In one study, male mice were injected with 2.5 mg/kg THC IP, 0.3 mg/kg CP 55,940 IP, or subjected to a 30 min acute restraint stress, and each of these challenges by themselves produced barely detectable c-fos immediate early gene expression in the central nucleus of the amygdala (CeA; Patel et al., 2005). However, when cannabinoids were administered before the stressor, there was robust c-Fos expression in the CeA, which suggests that the effects of cannabinoids and stress on CeA activation may be synergistic (Patel et al., 2005).

A much larger body of literature that has been thoroughly reviewed by others (e.g., see Petrie et al., 2021) has demonstrated acute cannabinoid-induced alterations in behavioral stress reactivity using preclinical assays designed to evaluate anxiety-like behavior. This literature generally indicates important dose-dependent effects of cannabinoids that is in line with what has been observed in human studies. Specifically, preclinical studies indicate that THC administration reduces anxiety-like behavior in rodents at low doses. Several studies have shown that 0.075 – 1 mg/kg THC administered via intraperitoneal (IP) injection increases the amount of time spent exploring the open arms of the elevated plus maze (EPM) in male rats (Bambico et al., 2012; Braida et al., 2007; Rubino et al., 2007) and male mice (Berrendero & Maldonado, 2002). This range of doses also reduced anxiety-like behaviors in male rodents the light-dark box (Bambico et al., 2012; Berrendero & Maldonado, 2002) by increasing time spent in a lit rather than dark compartment. Similarly, 0.3 mg/kg THC decreased anxiety-like behaviors in the light-dark box in male mice, while 5 mg/kg THC increased anxiety-like behavior in this assay (Valjent et al., 2002). Using a higher dose range of THC (1–10 mg/kg), Patel and Hillard (2006) demonstrated that IP THC administration decreased open arm exploration in the EPM in male mice (Patel & Hillard, 2006). Finally, Onaivi et al. (1990) administered 0.3 – 10 mg/kg THC IP to male rats and 1 – 20 mg/kg to male mice and found increased anxiety-like behavior in the EPM for rats injected with 1 – 10 mg/kg THC and mice that received 10 and 20 mg/kg THC.

Similar dose-dependent effects have been observed following acute synthetic cannabinoid administration. For example, while Patel and Hillard (2006) observed a robust decrease in anxiety-like behavior in male mice following 10 and 30 μg/kg injections of CP 55,940, Marín et al. (2003) showed that a 10 μg/kg dose of CP 55,940 had no effect in the EPM, while a 75 μg/kg dose significantly increased anxiety-like behavior in male rats. Accordingly, the synthetic CB1 receptor agonist WIN 55,212–2 increased the percentage of time spent in open arm exploration in the EPM at 1 and 3 mg/kg, whereas 10 mg/kg had the opposite effect in male mice (Patel & Hillard, 2006). Moreover, Marco et al. (2004) showed that a low dose of CP 55,940 (1 μg/kg) increased open arm exploration in the EPM, while a high dose (50 μg/kg) decreased open arm exploration and decreased head-dipping duration in the hole board test in male rats.

These dose-dependent biphasic effects of cannabinoids on anxiety-like behavior may be attributed to differential CB1 receptor activation on forebrain glutamatergic vs. GABAergic neurons. Rey and colleagues have elegantly shown that conditional deletion of CB1 receptors on cortical glutamatergic neurons abolishes the anxiolytic effect of low-dose CP55,940 administration, while conditional deletion of CB1 receptors on forebrain GABA receptors abolishes the anxiogenic effect of high-dose CP55,940 administration (Rey et al., 2012). In the BLA, CB1 receptors are expressed on both glutamatergic and GABAergic neurons, and activation of CB1 receptors specifically on BLA glutamatergic neurons dampens BLA activity and decreases anxiety-like behavior in male mice (Shonesy et al., 2014; Bedse et al., 2018). Thus, activation of CB1 receptors on glutamatergic terminals in the BLA likely contributes to its ability to decrease anxiety-like behavior. On the other hand, administration of high doses of cannabinoids activates CB1 receptors on GABAergic terminals in the BLA which can lead to long-term depression of inhibitory transmission (Azad et al., 2003; Di et al., 2016), thereby further contributing to increased BLA excitability that underlies stress-induced anxiety.

While these studies generally indicate consistent effects of acute cannabinoid exposure in preclinical assays of emotional behavior in male rodents, it should be noted that interpretation of these behavioral outcomes (e.g., EPM exploration) as a direct reflection of complex human traits (e.g., anxiety) can often be misleading and should be interpreted with caution. In addition to altering the stress response, cannabinoids also reliably depress locomotor activity, especially at higher doses, and can alter risk-taking behavior in manner that is distinct from its effects on brain systems that regulate anxiety. For instance, at 10 and 15 mg/kg, THC has been shown to decrease locomotor activity in mice and adolescent rats, respectively, while 5 mg/kg was sufficient to decrease locomotor activity in other studies in mice (Harte-Hargrove & Dow-Edwards, 2012; Valjent et al., 2002). Further, 25 and 50 μg/kg HU-210 led to severe locomotor impairments in stressed mice (Kinden & Zhang, 2015) and 0.3 mg/kg CP 55,940 increased stereotyped behavior in mice and compromised interpretation of EPM results for that dose (Patel & Hillard, 2006). Therefore, researchers assessing anxiety-like behavior in rodents should always strive to measure multiple behavioral outcomes in order to rule out potential effects in other domains that could indirectly influence their endpoint of interest. Finally, researchers assessing these effects should include female rodents as none of the reviewed studies included female rodents and therefore potential sex differences in these effects are entirely unknown.

3.3.3. Summary of Acute Effects of Cannabinoids on Stress Reactivity

Taken together, studies examining effects of cannabinoid administration on stress reactivity suggest dose-dependent effects (see Table 1). While low doses of THC reduce distress ratings (Childs et al., 2017), emotional arousal (Phan et al., 2008), and amygdala reactivity (Phan et al., 2008; Rabinak et al., 2020) in humans, high doses increase distress ratings (Childs et al., 2017). This is consistent with studies described above that indicate dose-dependent biphasic effects of cannabinoids on anxiety-like behavior in male rodents. Future research should determine what range of doses are most likely to produce changes in stress reactivity in female rodents and whether these effects differ according to route of administration in either sex. Finally, future research is needed to further examine cortisol reactivity following acute exposure to cannabinoids and stress, as no previous rodent studies on this topic have been conducted and only one study (Childs et al., 2017) has examined these possible effects in humans and it relied on oral THC administration, which has not been demonstrated to have effects on cortisol concentrations following acute exposure.

