11.1 Introduction

Preparations from Cannabis sativa, the hemp plant, have been used for centuries for both medicinal and recreational purposes (Howlett et al. 2002; Mackie 2006). Isolation of the active components of the plant, called cannabinoids, in 1960s, as much as subsequent cloning of cannabinoid receptors, discovery of their endogenous ligands and development of synthetic cannabinoids contributed to an intensive burst of cannabinoid research. Along with our expanding comprehension of mechanisms of cannabinoids action, targeting cannabinoid signaling for therapeutic purposes has inevitably emerged as an interesting area of scientific and clinical investigations.

One of the most extensively studied applications of cannabinoids is their potential use as anti-cancer agents. Anti-proliferative effects of cannabinoids have been reported in various cultured cancer cells, including neural, breast, prostate, skin, thyroid cancer cells and leukemia cells. Several studies demonstrated anti-tumor activity of cannabinoids in animal models (Guzman 2003). Since the first publication by Sanchez and co-workers in 1998, providing evidence that cannabinoids are effective in inducing glioma cell death, a growing interest of several groups, including ours, has been focused on understanding of molecular mechanisms of cannabinoid signaling in glioma cells and therapeutic potential of cannabinoids in glial tumors (Sanchez et al. 1998).

Signaling pathways and intracellular processes underlying cannabinoid action on glioma cells are reviewed here. Due to the number of studies carried out in the recent years on mechanisms of anti-proliferative effects of cannabinoids, these mechanisms are described in details. This chapter includes a brief overview of the endocannabinoid system, canonical signal transduction pathways coupled to the activation of cannabinoid receptors as well as data on alterations in the endocannabinoid system in gliomas, which may be of importance for tumor pathobiology and patient prognosis. We provide a comprehensive up to date summary of reported changes in the cannabinoid receptor expression, the endocannabinoid levels and activity of the enzymes involved in the endocannabinoid metabolism in these tumors. We also describe the latest findings on cannabinoid action in the tumor microenvironment. Beyond blocking of tumor cells proliferation cannabinoids were shown to inhibit angiogenesis as well as invasiveness and the stem cell-like properties of neoplastic cells in glioma tumors. We also report on current scientific and clinical data relevant to the use of cannabinoids in treatments of glioblastomas.

11.2 Cannabinoids and Their Receptors

Cannabinoids are a group of structurally heterogeneous but pharmacologically related compounds classified into three subtypes: plant-derived, synthetic and endogenous cannabinoids (Fig. 11.1). Plant-derived cannabinoids (phytocannabinoids) are uniquely found in the cannabis plant. Although the pharmacology of the majority of them is unknown, (−)-trans-Δ9-tetrahydrocannabinol (Δ9-THC) is recognized as the most potent out of approximately 70 identified phytocannabinoids. Various modifications of the chemical structure of natural cannabinoids led to generation of a still growing set of synthetic cannabinoids. Exogenous cannabinoids mimic the action of endogenous compounds, known as endocannabinoids, ubiquitously produced in both vertebrate and invertebrate tissues.

Fig. 11.1
figure 1

Chemical structures of cannabinoids. Plant-derived Δ9-THC ((−)-trans-Δ9-tetrahydrocannabinol), endocannabinoids: anandamide (N-arachidonoylethanolamine) and 2-arachidonoylglycerol, synthetic cannabinoids JWH133, selective for non-psychoactive CB2 receptor ((6aR,10aR)-3-(1,1-Dimethylbutyl)-6a,7,10,10a-tetrahydro-6,6,9-trimethyl-6H-dibenzo[b,d] pyran), and WIN55,212-2, a CB1/CB2 receptor agonist ((R)-(+)-[2,3-Dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenylmethanone)

Full size image

Two arachidonic acid derivatives, arachidonoylethanolamide (anandamide, AEA) and 2-arachidonoylglycerol (2-AG) are the best characterized endogenous cannabinoids, although some other amides and esters of long chain polyunsaturated fatty acids also exhibit cannabimimetic properties. Endocannabinoids are produced as rapidly inactivated lipid mediators and their levels are strictly controlled by a transporter system and hydrolyzing enzymes (Fig. 11.2). Biosynthesis of endocannabinoids is activated “on demand”. AEA is produced from lipid membrane precursors upon the stimuli that increase intracellular concentrations of calcium. Increased calcium levels activate N-acyltransferase, the enzyme generating an AEA precursor N-arachidonylphosphatidylethanolamide (NAPE), which is then hydrolyzed by NAPE phospholipase D. Separate mechanisms have been described for 2-AG, which is most likely produced through the phospholipase C/diacylglycerol lipase pathway. Contribution of the alternative pathways for endocannabinoid synthesis, which were delineated in other tissues, to the brain pool of AEA and 2-AG remains to be elucidated (Lu and Mackie 2016). Endocannabinoid production and release by glial cells may underlie the neuroprotective properties of cannabinoids in experimental models, however its physiological significance is still unknown.

