Cannabinoids are a class of hydrophobic molecules that typically bind to G-protein-coupled cannabinoid receptors, cannabinoid receptor type 1 (CB1R) or cannabinoid receptor type 2 (CB2R). Endogenously synthesized cannabinoids (endocannabinoids), cannabinoid receptors, and enzymes involved in cannabinoid formation and breakdown are collectively known as the endocannabinoid system (ECS). This system has been shown to impact different neurological and immunological processes (Zhu, 2006; Tanasescu and Constantinescu, 2010). Additionally, numerous reports demonstrate that endocannabinoids and their putative receptors play a role in the development of various malignancies (Velasco et al., 2016). Many studies investigating the effects of cannabinoids on tumor progression have been carried out in cell culture. However, due to the complexity of the tumor microenvironment, these findings do not always translate to whole animal models. As such, the goal of this review was to compile and analyze the results of studies which evaluated effects of cannabinoids on cancer development in animal models (Table 1) and to examine current knowledge regarding the impact of cannabinoids on tumor growth in humans.

Animal tumor models have become a mainstay in preclinical cancer research due to their ability to produce fairly predictable tumor growth. The first tumor model used to study tumorigenesis was the chemical carcinogenesis model. This model employs the use of two known carcinogens, an initiator and promotor. Initiators are typically employed in a single dose to cause irreversible DNA damage. A promotor is given in repeated doses to increase cell proliferation and induce subsequent DNA damage (Kociba and Schwetz, 1982). Implantable or grafted tumor models are also commonly used. The major subtypes of these models are the (1) syngeneic (allograft) model, in which tumor cell lines of host origin are subcutaneously injected into immune-competent animals and (2) the xenograft model, where cells distinct from the implantation host are subcutaneously injected or grafted into immune-deficient animals (Teicher, 2006). A more technically challenging type of tumor grafting is that of the orthotopic model in which tumor cells or tissues are grafted directly into the organ or system of tumor origin (Peterson and Houghton, 2004). Patient derived xenografts offer high translational value and are created when tumor tissue of human patients is implanted into an immune-compromised animal (Jin et al., 2010). Genetically engineered mouse models (GEMMs) are also used to study carcinogenesis. In the GEMMs, transgenic mice which over-express oncogenes or dominant negative tumor-suppressor genes are often more susceptible to spontaneous tumor induction (Sharpless and Depinho, 2006). Knockout models in which the expression of specific genes is disabled in animals exposed to carcinogens also provide valuable information about mechanisms of tumor formation (Rosenberg et al., 2009). Each of these tumor models is distinct and possess advantages and disadvantages depending upon scientific and clinical goals.

It is estimated that more than 23,000 individuals will be newly diagnosed with central nervous system (CNS) cancers in 2016 (Siegel et al., 2016). There are many distinct neuronal and glial tumor types that develop in the brain. Gliomas are a CNS tumor subtype that are derived from glial tissue and account for approximately 80% of all primary malignant brain tumors (Ostrom et al., 2015). The prognosis for this neoplasm is poor with an average 5-year survival rate of 5.1%. As such, novel strategies are needed to impact current therapeutic approaches. The ECS has been examined in clinical and preclinical glioma models. Several studies show that glioma tissues from rodents and humans express functional ECS components including the CB1 and CB2 receptors (Sánchez et al., 2001; Moreno et al., 2014). Furthermore, CB2 expression was found to be elevated in high grade gliomas (Sánchez et al., 2001; Ellert-Miklaszewska et al., 2007) Hence, a series of informative investigations primarily conducted by the Guzman group, have examined the effects of cannabinoids on glioma cell survival with the goal of assessing the feasibility of developing these agents as novel anti-neoplastics.

