Investigating endocannabinoid system and subjective responses to exercise in cancer patients: Rationale and future research directions

Abstract

Over forty years of evidence supports the integration of exercise therapy in cancer care. However, most cancer patients remain insufficiently active due in part to subjectively reported treatment-related side effects (and late effects) including fatigue, pain, appetite dysregulation, insomnia, cognitive impairment, depression, anxiety, low self-efficacy, and poor motivation. Many of these symptoms can be mitigated with exercise. However, the biological mechanisms by which exercise attenuates these cancer treatment-related side effects remain to be elucidated. This article presents a rationale for the investigation of endocannabinoid system (ECS) responses to exercise in cancer patients. We provide an overview of the ECS and preliminary evidence of ECS dysfunction induced by cancer, its risk factors (comorbidities) and cancer treatment. Further, we present a brief review of evidence from non-cancer cohorts demonstrating that acute (single bout) and chronic (>12 week) exercise can induce changes in circulating endocannabinoids (e.g. N-arachidonoylethanolamine (AEA or anandamide), 2-arachidonoylglycerol (2-AG) and related biogenic lipids). These changes are consistently accompanied by improvements in many subjectively reported, affective (mood) states (i.e. psychological outcomes) including sense of well-being, euphoria, vigour, anxiety, depression, fatigue, confusion, tension, mood disturbance, and pain. Given the substantial overlap between these subjective outcomes and the adverse effects that commonly arise as a consequence of cancer treatment, we clarify avenues for future research directed at improving our understanding of how cancer treatments negatively affect the ECS and patient symptomology, and how exercise may biologically mitigate these sequelae.

Graphical abstract

All images obtained under license from Adobe Illustrator.

Key Points.

• •This article articulates pathways for the investigation of endocannabinoid system responses to exercise in cancer patients;
• •The endocannabinoid system may be impacted by cancer, its risk factors and treatment-related sequelae;
• •Studies in non-cancer cohorts suggest that the endocannabinoid system may play a role in mediating the subjective, psychological (e.g. mood-related) benefits of acute and chronic exercise;
• •The aforementioned studies provide a rationale for investigating the potential biological link between endocannabinoid system responses and the common, subjectively reported side effects of cancer treatment (e.g. fatigue, pain, depression, low motivation, etc.)

1. Introduction

Clinical trials since the early 1980's have shown that professionally-directed exercise programs involving moderate-to-vigorous aerobic and resistance training can significantly improve clinical and health outcomes in cancer patients, including many treatment-related adverse effects and overall quality of life [[1], [2], [3], [4]]. Exercise can also reduce cancer risk and recurrence [5,6]. This evidence has culminated in a landmark position statement calling for exercise to be embedded in cancer care [7].

The integration of exercise services into cancer care is being met with many practical challenges [8]. The vast majority (∼80%) of cancer patients do not meet minimum physical activity guidelines [9]. Many are challenged by subjectively reported, treatment-related side effects (and late effects) including persisting and pervasive fatigue, pain, appetite dysregulation, insomnia, cognitive impairment, depression, anxiety, low self-efficacy, and poor motivation [3]. Lingering symptoms can act as barriers to exercise participation while, ironically, the evidence also suggests that appropriate exercise prescriptions can successfully treat these conditions [2]. At present, the biological mechanisms by which exercise attenuates many cancer treatment-related side effects are not known, and remain to be elucidated. Such research is a necessary step to achieve a better understanding of the key effects of exercise in cancer populations.

The endocannabinoid system (ECS) regulates many systemic physiological processes in the human body, including general homeostasis and metabolism, immune system responses, cognition, memory, appetite, emotions, mood, and perception (e.g. the subjective perception of pleasure and pain) [[10], [11], [12]]. Dysregulation of the ECS (e.g changes in receptor expression and signalling) has been documented in many types of cancer [13] suggesting that the ECS may be a target for anti-cancer therapies [14] and the mitigation of cancer treatment-related sequelae [15]. Recent systematic reviews [16,17] of studies involving healthy adults, elite athletes, and in some specific chronic conditions have demonstrated measured changes in circulating endocannabinoids (eCB) following acute (single bout) and chronic exercise (>12 weeks). These adaptations have been consistently accompanied by improvements in many subjectively reported affective (mood) states (e.g. psychological outcomes) including sense of well-being, euphoria, vigour, anxiety, depression, fatigue, confusion, tension, mood disturbance, and pain. Moreover, a few mechanistic studies suggest that the ECS may play a more central role in mediating the acute mood-related benefits of exercise than endogenous opiates (endorphins) [18,19]. There is substantial overlap between these psychological states and many of the adverse and late effects of cancer treatment. However, to date, there has been no investigation of ECS responses to exercise in any cancer population. Given that the ECS may play a role in mediating the subjective benefits of exercise in non-cancer cohorts, there is a need to consider such avenues of research pertinent to cancer patients.

This article presents a rationale for the investigation of ECS responses to exercise in cancer patients in an attempt to develop a more comprehensive biological understanding of the common sequelae of cancer treatment and their attenuation with exercise. Accordingly, an overview of the ECS including function and preliminary evidence of dysfunction in cancer populations is provided, followed by a brief review of studies investigating ECS and subjective responses to acute and chronic exercise in non-cancer cohorts. This article attempts to clarify avenues for future research directed at improving our understanding of how exercise mitigates common adverse effects of cancer treatment.

1.1. An overview of the ECS

A detailed historical timeline of the research on the ECS has been presented in several articles [[20], [21], [22], [23]]. The research began in 1940 with the isolation of two cannabinoids derived from the cannabis plant (phytocannabinoids) named cannabinol and cannabidiol (CBD) [21]. Over 20 years later, in 1964, delta-9-tetrahydrocannabinol (Δ9-THC), the psychoactive component of the cannabis plant was isolated by the Israeli chemist, Raphael Mechoulam [21]. The research advanced rapidly from the late 1980's with the discovery of the first cannabinoid receptor, cannabinoid receptor type 1 (CB₁) [21]. In 1992, the first eCB was identified and named N-arachidonoylethanolamine (AEA) and later termed anandamide after the Sanskrit word for bliss (ananda) [24]. A second receptor, cannabinoid receptor type 2 (CB₂), was cloned in 1993 [21], while another major eCB, 2-arachidonoylglycerol (2-AG), was identified in 1995 [25]. Further discovery of non-endocannabinoid N-acylethanolamines (e.g. palmitoylethanolamide (PEA), stearoylethanolamide (SEA), oleoylethanolamide (OEA)), and pertinent metabolic enzymes have contributed to a better understanding of the structure and function of the ECS [26]. The ECS is currently defined as a complex network consisting of G-protein coupled cannabinoid receptors (e.g. CB₁ and CB₂), eCB ligands (e.g. AEA and 2-AG) [10], the enzymes involved in the synthesis, degradation and transport of eCB, and the cells and neurological pathways involved in eCB signalling [26]. The precursors for AEA and 2-AG are derived from arachidonic acid (AA), a polyunsaturated fatty acid which is present in all cell membranes in humans [27,28].