Fig. 11.2
figure 2

The endocannabinoid system. Two arachidonic acid derivatives, anandamide (AEA) and 2-arachidonoyl glycerol (2-AG), are produced by cleavage of lipid precursors in response to increased calcium levels or receptor stimulation. AEA is generated from N-arachidonoylphosphatidylethanolamine (NAPE). NAPE originates from transfer of arachidonic acid from the sn-1 position of 1,2-sn-diarachidonoylphosphatidylcholine (PC) to phosphatidylethanolamine (PE), catalyzed by a Ca2+-dependent N-acyltransferase (NAT). NAPE is cleaved by phospholipase D (PLD) to release AEA and phosphatidic acid. 2-AG is synthesized in two steps via generation of 1-acyl-2-arachidonoylglycerol (diacylglycerol, or DAG) form phospholipids (such as phosphatidylinositol-4,5-bisphosphate PIP2) by phospholipase C (PLC) and subsequent hydrolysis of DAG by a diacylglycerol lipase (DGL). AEA and 2-AG are rapidly removed from the extracellular space possibly through a common purported high-affinity transporter (AMT), a carrier possibly working in both inward and outward directions. Once taken up by cells, AEA is a substrate for the fatty acid amide hydrolase (FAAH), which breaks the amide bond and releases arachidonic acid (AA) and ethanolamine (EtNH2). 2-AG is primarily degraded by a specific monoacylglycerol lipase (MGL). Both AEA and 2-AG bind to and activate CB1 and CB2 receptors; however, AEA is a weaker agonist than 2-AG at CB1 and is only a partial agonist at CB2. The transient receptor potential cation channel subfamily V member 1 (TRPV1) is another key molecular target of AEA, but importantly not of 2-AG. The binding site of TRPV1 receptors for AEA is on an intracellular domain. All elements of the ECS are located in the plasma membrane except from FAAH, which is bound to intracellular membranes, and MGL, which is cytosolic

Full size image

Although endocannabinoids, lipophilic in nature, can freely cross cell membranes, evidence suggests the existence of mechanisms facilitating endocannabinoid internalization. Endocannabinoids are transported into cells by a purported high-affinity membrane transporter to undergo enzymatic hydrolysis. Two degradative enzymes for endocannabinoids have been best described so far: the fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MGL), responsible mainly for AEA and 2-AG degradation, respectively (Fig. 11.2). FAAH and MGL exhibit a wide distribution in different structures of the rat and human central nervous system, but are mostly confined to neurons. FAAH is primarily a postsynaptic enzyme, whereas MGL localizes to presynaptic terminals. There are conflicting data on the expression of endocannabinoid metabolizing enzymes in glial cells in the normal brain. FAAH is increased in hypertrophic astrocytes surrounding AD plaques or inflammatory infiltrates but is not detected in other types of glial cells (Pazos et al. 2005). MGL can be found in oligodendrocytes at different developmental stages, however in cultured oligodendrocyte progenitor cells (OPCs) MGL expression was lower than during maturation stages, suggesting that the increased levels of 2-AG are required for OPCs differentiation (Gomez et al. 2010). Interestingly, an alternative pathway for AEA degradation is via oxidation by cyclooxygenase 2 (COX-2). The differences in structure between arachidonic acid and anandamide are sufficient to allow the development of COX-2 inhibitors that inhibit anandamide oxidation without affecting prostaglandin formation. Furthermore, COX-2 is reasonably selective for anandamide over other acyl ethanolamides, so its inhibition offers a more selective way to increase anandamide compared with inhibition of FAAH (Hermanson et al. 2014). 2-AG degradation can be also mediated by alpha/beta domain containing hydrolase 6 (ABHD6) and alpha/beta domain containing hydrolase 12 (ABHD12) (Lu and Mackie 2016).

Cannabinoids elicit a wide range of central and peripheral effects, which are mediated mostly through cannabinoid receptors (Howlett et al. 2002). There are two types of specific seven-transmembrane, Gi/o-protein-coupled receptors cloned so far, called CB1 and CB2 (Fig. 11.2, Fig. 11.3), although an existence of additional cannabinoid-binding receptors has been suggested (Howlett et al. 2002; Stella 2004). CB1 and CB2 differ in their predicted amino acid sequence, tissue distribution and expression pattern. Overall homology between CB1 and CB2 is remarkably low (e.g. 44% for human and 68% for murine receptors) with significant disparities in the domains, which interact with G proteins and effector proteins. Thus, quite expectedly two cannabinoid receptors were shown to play distinct physiological roles and share only some common signaling mechanisms.

Fig. 11.3
figure 3

Mechanisms of anti-tumoral action of cannabinoids. Increased ceramide synthesis de novo via induction of serine palmitoyltransferase (SPT) plays a central role in cannabinoid-induced cell death. Signals from CB1 via adaptor protein FAN (factor associated with neutral sphingomyelinase activation) trigger also the ceramide production from sphingomyelin breakdown, catalysed by neutral sphingomyelinase (SMase). Stimulation of cannabinoids receptors leads to inhibition of adenylyl cyclase (AC), reduction of cAMP levels and decreased protein kinase A (PKA) activity. Inhibition of PKA and pro-survival pathways (Akt and ERK signaling) stimulates translocation of Bad to the outer mitochondrial membrane and its pro-apoptotic function. Interaction between Bad and Bcl-2 triggers a decrease of the mitochondrial membrane potential (ΔΨ) and release of pro-apoptotic factors (such as cytochrome c) to the cytosol, where apoptosis is executed by caspase cascade. Alternatively, induction of apoptosis by cannabinoids can be mediated by ER (endoplasmic reticulum)-stress and autophagy. Cannabinoid-induced phosphorylation of eIF2α (eukaryotic translation initiation factor 2α) and subsequent up-regulation of the stress-regulated protein p8 and ER-stress-related downstream targets: ATF4 (activating transcription factor 4), CHOP (the C/EBP-homologous protein) and TRB3 (tribbles homologue 3) leads to inhibition of Akt, an upstream activator of mTORC1. Decreased activity of Akt/mTORC1 pathway contributes to initiation of autophagy, that precedes apoptosis of glioma cell

Full size image

Many of the effects of cannabinoids rely on the CB1 receptor activation. CB1 is particularly abundant in discrete areas of the brain, such as cortex, hippocampus, basal ganglia and cerebellum, as well as in peripheral nerve terminals, where it mediates inhibition of neurotransmitter release and is involved in the control of motor activity, memory, cognition, appetite and sensory perception. CB1 is also present in some extra-neural sites, such as testis, uterus, vascular endothelium, eye, spleen, and tonsils, controlling processes such as vascular tone, intraocular pressure and immune response (Howlett et al. 2002). Glial cells have been shown to express CB1 receptors, although their precise function in astrocytes, oligodendrocytes and microglia has been only partly unveiled.

