Cancer, resulting from abnormal and uncontrolled proliferation of cells in the body, is the leading cause of death in humans. Globally, it caused nearly 20 million deaths in 2020 (https://www.who.int/news-room/fact-sheets/detail/cancer; accessed on 7 June 2024). There exist more than 100 types of cancer affecting different tissues and cell types, and resulting from different etiology and carcinogenetic processes. A wide variety of chemotherapeutic drugs have been developed with variable efficacies in different types of cancers. Due to cancers’ remarkable ability to develop resistance to anti-cancer chemotherapy as well as to evade body’s anti-cancer immune responses, search for novel, more effective and safer anti-cancer drugs is ever ongoing. In this regard, immune checkpoint inhibitors (ICIs), also known as immune checkpoint blockers (ICBs), represent a novel class of anti-cancer therapeutics that aim at invigorating anti-cancer immune responses in cancer patients. Immune cells express different checkpoints to inhibit body harm that may be caused by an overactive immune response (2, 3). However, the inhibition of the checkpoints in cancer patients is accompanied by many moderate to severe immune-related adverse events (ir-AE). To overcome these adverse events and improve the tolerability of ICIs, clinicians/researchers are resorting to the use of cannabinoids. It is noteworthy that cancer patients are also often prescribed cannabinoids for relief from cancer/chemotherapy-associated pain, anxiety, nausea, anorexia and insomnia. Furthermore, current legalization of medicinal cannabis in many countries in the World as well in many states in the USA has provided impetus to its use for both medicinal and recreational purposes. This review is focused to understand pros and cons of the use of cannabinoids in cancer patients undergoing ICI immunotherapy. To understand the complex interplay between ICIs and cannabinoids, we will first describe different immune checkpoints, their ligands, relevant ICIs and their potential mechanisms of action for activating anti-cancer immunity. Thereafter, we will review cannabinoids and their known receptors through which they exert their effects on immune and nonimmune cells as well as on cancer cells in the human body. We will also present pre-clinical and clinical evidence regarding the interaction of ICIs and cannabinoids in the tumor bearing animal models and cancer patients. Finally, we will discuss various factors that are known to regulate effects of cannabinoids on immunotherapy with ICIs in these patients.

ICIs have been hailed as game changer in the treatment of cancer; they have drastically changed the landscape of anti-cancer therapy (2, 3). ICIs-based immunotherapies have revolutionized therapies for solid tumors including melanoma, non-small cell lung cancer (NSCLC), renal cell carcinoma (RCC) and breast cancer, to mention a few (2, 3). In the context of cancer immunotherapy, immune checkpoints refer to key molecules that negatively regulate activation and effector functions of immune cells. They include receptors or co-receptors that are expressed on immune cells (T, B, Natural Killer and dendritic cells, and macrophages) and, upon engagement with their respective ligands or counter-receptors, negatively regulate the immune cell functions such as cytotoxicity, phagocytosis, production of cytokines and chemokines, etc. Additionally, they also promote the development and function of immunosuppressive CD4⁺ T regulatory cells (Treg) cells and myeloid-derived suppressor cells (MDSC; pathologically activated monocytes and neutrophils) (4). Under physiological conditions, they are not expressed in immune cells, however, their expression occurs upon activation and persistent antigenic stimulation. They serve as homeostatic mechanisms to protect the host from harmful tissue destructive effects of an activated and over-functional immune system. They negatively regulate effector functions of immune cells, and the immune cells expressing these checkpoints are referred to as “exhausted”. Cancer cells exploit these homeostatic mechanisms, and express ligands for the immune checkpoints in order to suppress and evade anti-cancer immune responses. Furthermore, immune and other cells present in the tumor microenvironment (TME) such as Bregs, Tregs, MDSC, cancer-associated macrophages and fibroblasts express ligands/counter receptors for the checkpoints.