Numerous studies indicate that Δ⁹-tetrahydrocannabinol (Δ⁹-THC) is a potent inducer of glioma death in vivo. Utilizing a C6 glioma cell xenograft model, it was demonstrated that tumor growth was significantly reduced by Δ⁹-THC (Galve-Roperh et al., 2000). Subsequent to orthotopic implantation of the C6 cells the overall survival of Δ⁹-THC treated animals was also increased. In vitro studies with the C6 cell line suggested that the Δ⁹-THC-induced reduction in tumor growth occurred as a consequence of increased ceramide synthesis and Erk activation. Another study showed that Δ⁹-THC reduced the growth of U87MG glioma xenografts (Salazar et al., 2009). Experimentation in the U87MG cell line determined that ceramide accumulation and the induction of ER stress proteins p8, CHOP10, and TRB3, were required for Δ⁹-THC-induced death. The authors further demonstrated the importance of p8 in tumor growth by utilizing transformed mouse embryo fibroblasts (MEF) in which tumor development was reduced in p8^+/+ animals treated with Δ⁹-THC compared to vehicle treated animals. In contrast, no significant difference in tumor growth was observed in Δ⁹-THC or vehicle treated p8^−/− mice suggesting that p8 was needed for Δ⁹-THC-mediated tumor death. In a subsequent study by this group, endoplasmic reticulum stress-mediated autophagy and cell death was observed in Δ⁹-THC treated glioma. Specifically, Δ⁹-THC caused an increase in ER stress protein, TRB3, autophagy protein, LC3, and apoptosis as well as a decrease in mTORC protein, p-S6 in U87MG xenograft tumors (Salazar et al., 2009). This pattern of protein expression was also observed in glioma biopsies from 2 patients with glioblastoma multiforme treated via intracranial administration with Δ⁹-THC. In vitro experiments with the U87MG cell line suggested that Δ⁹-THC mediated activation of TRB3 was a critical link between the ER stress and autophagy pathways as TRB3 activation was needed for inactivation of Akt and the mTORC1 complex. Other reports indicated that the administration of Δ⁹-THC improves the therapeutic outcome of the clinically utilized glioma anti-neoplastic, temezolamide (TMZ) (Torres et al., 2011). In this study, Δ⁹-THC significantly reduced the growth of U87MG xenograft tumors treated with TMZ when compared to TMZ or Δ⁹-THC alone. Similarly, administration of Δ⁹-THC + cannabidiol (CBD) or Δ⁹-THC + CBD + TMZ significantly reduced tumorigenesis compared to the individual drugs (Torres et al., 2011). The combined drug regimen was associated with prevalent LC3 expression and cell death in U87MG tumors. The mechanism of cell death was also examined in cultured U87MG cells where inhibition of ceramide synthesis and autophagy prevented cellular cytotoxicity (Torres et al., 2011). Since each of these studies also determined that the cytotoxic effect of Δ⁹-THC was dependent on cannabinoid receptors, the findings of these reports indicate that ceramide synthesis, ER stress, autophagy and cannabinoid receptor activation are likely shared mechanism of Δ⁹-THC-induced glioma death.

Cannabinoid regulation of matrix remodeling proteins may also play a role in its anti-tumor activity. Matrix metalloproteinases (MMP) can promote or prevent tumor invasion and metastasis by degrading extracellular matrix and other cell-associated proteins (Lukaszewicz-Zajac et al., 2014). In addition, the effects of MMPs are counteracted by tissue inhibitors of metalloproteinases (TIMPs). MMPs/TIMPs also directly regulate tumor growth, tumor apoptosis and angiogenesis. An examination of the effect of Δ⁹-THC on C6.9 cell xenograft growth showed that animals treated with Δ⁹-THC contained smaller tumors with reduced MMP-2 expression in contrast control group animals (Blázquez et al., 2008a). Δ⁹-THC also reduced tumor growth and the expression of MMP-2 in tumors of Δ⁹-THC-treated patients. Cell culture studies with C6.9 cells confirmed that MMP-2 expression and cell viability were decreased as a result of Δ⁹-THC exposure (Blázquez et al., 2008a). A different study by this research group also indicated selective participation of TIMP-1 in the anti-neoplastic activity of Δ⁹-THC (Blázquez et al., 2008b). Similar to MMPs, TIMPs have both a positive and negative impact on tumor growth (Lukaszewicz-Zajac et al., 2014). In C6.9 xenografts, Δ⁹-THC decreased tumor development and tumoral TIMP-1 but had little impact on TIMP-2 or TIMP-3 expression. Selective TIMP-1 downregulation was also observed in patient tumor samples subsequent to Δ⁹-THC administration. C6.9 cell culture studies showed that Δ⁹-THC decreased TIMP-1 expression by increasing ER stress and ceramide synthesis and also demonstrated that Δ⁹-THC reduced C6.9 cell migration (Lukaszewicz-Zajac et al., 2014). These studies suggest that downregulation of MMP-2 and TIMP-1 is important for Δ⁹-THC regulation of tumor size, however additional work is needed to reveal whether the reduction in these matrix modulating proteins also alters the metastatic and invasive behavior of tumor cells in vivo.

An investigation of the effects of Δ⁹-THC on human glioblastoma-T98G xenografts demonstrated concentration-dependent effects with low Δ⁹-THC concentrations causing a modest increase in tumor growth and high Δ⁹-THC concentrations promoting tumor cell death (Sánchez et al., 2001). Experiments in cultured T98G cells suggested that this effect was mediated by the formation of a functional dimeric complex between GPR55 and CB2 receptors such that low concentrations of Δ⁹-THC activated the CB2 receptor and Erk expression thereby promoting tumor cell growth. However, in the presence of high Δ⁹-THC concentrations, Δ⁹-THC inhibited GPR55 which blocked the CB2 receptor and Erk activation thereby inhibiting tumor cell growth. This finding is consistent with reports showing that the putative G-protein-coupled cannabinoid receptor, GPR55, is overexpressed in cancers and promotes carcinogenesis (Falasca and Ferro, 2016).