By contrast, the CB2 receptor is predominantly expressed in cells and organs of the immune system (Howlett et al. 2002). The role of peripheral CB2 receptor activation under physiologic conditions is not well defined. CB2 signaling is involved in B-cell differentiation and modulation of immune response. Increased levels of CB2 are reported in tissues during development, inflammation, injury and cancer, revealing a critical role for the CB2 receptor in regulating these processes (Howlett et al. 2002). The CB2 receptor was believed to be absent from healthy brain, however, its expression has been detected in microglia - brain macrophages (Stella 2004; Gong et al. 2006), as well as in a small subpopulation of neurons (Gong et al. 2006; Van Sickle et al. 2005). Animal experiments show that CB2-selective agonists do not induce widespread psychoactive effects attributed to activation of the CB1 receptor (Guzman 2003; Valenzano et al. 2005).

Extensive molecular studies have demonstrated that activation of the CB1 and CB2 cannabinoid receptors upon agonist binding is canonically linked to inhibition of adenylyl cyclase via the α-subunit of Gi/o-protein. The consequent decrease in cyclic AMP (cAMP) production leads to down-regulation of protein kinase A (PKA) and impedes PKA – dependent signaling. The CB1 receptor, acting as a guardian on presynaptic membranes, is coupled to ion channels, inducing for example inhibition of voltage-gated L, N- and P/Q Ca2+ channels and activation of G-protein activated inwardly rectifying K+ channels (Howlett et al. 2002). Cannabionoids have been also reported to affect several pathways that are more directly involved in the control of cell proliferation, differentiation and survival. Depending on a cell type and treatment conditions, signaling via cannabinoid receptors is linked to activation or inhibition of phosphatidylinositol 3-kinase (PI3K)/Akt pathway. Stimulation of either CB1 or CB2 receptor results in activation of three classes of mitogen-activated protein kinases (MAPKs): extracellular signal-regulated kinase (ERK), c-Jun N-teriminal kinase (JNK) and p38 MAPK. Cannabinoid receptors are also coupled to activation of phospholipase C and a subsequent release of Ca2+ from inositol-1,4,5-trisphosphate (IP3)-sensitive stores, as well as to modulation of the sphingomyelin cycle (Howlett et al. 2002; Guzman 2003).

There is an increasing number of data suggesting that additional receptors act as targets for cannabinoids. Peroxisome proliferator-activated receptors (PPARs), G-protein coupled receptor 55 (GPR55), transient receptor potential cation channel subfamily V member 1 (TRPV1, also known as capsaicin or vanilloid receptor) and member 2 (TRPV2) have been shown to be activated by cannabinoids, including Δ9-THC, another phytocannabinoid – cannabidiol (CBD), and endogenous AEA. However, the precise role of these receptors in cannabinoid signaling is a subject of ongoing research (Mackie and Stella 2006; Lu and Mackie 2016).

11.3 Cannabinoid System in Gliomas

Endocannabinoids, their receptors and specific machinery involved in biosynthesis, uptake and degradation constitute the endocannabinoid system (ECS). A growing array of data suggests that alterations of a balance in the cannabinoid system between the levels of endogenous ligands and their receptors occur during malignant transformation in various types of cancer, including gliomas (Table 11.1). While non-transformed astrocytes express only the CB1 cannabinoid receptor, both types of functional cannabinoid receptors have been found in several established human glioblastoma cell lines, as well as in primary cultures derived from the most malignant brain tumor, glioblastoma multiforme (GBM) (Howlett et al. 2002; Galve-Roperh et al. 2000; Sanchez et al. 2001). Immunohistochemical analysis of low and high grade human glioma surgical specimens revealed increased CB2 receptor expression in tumor cells, invading microglia/macrophages and endothelial cells of the tumor blood vessels, as compared to non-tumor brain samples (Schley et al. 2009; Ellert-Miklaszewska et al. 2007; Sanchez et al. 2001; Wu et al. 2012). We detected the presence of CB2 receptors in all analyzed biopsies of astrocytomas and glioblastomas. The proportion of malignant tumors expressing high levels of CB2 (10 out of 16, 62.5%) was over twofold higher than that seen in the tumors of lower grade (7 out of 29, 24%) (Ellert-Miklaszewska et al. 2007). Thus, the extent of CB2 expression correlated with the tumor malignancy grade. Interestingly, some benign pediatric astrocytic tumors, such as subependymal giant cell astrocytoma (SEGA), which may occasionally cause mortality owing to progressive growth in some patients, also displayed high CB2-imunoreactivity. Moreover, as observed by Sanchez et al., CB2 receptor immunoreactivity markedly prevailed over detected CB1 receptor levels in grade IV astrocytomas (Sanchez et al. 2001). According to some studies the levels of CB1 receptor expression in tumor and tumor-associated endothelial cells were not significantly different from the control tissue and showed no dependence on tumor grade (Sanchez et al. 2001; Schley et al. 2009; Held-Feindt et al. 2006). However, the data on CB1 receptor expression are not consistent. De Jesus et al. (2010) showed the decreased CB1 receptor expression in tumor samples, in contrast to others that presented unchanged (Held-Feindt et al. 2006) or even increased CB1 receptor expression in high grade glioblastomas as compared to low grade gliomas and non-tumoral brain tissue (Wu et al. 2012). Of note, the pediatric low grade gliomas, which remained stable or that presented spontaneous involution after sub-total surgical resection, showed significantly higher CB1 receptor expression at the time of diagnosis. The authors hypothesize that high expression levels of CB1 receptor provide tumor susceptibility to the antitumor effects of circulating endocannabinoids like anandamide, resulting in tumor involution (Sredni et al. 2016).