The checkpoints include Cytotoxic T Lymphocyte Antigen (CTLA)-4, Programmed Death (PD)-1, Lymphocyte-Activation Gene (LAG)-3, T-Cell Immunoglobulin and Mucin-domain containing (TIM-3), V-domain Immunoglobulin Suppressor of T cell Activation (VISTA) and others (listed in Table 1 ) (3, 5). The immune checkpoints are currently being targeted through monoclonal antibodies (mAb) each of which targets a specific checkpoint. The mAb are humanized i.e., all or most of their murine sequences, except for their checkpoint binding sites, are replaced by the corresponding human sequences. The humanization reduces their antigenicity when administered to cancer patients. The mAb of IgG4 isotype lack FcR-mediated effector functions and work by inhibiting interaction of the targeted checkpoint with its ligand/counter-receptor through steric hinderance. Thus, through inhibiting interaction between a checkpoint and its ligand and/or by eliminating checkpoint or its counter-receptor bearing immunosuppressive immune cells and cancer cells, these humanized antibodies release the immune cells from inhibitory effects of the checkpoint, and augment their anti-tumor effector functions. The mAb of IG1 isotype, in addition to inhibiting a checkpoint-ligand interaction, can also mediate killing of the checkpoint ligand-positive cancer cells through antibody-dependent cell-mediated cytotoxicity (ADCC) by CD16+ NK cells and non-classical CD16+ monocytes. Alternately, these antibodies can also induce killing of the cancer cells through complement activation. They could also induce phagocytosis of the cells through antibody-dependent cellular phagocytosis (ADCP) by macrophages. The humanized antibodies are commonly referred to as immune checkpoint inhibitors (ICIs) or immune checkpoint blockers (ICBs). A list of currently targeted inhibitory immune checkpoints, their ligands/counter-receptors as well as their specific ICIs (approved and/or under clinical trials) is provided in Table 1 .

The first ICI approved by FDA in 2011 for advanced melanoma patients was Ipilumab (6), a humanized mAb of IgG1 isotype that targets CTLA-4. CTLA-4 is a CD28 homolog that binds B7 (CD80 and CD86) molecules expressed constitutively on the surface of antigen presenting cells (APC) with much higher affinity and avidity than CD28, a constitutively expressed co-stimulatory molecule on T cells ( Figure 1 ). CTLA-4 expression is intrinsically linked with T cell activation through T cell receptor (TCR), and upon this activation, CTLA-4 is rapidly mobilized from its intracellular vesicular stores to the cell surface. The expression peaks in 2-3 days following TCR engagement, and correlates with the strength of the TCR activation. It supersedes CD28 for binding with B7 molecules and thus limits/deprives T cells from CD28-induced co-stimulatory signaling via activation of PI3K and AKT (5). ICIs targeting CTLA-4 inhibit CTLA-4/B7 interactions allowing CD28/B7 interactions. This amplifies TCR-mediated T cell activation. As immunosuppressive Tregs constitutively express high levels of CTLA-4, CTLA-4-blocking antibodies ( Table 1 ) also inhibit as well as eliminate Tregs through ADCC and ADCP. These antibodies are less effective in aged patients as they have reduced numbers of Tregs. Furthermore, anti-CTLA-4 antibodies reduce expression of CD80 and CD86 on interacting APC through trogocytosis promoting immunosuppression. Soon it was discovered that monotherapy with anti-CTLA-4 ICIs leads to compensatory expression of other inhibitory checkpoints such as PD-1 and others (see below) resulting in resistance to the ICI.

PD-1, as well as its ligand, PDL-1, is now the most widely targeted checkpoint in cancer patients. Several humanized mAb for targeting PD-1 have been developed. For example, Nivolumab, Pembrolizumab, Cemiplimab and Dostarlimab target PD-1 and inhibit their interaction with PDL-1 expressed on cancer cells as well as on non-cancer cells in the tumor microenvironment (TME). All these antibodies are of IgG4 isotype, lack FcR-mediated effector functions and are meant to inhibit PD-1 interaction with PDL-1. Interestingly, anti-PD-1 ICIs are more effective in cancer patients that express PDL-1 in the TME. Such cancer patients have anti-cancer immune responses and are said to have “hot” tumors. PD-1-specific ICIs are more effective than the CTLA-4-specific ones, and are comparatively less toxic. The PD-1/PDL-1 interaction attenuates TCR-mediated signaling through activating SH-2 domain-containing protein tyrosine phosphatase (SHP)-2, which dephosphorylates several key molecules implicated in TCR-mediated signaling. Furthermore, it also blocks CD28/B7-mediated co-stimulatory signals. Anti-CTLA-4 antibodies mainly lead to the expansion of PD-1+ ICOS+T-bet+ CD4+ T cells, whereas anti-PD-1 antibodies primarily result in the expansion of CD8+ T cells. PD-1 ligation on T cells promotes fatty acid oxidation (FAO) but attenuates glycolysis, whereas CTLA-4 ligation leads to attenuation of glycolysis without affecting FAO. The blockade of the two inhibitory checkpoints reverses the metabolic changes in exhausted T cells and reinvigorates them (5, 7). Furthermore, it decreases threshold of T cell activation permitting activation of low affinity TCR resulting in so called “epitope spreading”. The low affinity TCR-bearing T cells recognize neo-antigens expressed by cancer cells. Not surprisingly, combined immunotherapy with anti-CTLA-4 and anti-PD-1, due to their synergism, gives better responses in cancer patients. The synergism also results, at least in part, from the fact that anti-CTLA-4 antibodies target circulating T cells while anti-PD-1 antibodies target tumor infiltrated T cells. The combined immunotherapy also gives better responses in PDL-1 negative tumors. The PD-1 ligand/counter-receptor, PDL-1, has also been targeted by humanized mAbs of IgG1 such as Avelumab and Atezolizumab ( Table 1 ). In addition to inhibiting PD-1/PDL-1 interaction through steric hindrance, anti-PDL-1 antibodies kill PDL-1-expressing cancer cells as well as non-cancer cells in the TME through ADCC and ADCP. As PDL-1-mediated signaling promotes aerobic glycolysis as the primary energy source, and promote cell proliferation through PIK-3/Akt/mTOR pathway, its blockage puts metabolic constraints in cancer cells (5). Not surprisingly, PDL-1 targeting ICIs show efficacy in only in PDL-1+ cancers (5, 7).