4. Effects of Chronic Cannabis Use on the Stress Response

4.1. Endocannabinoid Alterations After Chronic Cannabinoid Exposure

Chronic cannabis use is known to cause adaptations in the ECB system, which could directly interfere with HPA axis functioning and alter stress responses. For instance, several studies in rodents and humans have shown that chronic THC exposure is associated with reduced CB1 receptor density throughout the brain. In rodents, chronic THC administration leads to CB1 receptor desensitization and downregulation (Burston et al., 2010; Freels et al., 2020) and decreases CB1 receptor binding (Di Marzo et al., 2000; González et al., 2004; Romero et al., 1997, 1998). There have been similar findings in humans, as male chronic cannabis users have 15–20% lower CB1 receptor availability compared to controls (Ceccarini et al., 2015; D’Souza et al., 2016; Hirvonen et al., 2012), and females with cannabis use disorder have recently been shown to have decreased CB1 receptor availability in specific brain regions (hippocampus, amygdala, cingulate, and insula) compared to female controls (Spindle et al., 2021). Some research also suggests that chronic administration of THC to adolescent rodents has sex-dependent effects. While THC led to similar levels of CB1 receptor downregulation, adolescent female rats had greater CB1 receptor desensitization than males (Burston et al., 2010). Thus, the adolescent female brain may be particularly vulnerable to disruption of CB1 receptor signaling by chronic THC exposure. Nevertheless, these effects appear to be reversible, as one study was unable to find differences in CB1 receptor availability between users and non-users after only two days of monitored abstinence (D’Souza et al., 2016) and another reported that CB1 receptor availability in cannabis users had partially returned to normal after four weeks of monitored abstinence (Hirvonen et al., 2012).

In addition to affecting CB1 receptor density, chronic cannabinoid administration can also alter ECB content. In rodents, daily THC treatment (10 mg/kg for 8 days) was associated with increased AEA content in the limbic forebrain and higher concentrations of 2-AG in the cerebellum, brainstem, and hippocampus (Di Marzo et al., 2000; González et al., 2004). They also found decreased 2-AG and AEA in the striatum and decreased AEA in the midbrain and diencephalon (Di Marzo et al., 2000) and administration of the CB1 receptor inverse agonist SR141716A reversed these changes (González et al., 2004). One study in humans also revealed altered ECB content in heavy chronic cannabis users. Specifically, Morgan et al. (2013) found that heavy chronic cannabis users had significantly higher serum 2-AG concentrations and lower cerebrospinal fluid AEA levels compared to infrequent cannabis users, although neither of these groups differed from non-users. Altogether, these alterations in CB1 receptor density and ECB production could have prolonged effects on the HPA axis stress response in chronic cannabis users, even under non-intoxicated conditions.

4.2. Effects of Chronic Cannabis Use on the Cortisol Awakening Response in Humans

Only a few studies have measured the cortisol awakening response (CAR) in chronic cannabis users, but these studies have consistently found either a blunted CAR or flattened diurnal cortisol slope, both of which indicate a disruption in HPA axis functioning. The first was part of a longitudinal prospective cohort study of Dutch early adolescents followed over two years (Huizink et al., 2006). Subsets of participants were divided into an early onset cannabis user group (9–12), a later onset cannabis user group (13–14), and a nonuser group. The early onset group had a significantly blunted CAR and elevated evening cortisol levels compared to nonusers and late onset cannabis users (Huizink et al., 2006). However, both early- and later-onset users had only used cannabis a few times in their lives (53% of early users and 65% of later onset group had used cannabis only 1–3 times). Therefore, it seems unlikely that these alterations are a consequence of cannabis use, and a blunted CAR in adolescence might instead predict early-onset cannabis use (Huizink et al., 2006).

The second study to evaluate CAR compared waking cortisol concentrations among patients with schizophrenia with or without symptom onset after cannabis use and controls (Monteleone et al., 2014). They found that patients with schizophrenia symptom onset after initiation of cannabis use had a flattened CAR compared to control participants, but there were no significant differences between controls and patients with schizophrenia onset that was not preceded by cannabis use (Monteleone et al., 2014). These findings indicate that cannabis use could be related to cortisol dysregulation in schizophrenia by either contributing to, or being predicted by, cortisol dysregulation in this population.

Finally, Labad et al. (2020) measured CAR and diurnal cortisol patterns in cannabis-using and non-using outpatients with recent-onset psychosis and healthy controls. CAR did not significantly differ between current cannabis users and non-users (regardless of psychosis outpatient status), and the results of a dexamethasone suppression test also did not differ between these groups (Labad et al., 2020). However, by comparing the rate of decline of cortisol levels from morning to evening, they found that cannabis users had a flattened diurnal cortisol slope compared to non-users, which indicates HPA axis dysfunction (Labad et al., 2020). The authors also attempted to examine sex differences in CAR and diurnal cortisol slope, but no significant differences were noted. The lack of a difference in CAR could also be related to the small sample size of cannabis users in this study (n = 19 cannabis users vs. n = 84 non-users), diminishing their power to detect significant effects.

The results of these studies generally point to a blunted CAR and flattened diurnal cortisol slope in chronic cannabis users, which suggests circadian dysregulation of the HPA axis. However, future studies that examine heavy regular cannabis users without a psychotic disorder are needed, since schizophrenia has been associated with HPA axis dysregulation in the absence of cannabis use (see Shah & Malla, 2015 for a review). It is also presently unclear whether a blunted CAR is a product of chronic cannabis use or is rather a predictor of cannabis use and longitudinal studies that track cannabis use and daily cortisol rhythms over time would be better suited for clarifying the nature of this relationship. Finally, while all three of the reviewed studies included female participants, sex differences in the effects of chronic cannabinoid exposure on CAR or diurnal cortisol slopes were only examined in one study (Labad et al., 2020) and should be the subject of future investigations.

4.3. Effects of Chronic Cannabinoid Exposure on Basal Stress Hormone Levels

4.3.1. Basal ACTH and Cortisol Concentrations in Humans

In contrast to findings of a lower CAR, a handful of studies have found that regular cannabis users have elevated basal ACTH (Somaini et al., 2012) and cortisol concentrations (Carol et al., 2017; Chao et al., 2018; Huizink et al., 2006; King et al., 2011; Monteleone et al., 2014; Somaini et al., 2012) compared to non-users. These effects appear to be long-lasting, as ACTH and cortisol has been shown to remain elevated for up to six months after initiation of abstinence (Somaini et al., 2012). Cannabis users in many of these studies reported using cannabis almost daily (Chao et al., 2018; King et al., 2011; Somaini et al., 2012) or at least weekly (Monteleone et al., 2014). One of these studies included participants who were at “ultra-high risk” for developing psychosis and found that this group had elevated basal cortisol compared to healthy controls regardless of cannabis use status (Carol et al., 2017). However, individuals at ultra-high risk for psychosis who regularly used cannabis had higher basal cortisol compared to those that did not use cannabis (Carol et al., 2017). Another study found elevated salivary cortisol concentrations in daily cannabis users who had experienced trauma compared to those who had not, although this study did not include a comparison group of non-users (Chao et al., 2018). Notably, sex differences were examined in this study but no sex differences in salivary cortisol were noted (Chao et al., 2018). While the aforementioned studies found evidence that chronic cannabis users have elevated basal cortisol, several others found no differences in basal cortisol concentrations of cannabis users and non-users (Block et al., 1991; Cloak et al., 2015; Cuttler et al., 2017; Lisano et al., 2019, 2020; Petrowski & Conrad, 2019).