Most clinically available cancer chemotherapeutic agents demonstrate significant toxicity toward cancer and normal cells. This lack of selectivity often produces undesirable side effects that limit its therapeutic utility. Δ⁹-THC preferentially induced cell death in tumorigenic compared to non-tumorigenic astrocytes (Salazar et al., 2009). Moreover, in non-tumor bearing rats, Δ⁹-THC (2500 μg) exhibited limited toxicity and did not affect animal survival, health, behavior or hematopoietic parameters (Galve-Roperh et al., 2000). These findings suggest that Δ⁹-THC will cause limited adverse effects thereby making this compound an attractive candidate for therapeutic use. Furthermore, the anti-emetic, appetite stimulating and analgesic effects of Δ⁹-THC and other cannabinoids can provide relief for important co-morbidities commonly associated with cancer and cancer chemotherapy.

Cannabidiol (CBD) is a non-psychotropic phytocannabinoid whose effect on tumor growth is being extensively explored. CBD significantly reduced U87MG xenograft growth when compared to vehicle control animals (Massi et al., 2004). An in vitro examination of the mechanism of action suggested that the induction of reactive oxygen species (ROS) and activation of the CB2 receptor was involved in the anti-tumor effect. However, unlike Δ⁹-THC, CBD toxicity did not appear to be regulated by CB1 or ceramide. Furthermore, in an orthotopic and subcutaneous tumor model, CBD significantly decreased U251 cell tumor growth (Soroceanu et al., 2013). In tumors generated by both methods, the expression of the proliferation marker, Ki67, and inhibitor of DNA binding (Id-1) protein were reduced. Id-1 is a transcription factor that prevents tumor cell differentiation and its expression is elevated in highly aggressive tumors (Ling et al., 2006). Cell culture studies showed that CBD inhibited Id-1 expression which occurred coincident with a reduction in U251 cell invasiveness (Soroceanu et al., 2013). A follow up study by this group demonstrated that CBD inhibited tumor growth and increased the survival of animals orthotopically implanted with primary 3832 and 387 glioblastoma cells (Singer et al., 2015). The reduction in tumor size was accompanied by decreased expression of Id-1 and Sox-2 indicating a reduction in tumor aggressiveness. Continued administration of CBD in this model resulted in reduced responsiveness to CBD, increased antioxidant response gene expression and upregulated expression of mesenchymal markers in the tumors. Investigation of the in vitro mechanism suggested that resistance to CBD was primarily caused by activation of SCL7A11 (a transcriptional target of Nrf2) antioxidant signaling as inhibition of this target decreased tumor cell survival and augmented the antitumor activity of CBD. A different study revealed that CBD potentiated the killing effect of ionizing radiation (IR), a common clinical therapy for glioblastoma. Using an orthotopic xenograft model, CBD + Δ⁹-THC + IR significantly decreased the size of GL261 gliomas compared to CBD + Δ⁹-THC or IR alone (Scott et al., 2014). Increased TUNEL positive cells in tumors isolated from animals exposed to CBD + Δ⁹-THC + IR implicated cell death as the mechanism of tumor growth inhibition. Thus, the antitumor activity of Δ⁹-THC and CBD has been demonstrated in different in vivo tumor models and the diverse pathways by which these agents elicit death appear to favor co-administration with other chemotherapeutic agents.

Numerous studies suggest that synthetic cannabinoids such as JWH-133, WIN 55,212-2 and HU-210 are promising chemotherapeutic agents for treatment of gliomas. In murine C6 xenografts, the growth of glioma solid tumors was significantly inhibited by JWH-133 compared to control group animals (Sánchez et al., 2001). This effect was found to be mediated by the CB2 but not the CB1 receptor. In the same animal model, JWH-133 also inhibited the growth of human glioma cells. Examination of the mechanism of cytotoxicity in cultured C6 cells showed that ceramide synthesis was needed for JWH-133-mediated tumor cell death. Later studies by this group validated the in vitro mechanism by demonstrating that blockade of ceramide synthesis in vivo decreased the anti-tumor activity of JWH-133 (Blázquez et al., 2008a,b). These studies also revealed that ceramide was involved in JWH-133-induced suppression of the invasion associated proteins, MMP-2 and TIMP-1. The combined reports suggest that the induction of ceramide synthesis by JWH-133 causes a reduction tumor cell survival and invasion.