The pathophysiology of immune checkpoints is complex. Certain chemotherapeutic agents can induce expression of PDL-1 in cancer cells (8). In addition, oncogenic signals such as epidermal growth factor receptor (EGFR) activation, certain and TH1 cytokines such as IFN-γ and TNF-α also induce its expression in cancer cells as well as in non-cancer cells (such as fibroblasts and myeloid cells) in the TME. PDL-1 expressing macrophages inhibit infiltration of T cells in the TME (9). In contrast to TH1 cytokines, TH2 ones induce expression of PDL-2 on APC; not much is known about its significance in the checkpoint immunotherapy. Concerning, PD-1, persistent antigenic stimulation, IFN-α and -β and TLR stimulation induce expression of PD-1 in T cells and other immune cells such as macrophages, NK and B cells (10, 11). Other than being expressed in immune cells, CTLA-4 and PD-1 are also expressed in cancer cells. In vitro, anti-CTLA-4 antibody activates EGFR pathway in CTLA-4 expressing cancer cells, induces cell proliferation as well as the expression of PDL-1 in the cancer cells. Tumor-intrinsic PD-1 is also expressed in a subset of PDL-1⁺ cells in a broad range of cancer types; the two act in cis (when present on the same cancer cell) and suppress cancer cell proliferation in vitro in the absence of adaptive immunity. However, the interaction also makes the cancer cells resistant to the ICIs (12). In addition to their membrane-bound forms, the checkpoints such as CTLA-4, PD-1, PDL-1 and others are also expressed in soluble (s) forms (13, 14), which result either from their mRNA splice variants or from the shedding of their membrane-bound forms through proteolytic cleavage by matrix metalloproteases (MMP)-7, 9 and 13, as well as by the ADAM (a disintegrin and metalloproteinase) family member-10 and -17. The soluble forms bear implications for ICI immunotherapy. For example, sCTLA-4 binds B7 molecules (B7-1/CD80 and B7-2/CD86) on antigen presenting cells (APC) and prevents their interaction with the T cell co-stimulatory molecule CD28. In a murine model of cervical adenocarcinoma, sCTLA-4 was shown to attenuate CD8+ T cells and promote cancer progression (14). On the other hand, sPD-1 blocks PD-1/PDL-1 interactions (acts as an immune checkpoint blocker); higher plasma concentrations of sPD-1 in untreated anal and pancreatic ductal adenocarcinoma patients correlate with better prognosis (11). Unlike sPD-1, sPDL-1 can bind with PD-1 on immune cells and inactivate them; increased levels of sPDL-1 in cancer patients correlate with worse prognosis. It acts as a decoy for anti-PDL-1 antibodies necessitating higher doses of ICIs (11, 15).

It is noteworthy that PD-1 is a marker of cell activation, and only a subset of PD-1+ T cells are exhausted. The exhausted T cells also express other inhibitory immune checkpoints most notably LAG-3 and TIM-3. Accordingly, in addition to CTLA-4 and PD-1/PDL-1, humanized antibodies for these and several other checkpoints have been developed and are in various stages in clinical development. The checkpoints include LAG-3, TIM-3, TIGIT, ILT2/4, VISTA, NKG2A and CD24; see Table 1 for their expression, ligands/counter receptors, blocking antibodies and potential mechanisms of action. Many studies have been undertaken to investigate dual targeting of PD-1/PDL-1 with LAG-3, TIM-3 or TIGIT with better and less toxic side effects as compared with the combination of PD-1/PDL-1 with CTLA-4 (16). Furthermore, antibodies have been developed that simultaneously target two different checkpoints. For example, Cadonilimab is a tetravalent bispecific antibody of IgG1 isotype that targets CTLA-4 and PD-1 simultaneously (17). It has been approved for advanced or relapsed cervical, gastric and gastro-esophageal junction cancers. In addition, Lomvastomig targets both PD-1 and TIM-3, and KN046 targets CTLA-4 and PDL-1 (16, 18). Studies in murine models have shown that simultaneous targeting of two or more different checkpoints is more effective in regressing cancers. However, they are also more toxic.