4.3.2. Basal Corticosterone Concentrations in Rodents

Only a few animal studies have measured basal CORT in rodents chronically exposed to cannabinoids. Two studies found no difference in basal CORT concentrations after chronic treatment with CP 55,940 in male and female rats (Biscaia et al., 2003; Llorente-Berzal et al., 2011), while another found no difference in males after chronic treatment with HU-210 (Hill & Gorzalka, 2006). Only one study has detected evidence of increased basal CORT in both male and female rats after self-administration of vaporized cannabis extracts for 30 days, and basal CORT was higher in females than males (Glodosky et al., 2020). However, this increase in basal CORT was also observed in rats that self-administered the vehicle compound not containing cannabis extract (Glodosky et al., 2020). Therefore, the observed elevation in CORT cannot be attributed specifically to the effects of chronic cannabis use per se and it remains possible that the effects of chronic cannabis exposure on basal stress hormone concentrations are due to some extraneous variable associated with cannabis use (e.g., inhalation).

4.3.3. Summary of Effects of Chronic Cannabinoid Exposure on Basal Stress Hormone Levels

As further illustrated in Table 2, there is mixed evidence that chronic cannabinoid exposure is associated with elevations in basal stress hormone concentrations, with seven human studies detecting such effects and six human studies failing to replicate these effects. Part of this discrepancy might be explained by lower levels of cannabis use in participants in studies with null findings compared to those that detected elevated basal cortisol, including weekly use or less (Lisano et al., 2019, 2020; Petrowski & Conrad, 2019), but this was not always the case (Block et al., 1991; Cloak et al., 2015; Cuttler et al., 2017). Some did find that cannabis users had higher levels of subjective stress or anxiety than non-users, despite finding no differences in basal cortisol concentrations (Cloak et al., 2015; Cuttler et al., 2017) which could indicate a disconnect between psychological and physiological stress (Cuttler et al., 2017). Other studies were restricted to physically active participants (Lisano et al., 2019, 2020), which could potentially help to mitigate HPA axis dysregulation. Nevertheless, evidence from animal models that chronic cannabinoid exposure alters basal stress hormones have consistently demonstrated no effects relative to control conditions. Given this pattern of results, it is possible that elevated basal stress hormones represent a risk factor for cannabis use, or that extraneous variables in human studies (e.g., psychosis, trauma, inhalation) are driving the detected associations. In line with this possibility, Somaini et al. (2012) found that both active cannabis users and abstinent users had elevated basal stress hormone concentrations compared to control participants (Somaini et al., 2012).

Table 2.