The effect of synthetic cannabinoids on angiogenesis was also examined. JWH-133 reduced the growth of C6 gliomas that was accompanied by a decrease in tumoral, pro-angiogenic VEGF and VEGFR (Blázquez et al., 2004). It was determined that VEGFR levels were also downregulated in glioma biopsies from patients treated with Δ⁹-THC. In vitro studies with C6 cells demonstrated that both JWH-133 and WIN 55,212-2 decreased VEGFR2 phosphorylation and VEGF levels by increasing ceramide production. The reduction in VEGF signaling by the cannabinoids was also dependent on CB1 and CB2 receptors. A more detailed examination of the effect of cannabinoids on angiogenesis showed that JWH-133 decreased the permeability of tumor localized blood vessels as well as the growth of C6 and patient-derived glioma cells in xenograft tumor studies (Blázquez et al., 2003). Consistent with these results, the mRNA expression of the angiogenesis and metastasis-promoting proteins VEGF, Ang2 and MMP-2 were decreased by JWH-133. Furthermore, in vitro studies using HUVEC cells demonstrated that JWH-133 inhibited endothelial cell migration and initiated death that was reversed by antagonism of CB1 and CB2 receptors. The findings of these reports indicate that cannabinoid prevention of angiogenesis is reliant on ceramide production and cannabinoid receptor activation.

Accumulating evidence suggests that a subpopulation of brain tumors originate from pluripotent stem-like cells (Singh et al., 2004). These undifferentiated cells are often associated with poor prognosis and highly aggressive tumor phenotypes whereas well-differentiated glial cells are typically less aggressive (Wan et al., 2011). As such, the stem-like population of U87MG cells and human gliomas were treated with JWH-133 or HU-210 and then subcutaneously implanted to examine the effect of cannabinoids on this tumor cell subtype (Aguado et al., 2007). Cannabinoid treated cells formed smaller tumors which contained low levels of stem-like stage proteins such as nestin and high levels of differentiation-associated proteins including β-tubulin III and S-100β. Consistent with these findings, JWH-133 and HU-210 decreased the expression of nestin and increased the expression of GFAP, S-100β and β-tubulin III in cultured stem-like glioma cells. Hence it appears that the anti-neoplastic activity of the synthetic cannabinoids may be attributed to their ability to prevent tumor cell survival, invasion, and angiogenesis while promoting tumor cell differentiation.

KM-233 is a synthetic cannabinoid that signals through CB1 and CB2 receptors and is a derivative of Δ⁸-THC (Krishnamurthy et al., 2008). The anti-neoplastic activity of KM-233 was examined in an orthotopic U87MG cell tumor protocol where a significant reduction in tumor size occurred (Gurley et al., 2012). These results were verified in a xenograft tumor study using patient-derived glioblastoma multiforme tissue. In vitro studies showed that KM-233 induced U87MG death was prevented by antagonizing the CB1 receptor. A subsequent study by this group confirmed the anti-tumor activity of KM-233 against U87MG xenograft tumors and also demonstrated that KM-233 produced minimal damage to neuronal tissue within the therapeutic dosage range (Duntsch et al., 2006). As such, KM-233 may be an effective agent against glioma with an acceptable safety profile similar to other cannabinoids.

The bulk of our knowledge about the effect of cannabinoids on cancer growth in in vivo models is derived from the aforementioned glioma studies. Much of this work demonstrated potent antitumor activity, confirmed in vitro mechanisms of cannabinoid antitumor action and provided evidence that these molecules were minimally toxic and therefore paved the way for execution of a clinical study with a small number of glioma patients (Clinical trial ID# NCT01812603). However, additional experimentation in animals is needed to define the pharmacokinetic properties of cannabinoids before large scale human studies are conducted. It will be particularly important to examine drug accumulation in tumors and drug disposition in body compartments subsequent to drug administration via different routes as most of the glioma studies conducted thus far employ intra/peri-tumor drug delivery. Such investigations may identify more clinically convenient routes of administration, establish the extent of drug stability and metabolism while providing clues about potential adverse effects of these agents.

Digestive system cancers include colon, stomach, liver, and pancreatic cancer. In the United States, these combined malignancies account for the highest incidence of cancer besides skin cancer and are the second leading cause of death for both sexes. In 2016, an estimated 304,930 new cases of digestive system cancer will be diagnosed in the US with 153,030 estimated deaths (Siegel et al., 2016).