Detailed overviews of ECS structure and function have been provided elsewhere [13,26]. In brief, AEA and 2-AG have affinity for both dominant cannabinoid receptor subtypes (CB₁ and CB₂) [29]. AEA is a high-affinity, partial agonist of both receptors while 2-AG is a moderate-to-low affinity full agonist of both receptors [12,27,30]. CB₁ receptors are the predominant cannabinoid subtype in the central nervous system (CNS), with significant distribution in the neocortex, hippocampus, basal ganglia, cerebellum, olfactory bulb; CB₁ receptors are also found in the spinal cord, peripheral nervous system, gastrointestinal system including the liver, and peripheral tissues, including white and brown adipose tissues, liver, the myocardium and skeletal muscles [12]. CB₂ receptors are located in the CNS immune cells (astrocytes and microglia) and peripheral immune cells (lymphocytes and mast cells) with moderate expression in other peripheral tissues (liver, bone, adipose), the cardiovascular system and gastrointestinal tract [12]. AEA and 2-AG play a major role as retrograde neurotransmitters in the CNS, although, 2-AG exists at much higher (1000x) concentrations in the brain compared to the rest of the CNS [12,30]. These eCB are synthesized ‘on demand’ in post-synaptic nerve terminals, mediated in response to sustained synaptic activity via increased intracellular Ca²⁺ [31]. The synthesis of both eCB is facilitated by such enzymes such as diacylglycerol lipase and phospholipase D [30]. The eCB are then released from the postsynaptic neuron to bind to and activate the CB₁ receptor on the presynaptic neuron, which in turn results in the inhibition of synaptic transmission [29,30,32]. The CB₁ receptor also binds the phytocannabinoid Δ9-THC and this mechanism of action, i.e. acting as “a brake” on neural transmission, may at least partially explain the analgesic [33,34] and psychotropic effects of Δ9-THC [35,36]. While the CB₁ receptors appear to play a predominant and major role in the inhibition of neurotransmission, CB₂ receptors appear to have a more central role in regulating inflammatory and anti-inflammatory responses through their effects on the immune system [11,37]. Notably, the ECS receptors are widely expressed in the limbic system explaining the key role of this system in the regulation of emotions and perception [38].

1.2. ECS dysfunction in cancer, cancer-related risk factors and treatment-related sequelae

There is early evidence that cancer patients may be affected by dysfunction of the ECS. The ECS may be altered by cancer itself, underlying risk factors, comorbidities and by the side effects of cancer treatment.

The ECS may be involved in the genesis of specific cancers, including gastrointestinal, gynaecological, lung, breast, prostate, pancreatic, haematological, melanoma, thyroid, thoracic and brain cancer [13,26,39,40]. This evidence suggests that cannabinoid receptor over- or under-expression, and related changes in receptor activation and heteromerization alter ECS signalling and can potentially play a role in increasing the tumorigenic potential of the cell [13,[41], [42], [43]]. For example, altered CB₁ and/or CB₂ expression has been noted in a number of tumours (e.g. breast, brain, endometrial, pancreatic and colon) when compared to normal tissue suggesting a possible role of the ECS in tumour growth [43]. Endogenous ligands AEA and/or 2-AG have also been documented to be elevated in many tumours, accompanied by increases or decreases in degrading enzymes [43]. For example, multiple-fold increases in 2-AG and/or AEA have been noted in colorectal and prostate carcinomas [13,44]. There is also a correlation between ECS expression and cancer prognosis and survival, which differs according to cancer type [43]. In contrast, preclinical studies have shown that the administration of exogenous phyto- and endocannabinoids can inhibit the growth of several kinds of tumours via anti-proliferative, anti-invasive and anti-angiogenic activity, with data in humans currently emerging [43]. Accordingly, the ECS has been identified as a potential target for cancer therapies with both preclinical and clinical research activity advancing rapidly [39,43].

Dysfunction of the ECS has been noted in metabolic conditions that increase cancer risk. Obesity is a well-established risk factor for many cancers [45] and circulating AEA and 2-AG have been shown to be significantly elevated in obese adults [46]. Elevated circulating AEA and 2-AG both correlate positively with visceral fat deposits and insulin resistance [46] indicating that an overactive ECS may be involved in mediating immune system and inflammatory responses in metabolic disease [47]. Interestingly, a 1-year lifestyle intervention demonstrated a significant reduction of circulating 2-AG and AEA in viscerally obese men [48]. These changes were associated with a significant reduction of visceral fat, systolic and diastolic blood pressure, triglyceride, insulin resistance and C-reactive protein [48].

Psychological symptoms commonly experienced by people diagnosed with cancer may also be related to ECS dysfunction [49]. Preclinical studies have shown that dysregulation of ECS signalling can lead to emotional disorders including fear, anxiety and stress [50]. The ECS is also disrupted in related conditions including insomnia [51] and pain [52]. Exposure to chronic stress downregulates CB₁ signalling in brain regions involved in emotional regulation and mood [50]. Dysfunction and deficiency of both CB₁ and CB₂ receptors in the CNS have been implicated in depression and closely related disorders including drug and alcohol addiction [53]. Studies involving humans have shown maladaptation of the ECS, particularly dysfunction of CB₁ in the CNS [49]. CB₁ has been shown to be instrumental in mediating mood, with impairment potentially playing a mechanistic role in neuropsychiatric disorders [54]. Clinical research also suggests that eCB tone is altered in individuals with post-traumatic stress disorder (PTSD) and/or depression [49] and that the ECS may be heavily implicated in the aetiology of these conditions through a variety of pathways including the hypothalamus-pituitary-adrenal (HPA) axis and other neurotransmitters such as Gamma-aminobutyric acid (GABA) [55,56].

This early evidence indicates that ECS dysfunction may be induced or exacerbated by cancer, its risk factors, comorbidities and some side effects of treatment. However, there are numerous questions that remain to be investigated. For example, we have found no data on the impact of specific cancer treatments, and the side effects which they commonly induce, on ECS-related outcomes in humans. The impact of the total treatment and total symptom burden across various patient groups also remains to be investigated.

1.3. ECS and psychological responses to exercise

To our knowledge, there has been no clinical trial conducted to investigate ECS adaptations to exercise in any cancer population to date. Therefore, we conducted a brief review of the extant literature using MEDLINE (Ovid, Wolters Kluwer), with no date restrictions up to April 2024, to elucidate the effect of exercise intervention on the ECS and subjectively-reported outcomes in other cohorts to inform the development of future trials. We included studies conducted in humans only, investigating acute (single bout) or chronic exercise interventions, and collecting one or more ECS outcomes and one or more subjective mood, symptom- or psychological outcomes pre and post exercise. Eighteen studies meet our criteria (Table 1, Table 2).