Effects of chronic cannabinoid exposure on stress-related endpoints

Study	Species	Sample Size	Sample Characteristics	Drug Use/Administratio n	Study Design	Measurements	Findings
CAR
Huizink et al., 2006	Human	1768 (707 F)	Dutch adolescents (age 10–12 at assessment)	Early (9–12), late (13–14), or no cannabis use onset	Prospective cohort study (TRAILS)	Salivary cortisol measured at awakening, 30 min later, and 8 pm	Lower CAR in early onset group compared to late onset and controls
Labad et al., 2020	Human	103 (43 F)	Recent-onset psychosis outpatient non-users (n = 84) and cannabis users (n = 19)	At least weekly cannabis use in users	Quasi-experimental	Salivary CAR, diurnal cortisol slope, dexamethasone suppression test	No CAR diff Flattened diurnal cortisol slope for cannabis users
Monteleone et al., 2014	Human	43 (11 F)	Schizophrenia preceded by cannabis use (n = 16) or not (n = 12)	Cannabis use occurred 0.5 to 10 years before schizophrenia onset	Quasi-experimental	Salivary cortisol measured at awakening and 15, 30, 60 min post-awakening two weeks apart	Lower CAR for those with schizophrenia preceded by cannabis use at both measurements
Basal Stress Hormone Concentrations
Biscaia et al., 2003	Wistar Rat	48 (15 F)	Drug treatment PND 35–45; tested starting PND 75	0.4 mg/kg IP CP 55,940 daily	Experimental	Measured serum CORT following chronic CP 55,940 treatment	No basal CORT diff
Block et al., 1991	Human	149 (56 F)	Ages 18–42 Classified cannabis users as frequent (7+/week), moderate (5–6/week), or infrequent (1–4/week)	Frequent (n = 27), moderate (n = 18), and infrequent (n = 30) cannabis users	Quasi-Experimental	Analyzed morning and afternoon blood samples from cannabis users with radioimmunoassay	No basal cortisol diff in men or women
Carol et al., 2017	Human	75 (38 F)	Adolescence at ultra high-risk for psychosis with (n = 17) or without (n = 26) cannabis use	Cannabis use frequency M = 3 on AUS/DUS scale (0 = no use to 5 = almost daily)	Quasi-Experimental	Three saliva samples collected every 60 min between 8:45 am and 2 pm	Elevated basal cortisol in cannabis users and ultra high-risk adolescents
Chao et al., 2018	Human	125 (23 F)	Ages 18–50 No non-cannabis using controls	Current, heavy cannabis users (≥2 cannabis cigarettes/day, ≥4 days/week)	Quasi-Experimental	TSST conducted in afternoon; salivary cortisol samples collected at baseline and 0, 15, 30, and 90 min post stress	Elevated basal cortisol in cannabis users who reported trauma exposure
Cloak et al., 2015	Human	122 (55 F)	Youth (ages 13–23) with (n = 80) and without (n = 42) cannabis use	Light (4.1±0.4 days/week) and heavy (6.9±0.1 days/week) cannabis users	Quasi-Experimental	Collected salivary cortisol in late morning or afternoon	No basal cortisol diff
Cuttler et al., 2017	Human	82 (43 F)	Ages 20–64	Over half of cannabis users (n = 40) reported using more than once daily	Quasi-Experimental	Administered MAST; collected subjective stress ratings and salivary cortisol before, during, and after stress	No basal cortisol diff
Glodosky et al., 2020	Sprague Dawley Rat	104 (52 F)	n = 13/sex/condition	Self-administered vaporized cannabis extract (69.9% THC) 1 hr daily for 30 days	Experimental	Collected blood for plasma CORT before/after restraint stress pre-and post-cannabis self-administration	Elevated basal CORT following cannabis self-administrat ion Higher basal CORT in females than males
Hill & Gorzalka, 2006	Long Evans Rat	4–8/group (0 F)	No female subjects	Administered 5 or 100 μg IP HU-210 for 12 days	Experimental	Measured basal CORT following chronic treatment with HU-210	No basal CORT diff
Huizink et al., 2006	Human	1768	Dutch adolescents (age 10–12 at assessment)	Early (9–12), Late (13–14), or no cannabis use onset	Prospective cohort study (TRAILS)	Salivary cortisol measured at awakening, 30 min later, and 8 pm	Elevated basal cortisol in early and late onset cannabis users (8 pm)
King et al., 2011	Human	60 (28 F)	Ages 18–45	Cannabis users (n = 30) used 6–7 days/week for 1 year	Quasi-Experimental	Measured salivary cortisol between 10:30 and 11:30 am	Elevated basal cortisol in cannabis users
Lisano et al., 2019	Human	24 (0 F)	Ages 19–39 Physically active sample No female participants	Cannabis users (n = 12) used at least 1/week for past 6 months	Quasi-Experimental	Measured serum cortisol following 12-hour fast	No basal cortisol diff
Lisano et al., 2020	Human	30 (10 F)	Ages 19–39 Physically active sample	Cannabis users (n = 15) used at least 1/week for past 6 months	Quasi-Experimental	Measured serum cortisol following 12-hour fast between 7–9 am	No basal cortisol diff
Llorente-Berzal et al., 2011	Wistar Rat	100 (50 F)	n = 12–16/group	Chronic treatment with 0.4 mg/kg IP CP 55,940 from PND 28–42	Experimental	Measured plasma ACTH and CORT after PPI and EPM following chronic CP 55,940 treatment	No basal ACTH or CORT diff
Monteleone et al., 2014	Human	43 (11 F)	Schizophrenia preceded by cannabis use (n = 16) or not (n = 12)	Cannabis use occurred 0.5 to 10 years before schizophrenia onset	Quasi-experimental	Salivary cortisol measured at awakening and 15, 30, 60 min post-awakening two weeks apart	Elevated basal cortisol for those with schizophre nia preceded by cannabis use
Petrowski & Conrad, 2019	Human	21 (0 F)	Matched groups with panic disorder, cannabis-induced panic disorder, or controls (n = 7/group) No female participants	Cannabis-induced panic disorder group reported rare to occasional cannabis use (>2x/week)	Quasi-Experimental	Administered TSST between 3–6 pm and collected salivary cortisol at baseline, 0, 10, 20, 30, 40, and 50 min post stress	No basal cortisol diff
Somaini et al., 2012	Human	42 (10 F)	Ages 19–32 Healthy controls recruited from hospital staff and university students	n = 28 cannabis users seeking treatment; reported using cannabis for 3–14 years without abstinence	Quasi-Experimental	Half of cannabis users randomly assigned to group A (active users) or B (6-month abstinent users); viewed pleasant/unpleasant images and blood samples were collected to measure serum ACTH and cortisol	Elevated basal cortisol and ACTH in cannabis users, particularly among active users
Stress Reactivity
Chao et al., 2018	Human	125 (23 F)	Ages 18–50 No non-cannabis using controls	Current, heavy cannabis users (≥2 cannabis cigarettes/day, ≥4 days/week)	Quasi-Experimental	TSST conducted in afternoon; salivary cortisol samples collected at baseline and 0, 15, 30, and 90 min post stress	Elevated cortisol in cannabis users who reported trauma exposure compared to non-trauma exposed
Cloak et al., 2015	Human	122 (55 F)	Youth (ages 13–23) with (n = 80) and	Light (4.1±0.4 days/week) and heavy (6.9±0.1	Quasi-Experimental	Collected salivary cortisol in late	No cortisol stress reactivity
			without (n = 42) cannabis use	days/week) cannabis users		morning or afternoon	diff (cortisol decreased following the stressor for all groups)
Cornelius et al., 2010	Human	6 (1 F)	Ages 19–24 Youth with comorbid cannabis dependence and major depression in drug trial of fluoxetine	Participants reported using 1–7 cannabis blunts/day at baseline	Quasi-Experimental	Administered the threat-related amygdala reactivity task during fMRI scans at baseline and 12 weeks later and measured amygdala activation	Increased amygdala reactivity following decreased cannabis use (n = 5) and decreased reactivity following increased use (n = 1)
Cuttler et al., 2017	Human	82 (43 F)	Ages 20–64	Over half of cannabis users (n = 40) reported using more than once daily	Quasi-Experimental	Administered MAST; collected subjective stress ratings and salivary cortisol before, during, and after stress	Subjective and cortisol stress responses were blunted in cannabis users
DeAngelis & al’Absi, 2020	Human	79 (33 F)	Ages 19–66	Cannabis users (n = 45) used at least 15 days/month in the past year	Quasi-Experimental	Measured subjective and affective responses to stress before, during, and after the TSST	Blunted positive affect, state stress, and state anxiety in response to stress in cannabis users
Glodosky et al., 2020	Sprague Dawley Rat	104 (52 F)	n = 13/sex/condition	Self-administered 75, 150, or 300 mg/ml vaporized cannabis extract (69.9% THC) 1 hour per day for 30 days	Experimental	Collected blood to measure CORT before/after restraint stress pre-and post-cannabis self-administration	Blunted CORT stress reactivity in female rats but not males following cannabis exposure
Gruber et al., 2009	Human	30 (2 F)	Matched chronic heavy cannabis users (n = 15) with non-using controls on age, sex, and education	Cannabis users smoked at least: 3000 joints in their lifetime and four of the past seven days	Quasi-Experimental	Viewed masked affective stimuli during fMRI scans	Decreased activity in amygdala, anterior cingulate cortex in cannabis users when viewing masked faces
Hill & Gorzalka, 2006	Long Evans Rat	4–8/group (0 F)	No female subjects	Administered 5 or 100 μg IP HU-210 for 12 days	Experimental	Measured c-Fos induction and CORT reactivity to restraint stress following chronic HU-210 treatment	High dose sensitized CORT response to restraint stress and potentiated c-Fos induction in the CeA
Lee et al., 2014	Sprague Dawley Rat	25 (12 F)	n = 6–7/sex/condition	Administered escalating doses of HU-210 (25, 50, 100 μg/kg) from PND 35–46	Experimental	Measured CORT reactivity to 30 min acute restraint stress (PND 75)	Adolescent HU-210 administration elevated CORT stress response; this was more pronounce d in males than females
Llorente-Berzal et al., 2011	Wistar Rat	100 (50 F)	n = 12–16/group	Chronic treatment with 0.4 mg/kg IP CP 55,940 from PND 28–42	Experimental	Measured plasma ACTH and CORT after PPI and EPM following chronic CP 55,940 treatment	Elevated CORT, ACTH following PPI in males only; impaired PPI in females
Ma et al., 2020	Human	46 (13 F)	Ages 22–35	Cannabis users (n = 23) had lifetime diagnosis of cannabis dependence	Quasi-Experimental	Performed effective (directional) connectivity analysis using fMRI while participants performed an emotional face-matching task	Greater effective connectivity from eft amygdala to hypothalamus and right amygdala to bilateral fusiform gyri
McRae-Clark et al., 2011	Human	87 (29 F)	No non-cannabis-using controls	Participants met DSM-IV criteria for cannabis dependence	Quasi-Experimental	Measured subjective and endocrine (plasma ACTH, cortisol) stress responses before and after the TSST; test sessions began at 1 pm	Stress elevated ACTH and cortisol following stressor in cannabis-dependent participants
Mizrahi et al., 2014	Human	24 (11 F)	Participants at clinical high risk for schizophrenia with (n = 12) or without (n = 12) cannabis use	11 of 12 cannabis users met criteria for cannabis dependence and used daily for at least 2 years	Quasi-Experimental	Performed sensorimotor task with stress condition while undergoing PET scans and measured salivary cortisol every 12 min	No cortisol stress reactivity diff
Petrowski & Conrad, 2019	Human	21 (0 F)	Matched groups with panic disorder, cannabis-induced panic disorder, or controls (n = 7/group) No female participants	Cannabis-induced panic disorder reported rare to occasional cannabis use (>2x/week)	Quasi-Experimental	Administered TSST between 3–6 pm and collected salivary cortisol at baseline, 0, 10, 20, 30, 40, and 50 min post stress	No cortisol stress reactivity diff
Somaini et al., 2012	Human	42 (10 F)	Ages 19–32 Healthy controls recruited from hospital staff and university students	n = 28 cannabis users seeking treatment; reported using cannabis for 3–14 years without abstinence	Quasi-Experimental	Half of cannabis users randomly assigned to group A (active users) or B (6-month abstinent users); viewed pleasant/unpleasant images and blood samples were collected to measure serum ACTH and cortisol	Blunted endocrine, subjective, affective stress responses in active users; increased endocrine reactivity following abstinence compared to active users
Tull et al., 2016	Human	202 (100 F)	Ages 18–60 Patients with and without PTSD admitted to a residential SUD treatment facility	Cannabis-using participants (n = 59) met DSM-IV criteria for cannabis dependence	Quasi-Experimental	Presented an individualized trauma script to participants and assessed negative affect before and after (emotional reactivity), and collected salivary cortisol before and 20 min post-stress	Blunted subjective emotional stress reactivity in participants with cannabis dependence No cortisol stress reactivity diff
van Leeuwen et al., 2011	Human	591 (290 F)	Dutch adolescents (mean age = 16.10) in longitudinal study	n = 204 lifetime cannabis users n = 90 repeated cannabis users (5+ occasions in past year)	Prospective cohort study (TRAILS)	Administered a modified TSST and collected saliva at four timepoints before, during, and after to measure cortisol	Blunted cortisol stress response in lifetime cannabis users and more blunted response in repeated cannabis users