ICI immunotherapy has been used as a stand-alone and/or as an adjunctive therapy (in combination with chemotherapy) in a wide variety of cancers including melanoma, breast cancer, NSCLC, RCC, colorectal cancer (CRC), prostate, pancreatic and breast cancers as well as in hematological malignancies with variable response rates (RR). Depending upon the type of the cancer, the RR may vary between 20-50%. The responses have been in general better in advanced stage metastatic cancers. Unfortunately, only a small subset of patients responds to ICIs in each cancer type. Exact reasons for a low response are not known. It has been observed that tumors with higher mutational burden, microsatellite instability, inflamed stroma as well as increased infiltration of immune cells respond better to ICIs. Furthermore, altered signaling pathways and genetic polymorphism in checkpoint genes such as CTLA-4, PDCD (which encodes PD-1) and PDL-1 may play a role in variable responses to ICIs in cancer patients (19).

A major limitation of ICI immunotherapy is that it is accompanied by moderate to severe immune-related adverse events (ir-AE) including nausea, vomiting, dysphagia, skin rashes, cytokine release syndrome (CRS), dermatitis, colitis, hepatitis, hypophysitis, thyroiditis/hypothyroidism, myocarditis and neurological disease ( Figure 1 ). The ir-AE are different from those caused by chemotherapy. They are likely to result from the above normal activation of T cells due to a lowered threshold of TCR activation. Consequently, there occurs activation of T cells with low affinity TCR that would otherwise not be activated. Furthermore, an early increase in clonal proliferation of PD-1+ CD21-low B cells was observed and correlated with ir-AE in cancer patients following anti-PD-1 and anti-CTLA-4 therapies (5, 20). It is noteworthy that mice deficient in CTLA-4, PD-1 or PD-L1 develop various lymphoproliferative disorders and autoimmune manifestations (21). Not surprisingly, in their clinical presentation, ir-AE resemble autoimmune diseases. They may be acute or chronic, and may persist even after cessation of the immunotherapy. In up to 1.23% of the patients, the ir-AE could be fatal (22, 23). These events often lead to discontinuation of ICI immunotherapy. Irony is that effectiveness of the ICI immunotherapy correlates with their toxicity. To counter these adverse events, and to increase tolerability of ICIs, steroidal anti-inflammatory drugs (SAID) such as prednisone, and non-steroidal anti-inflammatory drugs (NSAID) such as Ibuprofen, intravenous immunoglobulins (IVIG) and TNF-α inhibitors, etc., have been used (24–26). However, these medications were found to be associated with worse clinical outcomes in ICI therapies. More recently, researchers have resorted to using cannabinoids instead of SAIDS or NSAIDS to limit ir-AE and increase tolerability of ICIs.

Cannabinoids refer to 21-carbon terpenophenolic background compounds and their derivatives isolated from the plant Cannabis sativa (C. sativa), commonly known as Marijuana. They were named cannabinoids signifying their origin from the cannabis plant. The plant has been used for medicinal and recreational purposes since millennia. To date, more than 125 different cannabinoids have been identified in this plant. The cannabinoids from the cannabis plant are more specifically phytocannabinoids. Other cannabinoids include endocannabinoids, which are produced in our bodies and synthetic cannabinoids, which include various structurally diverse cannabimimetic compounds that are synthesized in laboratories. Other names used for phytocannabinoids are natural cannabinoids or exocannabinoids (27–29). Two most studied phytocannabinoids are Δ9-tetrahydrocannabinol (THC) and cannabidiol (CBD). They are derived from their precursor molecule, cannabigerolic acid (CBGA), in cannbis. About 95% of the cannabinoids in cannabis plants exist in their acid forms such as THC-acid (THCA), cannabidiolic acid (CBDA) and cannabichromenic acid (CBCA), which are synthesized from CBGA through the action of cannabinoid-specific synthases. The acidic forms protect cannabis plants acting as anti-oxidants, insecticides and microbicides. The acidic forms are non-enzymatically decarboxylated to yield neutral cannabinoids THC, CBD, CBC and CBG (from CBGA) when exposed to high temperatures, light and oxygen as well as prolonged storage (29, 30). It may be relevant to mention here that THCA is converted to THC upon smoking or vaping cannabis. Furthermore, CBD found as nano-emulsion in energy drinks is converted to psychoactive THC under acidic conditions of the stomach (31).