Table 1.

Studies evaluating endocannabinoid and subjective responses to acute exercise.

Author/Year/Country/Design	Sample (n)	Population	Exercise Intervention or Conditions	Summary of key results
Studies in apparently healthy adults
Sirotiak et al. [57] (2023) USA Controlled trial, Trained vs. untrained	32	Healthy adults (13M, 19F) with (n ​= ​16) and without (n ​= ​16) resistance training experience in past 3 months Age: Trained group (21.7 ​± ​2.5 ​y) Untrained group (22.6 ​± ​3.3 ​y)	All participants completed 3 sets of 6 machine-based, resistance training exercises (leg press, leg curl, leg extension, chest press, lat pull-down, abdominal curl). Intensity (loading): set 1 (16 reps at 50% 1RM), set 2 (12 reps at 70% 1RM), set 3 (8 reps at 80% 1RM)	•Plasma AEA, OEA, and PEA decreased in both groups (time effects) (all p ​< ​0.038) •No significant change in 2-AG in either group •No group x time differences or sex x time (male vs. female) differences in any eCB or related lipid •Positive affect and vigour increased, while total mood disturbance, confusion, anger and depression decreased in both groups (time effects) (all p ​< ​0.036) •No change in tension or fatigue •No group x time effects for any psychological variables
Siebers et al. [18] (2021) Germany RCT, counterbalanced	63	Healthy adults (31M, 32F) performing exercise more than twice per week Age: Naltrexone group: 28.1 ​± ​1.1 ​y Placebo group: 26.5 ​± ​1.0 ​y	Patients randomised (1:1) to receive naltrexone (opioid agonist) or placebo and performed two conditions: 1.Walk on treadmill <50% HRmax, 50 ​min 2.Run on treadmill 70–85% HRmax, 50 ​min (including 5 ​min warm-up)	•Plasma AEA, 2-AG, AA and PEA increased in both conditions combined (time effects) (all p ​< ​0.001) •Increases in AEA, 2-AG, AA and PEA were higher in the run condition (condition∗time effect) (all p ​< ​0.004) •Increases in euphoria (p ​< ​0.001) and decreases in anxiety (p ​= ​0.24) were greater in run versus walk •Naltrexone (opioid blockade) had no effect on the above outcomes, and did not affect experience of “Runner's high”
Hughes et al. [62] (2021) United Kingdom Within-subjects crossover (randomized)	12	Recreationally active men Age: 27 ​± ​6 ​y	Aerobic exercise (exercise cycle) for 20 ​min performed under four conditions: 1.Low intensity (at 40% VO2max) 2.Low intensity plus 40% limb occlusion pressure (manual blood flow restriction applied bilaterally) 3.Low intensity protocol plus 80% limb occlusion pressure (manual blood flow restriction applied bilaterally) 4.Moderate intensity (at 70% VO2max)	•Plasma 2-AG increased significantly post exercise in Conditions 2–4 with the largest increases noted in Condition 4 (24.45 ​± ​18.12 to 33.35 ​± ​20.24 ​ng/mL, p ​< ​0.05) •Pressure pain threshold increased at 3 of four sites in Condition 4 versus condition 1.
Hughes & Paterson [61] (2020) United Kingdom Within-subjects crossover (randomized)	12	Recreationally active adults (10M, 2F) Age: 27 ​± ​6 ​y	Unilateral leg press exercise (dominant leg) performed under four conditions: 1.Light load (4 sets at 30%1RM) 2.Light load protocol plus 40% limb occlusion pressure (manual blood flow restriction applied unilaterally) 3.Light load protocol plus 80% limb occlusion pressure (manual blood flow restriction applied unilaterally) 4.Moderate load (4 sets at 70% 1RM)	•Pressure pain thresholds at all sites post exercise increased in all four conditions, with Condition 2–4 producing significantly greater change than Condition 1 (all p ​< ​0.05) •Plasma 2-AG did not change in any condition post exercise but was found to be significantly lower 24 ​h post exercise in Condition 4 (p ​< ​0.05)
Stone et al. [64] (2018) United Kingdom Within-subjects crossover (non-randomized)	9	Apparently healthy post-menopausal women recruited from a local choir Mean age: 61 y	1.Dance exercise class to upbeat music, 35 min (including 5 min warm-up) 2.Supervised quiet reading to classical music, 30 min 3.Choir practice (singing), 30 min 4.Spin class (exercise cycle), 35 min (including 5 min warm-up)	•Dancing had no effect on plasma eCB (AEA, 2-AG, PEA, OEA) but decreased negative mood and emotions (p < 0.01) •Spin class (exercise cycle) increased OEA only (+26%; p < 0.05), but had no effect on mood/emotions
Brellenthin et al. [63] (2017) USA Controlled, counterbalanced (recruited and stratified by low/mod/high activity level)	36	Healthy adults (18M, 18F) with low, moderate and high physical activity levels (n ​= ​12 per group) Age: 21 ​± ​4 ​y	1.Prescribed exercise: Treadmill at 70–75% VO2max, 45 ​min (plus 10 ​min warm-up and 5 ​min cool-down) 2.Preferred exercise: Treadmill at self-selected intensity and duration	•Plasma 2-AG and PEA increased in both conditions (time effect) (both p ​< ​0.01). •Plasma AEA (p ​< ​0.05) and OEA (p ​< ​0.01) increased in the ‘prescribed’ versus preferred exercise condition (condition∗time effect) •Tension, depression, and anger decreased, while vigour increased in both conditions combined (time effect) (all p ​< ​0.05) •Confusion, total mood disturbance and state anxiety decreased in the ‘preferred’ condition versus the prescribed condition (condition∗time effect) (all p ​< ​0.05) •Increases in 2-AG associated with decreases in tension, depression and total mood disturbance in ‘preferred’ condition (all p ​< ​0.05) •Increases in AEA associated with increases in vigour in the prescribed condition (p ​< ​0.05) •Level of physical activity of participants did not influence any endocannabinoid or mood outcomes
Crombie et al. [59] (2018) Koltyn et al. [60] (2014) USA Within-subjects crossover	58	Healthy adults (29M, 29F) without a history of major medical conditions Age: 21 ​± ​3 ​y	Exercise protocol: Isometric hand grip exercise for 3 ​min at 25% maximal voluntary contraction 1.Exercise protocol performed after ingesting naltrexone (opioid agonist) 2.Exercise protocol performed after ingesting placebo	•Plasma AEA, 2-AG, OEA, and PEA increased under both naltrexone and placebo conditions (time effect) (p ​< ​0.05) •Plasma AEA (p ​= ​0.001) and OEA (p ​= ​0.009) increased in the placebo condition versus naltrexone (condition∗time effect) •Pressure pain threshold increased post-exercise under both naltrexone and placebo conditions (p ​< ​0.05); no condition∗time effect •Pain rating decreased significantly post-exercise; no condition∗time effect (p ​< ​0.05).
Studies in adults with chronic health issues
Botsford et al. [56] (2023) USA Controlled trial, PTSD vs. non-PTSD	98	Women with (n ​= ​42) or without (n ​= ​56) PTSD Age: PTSD group: 28.9 ​± ​7.6 ​y Control group: 22.5 ​± ​4.9 ​y	Data pooled from prior studies conducted by this research group: Treadmill walking or running, at 70–75% HRmax, RPE ​= ​12–15, 30–45 ​min (plus 5min each warm-up/cool-down)	•Plasma AEA increased in both groups (time effect) (p ​= ​0.002); no group∗time effect •Plasma 2-AG increased more in the control (non-PTSD) group versus the PTSD group (group∗time effect) (p ​= ​0.025) •No correlations between changes in eCB (AEA or 2-AG) and changes in mood states