4.4. Stress Reactivity

4.4.1. Stress Reactivity in Humans

In addition to findings that chronic cannabis use is associated with alterations in stress hormones under no stress conditions, there is also emerging evidence of blunted reactivity to acute stress in human cannabis users. For instance, several studies have found attenuated stress responses in male and female chronic cannabis users, including blunted endocrine (Cuttler et al., 2017; Somaini et al., 2012; van Leeuwen et al., 2011), subjective and affective responses (Cuttler et al., 2017; DeAngelis & al’Absi, 2020; Somaini et al., 2012; Tull et al., 2016) and amygdala reactivity (Cornelius et al., 2010; Gruber et al., 2009). More specifically, Cuttler et al. (2017) found blunted cortisol and subjective stress responses in chronic cannabis users compared to non-users in response to a multidimensional stressor with physiological and psychosocial stress components (Maastricht Acute Stress Test; Smeets et al., 2012). While the change in cortisol from before to after the stressor correlated with subjective stress ratings in non-users, this relationship was absent for cannabis users, suggesting a disconnect between subjective and endocrine stress responses (Cuttler et al., 2017).

Similarly, others have found a blunted cortisol response in adolescents who reported repeated cannabis use in the past year or lifetime cannabis use compared to non-users and lifetime tobacco users (van Leeuwen et al., 2011). In response to the stressor, repeated cannabis users also had lower cortisol reactivity than lifetime cannabis and tobacco users, and lifetime cannabis users had lower reactivity than non-using controls and lifetime tobacco users. However, repeated use was defined as using cannabis at least five times in the past year, so it is unlikely that this level of use led to alterations in stress responsivity and may instead be evidence that blunted stress reactivity is a risk factor for cannabis use (van Leeuwen et al., 2011).

Somaini et al. (2012) also found that active cannabis users displayed decreased ACTH, cortisol, and anxiety in response to unpleasant images compared to non-users and 6-month abstinent cannabis users. Similarly, DeAngelis and al’Absi (2020) found blunted affective responses to the TSST in daily cannabis users. More specifically, they found that regular cannabis use was associated with smaller increases in positive affect, state stress, and state anxiety compared to non-using controls (DeAngelis & al’Absi, 2020). Finally, another study found blunted subjective emotional reactivity to a personalized trauma cue between PTSD patients who were cannabis-dependent and non-dependent, although there were no significant differences in cortisol reactivity (Tull et al., 2016). The authors note, however, that this study was conducted with patients from an in-patient substance use disorder treatment facility, so there is a high risk of confounds with other drug use and psychiatric disorders (Tull et al., 2016).

In chronic users, abstinence from cannabis is associated with increased amygdala reactivity to stress. Six months after baseline scans were conducted, participants that decreased their level of cannabis use had a relative increase in amygdala reactivity to threatening faces, while increased cannabis use was associated with decreased reactivity (Cornelius et al., 2010). These effects on amygdala reactivity should be interpreted with caution, however, due to small sample sizes in this study (n = 5 participants decreased cannabis use while n = 1 increased use) (Cornelius et al., 2010). However, in a separate study, six months of cannabis abstinence was also associated with increased cortisol and ACTH reactivity to stress compared to active users when viewing unpleasant images (Somaini et al., 2012). Together, these findings suggest partial recovery of amygdala reactivity and the neuroendocrine stress response after prolonged abstinence, which may be partially due to the recovery of CB1 receptor density.

However, it should be noted that others have found no significant differences in stress reactivity between cannabis users and non-users (Cloak et al., 2015; Mizrahi et al., 2014; Petrowski & Conrad, 2019). In one study, Cloak et al. (2015) found no significant differences in stress reactivity between cannabis users and non-users following the TSST. However, they also reported that there were no significant differences in cortisol levels from before to after the TSST in non-users and the only difference was a general decrease in cortisol across the three timepoints (i.e., pre-stress, post-stress, one-hour post-stress) indicating that the stressor was not particularly effective (Cloak et al., 2015). Petrowski and Conrad (2019) found that participants with panic disorder had significantly lower cortisol reactivity than healthy controls, while cortisol reactivity of those with cannabis-induced panic disorder was not significantly different from those with panic disorder unrelated to cannabis use or healthy controls (Petrowski & Conrad, 2019). Finally, Mizrahi et al. (2014) found no differences in changes in salivary cortisol following exposure to an acute stressor in participants at high risk for schizophrenia with versus without concurrent cannabis use.

Two other studies have found stress hormone elevations in response to a stressor in cannabis users. McRae-Clark et al. (2011) found that cannabis users had significantly higher cortisol and ACTH after a stressor compared to before the stressor, and Chao et al. (2018) found that cannabis users who had been exposed to trauma had significantly higher cortisol concentrations before, during, and after a stress challenge compared to non-trauma exposed cannabis users, which did not vary as a function of sex. However, neither of these studies had a non-cannabis using control group for comparison which limits the ability to determine whether cannabis users’ stress reactivity is dampened relative to non-users.