Relative to THC and CBD, CBC, CBG, cannabinol (CBN), cannabidivarin (CBDV), tetrahydrocannabivarin (THCV), hexahydrocannabinol (HHC) and Δ-8 tetrahydrocannabinol (Δ-8 THC) are also found in minor or trace amounts in cannabis (30, 32). Of these, CBN is an oxygenation product of THC; CBDV and THCV have propyl side chains instead of usual pentyl ones in CBD and THC from whom they are derived; HHC is a hydrogenated derivative of THC, and Δ-8 THC is an isomer of THC with a double bond beginning at position 8, instead of position 9. In addition, Abnormal-CBD (Ab-CBD) is a synthetic regio-isomer of CBD, and is found as an impurity when cannabinoids are synthesized in laboratory (33). Of the cannabinoids, THC exerts its psychoactive effects mainly through its ability to activate the first discovered cannabinoid receptor (CB1), and is also responsible for the psychoactive effects of cannabis. Some other cannabinoids such as HHC, Δ-8 THC and CBN also have psychoactive effects, however they are less potent than THC in equivalent doses. On the other hand, CBD and CBG lack psychoactive effects. In fact, they counter some of THC’s psychoactive effects (30). In vitro, Δ-8 THC is synthesized from CBD, and is illegally marketed as it is not specifically prohibited in the 2018 Farm Bill in the USA (31).

The genus cannabis contains three species, C. sativa, C. indica and C. ruderalis representing low, high and intermediate levels of THC-containing species, respectively. Most strains or cultivars currently grown are hybrid species. They are grown to obtain fiber, seeds (rich in unsaturated fatty acids for edible oil) or for recreational/medicinal purposes. Those used for obtaining fiber and food are legally required to contain < 0.03% THC (by dry weight) and are called ‘hemp’; others (drug type) contain as much as 30% THC. Cannabis strains or cultivars are selectively bred for specific traits, for example for their THC, CBD, CBC, unsaturated oils or fiber contents (34, 35). It is noteworthy that in addition to cannabinoids, cannabis plants also contain a diverse array of many other compounds including terpenoids (e.g., β-caryophyllene), flavonoids (e.g., quercetin), phenols and other phytochemicals. Current trend is to define cannabis cultivars based upon their biochemical constituents into “chemovars” (34).

Several cannabinoid preparations are currently available commercially. Two synthetic THC analogs (Nabilone and Dronabinol) have been approved by FDA for chemotherapy-induced nausea, anorexia and vomiting in cancer patients (36). Dronabinol is also used in people living with HIV (PLWH) for improving appetite and preventing weight loss. The THC preparations are also used for other therapeutic applications including glaucoma, migraine, headache, anorexia, spasticity, anxiety, and pain. The second major phytocannabinoid, CBD, is well known for its anti-convulsive and neuroprotective effects. Nabiximols (Sativex) is a cannabis extract containing THC and CBD in 1:1 molar ratio and is used as buccal spray for alleviating neuropathic pain, multiple sclerosis spasticity and overactive bladder. On the other hand, Epidiolex, contains purified CBD from cannabis plants and is approved for controlling seizures in therapy-resistant childhood epilepsy as well as in a rare and severe form of epilepsy, namely Lennox-Gastaut syndrome or Dravet syndrome. CBD has great therapeutic potential due to its antipsychotic, antidepressant, anxiolytic, anti-inflammatory and analgesic effects (37). It is non psychoactive, has little potential for abuse and is often the preferred choice for cannabis-based therapies. Another preparation, Spectrum Yellow Oil contains CBD (20 mg), THC (0.9 mg) and CBC (1.1 mg) per ml and a candidate for treating neuroinflammatory conditions (38).