Crombie et al. [19] (2021) USA RCT	35	Women, aged 21–50 with a current diagnosis of PTSD Age: Moderate-intensity group (31.8 ​± ​7.9 ​y) Low-intensity group (29.6 ​± ​5.9 ​y)	Participants were randomized to perform one of the following conditions after fear acquisition and extinction training. 1.Walking or running on treadmill at 70–75% HRmax for 30 ​min (plus 5 ​min each of warm-up and cool-down) 2.Walking on treadmill at 40–55% HRmax for 30 ​min (plus 5 ​min each of warm-up and cool-down)	•Trend toward a larger increase of plasma AEA from pre to post exercise in the moderate- versus low-intensity group (group∗time effect, p ​= ​0.054) •No significant change in 2-AG between groups •Threat expectancy ratings were reduced in the moderate-intensity exercise group versus the control group (group∗time effect) (p ​= ​0.032) •AEA found to mediate the relationship between moderate-intensity aerobic exercise and reduced threat expectancy ratings (95% CI: −0.623 to −0.005)
Crombie et al. [66] (2021) USA Controlled, counterbalanced	40	Women with PTSD (n ​= ​14) Trauma-exposed women without PTSD (n ​= ​14) Trauma-free women controls (n ​= ​12) Ages: PTSD: 27.6 ​± ​6.6 ​y Trauma no PTSD: 23 ​± ​5.4 ​y Trauma-free control: 21.4 ​± ​3.9 ​y	1.Treadmill walk or run at 70–75% HRmax, 30 ​min (plus 5 ​min each warm-up/cool-down) 2.Quiet rest session: sit quietly in a sound-dampened room for 40 ​min	•Plasma AEA, 2-AG, OEA and PEA increased in the exercise versus quiet rest condition in the total cohort (condition∗time effect, all p ​< ​0.05). No other interaction effects on these outcomes •Anxiety (p ​< ​0.001) and fear (p ​= ​0.007) ratings to unpredictable and predictable threats (electric shock administration) were decreased in the exercise versus quiet rest condition (condition∗time effects, both p ​< ​0.07). No other interaction effects on these outcomes •Fatigue, confusion and total mood disturbance decreased, and positive affect increased in the exercise versus quiet rest condition (condition ∗ time effect) (all p ​< ​0.003) •PTSD group experienced decreased depression after exercise (group ∗ condition ∗ time effect; p ​< ​0.001) •Increases in 2-AG associated with decreases in anxiety (p ​= ​0.023) and fear (p ​= ​0.031) in the total cohort in the exercise condition versus control condition
Brellenthin et al. [68] (2021) USA RCT	21	Young adults (12M, 9F) with substance abuse disorder and no other comorbidities Ages: Exercise group: 35.1 ​± ​10.2 ​y Control group: 35.0 ​± ​7.1 ​y	Exercise group: treadmill walking, at 70–75% HRmax, RPE ​= ​12–15, 30 ​min (plus 5-min warm-up/5-min cool-down), 3 sessions/week, 6 weeks Control group: usual care (quiet rest, 1 session/week, 30 ​min) NB: Blood samples were collected 5 ​min before and after exercise or quiet rest on the first (baseline) and last (week 6) study visits	•Plasma AEA levels increased in the exercise group versus control group (group ∗ time effect; p ​= ​0.002); no change in 2-AG •Vigor increased acutely in the exercise group (group ∗ time effect, p ​< ​0.0001) •No change in substance use, cravings, or psychological outcomes (i.e. self-efficacy to abstain from drug use, perceived stress, depression, and anxiety)
Meyer et al. [69] (2019) USA Within-subjects crossover (randomized)	17	Women with major depressive disorder Age: 40.8 ​± ​14.8 ​y	1.Cycle ergometer at moderate intensity (RPE ​= ​13), 20 ​min (plus 5 ​min warm-up and 5 ​min cool-down procedure) 2.Cycle ergometer at self-selected (preferred) intensity, 30 ​min (including 5 ​min warm-up and 5 ​min cool-down procedure)	•Plasma AEA and OEA increased in the moderate intensity condition (p ​< ​0.05), but no change noted in 2-AG, PEA or 2-OG •No change in AEA, OEA, 2-AG, PEA or 2-OG in the self-selected intensity condition •All measures of the POMS (i.e. total mood disturbance, tension, depression, anger, vigour, fatigue, confusion) and state anxiety improved in the moderate-intensity condition (all p ​< ​0.003) •All above measures except anger and state anxiety also improved in the self-selected condition (p ​< ​0.004) •Correlations in the moderate-intensity condition: Increases in AEA associated with reductions in state anxiety, depression, confusion, fatigue, total mood disturbance at 10 ​min post exercise. All effects except fatigue remained significant at 30 ​min post exercise. Increases in 2-AG associated with reductions in depression and confusion at 10 ​min post exercise. Effect for confusion remained significant at 30 ​min post exercise; •No correlations between eCB and perception variables in the self-selected condition
Crombie et al. [65] (2019) USA Controlled, counterbalanced	20	Adults with current diagnosis of PTSD (n ​= ​10; 2M, 8F) and healthy adult controls (n ​= ​10; 3M, 7F) Age: 23.7 ​± ​7.2 ​y	1.Treadmill walk (inclined) or run at 70–75% HRmax, 30 ​min (plus 5 ​min warm-up and 5-min cool-down) 2.Trier Social Stress Test (TSST), a reliable and valid test for inducing acute biological stress; 40 ​min (including 10 ​min preparation period)	•Plasma AEA and OEA increased in both the PTSD group and healthy group in the exercise condition (time effect) (both p ​< ​0.001) •Plasma 2-AG increased in the healthy group versus the PTSD group in the exercise condition (group ∗ time effect; p ​< ​0.05) •Tension, depression, fatigue, confusion, total mood disturbance, state anxiety and negative affect decreased in the PTSD group versus the healthy group in the exercise condition (group∗time effect, all p ​< ​0.05) •Vigour (p ​= ​0.023) and positive affect (p ​= ​0.038) increased in both groups in the exercise condition (time effect)
Crombie et al. [67] (2018) USA Controlled, counterbalanced		Adults with current diagnosis of PTSD (n ​= ​12; 3M, 9F) Age: 22.0 ​± ​4.7 ​y healthy adult controls (n ​= ​12; 3M, 9F) Age: 26.1 ​± ​6.7 ​y	Treadmill walk (inclined) or run at 70–75% HRmax, 30 ​min (plus 10 ​min warm-up and 5-min cool-down)	•Plasma 2-AG, AEA and OEA increased in both the PTSD group and healthy group in the exercise condition (time effect) (both p ​< ​0.05); no change in PEA, 2-OG •No group ∗ time interaction effects for 2-AG, AEA, OEA, PEA and 2-OG. However, effect size calculations showed larger responses in healthy controls versus the PTSD group •Vigour and positive affect increased in both groups (time effect, both <0.02) •Group ∗ time interaction effects favouring larger change in PTSD group were noted for state anxiety, negative affect, tension, fatigue, confusion, depression and total mood disturbance (all p ​< ​0.05) •Pain was reduced in both groups (time effect, p ​= ​0.001) •Increase in 2-AG correlated with reduction in negative affect in the PTSD group (p ​= ​0.04). No other associations noted.
Stensson et al. [73] (2019) Sweden Controlled trial	32	Adults with (n ​= ​5M, 16F) with chronic neck and shoulder pain (CNSP) and healthy adult controls (n ​= ​11, 5M, 6F) Ages: CSNP: 50.8 ​± ​12.9 ​y Control: 37.7 ​± ​15.9 ​y	Dynamic, loaded arm cycling at 25 laps/minute, 30 ​min Women: starting load 100g, increased to 300g at 10 ​min, increased to 500g at 20 ​min and maintained for final 10 ​min Men: starting load 200g, increased to 400g at 10 ​min, increased to 600g at 20 ​min and maintained for final 10 ​min	•Plasma AEA increased non-significantly in the CNSP group (0.84 ​± ​0.09 to 0.92 ​± ​0.09 ​nM; p ​= ​0.06) and decreased in the healthy control group (0.85 ​± ​0.14 to 0.63 ​± ​0.13 ​nM; p ​= ​0.04) •No change in OEA, PEA, SEA or 2-AG in either group •No change in pain scores in either group