Breast cancer is the most commonly diagnosed cancer in women and the second leading cause of female cancer deaths in the United States (Siegel et al., 2016). Risk factors for developing this type of cancer include: being of the female sex, having a family history of this disease, obesity, and menopause-associated hormone replacement (Kaminska et al., 2015). Recent research suggested that cannabinoids and the ECS played a role in breast cancer progression and may possess therapeutic potential (Caffarel et al., 2012).

Breast cancer can be divided into three sub-types: hormone sensitive breast cancer, which expresses estrogen receptors (ER) and/or progesterone receptors (PR); HER2-positive breast cancer, which expresses human epidermal growth factor receptor (ErbB2, a member of the EGFR family); and triple negative breast cancer (TNBC), a type of breast cancer which is difficult to treat due to the lack of ER, PR, and HER2 expression. Cannabinoid receptor expression differs between these breast cancer sub-types. CB1 receptors were reported to be present in 28% of all breast carcinoma patients, with 14% being HER2 positive. On the other hand CB2 receptors were found in 72% of breast carcinomas and 91% of these tumors were HER2 positive, suggesting a possible link between HER2 and CB2 expression (Qamri et al., 2009; Caffarel et al., 2010; Pérez-Gómez et al., 2015). Of interest, a correlation between CB2 receptor expression and breast cancer aggressiveness has been suggested (Caffarel et al., 2006). As such, the effects of cannabinoids on breast tumor growth and metastasis using in vivo models has helped elucidate their therapeutic and prognostic potential.

Phytocannabinoids such as Δ⁹-THC and CBD induce both breast cancer regression and progression in vivo, indicating that a complex interaction between these compounds and cancer exists. CBD was shown to decrease tumor growth in multiple breast cancer models including: two orthotopic models, one genetically engineered mouse model (GEMM) and one TNBC xenograft model (Ligresti et al., 2006; McAllister et al., 2011; Murase et al., 2014; Elbaz et al., 2015). CBD significantly decreased the incidence of lung metastasis in TNBC xenograft metastatic and orthotopic mouse models. Furthermore, two of these studies which used genetically engineered and xenograft models proposed that CBD elicited these effects by inhibiting EGFR activation, cytokine secretion, and Akt expression (McAllister et al., 2011; Elbaz et al., 2015). The CBD derivative, O-1663, was also shown to decrease TNBC xenograft growth through a mechanism requiring the generation of oxidative stress (Murase et al., 2014). Consistent with the findings for CBD, Δ⁹-THC impaired tumor growth, multiplicity and metastasis in both HER2 positive MMTV-neu mice (a GEMM) and HER2 positive xenografted mice (Caffarel et al., 2010). This group also found that downregulation of pro-tumorigenic Akt was involved in the anti-proliferative effects of Δ⁹-THC (Caffarel et al., 2010). In contrast, phytocannabinoids also promoted tumor development. The Nagarkatti group found that mice implanted with 4T1 cells that were treated with Δ⁹-THC displayed greater tumorigenicity and metastatic potential by inhibiting host anti-tumor immune responses (McKallip et al., 2005). Indeed, phytocannabinoid-induced immune suppression has been reported to be mediated by the CB2 receptor in similar tumor models (Zhu et al., 2000).

Synthetic cannabinoid receptor agonists have also been utilized to elucidate the role of CB1 and CB2 receptors in breast cancer growth. Activation of CB1/CB2 receptors using WIN-55,212-2 led to significant decreases in tumor growth and lung metastasis in a TNBC xenograft and a genetically engineered MMTV-PyTV (mouse mammary tumor virus encoding polyoma virus middle T antigen) mouse model (Qamri et al., 2009). Furthermore, selective CB2 receptor activation with JWH-133 inhibited the growth of tumors and decreased metastasis in MMTV-PyTV mice, HER2 positive MMTV-neu mice and MDA-MB231-derived xenografts (Qamri et al., 2009; Caffarel et al., 2010; Morales et al., 2015). Follow-up in vitro studies performed in MDA-MB231, MDA-MB468 and N202.1 cells also suggested that CB2 receptor activation by synthetic cannabinoids such as JWH-133 downregulated pro-tumorigenic prostaglandin E₂ (PGE₂) or Akt similar to phytocannabinoids (Qamri et al., 2009; Caffarel et al., 2010). Another CB2 receptor agonist, JWH-105, was found to inhibit tumor growth in a NT2.5-orthotopic mouse model and the MMTV-PyMT model and this observation corresponded to decreased pro-metastatic phospho-CXCR4 and phospho-ERK levels in vivo and in vitro (Nasser et al., 2011). Additionally, activation of CB2 receptors by a similar compound, JWH-015, was also found to decrease tumor growth and weight in both, ER- positive and ER- negative orthotopic models (Elbaz et al., 2016). JWH-015 treated mice exhibited decreased EGFR and IGF-1R signaling and Akt expression (Elbaz et al., 2016). Finally, the synthetic and metabolically stable analog of AEA, 2-methyl-2'F-anandamide (Met-F-AEA), was also shown to decrease lung metastasis in TSA-E1 cell-derived xenografts (Grimaldi et al., 2006). For the most part, in vivo studies indicate that phyto-, synthetic- and endo- cannabinoids suppress tumor development in hormone receptor positive, EGFR positive and triple negative breast cancer animal models. However, additional animal studies are needed to understand how cannabinoid regulation of the immune system impacts breast tumor growth.