Finally, there have also been two fMRI studies that measured stress reactivity in response to images of emotional faces. One found that chronic cannabis users had decreased activity in the amygdala and anterior cingulate cortex while viewing masked affective faces (i.e., happy, fearful) while non-users had increased activity in these regions (Gruber et al., 2009). The other compared effective connectivity between control subjects and participants with cannabis use disorder while they performed an emotional face matching task with fearful and angry faces or neutral stimuli (Ma et al., 2020). While the groups were no different in task performance, the researchers found greater effective connectivity in participants with cannabis use disorder between the left amygdala to hypothalamus and right amygdala to bilateral fusiform gyri, both of which were positively correlated with scores on a perceived stress scale. There was also enhanced connectivity from the left ventrolateral PFC to the bilateral fusiform gyri, which was negatively correlated with perceived stress scores. They suggest that the first two effective connectivity differences might be connections underlying the increased risk of stress-related disorders in those with cannabis use disorder, while the third could instead represent a protective mechanism (Ma et al., 2020). Clearly, more neuroimaging studies are needed to evaluate alterations in brain activation in response to emotional stimuli in chronic cannabis users.

4.4.2. Stress Reactivity in Rodents

Only a few studies have measured stress reactivity in rodents chronically exposed to cannabinoids. Specifically, we recently conducted a study to examine whether chronic cannabis vapor self-administration causes alterations in stress reactivity in male and female rats (Glodosky et al., 2020). We found that female rats had a blunted CORT response to acute restraint stress after 30 days of self-administration of vaporized cannabis extracts. Notably, both sexes and three different concentrations of cannabis vapor (75, 150, 300 mg/ml) were used in this study, but the significant effect was only observed in female rats that self-administered the medium concentration of cannabis extract. This was likely because this concentration was the most readily self-administered and produced the highest concentrations of plasma THC in this study (Glodosky et al., 2020). This is significant in that it provides evidence that chronic cannabis exposure can cause diminished stress reactivity in rodents rather than blunted stress reactivity representing a risk factor for cannabis use, as stress reactivity before cannabis exposure was not significantly correlated with rates of self-administration. These results also indicate potential sex differences in chronic cannabis-induced stress alterations because an attenuated stress response was only found in female rats. However, females self-administered significantly more cannabis extract than males in this study, which could also be responsible for this effect (Glodosky et al., 2020).

While this study was the first to detect blunted stress reactivity that has been observed in human cannabis users, it should be noted that others have found heightened stress responses following chronic synthetic CB1 receptor agonist treatment. Hill and Gorzalka (2006) administered a high (100 μg/kg) or low (5 μg/kg) dose of HU-210 IP to male rats for 12 days. After a 30 min acute restraint stress, rats that received the high dose had significantly higher CORT concentrations compared to those that received the low dose or vehicle. The high dose of HU-210 was also associated with potentiated c-Fos induction in the CeA (Hill & Gorzalka, 2006). Chronic CP 55,940 administration also led to elevated CORT and ACTH concentrations in males, but not in females, following stimulus presentation in a prepulse inhibition task (Llorente-Berzal et al., 2011). Additionally, escalating doses of HU-210 during adolescence (PND 35–46; 25, 50, 100 μg/kg) resulted in higher peak CORT following a 30 min restraint stress in adulthood (PND 70), and this effect was more pronounced in male rats than females (Lee et al., 2014).

4.4.3. Summary of Effects of Chronic Cannabinoid Exposure on Stress Reactivity

As detailed above and further summarized in Table 2, a handful of human studies have found attenuated stress responses in male and female chronic cannabis users, including blunted endocrine responses (Cuttler et al., 2017; Somaini et al., 2012; van Leeuwen et al., 2011), subjective stress ratings and affective responses (Cuttler et al., 2017; DeAngelis & al’Absi, 2020; Somaini et al., 2012; Tull et al., 2016), and amygdala reactivity (Cornelius et al., 2010; Grueber et al., 2009). While others have failed to replicate findings of blunted cortisol reactivity in humans, they relied on clinical populations (Mizrahi et al., 2014; Petrowski & Conrad, 2019), had a failed stress manipulation (Cloak et al., 2015), or did not include a group of non-users for comparison (Chao et al., 2018; McRae-Clark et al., 2011).

One animal study has corroborated findings of blunted CORT reactivity in humans following chronic cannabis exposure (Glodosky et al., 2020), with several others indicating heightened stress reactivity following chronic administration of synthetic cannabinoids (Hill & Gorzalka, 2005; Lee et al., 2014; Llorente-Berzal et al., 2011). Importantly, Glodosky et al. (2020) only detected evidence of blunted CORT reactivity in female rodents, while studies that have detected augmented reactivity have used males exclusively (Hill & Gorzalka, 2005; Llorente-Berzal et al., 2011) or detected more pronounced effects in males than females (Lee et al., 2014). This suggests that there may be important sex differences or differences in type of cannabinoid (synthetic cannabinoid vs. THC) chronically administered on cortisol reactivity that will require further investigation.

5. Discussion

5.1. General Summary

Cannabis is used to cope with stress more than any other drug (Green et al, 2003) and there is some evidence that stress ratings are substantially reduced from before to after cannabis use (Cuttler et al., 2018). Nevertheless, the effects of cannabinoid exposure on the stress response are not as straightforward as they may seem, with evidence of dose-, sex- and possibly drug-dependent effects. Specifically, in humans, low doses of THC decrease distress ratings (Childs et al., 2017), emotional arousal (Phan et al., 2008), and amygdala reactivity after a stressor (Phan et al., 2008; Rabinak et al., 2020), while high doses result in elevated distress ratings (Childs et al., 2017). While there is an absence of studies examining differences in neuroendocrine reactivity to a stressor following acute cannabinoid exposure in rodents, the dose-dependent biphasic effects of acute THC exposure found in humans are generally in agreement with preclinical rodent studies of anxiety-like behavior (see Petrie et al., 2021 for review). In line with work conducted by Rey et al. (2012), we posit that these dose-dependent effects are primarily attributed to differential CB1 receptor activation on glutamatergic vs. GABAergic neurons primarily within the amygdala. Low-dose administration of CB1R agonists activates CB1 receptors primarily on glutamatergic neurons in the BLA, thereby dampening excitability of these neurons that critically orchestrate stress-induced anxiety and negative affect (Shonesy et al., 2014; Bedse et al., 2018). Conversely, high-dose cannabinoid administration activates CB1 receptors on GABAergic terminals in the BLA which leads to long-term depression of inhibitory transmission (Azad et al., 2003; Di et al., 2016), and a resultant increase in BLA excitability.

Our review of the literature has further revealed that even in the absence of acute cannabis intoxication, chronic cannabis users exhibit blunted endocrine (Cuttler et al., 2017; Somaini et al., 2012; van Leeuwen et al., 2011), affective (Cuttler et al., 2017; DeAngelis & al’Absi, 2020; Somaini et al., 2012; Tull et al., 2016), and amygdala (Cornelius et al., 2010; Gruber et al., 2009) reactivity to stressful or emotional stimuli. Given the established role for AEA signaling in constraining BLA excitability and stress-induced anxiety in preclinical studies (Grey et al., 2015) and noted similarities between THC and AEA with respect to agonist efficacy and CB1 receptor binding affinity (Devane et al., 1988), it seems logical to speculate that in chronic cannabis users, residual circulating THC and its metabolites may act as an effective substitute for AEA under conditions of stress. Under these conditions, acute stress-induced recruitment of FAAH in the BLA would still result in AEA degradation, but exogenous THC and/or metabolites could still occupy the binding site so that the impact of AEA degradation on the excitability of glutamatergic BLA neurons would be diminished. This would explain why cannabis users exhibit dampened amygdala recruitment in response to emotionally laden stimuli and why this typically coincides with blunted neuroendocrine and affective responses to stressors. Alternately, the ability of chronic cannabinoid exposure to increase AEA content throughout the limbic forebrain may also help to limit activation of the stress response during exposure to stress. However, future studies are needed to systematically evaluate these hypotheses.