Abbreviations: 1RM ​= ​one-repetition maximum; 2-AG ​= ​2-arachidonoylglycerol; 2-OG ​= ​2-oleoylethanolamine; AEA ​= ​anandamide; AA ​= ​arachidonic acid; BMI ​= ​body mass index; BDNF ​= ​brain-derived neurotrophic factor; CNSP ​= ​chronic neck and shoulder pain; eCB ​= ​endocannabinoid; HRmax ​= ​maximum heart rate; NAL= Naltrexone; OEA ​= ​oleoylethanolamine; PEA ​= ​palmitoylethanolamide; SEA ​= ​stearoylethanolamide; RCT ​= ​randomized controlled trial; RPE ​= ​rating of perceived exertion; PTSD ​= ​post-traumatic stress disorder.

Table 2.

Studies evaluating endocannabinoid and subjective responses to chronic exercise training.

Author/Year/Country/Design	Sample (n)	Population	Intervention or Conditions	Summary of key results
Studies in apparently healthy adults
de Oliveira et al. [71] (2019) Brazil RCT	34	Healthy, physically inactive adults (12M, 18F) Age: 38 ​± ​11.5 ​y	Exercise group: Treadmill exercise at ventilatory threshold, 40 ​min (including warm-up, cool-down) 3 sessions/week, 12 weeks Control group: no exercise For the exercise group, all post-intervention blood samples were collected within 2–5 days of the last exercise session	•Plasma AEA decreased in the exercise group post-intervention (0.33 (95% CI 0.28–0.41) to 0.22 (95% CI 0.18–0.26) pmol/mL); time effect, p ​= ​0.004); no sig. change in AEA in the control group. •The exercise group decreased anxiety and anger (all time effects, p ​< ​0.05)
Studies in adults with chronic health issues
Stensson et al. [73] (2020) Sweden Uncontrolled trial	70	Adult women with fibromyalgia (n ​= ​37; Age: 50.1 ​± ​10.5 ​y) and 33 age/sex-matched healthy controls (n ​= ​33; age: 50.3 ​± ​12.8 ​y)	All participants engaged in supervised resistance training, 2 x 60-min sessions/week, 15 weeks. Exercises targeted most major muscle groups but mainly focussed on the lower body. Loads: 40% of 1RM with 15–20 repetitions in 1–2 sets during weeks 1–2, 60% of 1RM with 10–12 repetitions in 1–2 sets during weeks 3–5, 80% of 1RM with 5–8 repetitions in 1–2 sets during weeks 6–15 Blood samples were obtained within 1–7 days of the start of the intervention and between 1 and 7 days after the intervention	•Plasma AEA increased (0.26 ​± ​0.13 to 0.34 ​± ​0.18 ​nM; p ​= ​0.01) and SEA decreased (2.56 ​± ​1.18 to 2.05 ​± ​1.10 ​nM; time effect, p ​= ​0.03) in the fibromyalgia patients; no change in OEA, PEA, or 2-AG •Plasma SEA increased in the healthy controls (2.05 ​± ​0.83 to 2.57 ​± ​1.06 ​nM; time effect, p ​= ​0.03); trend toward increase in plasma AEA in healthy controls (p ​= ​0.14) •Pain, fibromyalgia symptoms, depression and fatigue decreased in the fibromyalgia patients (time effects, all p ​< ​0.008)
de Oliveira et al. [72] (2019) Brazil RCT	50	Adults with episodic migraine not taking any preventative drugs (n ​= ​25; 5M, 20F), and healthy adults (n ​= ​25; 4M, 21F) Age: 36.2 ​± ​10.9 ​y	Exercise group: Treadmill exercise at ventilatory threshold, 40 ​min (including warm-up, cool-down) 3 sessions/week, 12 weeks Control group: Waitlist control group (no exercise) Blood samples were collected within 2–5 days of the last exercise session OR 48 ​h after the last exercise session within the late follicular phase of the menstrual cycle	•Plasma AEA decreased with exercise in healthy adults (p ​< ​0.01) and adults with episodic migraines (p ​= ​0.040) (time effects) •Exercisers with migraines reduced days with migraine per month, frequency of migraine attacks and medication usage (all time effects, all p ​< ​0.05) •Exercisers with migraines improved 4 of 6 POMS measures (tension-anxiety, depression, anger, fatigue) in exercise group (all time effects, all p ​< ​0.05) •Healthy exercises and exercisers with migraines increased VO2peak and VO2 at ventilatory threshold (all time effects, all p ​< ​0.05) •Changes in AEA inversely correlated with attendance in all exercisers (p ​= ​0.02) and migraine exercisers (p ​= ​0.029) (i.e. greater attendance correlated with greater reduction in AEA) •Increases in AEA correlated with increases in medication usage in the migraine group (r ​= ​0.68; p ​= ​0.03)