Table 11.1 Alterations of the major components of the cannabinoid system in gliomas
Full size table

A link between CB2 expression and malignancy grade of the tumor has been reported also in prostate, breast and pancreatic cancer, and the level of CB2 expression in transformed cells was higher than in the respective normal tissue (Caffarel et al. 2006; Sarfaraz et al. 2005; Carracedo et al. 2006a). In examined tumors, CB2 receptors were usually located in the areas of intense tissue proliferation and invading cells. The enhancement of cannabinoid receptor expression in malignant versus healthy tissues might suggest a possible role of the endocannabinoid system in the tonic suppression of cell divisions and cancer growth. In support of this hypothesis, inhibitors of endocannabinoid transport or degradation (VDM-11 and AA-5-HT) have been shown to inhibit tumor growth and progression in some types of cancer by enhancing the levels of endocannabinoids in the cells (Izzo et al. 2008).

Alterations in the endocannabinoid system in tumors refer also to endogenous ligands, specific enzymes involved in their biosynthesis and degradation, as well as other putative target receptors (Maccarrone et al. 2000; Wu et al. 2012; Petersen et al. 2005). Table 11.1 summarizes the reported levels of the two major endocannabinoids human glioblastoma specimens compared to non-tumor brain tissues. The alterations of the endocannabinoid levels were consistent with the changes in the expressions and activities of the enzymes responsible for their biosynthesis and degradation (Wu et al. 2012; Petersen et al. 2005). Comparing to the non-tumor tissues, mRNA expression and the enzymatic activity of NAPE-PLD, the enzyme responsible for anandamide biosynthesis, decreased in tumor tissues. It was accompanied by the reduction of the anandamide hydrolyzing enzyme FAAH’s mRNA expression and enzymatic activity in low- and high-grade glioma samples (by 30% and 60%, respectively), which might be a result of a negative feedback induced by low levels of the substrate. mRNA expression and activity of MGL, the 2-AG hydrolyzing enzyme, decreased in glioma tissues comparing to that in the non-tumor tissues, whereas there was no difference in the expression levels of DGL-a, the 2-AG generating enzyme, between the groups (Wu et al. 2012).

GPR55, which can be engaged by some cannabinoids, appears to be up-regulated in cancer-derived cell lines and plays a pivotal role in tumor cells proliferation (Andradas et al. 2011). Another putative cannabinoid target, TRPV1, was also found upregulated in high grade human astrocytomas as compared to non-neoplastic brains (Stock et al. 2012).

11.4 Action of Cannabinoids in Glioma Cells

Cannabinoids induce significant inhibition of cell growth in tumor cells, due to modulation of proteins and nuclear factors involved in the control of cell survival, transformation and cell death. The programmed cell death of glioma cells after cannabinoid treatment was first described by Manuel Guzman and his co-workers (Sanchez et al. 1998). They showed that Δ9-THC is able to inhibit growth of rat C6 glioma cells in vitro and induce cell death with features typical for apoptosis, a programmed cell death process (Sanchez et al. 1998). We reported an apoptotic death triggered by a mixed CB1/CB2 synthetic agonist WIN 55,212–2 in rat glioma cells (Ellert-Miklaszewska et al. 2005). Following studies (Salazar et al. 2009; Gomez del Pulgar et al. 2002) and our own observations show effectiveness of Δ9-THC, WIN55,212–2 and a CB2-selective synthetic cannabinoid JWH133 in inducing apoptosis of cultured human glioblastoma cells and tumor-derived primary cultures (unpublished).

Cannabinoids exert anti-tumor effects in vivo leading to a significant regression of malignant gliomas in cannabinoid-treated animals (Duntsch et al. 2006; Massi et al. 2004; Galve-Roperh et al. 2000; Sanchez et al. 2001). Local administration of Δ9-THC or the synthetic cannabinoid, WIN55,212-2, reduced the size of tumors generated by intracranial inoculation of C6 glioma cells in rats, leading to complete eradication of gliomas and increased survival in one third of the treated rats (Galve-Roperh et al. 2000). Studies performed in mouse xenograft models with intratumoral and intraperitoneal drug administration demonstrated that non-psychoactive phytocannabinoid CBD (Massi et al. 2004), a CB2-selective agonist JWH133 (Sanchez et al. 2001) or a novel synthetic cannabinoid KM-233 (Duntsch et al. 2006) blocked the proliferation of human astrocytoma cells implanted subcutaneously in the flank of immune-deficient mice. Our preliminary data suggest that systemic cannabinoid administration can effectively hamper intracranial tumor growth in rats.

11.4.1 Mechanism of Cannabinoids Pro-apoptotic Action – Inhibition of Pro-survival Pathways

Several events and signal transduction pathways triggered by stimulation of the CB1 and CB2 receptors have already been described to participate in the cannabinoid-induced cell death in various tumor cells (Guzman 2003; Guzman et al. 2001). They include inhibition of PKA, activation of MAPK, superoxide generation, and a strong increase in intracellular calcium concentration, as well as the best elaborated alterations in sphingolipid metabolism (Howlett et al. 2002; Velasco et al. 2007).