Studies on THC’s psychoactive effects resulted in the identification of the first cannabinoid receptor (CB1) in human and rat cells in 1988 (39). Search for endogenous ligands of CB1 led to the discovery of endocannabinoids, the cannabinoids produced in our bodies. Two most studied endocannabinoids include N-arachidonoylethanolamine (AEA; also known as anandamide meaning bliss in Sanskrit) and 2-arachidonoylglycerol (2-AG). They bind and activate several canonical and non-canonical cannabinoid receptors and exert their biological effects ( Table 2 and discussed below). A variety of tissues in the body including brain, muscle, fatty tissue, spleen, liver and pancreas as well as immune and non-immune cells produce small quantities of endocannabinoids, and can be measured in the circulation (40). Endocannabinoids are synthesized in the body from arachidonic acid (AA) and phospholipids, important constituents of the plasma membrane ( Figure 2 ). The endocannabinoid system (ECS) maintains physiological homeostasis in the body by interacting with several neurotransmitters (acetylcholine, norepinephrine, glutamate, γ-aminobutyric acid, dopamine, serotonin and endorphins) and immune system. The system, in addition to endocannabinoids, comprises enzymes that synthesize endocannabinoids, enzymes that degrade endocannabinoids in the body as well as several receptors through which endocannabinoids mediate their biological effects (41). In addition to two main endocannabinoids, AEA and 2-AG, a large number of cannabinoid-like lipid mediators have been identified in the body that are related to the N-acylethanolamines (NAE) or Acylglycerol (AG) families. They include 2-arachidonyl glyceryl ether (Noladin ether), O-arachidonyl ethanolamine (Virodhamine), N-arachidonyl dopamine (NADA), N-palmitoyl ethanolamide (PEA), oleoylethanolamine (OEA), 1-palmitoylglycerol (1-PG) and 2-palmitoylglycerol (2-PG), etc. Endocannabinoids have very short half-life, are rapidly taken up by cells through transporters and are rapidly degraded. Two enzymes, fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL) hydrolyze and degrade AEA and 2-AG in post-synaptic and pre-synaptic neurons, respectively. FAAH is the rate limiting enzyme for degrading and liberating AA from AEA and related endocannabinoids of the NAE family. The inhibition of FAAH results in increased concentrations of AEA and suppresses inflammation in a mouse model of colitis (42). Several FAAH selective inhibitors have been developed (43). In addition, MAGL-specific inhibitors have also been developed; they degrade 2-AG and other endocannabinoids of the AG family (44). Importantly, inhibitors of the endocannabinoid degrading enzymes are being researched for potential clinical applications. Interestingly, phytocannabinoids can also regulate endocannabinoids in the body. Being lipophilic substances, phytocannabinoids permeate through plasma membrane and are chaperoned by intracellular fatty acid binding proteins (FABP)-3, 5 and 6. The FABP transport them to FAAH for degradation. The phytocannabinoids, THC and CBD, compete with endocannabinoids for binding with the FABP, and decrease FAAH-mediated degradation of AEA and related endocannabinoids. Recently, CBC was shown to inhibit MAGL and the cellular re-uptake of endocannabinoids (38). Normally, the plasma concentrations of AEA and its congeners relative to 2-AG and its congeners are higher in adult humans but the balance may change in disease conditions (45).

As discussed above, phytocannabinoids indirectly increase endocannabinoid tone in humans (46). However, it has to be noted that there is an intricate cross talk between endocannabinoid and prostanoid systems. The hydrolysis of endocannabinoids yields AA, a substrate for cyclooxygenase (COX) enzymes. The enzymes also metabolize EAE and 2-AG into prostaglandin-ethanolamides (PG-EA) and prostaglandin-glycerol esters (PG-G), respectively. The COX-generated metabolites such as PGD-2, PGE2, PGI-2, PGF2a-EA, and PGE2-G exert pro-inflammatory and pro-algesic effects countering some of cannabinoids’ biological effects (47). It follows that concurrent targeting of COX enzymes or PG synthases would augment beneficial effects of phytocannabinoids, FAAH and MAGL inhibitors concerning nociception and inflammation.

Typically, endocannabinoids are not produced constitutively in the body. Immune cells such as T and B cells, platelets and macrophages produce and release them upon activation (40). In the brain, endocannabinoids are produced on demand when certain post-synaptic receptors such as glutamate or acetylcholine are activated (28). The released endocannabinoids then act on pre-synaptic neurons in a retrograde fashion and generally inhibit the release of neurotransmitters such as dopamine, gamma-amino butyric acid, glutamate, serotonin, opioids and noradrenaline, etc. In addition, endocannabinoids are tonically produced by hippocampal neurons in response to hunger (28).

Cannabinoids are mainly consumed orally, inhaled (smoked or vaped) or used as oromucosal (sublingual or inside cheeks) spray (48). The bioavailability of the consumed cannabinoids varies greatly depending upon their mode of consumption. When inhaled, cannabinoids enter blood stream through lungs within a few minutes. The bioavailability of THC and CBD via this route is 10-35% and 11-45%, respectively. Through vaping, cannabinoids can be inhaled without burning. During vaping, volatile cannabinoids are vaporized and inhaled when hot air passes through cannabis. Vaporization, like smoking, leads to higher availability of cannabinoids. Users can experience psychotropic effects within 15-30 minutes; the effects lasting for 2-3 hours. However, their bioavailability upon ingestion is only 4-10% for THC and 6% for CBD (49). In most of the experimental studies, cannabinoids are administered orally and no assessment is made as to the plasmatic concentrations of these cannabinoids or their metabolites in the patients. Being highly lipophilic, cannabinoids accumulate in adipose tissues of the body and are released over prolonged periods of time. For this reason, chronic users of cannabinoids may experience less severe withdrawal symptoms upon their cessation (48).