Abbreviations: 1-AG ​= ​1-arachidonoylglycerol; 2-AG ​= ​2-arachidonoylglycerol; AEA ​= ​anandamide; BNDF ​= ​brain-derived neurotrophic factor; DEA ​= ​docosatetraenoylethanolamide; HRmax ​= ​maximum heart rate; OEA ​= ​oleoylethanolamine; PEA ​= ​palmitoylethanolamide; POMS=Profile of Mood States; RCT ​= ​randomized controlled trial; RPE ​= ​rating of perceived exertion; SEA ​= ​stearoylethanolamide

1.3.1. Acute exercise and the ECS

Seven studies have investigated ECS and affective state responses to a single (acute) bout of exercise in apparently healthy adults (Table 1). Four of these studies prescribed conventional forms of aerobic exercise (e.g. treadmill, spin class, cycling) while only one trial has prescribed conventional resistance exercise [57] according to established exercise prescription guidelines [58]. The other studies investigated isolated, isometric handgrip [59,60] and leg press exercise [61].

Of these studies, Siebers et al. [18] have conducted the largest randomized controlled trial (RCT). Sixty-four physically active men and women (aged 18–50 years) performed two exercise conditions in randomized order, separated by a 30-day washout period. The exercise conditions consisted of: (1) a 50 ​min run on a treadmill at 70–85% of maximal heart rate (moderate-to-vigorous intensity) and (2) a 50 ​min walk on a treadmill at <50% of maximal heart rate (low-intensity). The participants were also randomized (1:1) to receive 50 ​mg of an opioid receptor antagonist (naltrexone; Desitin Arzneimittel GmbH, Germany) or a placebo prior to each session. Blood samples, visual analog scales and an anxiety assessment task were completed immediately prior to and following each exercise condition and the participants were also asked if they experienced a ‘Runner's High’ (yes/no/I do not know) at the end of each exercise condition. Both the run and walk conditions significantly increased plasma AEA, 2-AG, AA and an eCB-like molecule, PEA, from immediately pre to post exercise. However, the running condition resulted in approximately 2-fold higher increases in these circulating ECS markers versus the low-intensity walking condition (all p ​< ​0.04; AEA: +91% vs. +48%, 2-AG: +43% vs. +14%, AA: +210% vs. +120%, PEA: +63% vs. +38%). The run condition also significantly increased levels of euphoria (p ​< ​0.001) and decreased anxiety (p ​= ​0.024) versus the lower intensity condition. Descriptive analysis showed that participants were more likely to experience a ‘Runner's High’ in the run versus walk condition. Notably, the outcomes were not influenced by opioid receptor blockade, suggesting that these adaptations occurred independent of the action of endorphins implicating alternative mechanisms related to the changes in mood.

Two additional studies in young adults have demonstrated no influence of opioid receptor blockade on ECS and mood adaptations to an unconventional exercise protocol (i.e. isometric handgrip exercise performed at a low intensity, i.e. 25% of maximal voluntary contraction performed for 3 ​min) [59,60]. These studies noted significant increases in AEA, 2-AG, PEA and OEA with a concomitant increase in pressure pain thresholds and reduction in pain ratings in both the placebo and naltrexone conditions. Increases in circulating 2-AG and pressure pain threshold have also been noted in another study [62] prescribing low-versus moderate-intensity exercise cycling combined with or without manual bilateral lower limb blood flow restriction. The largest increases in 2-AG were noted with the moderate-intensity protocol without limb occlusion (70% VO_2max) and pain pressure ratings across three of four anatomical sites (left and right quadriceps, and biceps) were significantly improved in this condition versus the low-intensity condition [62].

A study comparing a prescribed versus preferred intensity/duration treadmill condition in young men and women [63] noted an increase in circulating 2-AG and PEA coupled with an improvement in tension, depression, anger and vigour in both conditions (time effects, all p ​< ​0.05). The prescribed exercise condition induced significantly greater increases in plasma AEA (p ​< ​0.05) and OEA (p ​< ​0.01) versus the preferred condition. However, paradoxically, the preferred condition induced significantly greater reductions in confusion (p ​< ​0.01), total mood disturbance (p ​< ​0.001) and state anxiety (p ​< ​0.05). Increases in 2-AG were associated with reductions in tension, depression, and total mood disturbance in the preferred condition (all p ​< ​0.05), while increases in AEA were associated with increases in vigour in the prescribed condition (p ​< ​0.05). The authors also determined that the results did not differ across low, moderate, and highly active participants enrolled in the study.

Two small studies found no significant change in any circulating eCB in nine apparently healthy post-menopausal women following a 35 ​min dance class performed ‘to upbeat music,’ [64] or four sets of unilateral leg press performed with and without blood flow restriction in 12 recreationally-active young adults [61], despite an improvement of pressure pain thresholds and ‘negative mood and emotions’. However, one of these studies [64] also prescribed a cycling condition (35 ​min spin class of unspecified intensity) and noted a 26% increase in OEA but no improvement in AEA, 2-AG, PEA or any subjective mood outcome.

Only one study to date evaluated the effect of an acute resistance exercise session performed according to established exercise guidelines [57] Thirty-two women with and without prior resistance training experience performed three sets of six resistance training exercises at a moderate-to-high intensity (i.e. 50%, 70% and 80% of 1RM over three sets). In contrast to several aerobic exercise studies, plasma AEA, OEA and PEA significantly decreased with no change in 2-AG from pre-to post-exercise in both the trained and untrained participants. These adaptations were accompanied by increased positive affect and vigour and decreased total mood disturbance, confusion, anger and depression. Responses did not differ by training status or sex.

Eight studies enrolled adults with chronic health issues, including PTSD [19,56,[65], [66], [67]], substance abuse disorder [68], major depressive disorder [69] and chronic neck and shoulder pain [70] (Table 1). A healthy control group was included in five of these studies [56,[65], [66], [67],70].