Sanchez and coworkers showed that Δ9-THC was able to partially antagonize the forskolin-induced elevation of intracellular cAMP concentration, but did not affect basal cAMP levels in C6 glioma cells (Sanchez et al. 1997). As described in details in Sect. 3.5, glioma C6 cells are characterized by a very low constitutive level of cAMP, and rather increased, than decreased concentration of intracellular cAMP may be responsible for inhibition of cell proliferation. It suggest, that cannabinoid receptors may be coupled to inhibition of adenylyl cyclase in glioma cells, but down-regulation of cAMP levels is unlikely to play a role in the induction of apoptosis triggered by cannabinoids in these cells. Involvement of other signaling molecules or adaptor proteins (of still mostly unknown identity) in cannabinoid receptor signaling is a subject of ongoing studies.

The best characterized mechanism of cannabinoid-induced cell death of glioma cells involves sustained accumulation of pro-apoptotic sphingolipid ceramide (Fig. 11.3), which modulates signaling pathways crucial in the control of tumor cell growth and survival (Galve-Roperh et al. 2000; Sanchez et al. 2001). Activation of the CB1 receptor triggers two peaks of ceramide generation in glioma cells (Galve-Roperh et al. 2000; Gomez del Pulgar et al. 2002; Sanchez et al. 2001). Treatment with Δ9-THC or another CB1 receptor agonist produces a rapid release of ceramide via enzymatic hydrolysis of sphingomyelin from the cell membrane, catalyzed by neutral sphingomyelinase (Fig. 11.3). This effect is G-protein independent and involves the adaptor protein FAN (factor associated with neutral sphingomyelinase activation). The second ceramide peak is generated within hours or days after receptor activation and depends on increase of ceramide synthesis de novo via induction of serine palmitoyltransferase, a regulatory enzyme of sphingolipid biosynthesis (Gomez del Pulgar et al. 2002). Selective CB2 receptor agonists, such as JWH133, are supposed to stimulate only the ceramide synthesis process, which is sufficient to turn on the cell death program (Sanchez et al. 2001; Gomez del Pulgar et al. 2002). Thus, enhanced production of ceramide de novo is considered as an important event in cannabinoids-induced apoptosis. However, still little is known about the signaling pathways underlying the promotion of ceramide synthesis through cannabinoid receptor activation.

Galve-Roperh and co-workers postulated that the increased ceramide levels reported upon cannabinoid challenge led to prolonged activation of Raf-1/MEK/ERK signaling cascade and thus mediated glioma cell cycle arrest and cell death (Galve-Roperh et al. 2000). The same authors showed also that pharmacological inhibition of ceramide synthesis de novo prevented the inhibition of protein kinase B/Akt triggered by cannabinoids (Gomez del Pulgar et al. 2002). Our studies revealed that rather down-regulation of ERK activity, together with inhibition of PI3K/Akt pathway, contributed to rat C6 glioma cell death induced by WIN55,212-2 (Ellert-Miklaszewska et al. 2005). The serine/threonine protein kinase Akt, activated downstream of PI3K, as well as the Ras-activated Raf1/MEK/ERK pathway are widely recognized as key mediators of growth factor-promoted cell survival in gliomas (Kapoor and O’Rourke 2003). Both survival pathways converge on a small pro-apoptotic member of a Bcl-2 family of proteins, Bad. Bad is also a substrate for PKA, which is negatively linked to both cannabinoid receptors via canonical Gi/oα-mediated inhibition of adenylyl cyclase and subsequent decrease of cAMP levels. Phosphorylation of Bad by Akt, ERK and PKA retains the protein in the cytosol, where it is recognized by 14-3-3 regulatory proteins and sequestered (Zha et al. 1996). Otherwise, Bad translocates to mitochondria, and formation of heterodimers between non-phosphorylated Bad and anti-apoptotic proteins, such as Bcl-XL or Bcl-2, may result in a loss of integrity of the outer mitochondrial membrane (Zha et al. 1996). The release of cytochrome c and other pro-apoptotic proteins from mitochondria triggers the executive phase of programmed cell death. We proposed a mechanism, in which the decrease of mitogenic/pro-survival signaling evoked by the synthetic cannabinoid WIN55,212-2 promotes the pro-apoptotic function of Bad (Fig. 11.3). Accordingly, we demonstrated changes in Bad phosphorylation level followed by collapse of the mitochondrial membrane potential in C6 glioma cells treated with WIN55,212-2. These events preceded activation of caspase 9 by factors released from disrupted mitochondria, subsequent processing of effector caspases and finally oligonucleosomal DNA fragmentation.

Our further studies, as well as some published data suggest that human glioma cells treated with cannabinoids enter the suicide cell death using the same mitochondria-dependent pathway (Carracedo et al. 2006b). This mechanism contributes to the induction of apoptosis by cannabinoids also in other types of tumor cells (Velasco et al. 2007). However, involvement of an alternative, death receptor–dependent pathway in the apoptotic process triggered by these compounds cannot be ruled out.