Drug metabolizing enzymes, Cytochrome P 450 (CYP450) and uridine diphosphate (UDP)-glucuronosyltransferases (UGT) are implicated in the metabolism of drugs and xenobiotics including cannabinoids (50, 51). They are mainly expressed in the mitochondria and endoplasmic reticulum of the cells in the liver and to a lesser extent in the intestines, lungs and kidneys. CYP450 are a superfamily of 57 isozymes that metabolize drugs through phase I reactions (oxidation, reduction and hydrolysis); of these, three subfamilies CYP1, 2 and 3 playing a major role in humans. UGT metabolize drugs and/or their phase I metabolites mainly through glucuronidation. The metabolic pathways render drugs more hydrophilic for clearance through hepatobiliary and renal excretory systems. The major psychoactive cannabinoid THC is mainly metabolized by CYP2C9 and CYP2C19 to 11-OH-THC (as psychoactive as THC) and then oxidized to inactive 11-COOH-THC, which is further converted to 11-COOH-THC-glucuronide by UGT1A and UGT1A3, and excreted (51, 52). The other major non-psychoactive cannabinoid CBD is mainly hydroxylated at position 7 by CYP2C9 and CYP2C19 into 7-OH-CBD, which is biologically active like CBD. On the other hand, CYP3A4 hydroxylates CBD at positions other than 7, and converts it into inactive metabolites; chemical inhibition of the CYP3 increases 7-OH-CBD (53). Rifampicin, a CYP3A4-inducing antibiotic decreases concentrations of CBD and 7-OH-CBD whereas Ketoconazole, a CYP3A-inhibiting fungicide, increases CBD and 7-OH-CBD in the circulation. The 7-OH-CBD metabolites are further metabolized to inactive 7-COOH-CBD and are excreted from the body. Un-metabolized CBD is also excreted from the body. The CBD metabolism involves very minor contributions from UGT enzymes. A fraction of it as well as of its metabolites are glucuronidated by UGT1A9 and UGT2B7 and are eliminated from the body.

Cannabinoids exert their biological effects through canonical and non-canonical cannabinoid receptors (54) ( Table 2 ). Their canonical receptors include CB1 and CB2, whereas non-canonical ones include G-protein coupled receptor (GPR)-55, GPR-18, different ion channels and peroxisome proliferator-activated receptors (PPARs). Through these receptors, cannabinoids activate multiple cell signaling pathways. They are discussed below:

GPR are well known for their propensity to form heteromers with other GPR. Being GPR, CB1 and CB2 also form heterodimers with other GPR including human HER-2, CXCR4, adenosine receptor (A2_A) and dopamine receptor D2, etc. The heterodimers, in response to cannabinoids, result in altered signaling cascades. The hetero-dimerization affects important physiological and pathological processes. For example, CB1 forms heterodimers with adenosine receptor A_2A; motor depressant effects of a CB1 agonist were completely blocked by A2_A antagonists (116). CB2 forms heterodimers with HER2 and CXCR4. The hetero-dimerization with HER2 stabilizes the growth factor receptor and indicates poor prognosis for HER2+ cancers such as breast cancer. The dissociation of the heterodimer with CB2 agonist, THC, was shown to cause degradation of HER-2 and the cancer cell death (66). On the other hand, CB2 hetero-dimerization with CXCR4 affects migration of CXCR4+ hematopoietic cells. CB2 ligands inhibit migration, invasion and metastasis of CXCR4+ cancer cells (67, 117). CB1/D2 heteromers, when activated by CB1 agonists, induce signaling via Gαs, rather than Gαi protein, and activate adenylate cyclase and production of cAMP (118).

Overall, cannabinoids induce pain relieving, orexigenic, analgesic, anti-emetic, anxiolytic and anti-inflammatory effects through many of their canonical and non-canonical receptors. Many of these receptors also induce anti-tumor effects through their pro-apoptotic and metastasis inhibitory effects. However, cannabinoids also exert immunosuppressive effects. Furthermore, cancers have evolved strategies to evade anti-tumor effects of cannabinoids by upregulating pathways and/or accumulating variants of the receptors that promote their proliferation and spread. There are many studies that have investigated the effects of cannabinoids on different types of cancer. This review focuses at understanding the impact of cannabinoids on ICI immunotherapy in cancer patients. For the effects of cannabinoids on cancer, interested readers are referred to recent reviews (119–121).

ICIs are only effective in a subset of cancer types and patients and are often accompanied by ir-AE, which leads to suboptimal cancer therapies because of dose reductions or switches to traditional but less effective oncological treatments. There are preclinical and clinical studies which have shown the effect of cannabinoids on irAE.