Crombie et al. [66] conducted a study enrolling: (1) young women with PTSD, (2) young women experiencing a previous traumatic event but without a PTSD diagnosis and (3) a trauma-free, healthy control group. In the total cohort, 30 ​min of moderate-intensity exercise (plus 5 ​min each of warm-up and cool-down) on a treadmill was found to significantly increase plasma AEA, 2-AG, OEA and PEA versus an equivalent duration (40 ​min) of quiet rest. These adaptations were concomitant with improvements in anxiety, fear, fatigue, confusion, total mood disturbance and positive affect in the total cohort. Although effect size calculations revealed a greater magnitude of change for the control versus PTSD participants, no interaction effects were noted in these outcomes. However, the PTSD group did experience a significant reduction in depression from pre-to post-exercise versus the other groups (group ∗ condition ∗ time effect; p ​< ​0.001). In the total cohort, larger increases in 2-AG were associated with reductions in anxiety ratings to a predictable threat (p ​= ​0.023) and reductions in fear ratings to an unpredictable threat (p ​= ​0.031).

Another study by Crombie et al. [65] tested the same exercise protocol (i.e. moderate-intensity exercise on a treadmill for 30 ​min) in both young women and men with a diagnosis of PTSD versus a healthy control group. Plasma AEA and OEA increased in both groups in the exercise condition with no group ∗ time effects noted, while 2-AG increased in the healthy group versus the PTSD group in the exercise condition (i.e. no change in 2-AG in the PTSD group) (p ​< ​0.05). Tension, depression, fatigue, confusion, total mood disturbance, state anxiety and negative affect decreased significantly with exercise in the PTSD group versus the healthy controls (all p ​< ​0.05). Moreover, greater reductions in negative affect were significantly associated with greater increases in 2-AG in the PTSD group. Vigour and positive affect increased in both groups in the exercise condition (time effect only). Similar findings were noted in other studies of a PTSD cohort by the same authors, with larger effect size responses in eCB in the healthy control group versus the PTSD group [67]. Botsford et al. [56] showed that 2-AG was increased to a significantly greater extent in a healthy control group versus a PTSD group, though no difference in response was noted in AEA (Table 1).

Brellenthin et al. [68], using a similar exercise protocol in adults with substance use disorder, noted a significant increase in circulating AEA and no change in 2-AG. The increase in AEA was accompanied by an increase in vigour but no change in substance use, self-efficacy to abstain from drug use, perceived stress, depression, or anxiety. The data reported in this study were averaged across two acute exercise sessions (i.e. before after a 6-week exercise intervention).

Meyer et al. [69] enrolled an older cohort of women (mean age: 40.8 years) with a diagnosis of major depressive disorder without enrolling a healthy control group and determined that circulating AEA and OEA significantly increased following moderate-intensity exercise performed on a cycle ergometer. No change was noted in 2-AG, PEA or 2-oleoylethanolamine (2-OG). By contrast, when the participants performed the exercise session at a self-selected intensity, no changes were noted in any ECS marker. A range of mood outcomes as measured by the Profile of Mood States questionnaire did improve in both exercise conditions. Correlations revealed that increases in circulating AEA and 2-AG were associated with a range of positive mood adaptations at 10 ​min and 30 ​min post exercise in the moderate intensity exercise condition. However, no such relationships were detected in the self-selected condition. Overall, intense exercise increased AEA and OEA and improved mood outcomes whilst the self-directed exercise did not.

By contrast to these findings, a study prescribing upper body aerobic exercise (i.e. arm ergometry) in patients with chronic neck and shoulder pain and healthy controls [70] showed a non-significant increase in AEA in the patient group and a significant reduction in AEA in the control group. There was no improvement of any pain scores with the exercise protocol.

1.3.2. Chronic exercise and the ECS

We identified only three studies which have evaluated the effect of chronic exercise training on the ECS and mood/psychological outcomes (Table 2). Two of these studies were RCTs conducted by de Oliveira et al. [71,72] Both RCTs evaluated a 12-week aerobic training program compared to no exercise. One study enrolled healthy but physically inactive adults (n ​= ​34) [71] and the other [72] enrolled adults with episodic migraines (n ​= ​25) and a healthy (non-headache) group (n ​= ​25). The third trial enrolled women (n ​= ​37) with a diagnosis of fibromyalgia and a health comparison group (n ​= ​33) matched for age and body mass index [73] with all subjects receiving a 15-week resistance training intervention. Blood samples were collected under resting (basal) conditions before and after the exercise intervention in all studies.

AEA was the only ECS outcome measured in the two RCTs by de Oliveira et al. [71,72] The study limited to healthy inactive adults [71] noted a significant decrease in basal AEA concentrations in response to the aerobic exercise training program (p ​= ​0.004) and this was concomitant with a significant decrease in anxiety and anger (p ​< ​0.05). In the other study [72], participants both with and without episodic migraines who engaged in the aerobic training program also significantly reduced AEA over time (Table 2). In the migraine sufferers, this effect was accompanied by reduced days with, and frequency of, migraine attacks, and reduced usage of analgesic medication. The authors did not report on interaction effects between groups in either RCT.

By contrast to these findings [71,72], Stensson et al. [73] noted a significant increase in circulating AEA in patients with fibromyalgia participating in the resistance training intervention (p ​= ​0.01). The healthy control group experienced a non-significant decrease in AEA (p ​= ​0.14). SEA significantly decreased in patients with fibromyalgia and increased in the healthy group after the exercise program. No changes were noted in OEA, PEA or 2-AG in either group. Pain, fibromyalgia symptoms, depression and fatigue significantly decreased in the exercising patients with fibromyalgia (all p ​< ​0.008; Table 2).

2. Discussion and future research directions

Clinical trial evidence supports the inclusion of exercise therapy as standard practice in cancer care [7]. Yet most cancer patients remain insufficiently active due in part to subjectively reported treatment-related side effects (e.g. fatigue, pain, appetite dysregulation, insomnia, cognitive impairment, depression, anxiety, low self-efficacy, and poor motivation), many of which can be mitigated with exercise. The biological bases of these side effects and their attenuation with exercise are largely unknown. We have provided preliminary evidence that ECS dysfunction may be induced by cancer, its risk factors, and the common side effects of treatment. Moreover, data from our review in non-cancer cohorts suggests that exercise-induced improvement of affective states and other subjective outcomes, including pain thresholds, pain ratings, anxiety, depression, fear, tension, anger, confusion, migraine symptoms, mood disturbance, threat expectancy ratings, fatigue, positive affect, vigour, and euphoria are accompanied by, and may be influenced, by adaptation of the ECS (Table 1, Table 2). Evidence of potential ECS dysfunction in cancer patients coupled with the findings of these recent exercise studies provides a rationale for several lines of research in cancer populations.