Carcinoma of the prostate is the development of malignant cells within the male reproductive prostate gland. It is the most commonly diagnosed cancer in all men with an estimated 180,000 new cases in 2016 (Siegel et al., 2016). This cancer type is also the second leading cause of cancer deaths in men, which primarily appears to be a result of late detection. Risk factors for developing this disease are advanced age, family history, and race.

Cannabinoids and their receptors have shown promise as potential therapeutics in prostate cancer. CB1 and CB2 receptor expression was shown to be elevated in prostate cancer compared to normal prostate tissue and are suggested to be positively correlated with poor disease outcome (Sarfaraz et al., 2005; Chung et al., 2009). Prostate cancer progression depends on androgen activity and while the ECS may influence this signaling, the interplay between these systems in the context of cancer is poorly understood (Sánchez et al., 2003a,b). Although numerous studies have evaluated the effect of cannabinoids on prostate cancer in cell culture studies, relatively few reports have determined if these effects occur in in vivo animal models. Di Marzo's group found that cannabis extracts enriched with CBD effectively decreased tumor growth in androgen receptor positive LNCaP xenografts but potentiated tumor growth of androgen receptor negative DU-145 xenografts (De et al., 2013). Furthermore, in vitro experiments using PC-3, DU-145, 22RV1, and LNCaP cell lines in this study found that CBD decreased androgen receptor expression suggesting that crosstalk between the CBD and androgen signaling pathways may affect tumor growth. CBD also significantly increased the generation of ROS and pro-apoptotic CHOP10 expression in these four prostate cancer cell lines (De et al., 2013). Similarly, JWH-015 led to decreased tumor proliferation in PC-3 derived xenografts. These anti-proliferative effects were examined in vitro and were suggested to be mediated by CB2 induced synthesis of ceramide (Olea-Herrero et al., 2009). While these studies show a promising role for cannabinoids in targeting prostate cancer, more research is needed to definitively establish cannabinoid efficacy in vivo and the importance of androgen receptor signaling in this system.

Lung cancer is the most deadly neoplasm in both males and females in the United States with a death toll anticipated to exceed 150,000 individuals in 2016 (Siegel et al., 2016). The two major types of lung cancer are small-cell lung cancer (SCLC) and non-small-cell lung cancer (NSCLC). Risk factors for developing this disease are tobacco smoking, genetic predisposition, and exposure to radon gas. Survival figures from this disease powerfully demonstrate the need for novel interventions. As such, cannabinoids have been evaluated in vivo for their efficacy against lung cancer.

CB1 receptors were expressed in 24% human NSCLC tissues, while CB2 receptor expression was found in 55% of NSCLC, suggesting a role for CB-receptors in tumor development (Preet et al., 2011). Cannabinoids such as Δ⁹-THC, CBD and Met-AEA elicited different effects in lung cancer animal models. CBD administration to mice with A549 cell xenografts decreased tumor growth, invasion and metastasis (Ramer et al., 2010a,b, 2012). Follow up in vitro experiments using A549, H358, and H460 lung cancer cell lines suggested that the anti-metastatic effect of CBD was mediated by upregulation of intracellular adhesion molecule-1 (ICAM-1), an anti-metastatic protein which has been suggested to suppress tumor growth via an immuno-surveillance mechanism (Wolfram et al., 2000; Ramer et al., 2012). In another study, CBD exhibited anti-tumor effects that were reversed by co-administration of a PPARγ antagonist in vivo (Ramer et al., 2013). Similarly, Preet et.al. found that the phytocannabinoid, Δ⁹-THC, downregulated Akt in A549-derived xenografts, which corresponded with decreased tumor growth and metastasis (Preet et al., 2008). In contrast, Zhu et al. showed that Δ⁹-THC increased the tumorigenicity of 3LL lung cancer cells in allografted immunocompetent mice and demonstrated that CB2 mediated inhibition of anti-tumor lymphocyte activity was the primary mechanism for accelerated tumor growth (Zhu et al., 2000). Furthermore, these investigators and the Arevalo group found that Δ⁹-THC administration had no effect on tumor growth in xenografted immunodeficient mice (Zhu et al., 2000; McKallip et al., 2005). Stable analogs of AEA also showed differential effects on lung cancer progression in vivo. The Dubinett group found that met-AEA increased tumor growth in a xenograft and allograft murine model and that COX-2 inhibition abrogated these effects (Gardner et al., 2003). Conversely, Bifulco et al. found that met-F-AEA treatment decreased tumor metastasis in a 3LL cell-derived xenograft model (Bifulco et al., 2001). These studies suggest that the ECS plays a role in both lung cancer progression and suppression. Consequently, several of these studies indicate that anti-tumor immunity is suppressed by certain cannabinoids, exacerbating tumor growth in immune-competent animals. As such, exploring the interplay between the immune system, the ECS and lung cancer development may be critical in determining whether cannabinoids are suitable for lung cancer treatment.