11.4.2 The Role of ER Stress and Autophagy in Cannabinoid-Induced Cell Death

Different experimental approaches showed that the pro-apoptotic and tumor growth-inhibiting activity of cannabinoids relies on the accumulation of de novo-synthesized ceramide, an event that occurs in the endoplasmic reticulum (ER) and eventually leads to execution of cell death via apoptotic mitochondrial pathway. The mechanisms linking these two events have been revealed in details (Carracedo et al. 2006b; Salazar et al. 2009; Hernandez-Tiedra et al. 2016). Carracedo et al. showed that Δ9-THC treatment of glioma cells leads to up-regulation of the transcription co-activator p8 and its ER stress-related downstream targets: ATF4 (activating transcription factor 4), CHOP (the C/EBP-homologous protein) and TRB3 (pseudo-kinase tribbles homologue 3). Inhibition of ceramide synthesis de novo prevented Δ9-THC-induced p8, ATF4, CHOP and TRB3 up-regulation as well as ER dilation, and selective knockdown of ATF4 and TRB3 blocked cannabinoid-induced apoptosis in glioma cells. This indicated that ceramide accumulation is an early event in the cannabinoid-triggered ER stress and apoptosis in glioma cells (Carracedo et al. 2006b). Further studies by the same group implicated the role of eukaryotic translation initiation factor 2α (eIF2α) in cannabinoid-evoked ER stress. (Salazar et al. 2009). Activated eIF2α (after phosphorylation of Ser51 by a protein kinase-like endoplasmic reticulum kinase, PERK) is known to attenuate general protein translation, while enhancing the expression of several genes related to the ER stress response (Schroder and Kaufman 2005). Δ9-THC induced Ser51-phosphorylation of eIF2α, which was required for the up-regulation of the stress protein p8, as well as ATF4, CHOP and TRB3 (Salazar et al. 2009) and the subsequent glioma cell death (Fig. 11.3).

Furthermore, Salazar et al. observed that ER-stress-related stimulation of the p8/TRB3 pathway resulted in the induction of autophagy (Salazar et al. 2009). Autophagy is an evolutionarily conserved catabolic process, where a cell self-digests its cytoplasmic contents in newly formed vesicle structures called autophagosomes. Autophagy is activated in response to stress conditions, such as nutrient starvation, the unfolded protein response (ER stress) and hypoxia, both in normal cells and during cancer progression. Classically, autophagy promotes cell survival but it may also contribute to cell death. A detailed analysis indicated that Δ9-THC treatment led to formation of autophagosomes in human astrocytoma cell lines and primary cultures of human glioma cells (Salazar et al. 2009). This was associated with the inhibition of the Akt/mTORC1 (mammalial target of rapamycin, complex 1) axis, considered a key step in the early triggering of autophagy (Klionsky and Emr 2000). In fact, Δ9-THC-induced expression of TRB3 promoted the interaction of this ER stress-related protein with Akt. This led to decreased phosphorylation of Akt, as well as of its direct substrates TSC2 (tuberous sclerosis protein 2, tuberin) and PRAS40 (the proline-rich Akt substrate of 40 kilodaltons), which in turn resulted in mTORC1 inhibition (Salazar et al. 2009). Δ9-THC-treatment decreased phosphorylation of p70S6 kinase (a well-established mTORC1 downstream target) and its substrate phospho-S6 ribosomal protein. Altogether, Δ9-THC treatment triggered the following cascade of intracellular events: up-regulation of ER-stress-related TRB3, mTORC1 inhibition and induction of autophagy, which led to apoptotic glioma cell death (Fig. 11.3).

Recent studies additionally showed that changes in sphingolipid metabolism induced by Δ9-THC in glioma cells, including ceramide synthesis de novo, along with leading to the activation of ER stress/authophagy pathway, altered also the balance between ceramides and dihydroceramides in autophagosomes and autolysosomes. In turn, modified sphingolipid content promoted the permeabilization of the organellar membrane, the release of cathepsins to the cytoplasm and the subsequent activation of apoptotic cell death (Hernandez-Tiedra et al. 2016). This parallel mechanism seems to play a crucial role in determining the cell death-promoting (rather than protective) fate of authophagy stimulation by cannabinoids.

Administration of Δ9-THC to mice bearing human astrocytoma-derived tumors resulted in increased TRB3 expression, inhibition of mTOR signaling pathway, appearance of autophagy markers and caspase-3 activation. These findings indicate that cannabinoid promotes the apoptotic cell death through stimulation of ER stress and autophagy in human glioma cells and is essential for cannabinoid anti-tumoral action in vivo (Salazar et al. 2009).

11.4.3 Proapoptotic Actions Beyond the Cannabinoid Receptors

Some cannabinoids activate apoptosis independently of cannabinoid-receptor binding. Antiproliferative action of CBD likely involves the induction of oxidative stress through the generation of reactive oxygen species, which was linked to a later induction of apoptosis. This effect was inhibited by tocopherol, a potent antioxidant and interestingly was not observed in non-cancerous primary glial cells (Massi et al. 2006). Recently however, Scott et al. showed that CBD-induced increase in ROS production was accompanied by an upregulation of a number of genes belonging to the heat-shock protein (HSP) super-family, which diminished the cytotoxic effect of CBD. The authors proposed the combination of HSP inhibitors might enhance the anti-tumor effects of cannabinoids (Scott et al. 2015).

Anandamide promotes apoptosis through either the activation of vanilloid TRPV1 receptor (Maccarrone et al. 2000) or the accumulation of the pro-apoptotic sphingolipid ceramide mediated via CB1 or CB2 receptor activation (Galve-Roperh et al. 2000). All these results suggest that overall cannabinoids affect multiple cellular signaling pathways and thus have the potential to decrease cancer development. The mechanisms are however both cancer- and cannabinoid-specific. Defining the role of the cannabinoid receptors (CB1 and CB2) versus vanilloid or novel receptors in antitumoral action of these compounds is a matter of ongoing research. More recently, Moreno et al. described that GPR55 and CB2 can form heteromers in glioma and other cancer cells, in which the two receptors are highly abundant, and co-participate in the control of tumors growth (Moreno et al. 2014).

11.4.4 Effects of Cannabinoids on The Tumor Microenvironment in Malignant Gliomas

Cannabinoids have displayed a great potency in reducing glioma tumor growth in experimental animal models. Apparently their effectiveness in vivo is attributed not only to antiproliferative action against tumor cells but points to other cellular targets and additional mechanisms of action within the tumor microenvironment. Beyond affecting tumor cell survival cannabinoids impair tumor angiogenesis, glioma cells invasiveness and even malignant potential of glioma stem-like cells (Fig. 11.4) (Blazquez et al. 2003; Aguado et al. 2007; Nabissi et al. 2015; Singer et al. 2015; Soroceanu et al. 2013).