Studies of acute aerobic exercise in apparently healthy adults and those with specific health issues (Table 1) have shown significant increases in circulating AEA [18,56,59,60,63,[65], [66], [67], [68], [69]] and 2-AG [18,56,59,60,62,63,[65], [66], [67]]. Recent systematic reviews have concluded that AEA is more consistently increased by acute exercise than 2-AG, however the reason for these incongruent effects are not known and may relate to the differential roles and actions of these eCB and/or the exercise dosages applied [16,17]. Many acute studies have also noted changes in related biogenic lipids including, AA [18], PEA [18,59,63,66] and OEA [59,[63], [64], [65], [66], [67],69]. Conventional aerobic exercise modalities (i.e. exercise cycling or treadmill) performed at a moderate-to-vigorous intensity, as recommended by exercise prescription guidelines [58], appear to yield larger changes in these markers, whereas interventions involving unconventional exercise (i.e. arm cycling, unilateral single set resistance training) or exercise of an unspecified intensity (i.e. dance class or spin class) [64] have yielded inconsistent or null effects (Table 1). Paradoxically, the lone study on conventional acute resistance exercise [57] demonstrated significant reductions in AEA, OEA and PEA with concomitant improvement of some subjective mood states. These early findings suggest that the ECS may be differentially affected by these two distinct exercise modalities (i.e. aerobic and resistance exercise). Future studies involving acute exercise in cancer patients may consider prescribing conventional aerobic or resistance training protocols and evaluating dose–response effects on ECS outcomes and common treatment-related side effects (e.g. fatigue, pain, insomnia, anxiety, etc.).

All three studies involving chronic exercise training have applied interventions that align with established exercise guidelines [58], and have noted significant improvements in many subjective outcomes (Table 2). Two studies involving healthy adults and migraine sufferers noted a decrease in basal AEA secondary to 12 weeks of aerobic training on a treadmill [71,72]. The other study [73] noted a significant increase in basal AEA and decrease in SEA secondary to 15 weeks of resistance training in patients with fibromyalgia and the opposite effects in the healthy control group (Table 2). These differential responses are difficult to explain and could be attributed to the unique characteristics of the cohort being investigated and/or their response to the exercise modality prescribed. Elevated AEA and 2-AG are present in chronic inflammatory conditions including obesity and type 2 diabetes and hence a reduction in basal eCB concentration may be associated with better health status. These chronic conditions become more prevalent with increasing age and are associated with age-related muscle wasting (sarcopenia). Participants in the fibromyalgia study were older (∼50 ​y) and therefore may have been more impacted by a pro-inflammatory state. However, mean BMI was nearly identical across all three studies (∼26 ​kg/m²). It is worth noting that there are currently no established normative ranges for circulating eCB collected under basal (resting/fasting) conditions and this is perhaps an important undertaking for future research.

Several acute and chronic studies noted correlations between changes in circulating ECS markers and subjective adaptations (Table 1, Table 2) [19,63,66,67,69,72]. Moreover, two acute studies involving opioid (endorphin) receptor blockade resulted in no significant change in affective state responses, indicating that such effects were at least partially independent of the action of endorphins [18,59]. These data provide justification for exploring ECS outcomes as a potential biological mechanism by which subjectively reported, treatment-related sequelae can be mitigated with exercise in cancer patients. At present, the biological mechanisms linking ECS adaptation with changes in subjective, outcomes are unknown. Notably, all studies included in our review have limited their investigation to circulating eCB only, with no tissue samples and imaging technologies being utilised to date. Skeletal muscle and adipose tissue are heavily involved in mediating the systemic physiologic effects of exercise [74,75]. For example, Dalle & Koppo [76] have shown that 12 weeks of resistance training tended to increase CB₁ (+11%, p ​= ​0.055) and CB₂ (+37%, p ​= ​0.066) receptor expression in skeletal muscle in older adults (>65 years). While this study did not collect any subjective outcome measure, it is conceivable that ECS-related changes in both muscle and adipose tissue (genetic and genetic expression) may contribute to systemic biological effects, including adaptive neurological changes [46] which may be implicated in improving subjectively reported outcomes. The influence of potentially interrelated inflammatory (e.g. cytokines/myokines/adipokines) and immune system markers may also offer insight regarding such mechanistic pathways. There is, for example, a need to determine how each specific ECS marker contributes to the mitigation of treatment-related side effects in cancer patients. Exercise prescriptions in future studies need to be adequately defined in terms of modality, frequency, intensity, duration, and volume, to elucidate dose–response effects.

Two studies have explored preferred versus prescribed acute exercise and noted that prescribed intensity/duration facilitated larger increases in AEA and OEA versus self-selected (‘preferred’) intensity/duration exercise [63,69]. Cancer patients have a broad array of exercise preferences in terms of modality, intensity and duration, which may need to be considered to facilitate long-term engagement in exercise training [77]. Therefore, the effect of various ‘preferred’ prescriptions warrants investigation, particularly in relation to cancer treatment protocols which can negatively affect exercise tolerance and capacity and hence limit or inhibit participation in moderate-to-vigorous exercise, for example.

We have noted some evidence of differential ECS responses to exercise between healthy adults and those with a specific health condition. Crombie and colleagues [56,65,67] noted significantly larger increases (or larger effect sizes) in 2-AG, AEA, OEA in healthy controls versus the PTSD group [67]. The authors attributed the blunted response of the PTSD group to ECS dysfunction potentially caused by lowered eCB-CB₁ receptor signalling as a consequence of chronic trauma, a mechanism of action that has been noted in animal models [65]. It may be beneficial for future studies involving cancer cohorts to enrol a healthy control group matched for potential confounding variable such as age, sex, BMI, and key comorbidities including mental health status.

Cancer populations are heterogenous and hence future studies involving cancer patients should adequately define participant characteristics including stage/grade of cancer, treatments received and sequelae of treatment to determine how specific cancers and various treatment modalities affect the ECS and its response to acute and chronic exercise. Variations in cannabinoid receptor expression have been shown in different cancers [13,26,39] and so the type of cancer or histological pattern may have influential effects. Investigating ECS responses during various disease-modifying, anticancer treatments such as chemotherapy, targeted therapy or immunotherapy may inform how exercise training can be prescribed or recommended throughout the cancer continuum. Further, if such research can provide a clearer understanding of ECS-related dysfunction in cancer patients, avenues for the investigation of complementary therapies, including cannabinoid supplementation, in addition to exercise may be opened.

3. Conclusion

In summary, we have provided preliminary evidence that ECS dysfunction can be induced by cancer, its risk factors (comorbidities) and cancer treatment-related side effects. Moreover, our review presents evidence that there may be a relationship between ECS and subjective adaptations in response to acute and chronic exercise training in non-cancer cohorts. This evidence forms a rationale for the investigation of these outcomes in cancer patients in an effort to better understand how exercise mitigates treatment-related sequelae across various cancer populations and across the cancer continuum.

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Declaration of competing interest

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