Thyroid cancer is the most frequently diagnosed endocrine malignancy in the United States. It is estimated that more than 64,000 new cases will be identified and 1900 deaths will result from this disease (Siegel et al., 2016). Although the death rate is relatively low compared to other types of cancer, therapeutic options are needed for aggressive forms of this disease whose lethality is thought to be associated with the differentiation state of the tumor (Lee et al., 2016). Relatively few studies have examined the effect of cannabinoids on thyroid tumor development in vivo, but these reports showed that the ECS was expressed and was capable of modulating tumor growth (Portella et al., 2003; Bifulco et al., 2004; Shi et al., 2008). For example, the activity of a variety of ECS modulators was examined in KiMol thyroid cells which generate undifferentiated tumors in xenograft models (Bifulco et al., 2004). In this study, the endogenously synthesized cannabinoid, 2-AG reduced thyroid tumor development. Similarly, agents that inhibited the EMT (VDM-11) and blocked AEA hydrolysis (AA-5HT) prevented in vivo tumor growth. Also, the metabolically stable molecules, arvanil and met-F-AEA also produced a significant reduction in tumor growth. The results of this study suggested that manipulation of the EC system was a viable option to prevent propagation of thyroid tumor cells. Additional preclinical studies were conducted with met-F-AEA. Using the xenograft model with KiMol cells it was determined that met-F-AEA prevented thyroid tumor growth and this that effect was reversed by antagonism of the CB1 receptor (Bifulco et al., 2001). Other studies with KiMol cells confirmed the anti-tumor activity of met-F-AEA and also indicated that this agent regulated tumor angiogenesis (Portella et al., 2003). Specifically, tumors isolated from met-F-AEA treated animals contained reduced levels of VEGF and VEGF receptor compared to vehicle treated animals. Furthermore, the reduction in VEGF signaling could be prevented by blockade of CB1 receptor activity. In another study, the synthetic cannabinoid, JWH-133, was tested in the highly aggressive ARO cell tumor model (Shi et al., 2008). Xenograft tumor growth using ARO/CB2 cells (a CB2 receptor overexpressing thyroid cancer cell line) was inhibited in mice treated with JWH-133 compared to vehicle. Cell culture experimentation suggested that JWH-133 toxicity was mediated by the CB2 receptor in contrast to met-F-AEA which was reported to initiate thyroid carcinoma cell death via CB1 (Portella et al., 2003). The data collected thus far suggest that synthetic EC system modulators represent novel therapeutic targets for thyroid cancer. However, because of the scarcity of in vivo thyroid cancer studies with cannabinoids, more extensive evaluation is needed to confidently define the role of the ECS in this malignancy.

The epidermis is comprised of different cell types including keratinocytes and melanocytes, which are the source of non-melanoma skin cancer (NMSC) and melanoma, respectively. Skin cancer is the most common cancer in the United States with approximately 5.4 million new lesions (3.3 million cases) of NMSC and more than 76,000 new cases of melanoma diagnosed each year (American Cancer Society, 2016; Siegel et al., 2016). Melanoma is more deadly and aggressive than NMSC. It is estimated that over 10,000 patients will die from melanoma in 2016 (Siegel et al., 2016). Several in vitro and in vivo studies suggest that the ECS plays a vital role in melanoma and NMSC development and progression (reviewed in Soliman et al., 2016a). It has been reported that CB1 and CB2 receptors were expressed in human melanoma cell lines as well as in human cutaneous melanoma biopsies (Blázquez et al., 2006). In addition, circulating levels of the endocannabinoid, 2-AG, increased in metastatic B16F10 mouse melanoma as well as in patients with metastatic melanoma suggesting a link between the ECS and melanoma progression (Sailler et al., 2014). Furthermore, in human and murine normal keratinocytes and non-melanoma skin tumors (e.g., papilloma, basal cell carcinoma, squamous cell carcinoma) CB1 and CB2 receptors were expressed and were required for the development of NMSC in vivo (Casanova et al., 2003; Zheng et al., 2008).