Fig. 11.4
figure 4

Multimodal action of cannabinoids in the tumor microenvironment. Cannabinoids induce tumor cell death, block angiogenesis and invasiveness crucial for tumor progression and induce differentiation of cancer stem cells (CSC)

Full size image

Increased demand for oxygen and nutrients supply to proliferating cancer cells makes angiogenesis a critical factor for the progression of solid tumors and a popular target for oncologic therapies. Local administration of the CB2 selective cannabinoid JWH133 in a mouse flank inoculation model of glioma turned the vascular hyperplasia characteristic of actively growing tumors to a pattern of blood vessels characterized by small, differentiated and impermeable capillaries, thus proving anti-angiogenic potential of the cannabinoid (Blazquez et al. 2003). This was associated with a reduced expression of pro-angiogenic factors, such as vascular endothelial growth factor (VEGF) and angiopoietin-2 in cannabinoid treated tumors. Downregulation of the major vascularization factors, including VEGF, was observed following treatment with Δ9-THC, Met-F-AEA, WIN-55,212-2, and JWH-133 in various types of tumors (Blazquez et al. 2003; Pisanti et al. 2007). More importantly, intratumoral administration of Δ9-THC to two patients with glioblastoma multiforme (grade IV astrocytoma) decreased VEGF levels and VEGFR-2 activation in the tumors (Blazquez et al. 2004). Antiangiogenic activity of cannabinoids could be partly related to a direct influence of the cannabinoids on endothelial cells migration and survival (Blazquez et al. 2003). Similarly, Solinas et al. demonstrated that CBD inhibited endothelial cell proliferation, migration and sprouting in vitro, and inhibited angiogenesis in vivo (Solinas et al. 2012). Overall, cannabinoids appear to have consistent effects on the vascularization pathway, causing a decrease in tumor vascularization in in vivo models.

Infiltrative growth through the surrounding brain parenchyma is one of the hallmarks of malignant gliomas and a major cause of their inevitable recurrence. Several studies have referred so far to targeting of glioma cell migration and invasiveness by cannabinoid treatment. Δ9-THC or JWH133 decreased the activity and expression of matrix metalloproteinase-2 (MMP-2) and tissue inhibitor of metalloproteinase 1 (TIMP-1) in cultured glioma cells and in the tumors in glioma-bearing mice (Blazquez et al. 2003; Blazquez et al. 2008b). MMP-2 is a proteolytic enzyme that allows tissue breakdown and remodeling during angiogenesis and metastasis and its up-regulation is associated with a high risk of progression and poor prognosis of gliomas. Blocking the release of the extracellular matrix degrading enzyme by in vivo administered cannabinoids correlated with decreased tumor volumes (Blazquez et al. 2003; Blazquez et al. 2008b). TIMP-1 plays a critical role in the acquisition of migrating and invasive capacities by tumor cells. Decreased TIMP-1 expression was observed in biopsies from patients with recurrent glioblastoma multiforme tumors undergoing a clinical trial on Δ9-THC efficacy (Blazquez et al. 2008a).

Multiple mechanisms were also suggested for antineoplastic activity of CBD, however they are unlikely mediated by the cannabinoid receptors. CBD effectively inhibited glioma cell proliferation, migration and invasion and caused a decrease in the expression of a set of proteins specifically involved in these processes (Vaccani et al. 2005; Solinas et al. 2013). CBD also inhibited HIF-1a, the regulatory subunit of the hypoxia inducible transcription factor, which is responsible for orchestrating the adaptive transcriptional programs, inducing cell survival, motility and tumor angiogenesis under hypoxic conditions (Solinas et al. 2013).

Accumulating evidence supports the existence of a highly tumorigenic subpopulation of glioma stem-like cells (GSC). GSC retain self-renewal properties, display increased resistance to radiation and conventional chemotherapeutic drugs, and therefore contribute to tumor progression and recurrence (Cheng et al. 2011). Strategies to induce GSC differentiation and alleviate their malignant features have been recently included to treatment modalities as a promising new approach to improve the effectiveness of anticancer drugs. Aguado et al. (2007) showed that activation of CB1 and CB2 receptors using synthetic cannabinoids HU-210 and JWH133 promoted differentiation of GSC derived from GBM biopsies and glioma cell lines. Cannabinoid-treated GSC displayed decreased efficiency to initiate glioma formation in vivo (Aguado et al. 2007). Several potential mechanisms regulating the GSC differentiation upon CBD treatment have been also recently proposed, although they are non-CB receptor mediated. CBD significantly downregulated the expression of a transcription regulator Id1, which controls multiple tumor-promoting pathways in glioblastoma, including glioma cell invasiveness and stemness-related self-renewal. This mechanism was suggested to cause an inhibition of glioma progression upon CBD administration in vivo (Soroceanu et al. 2013). Further studies revealed that CBD-triggered downregulation of Id1 and other stem cell mediators, such as Sox2, is caused by increased ROS production upon cannabinoid treatment, as it was abrogated by co-treatment with antioxidants (Singer et al. 2015). Another study described a novel mechanism, in which CBD induced TRPV2-dependent autophagy and stimulated GSCs differentiation by upregulating the expression of acute myeloid leukemia 1a (Aml-1a) transcription factor. Aml-1a, in turn, bound to TRPV2 promoter and stimulated TRPV2 receptor expression (Nabissi et al. 2015). CBD activity led to inhibition of proliferation, promotion of GSC differentiation, and sensitization of glioma cells to the cytotoxic effects of alkylating agents. Multimode action of cannabinoids, including inhibition of gliomagenesis by blocking the potential of GSC to initiate tumor formation or recurrence may have important implications for the development of cannabinoid-based therapeutic strategies.