Only one rodent study has replicated the blunted stress response found in human chronic cannabis users (Glodosky et al., 2020), with a small number of other studies indicating that cannabinoid exposure increases stress reactivity. Specifically, animal studies have revealed increased activation of stress-related brain regions following acute (Patel et al., 2005) and chronic (Hill & Gorzalka, 2006) administration of a synthetic cannabinoid as well as elevations in ACTH (Llorente-Berzal et al., 2011) and CORT (Lee et al., 2014; Llorente-Berzal et al., 2011) following acute stress. While findings in human and rodent studies appear somewhat contradictory, there are several factors that could potentially explain these differences. First, differences may be related to the relative effects of synthetic cannabinoids compared to THC. Studies that found heightened stress reactivity used synthetic cannabinoids (i.e., CP 55,940, HU-210) that are full and long-lasting CB1 receptor agonists, whereas THC is a weak partial agonist. Thus, differences in the efficacy of these two types of drugs may be responsible for the differential effects. Second, one of the studies that found heightened stress reactivity was in animals treated during adolescence and tested for stress reactivity in adulthood (Lee et al., 2014). This introduces a potential confound in that cannabinoid exposure during adolescence, when the brain is still developing, could have different effects from exposure during adulthood. Third, the animal study that revealed blunted stress reactivity involved volitional exposure (i.e., self-administration), while those that have revealed elevated stress reactivity relied on forced injections which would be more stressful. Finally, there may be sex differences in the effects of cannabinoids on CORT reactivity, as Glodosky et al. (2020) only detected evidence of a blunted CORT response in female rodents (Glodosky et al., 2020); while studies that have detected elevated CORT reactivity have used male rodents exclusively (Hill & Gorzalka, 2005; Llorente-Berzal et al., 2011) or detected more pronounced effects in males than females (Lee et al., 2014).

Numerous studies in humans and animals have provided compelling evidence that acute cannabinoid administration increases concentrations of stress-related hormones (Androvicova et al., 2017; Cone et al., 1986; De Sousa Fernandes Perna et al., 2016; Jackson & Murphy, 1997; Kleinloog et al., 2012; Klumpers et al., 2012; Manzanares et al., 1999; Marín et al., 2003; McLaughlin et al., 2009; Ranganathan et al., 2009; Schramm-Sapyta et al., 2007; Weidenfeld et al., 1994). Pharmacological studies have provided important mechanistic insight into this phenomenon. Specifically, acute cannabinoid administration increases CORT concentrations via a mechanism involving activation of CRHR1 receptors (Jackson & Murphy, 1997) and recruitment of serotonergic and noradrenergic systems that in turn serve to indirectly activate the HPA axis (McLaughlin et al., 2009). However, it remains to be seen whether this phenomenon and the proposed mechanism of action is similar in females and whether it extends to other drugs (i.e., cannabis) and other routes of administration (i.e., intrapulmonary delivery).

Elevated basal cortisol has been detected in sober chronic cannabis users (Carol et al., 2017; Chao et al., 2018; Huizink et al., 2006; King et al., 2011; Monteleone et al., 2014; Somaini et al., 2012), although these effects have not been reliably detected (Block et al., 1991; Cloak et al., 2015; Cuttler et al., 2017; Lisano et al., 2019, 2020; Petrowski & Conrad, 2019), possibly for reasons described in section 4.3.3. In contrast, research in rodents has consistently found no differences in basal CORT between chronic cannabinoid- and vehicle-exposed animals (Biscaia et al., 2003; Hill & Gorzalka, 2006; Llorente-Berzal et al., 2011). As such, elevated basal stress hormones detected in humans could represent a risk factor for cannabis use rather than a consequence of such use.

Alternatively, unreliable findings of elevated basal cortisol in chronic cannabis users could be explained by a flattened diurnal cortisol slope (Labad et al., 2020). Stress hormones follow a regular circadian rhythm, increasing early in the active period of the day and peaking after awakening before decreasing throughout the rest of the day and reaching a nadir during sleep (Debono et al., 2009; Weitzman et al., 1971). Therefore, while stressors can trigger cortisol release, circulating concentrations are also intrinsically variable (Klerman et al., 2002). While basal, or resting, cortisol concentrations are commonly used to measure HPA axis functioning, a more reliable indicator is the CAR, or the increase of cortisol levels in the first hour after waking. The advantages of measuring the CAR rather than basal measurements taken during the day are the ability to avoid fluctuating cortisol levels and minimize the variability of individual differences in time of awakening (Clow et al., 2004; Pruessner et al., 1997). There is evidence of a blunted CAR and/or a flattened diurnal slope in chronic cannabis users compared to non-users (Huizink et al., 2006; Labad et al., 2020; Monteleone et al., 2014l). Thus, studies may have failed to find differences in basal stress hormone concentrations because cortisol was measured at midday when cortisol concentrations between users and non-users would be most similar (Block et al., 1991; Cloak et al., 2015; Hill & Gorzalka, 2006), or early in the morning when others have found a blunted CAR (Lisano et al., 2019, 2020). Therefore, further research repeatedly assessing stress-hormone concentrations across the day are needed to reconcile the mixed findings pertaining to basal stress hormone concentrations in chronic cannabis users. Nevertheless, we speculate that repeated exposure to exogenous cannabinoids may lead to a state of CB1 receptor downregulation/desensitization in brain areas that are known to coordinate HPA axis activity, which could subsequently impair the ability of the ECB system to effectively regulate circadian fluctuations in stress-related hormones.

5.2. Sex Differences

Despite substantial sex differences in the ECB system and HPA axis activity, only a few sex differences have emerged in this review, particularly in animal studies. First, chronic cannabinoid exposure appears to have effects on the ECB that differ by sex in rodents. While chronic THC administration leads to CB1 receptor downregulation in both sexes, adolescent female rats have displayed greater CB1 receptor desensitization than males (Burston et al., 2010). Second, female rats self-administer more cannabis vapor than male rats and they subsequently exhibit a greater increase in basal CORT following 30 days of cannabis self-administration than do males (Glodosky et al., 2020). Third, there is evidence that cannabinoids have sex-dependent effects on stress reactivity. While Glodosky et al. (2020) detected blunted stress reactivity in female rodents that self-administered a moderate dose of cannabis extract, others have detected an exaggerated CORT response following acute stress in male rodents chronically exposed to synthetic cannabinoids (Hill & Gorzalka, 2006; Llorente-Berzal et al., 2011), or have detected more pronounced effects of synthetic cannabinoids on stress reactivity in male rodents relative to female rodents (Lee et al., 2014). Similarly, the animal study that found exaggerated amygdala reactivity following acute administration of cannabinoids only used male mice (Patel et al., 2005). Human studies assessing stress reactivity have predominantly used mixed-sex samples but have largely failed to assess sex differences in stress reactivity following acute or chronic cannabis exposure. As such, it will be important to consider potential sex differences in future human studies.