Human studies that investigate the pharmacotherapeutic benefits associated with the use of cannabinoid ligands focus on reductions in pain, spasticity and cognitive deficits in a number of central and peripheral nervous system disorders (Velasco et al., 2016). To date only a few cannabis-based pharmacological agents are licensed for clinical use. In Europe, Sativex® is approved for treatment of spasticity associated with multiple sclerosis and in Canada it is additionally approved as an adjunct analgesic for cancer pain. The CB1 antagonist rimonabant (Acomplia®) was licensed for use in Europe to treat obesity and related conditions but was discontinued due to adverse effects (Fijal and Filip, 2016). In the United States and Europe, nabilone (Cesamet®) and dronabinol (Marinol®) are approved compounds for prevention of chemotherapy-induced nausea and vomiting. The emergence of preclinical studies that demonstrate the anti-tumor effects of cannabinoids are growing in number and have formed the basis of limited clinical studies which are beginning to shed light on the translational value of the preclinical work.

Guzman and colleagues were the first group to report a Phase I human study of the anti-metastasic effects of Δ⁹-THC. In this open-labeled two-part clinical trial, nine patients with recurrent glioblastoma who were refractory to surgery and radiotherapy were treated with intracranial Δ⁹-THC in combination with high-dose chemotherapeutic agent, temozolomide (Temodar®). Following a washout phase (a period in a clinical trial where patients receive no active medication), the safety and tolerability of Δ⁹-THC was then compared to placebo (Clinical trial ID# NCT01812603). The primary outcome measured was the incidence of adverse effects in patients receiving the combination of Δ⁹-THC and Temodar®. Secondary measures were the progression free survival at 6 months and overall survival. Additional study outcomes have not yet been reported. Δ⁹-THC was found to reduce tumor cell growth in vitro from 2 study participants and reduce tumor cell immunostaining in these same patients. Importantly, intracranial administration of Δ⁹-THC was found to be a safe and tolerable approach with no apparent psychoactive effects (Gúzman et al., 2006). In a 2011 case report by Foroughi et al. two children with septum pellucidum/forniceal policytic astrocytomas experienced tumor regression following craniotomies and partial excision. The tumor regression occurred during the time that cannabis was inhaled by the patients (Foroughi et al., 2011). In another study, a 14-year-old female Philadelphia chromosome positive patient who had been unsuccessfully treated with “traditional” therapy (i.e., chemotherapy, bone marrow transplant and radiation) for acute lymphoblastic leukemia experienced a dose-dependent management of the condition with orally administered cannabinoid extracts (Singh and Bali, 2013). A clear limitation of the current human studies evaluating the anti-cancer effects of cannabinoid compounds is the small patient size. To date, no study findings have been replicated in multiple cohorts. Moreover, the measured outcomes and study designs are often variable across trials, making it difficult to compare their findings. However, despite these challenges, the evidence generated in these human studies are still informative and, when taken together with the strong in vivo animal data demonstrating anti-tumor effects of cannabinoids, offer promise for a clinical role for cannabinoids in the eradication of tumors.

To gain an understanding regarding the translational and therapeutic potential of cannabinoids, this review focused on examining the overall efficacy of these molecules in animal and human studies. The majority of in vivo animal studies discussed here indicate that cannabinoids from plant, synthetic and endogenous origin are capable of effectively decreasing tumor growth and invasion (Table 1). Furthermore, clinical studies evaluating cannabinoid efficacy in human subjects are limited, yet these studies showed that cannabinoids may be safe and effective anti-neoplastics. Because large scale clinical trials are lacking, examining the activity of cannabinoids in clinically translatable animal models (such as PDX) should be the goal of future research as this approach may provide a more accurate assessment of the therapeutic potential of cannabinoids.

This review also sought to couple in vivo animal studies with in vitro experiments in order to examine proposed cannabinoid mechanisms of action and to identify specific cellular targets. The studies reviewed herein indicate that cannabinoids elicit activity through cannabinoid receptor dependent and independent pathways. Moreover, processes such as ceramide production, ER-stress, autophagy, angiogenesis and matrix remodeling also appear to regulate the anti-tumor activity of cannabinoids. Hence, these investigations shed light on the role of cannabinoids on tumor growth in vivo and may ultimately pave the way for the development of novel cannabinoid therapeutics for cancer treatment.

DL, ES, LG, and RV contributed to the conception and design of the study, the acquisition of data, and the analysis and interpretation of the data. All authors participated in drafting or revising the manuscript, and all authors approved the final version of the manuscript for submission.