Growing evidence suggests the modulating role of tumor-associated glial cells, especially microglia, on glioma progression (Li and Barres 2018). Microglia are the resident macrophages of the central nervous system. They participate in brain development, regulation of homeostasis and synaptic plasticity, as well as protect the brain from infections, metabolic disturbances or misfolded proteins. Upon stimulation by tumor-derived factors microglia become pro-invasive, anti-inflammatory cells that promote glioma invasion, immunosuppression and angiogenesis. Microglia express a functional endocannabinoid signaling system and the activation of CB receptors was shown to affect microglia behavior, including migration, proliferation, free radicals release and phagocytosis (Stella 2009). Of note, endocannabinoids drive the acquisition of the alternative phenotype in microglia, which counteracts the inflammatory activation and resembles the tumor-induced phenotype (Mecha et al. 2015). It is likely that stimuli related to anti-inflammatory and repair mechanisms activate 2-AG and AEA synthesis in order to promote the autocrine and paracrine activity of these lipid messengers, activating the CB1 or CB2 receptors and specific CB signaling cascades (Mecha et al. 2015). The impact of cannabinoids on tumor associated microglia and potentially astrocytes in the microenvironment of gliomas is still to be studied.

11.5 Therapeutic Potential of Targeting Cannabinoid Signaling in Gliomas

Standard chemotherapeutics are a double edged sword; they eliminate cancer cells but affect severely healthy cells in the body. Based on evidence from in vitro and in vivo preclinical studies, as well as from pilot clinical trials in patients with recurrent glioblastoma multiforme, cannabinoids appear to have a favorable safety profile and do not produce the generalized toxic effects as most conventional chemotherapeutic drugs (Galve-Roperh et al. 2000; Guzman 2003; McAllister et al. 2005; Ladin et al. 2016). Furthermore, cannabinoids promote survival of glial cells and neurons in different models of injury, suggesting that the anti-proliferative effect of cannabinoids is selective for brain tumor cells, while viability of normal brain cells remains unaffected or even favored by cannabinoid challenge (Guzman 2003; Molina-Holgado et al. 2002). Several mechanisms could be responsible for cannabinoid compounds targeting only the cancer cells, including differences in cannabinoid signaling in glioma and normal neural cells and selective over-expression of the CB2 receptor in tumor cells. In contrast to pro-apoptotic action of Δ9-THC and WIN55,12–2 on transformed glial cells, treatment of primary cultured astrocytes with these CB1/CB2-activating cannabinoids does not trigger ceramide generation de novo, induction of ER stress-related genes or apoptosis. In our studies administration of the CB2-selective agonist JWH133 was effective toward tumor cells, and it did not affect survival or morphology of normal astrocytes (unpublished). Apparently negligible CB2 receptor expression in the normal brain and its abundance in high grade gliomas seems to confer a relative safety of CB2-selective agonists for targeted glioma therapy. Moreover, the CB2 selective compounds are devoid of undesirable psychodysleptic side-effects, attributed to marijuana abuse, which are mediated by the CB1 receptor.

Due to genetic and epigenetic alterations malignant glioblastomas are highly resistant to radiation and chemotherapy. Mainstream therapeutic strategies for the management of all primary brain tumors are still mostly palliative, known to leave survivors with devastating neurological deficits and frequently with a high risk of the disease recurrence. The potency of synthetic cannabinoids to induce apoptosis in glioblastoma cells has been tested by us and others on several cell lines and primary cell cultures derived from biopsies of human tumors, which to some extent may reflect the heterogeneity of glioma molecular characteristics (Duntsch et al. 2006; Ellert-Miklaszewska et al. 2005; Galve-Roperh et al. 2000; Guzman 2003; Massi et al. 2004; McAllister et al. 2005; Sanchez et al. 2001). Thus, cannabinoids were able to override alterations of growth regulatory and apoptotic pathways, caused by common mutations reported in primary and secondary glioblastomas. Pro-apoptotic action of cannabinoids relies on the generation of ceramide and disruption of signaling pathways crucial for the regulation of cellular proliferation, differentiation or apoptosis (Ellert-Miklaszewska et al. 2005; Galve-Roperh et al. 2000; Gomez del Pulgar et al. 2002; Salazar et al. 2009). Their unique mechanism of action among standard oncology remedies justifies further research on their anti-tumoral properties. This also stimulates investigations on the effectiveness of combination of cannabinoids with standard GBM therapies to improve patients survival. Pretreating glioma cells with a mixture of Δ9-THC and CBD increased their sensitivity to irradiation. These in vitro results were recapitulated in an orthotopic murine glioma model, which showed dramatic reductions in tumor volumes, when both cannabinoids were combined with irradiation (Scott et al. 2014). Co-administration of Δ9-THC or Δ9-THC plus CBD (at a 1:1 ratio) with an alkylating agent temozolomide (TMZ) synergistically reduced the growth of subcutaneous gliomas upon local injection (Torres et al. 2011) and showed enhanced anti-tumor action in orthotopic models using patient-derived xenografts with oral drug delivery (Lopez-Valero et al. 2018). These pre-clinical observations have led to phase II clinical trial investigating a Δ9-THC:CBD mixture in combination with TMZ in GBM patients. Patients treated with TMZ plus Δ9-THC:CBD mixture presented nearly twice higher survival rates as compared to the control group that received TMZ only (Dumitru et al. 2018). A promising field to explore is to prolong the effectiveness of endocannabinoids by interference with their biosynthesis, uptake and breakdown. Therefore future studies on cannabinoid signaling system in gliomas are clearly needed.