ICIs are mainly administered intravenously and are distributed rapidly into various body tissues and fluids. However, their access to tumor microenvironment depends upon the latter’s composition. Being monoclonal antibodies, ICIs are catabolized by proteolytic enzymes in the body with half-lives ranging from several days to a few weeks (131). As ICIs are not metabolized in the body by CYP450 enzymes, cannabinoids are not likely to affect their pharmacokinetics. Thus, xenobiotics like cannabinoids have little capacity to affect pharmacokinetics (absorption, distribution, metabolism and excretion) of ICIs. Cannabinoids, however, could potentially modulate pharmacodynamics of ICIs. In CRC, CBD was shown to rewire TME, decreases alternate activation of macrophages and increase expression of IFN-γ and IFN-α. It also enhanced efficacy of anti-PD-1 ICIs through increased expression of PDL-1 (128). Furthermore, CBD was reported to stimulate the expression of PDL-1 through activating cGAS-STING pathway in triple negative breast cancer cells. It also enhanced efficacy of Atezolizumab, an anti-PDL-1 ICI (129). In contrast, CBD and/or THC were shown to reduce PDL-1 expression by pancreatic cancer and pancreatic stellate cells (130). These studies suggest cancer-type specific effects of cannabinoids on the expression of immune checkpoints. Clearly, further research is needed on this topic.

Despite anti-cancer effects of cannabinoids demonstrated in several in vitro studies and animal models, such beneficial effects have not been consistently observed in human patients (132). The discordance may be attributed to variations in cancer types, treatment regimens, patient’s genetics, and microbiota. The same variables may lead to divergent results in the studies examining the impact of cannabinoids in conjunction with ICIs. The variables are discussed below.

Cannabinoids may relieve pain, nausea, anorexia and anxiety, and improve quality of life in cancer patients. ICI immunotherapy, although very successful in a subset of cancer patients, is accompanied by moderate to severe ir-AE that necessitate discontinuation of the immunotherapy in many patients; the ir-AE may persist even after discontinuation of the therapy (1). Because of the anti-inflammatory and immunosuppressive effects of cannabinoids, researchers have contemplated their use in combination with ICI immunotherapy. Results from a few studies conducted so far strongly suggest that the use of medicinal cannabis in cancer patients attenuates many of the ir-AE associated with the use of ICIs and increase its tolerability. However, no significant beneficial effects on overall survival, progression free survival or cancer relapses were observed; rather some of the studies noted adverse effects of concurrent administration of cannabinoids with the immunotherapy on clinical outcomes (163). Because of cannabinoids’ well documented immunosuppressive effects mediated through CB2 (122), we propose considering these molecules as an inhibitory immune checkpoints. A simultaneous neutralization of this checkpoint may lead to better clinical outcomes in cancer patients receiving ICIs concurrent with cannabinoid treatment. In this regard, cannabinoids such as CBD and CBG, with little agonism for CB2, may be better therapeutic choices. They also lack psychotoxic effects of THC. Additional strategies e.g., the use of MAGL inhibitors that decrease lipogenesis and formation of lipid bilayers in cancer cells (164) may also be explored. They increase concentrations of 2-AG and their congeners in the body. However, they liberate AA which is metabolized by COX into pro-inflammatory prostaglandins (PG). Furthermore COX also metabolize 2-AG to produce PG-glycerols. These COX-generated lipid mediators counter some of the endocannabinoids/phytocannabinoids’ beneficial effects on cancer progression and hypo-algesia. A simultaneous use of cannabinoids and COX-inhibitors may yield better results and should be explored. Future studies should take into consideration CYP450 genotypes and haplotypes and cannabinoid-drug interactions for personalized cannabis-based therapies in cancer patients receiving cannabinoids along with ICIs. This may lead to rational knowledge-based regimens tailored to individual cancer patients. Despite several studies demonstrating anti-cancer effects of cannabinoids primarily in pre-clinical models, these effects were not confirmed in cancer patients. It is very difficult to reconcile results from different studies due to heterogeneous nature of the cancers studied, unstandardized cannabinoid preparations, populations of patients with already advanced disease stages, different genetic makeup of cancer patients with respect to their HLA genotypes and CYP450 genotypes and haplotypes, patients’ unique gut microbiota as well as different genetic and somatic variations occurring in the cannabinoid receptors and their signaling pathways in the cancer cells. Unfortunately, many of the studies that investigated the impact of cannabinoids on cancer progression were carried out in the absence of anti-cancer immune responses. Finally, purified and standardized cannabinoids such as CBD and CBG that lack psychoactive effects should be investigated should be investigated in cancer patients receiving ICI immunotherapy.