Research in humans has also found that women are more likely to use cannabis to cope with anxiety and report larger reductions in anxiety (but not stress) after using cannabis compared to men (Cuttler et al., 2016, 2018). Moreover, there is evidence of a ‘telescoping effect’ of cannabis dependence in women, where women progress from first use to problematic cannabis use more rapidly than men (Cooper & Craft, 2018), which resembles findings of higher rates of self-administration in rodent models (Glodosky et al., 2020). Nevertheless, more research will be needed to determine the nature of this relationship and whether the sex differences that have been found in animal studies translate to humans.

5.3. Limitations and Future Directions

While this review highlights several important findings from studies of acute and chronic cannabinoid exposure on stress responses, it also brings to light several issues that limit the generalizability of past findings. In the human literature, studies often use different definitions of chronic or frequent cannabis use. For example, some studies defined chronic cannabis use as using at least 6–7 days per week in the past year (e.g., King et al., 2011), while others included participants who only used cannabis once per week (e.g., Lisano et al., 2019). Therefore, the dose and frequency of cannabis use may contribute to some of the discrepant findings in the cannabis literature. Cannabis researchers should seek to standardize the definition of chronic cannabis use and adopt validated measures to assess duration, frequency, and quantity of cannabis use (e.g., the DFAQ-CU; Cuttler & Spradlin, 2017). Similarly, there are other demographic characteristics that often limit the generalizability of studies in this review. Some studies include participants with psychiatric conditions such as psychosis and PTSD that could confound the effects of cannabis, as these conditions are also associated with a dysregulated cortisol rhythm or HPA axis activity (Bruijnzeel & Gold, 2005; Shah & Malla, 2015). Finally, despite important sex differences in the ECB system and the neuroendocrine stress response, many studies have failed to include female participants and only a few of those that did include both sexes properly examined sex differences in the impacts of cannabis use on stress-related endpoints. Future research examining these sex differences is therefore needed and should consider estrous/menstrual cycle phase given their effects on stress-related hormones and stress reactivity (Bale & Epperson, 2015; Viau & Meaney, 1991).

Studies with animal models also have several limitations that should be addressed to better understand the effects of cannabis and increase the translational relevance of their results. Most animal studies use synthetic analogs of THC that are potent, sometimes long-lasting CB1 receptor agonists that do not fully recapitulate, and often vastly exceed, the effects of THC or broad-spectrum cannabis products (McLaughlin, 2018). Full dose ranges are also required to understand the effects of cannabinoids and should be complemented by directly comparing the effects of different cannabinoids on stress-related endpoints. Moreover, most animal studies use injected cannabinoids rather than the intrapulmonary route of administration that is overwhelmingly the most common route of administration among human cannabis users (McLaughlin, 2018; Sexton et al., 2016). Since IV administration of THC in humans is associated with adverse experiences (D’Souza et al., 2004), and IP administration is subject to first-pass metabolism, which leads to greater production of the potent 11-OH-THC metabolite (Huestis, 2007), it is becoming increasingly evident that models of intrapulmonary drug delivery are essential for generating robust, translationally relevant data that can provide insights into effects observed in human cannabis users. Finally, many animal studies still do not include female subjects. Until recently, experiments conducted in rodents frequently used only males, but cannabis has important sex-specific effects that must be considered when extrapolating to human populations.

Finally, one major outstanding question concerns the downstream consequences of these alterations in the stress response. While acute stress responses are typically adaptive, chronic stress exposure has several negative consequences. Chronic stress is associated with poorer health outcomes (National Institute of Mental Health, 2015), including memory impairment (Lupien et al., 1998; Reich et al., 2009), hypertension, and obesity (Whitworth et al., 2005). In rats, chronic stress leads to hippocampal dysfunction and accelerates biological markers of aging (McEwen, 1998). Most importantly, chronic stress can result in HPA axis dysfunction (McEwen, 1998; McEwen et al., 2016) and dysregulation of the HPA axis is one of the most common features of neuropsychiatric diseases (Bale & Epperson, 2015; Oyola & Handa, 2017; Yehuda et al., 1996). For instance, chronic stress is associated with increased levels of negative affect (McEwen, 1998; McEwen et al., 2016; Sotnikov et al., 2014) and greater vulnerability to depression and anxiety (American Psychiatric Association, 2013; Hovens et al., 2012). Thus, the ability of cannabis to dampen stress reactivity in humans could be beneficial in combating the cumulative wear-and-tear of daily stressors. On the other hand, a blunted cortisol response is associated with anxiety, depression (Zorn et al., 2017), and PTSD (Metz et al., 2020) in women. Thus, it remains possible that impaired engagement of the HPA axis in chronic cannabis users during periods of acute stress could instead contribute to the emergence of stress-related disorders in this population. In future studies it will be imperative to examine the downstream consequences of an altered HPA axis and whether this contributes to, or protects against, the high rates of stress-related neuropsychiatric symptoms experienced by chronic cannabis users.

5.4. Summary and Conclusions

In summary the research reviewed herein provides ample evidence of alterations in the stress response following both acute and chronic cannabinoid exposure. Specifically, there is strong evidence from human studies and rodent models that acute cannabinoid administration increases basal stress hormone concentrations. There is also more limited evidence from human (but not animal) studies that chronic cannabinoid exposure is associated with an elevated CAR, a flattened diurnal cortisol slope, and possibly elevated basal cortisol concentrations, although evidence indicates the latter may represent a risk factor, rather than a consequence, of cannabis use. Evidence from human studies and rodent models also appears to converge on the notion that acute cannabinoid exposure dose-dependently affects stress-reactivity and anxiety-related outcomes. Finally, evidence from human studies and one rodent study indicate that chronic cannabis exposure can dampen stress reactivity. Collectively these findings indicate that exposure to cannabinoids may elevate basal stress hormones while suppressing stress reactivity.

Nevertheless, more research on these effects and the mechanisms underlying them is needed, as daily cannabis use is becoming increasingly common and medical users frequently report using cannabis to manage stress and treat stress-related disorders such as anxiety and depression (Cuttler et al., 2018). There are also high rates of stress-related disorders in cannabis users and chronic cannabis use is associated with increased negative affect (Lev-Ran et al., 2014). While the direction of these relationships remains unclear, using cannabis to cope with stress is associated with more cannabis use problems (Lee et al., 2007; Simons et al., 2005), higher levels of depression (Bonn-Miller et al., 2014; Glodosky & Cuttler, 2020), and social anxiety (Buckner et al., 2006). Therefore, future studies should seek to clarify the downstream consequences of cannabinoid-related perturbations in the stress response on mental health outcomes.

Highlights:

• Acute cannabinoid administration increases stress hormone concentrations

• Cannabinoids have acute dose-dependent effects on stress reactivity and anxiety

• Chronic cannabis users have a blunted CAR and flattened diurnal cortisol slope

• Chronic cannabis exposure dampens stress reactivity in humans and female rats

• More research is needed on sex differences in cannabis’ effects on stress hormones