Cannabinoids As Potential Treatment for Chemotherapy-Induced Nausea and Vomiting

Despite the advent of classic anti-emetics, chemotherapy-induced nausea is still problematic, with vomiting being somewhat better managed in the clinic. If post-treatment nausea and vomiting are not properly controlled, anticipatory nausea—a conditioned response to the contextual cues associated with illness-inducing chemotherapy—can develop. Once it develops, anticipatory nausea is refractive to current anti-emetics, highlighting the need for alternative treatment options. One of the first documented medicinal uses of Δ9-tetrahydrocannabinol (Δ9-THC) was for the treatment of chemotherapy-induced nausea and vomiting (CINV), and recent evidence is accumulating to suggest a role for the endocannabinoid system in modulating CINV. Here, we review studies assessing the therapeutic potential of cannabinoids and manipulations of the endocannabinoid system in human patients and pre-clinical animal models of nausea and vomiting.


INTRODUCTION
Cannabis sativa has been used as a medicine for centuries (see Hanus and Mechoulam, 2005;Iversen, 2008). It was not until the 1970's that oncologists demonstrated that smoked cannabis attenuated chemotherapy-induced nausea and vomiting (CINV). Few clinical trials have compared the efficacy of cannabis-based medicines with the currently recommended anti-emetic regimen, or as an adjunct to this treatment. We review findings on the potential of exogenous cannabinoids and manipulations of the endogenous cannabinoid system to reduce acute and anticipatory CINV.

CHEMOTHERAPY-INDUCED NAUSEA AND VOMITING (CINV)
Chemotherapy patients experience acute nausea and vomiting (occurring up to 24 h post-treatment; Fiore and Gralla, 1984). If improperly managed, this post-treatment CINV can lead to anticipatory nausea and vomiting; a conditioned nausea response upon re-exposure to the chemotherapy clinic (Morrow, 1982). Current guidelines to manage highly emetogenic acute CINV recommend a three-drug regimen of the 5-hydroxytryptamine 3 (5-HT 3 ) receptor antagonist (such as ondansetron), along with dexamethasone, and a neurokinin 1 (NK 1 ) receptor antagonist (such as aprepitant) before beginning chemotherapy (Roila et al., 2010). Even with this standard treatment acute nausea is still problematic (no acute nausea reported in 66% of patients; Kim et al., 2015). None of these treatments are effective in reducing anticipatory nausea (e.g., Roscoe et al., 2000), with sedating benzodiazepines currently prescribed (Razavi et al., 1993;Malik et al., 1995). Therefore, nausea (acute and anticipatory) continues to be problematic.

CANNABINOIDS IN HUMAN CINV
Because current treatments cannot properly manage CINV, alternatives including constituents of the cannabis plant and modulation of the endogenous cannabinoid system, have been investigated.
The only published clinical trial assessing the effect of dronabinol on anticipatory nausea showed that dronabinol was ineffective, although most patients were receiving highly emetogenic chemotherapy regimens (Lane et al., 1991). Therefore, dronabinol may be effective in reducing anticipatory nausea developing from less emetogenic chemotherapy regimens.

Effect of Anandamide (AEA) and FAAH Inhibition on Vomiting
The endogenous cannabinoid, anandamide (AEA), produced and released on-demand, is rapidly degraded by fatty acid amide hydrolase (FAAH). As well, FAAH degrades other fatty acids including oleoylethanolamide (OEA) and palmitoylethanolamine (PEA), which act on peroxisome proliferator-activated receptor alpha (PPARα), instead of CB 1 or CB 2 receptors. Interestingly, Venkatesan et al. (2016) reported increased levels of serum OEA and PEA (with a trend toward increased AEA and 2-AG) while patients were experiencing cyclic vomiting. On the other hand, no differences in plasma AEA, OEA or PEA were detected in pregnant women experiencing hyperemesis gravidarum-severe nausea and vomiting (Gebeh et al., 2014).  In animal models, AEA (1, 2 mg/kg, i.p.) reduced M6Ginduced emesis in ferrets, an effect blocked by a transient receptor potential cation channel subfamily V member 1 (TRPV1) receptor antagonist (Sharkey et al., 2007) or AM251 (5 mg/kg, i.p.; Van Sickle et al., 2005). The FAAH inhibitor, URB597 (3, 5 mg/kg, i.p.) also reduced M6G-induced emesis in ferrets, an effect blocked by AM251 (5 mg/kg, i.p.) or a TRPV1 receptor antagonist (Van Sickle et al., 2005;Sharkey et al., 2007) but a PPARα antagonist was not evaluated. URB597 (0.9 mg/kg, i.p.) also reduced nicotine-induced vomiting in house musk shrews, an effect blocked by SR141716 (2.5 mg/kg, i.p.; Parker et al., 2009). These results suggest the anti-emetic effects of AEA and FAAH inhibition are mediated by activation of the CB 1 receptor. In ferrets, the TRPV1 receptor also plays a role, an effect not yet been evaluated in house musk shrews.
In comparison, administration of the FAAH inhibitors AA-5-HT (10 mg/kg, i.p.) or URB597 (20 mg/kg, i.p.) themselves induced emesis ; however 20 mg/kg of URB597 is a much higher dose than is typically given. These species-dependent effects of AEA in the modulation of emesis are puzzling, warranting further investigation.

CONDITIONED GAPING RE-CLINICAL MODELS OF NAUSEA IN RATS
Use of pre-clinical animal models has led to a good understanding of emesis neurobiology (Hornby, 2001), but the brain circuits mediating nausea are still not well characterized (Andrews and Horn, 2006). Such nausea circuitry may be more complex than that of emesis (see Kenward et al., 2015). Emesis is a gastrointestinal event controlled by structures within the brainstem (Hornby, 2001), whereas nausea is thought to require forebrain activation (Sanger and Andrews, 2006;Horn, 2008;Holmes et al., 2009). Although the visceral inputs from the gastrointestinal tract to the brain have been identified (Cechetto and Saper, 1987), it is unclear how these inputs are processed in the forebrain to produce nausea, largely due to the lack of reliable animal models of nausea. Here we describe current animal models of nausea. For a complete review of these models please refer to Sharkey et al. (2014).
To evaluate potential anti-nausea compounds, selective pre-clinical animal models are necessary. One such model is conditioned gaping in rats. Please refer to Table 2 for details regarding the effects of exogenous cannabinoids and manipulations of the endogenous cannabinoid system in rat models of conditioned gaping.

Acute Nausea-Induced Conditioned Gaping
Although rats cannot vomit, they display conditioned gaping reactions to a taste previously paired with an illness-inducing agent such as LiCl (Grill and Norgren, 1978). Only emetic drugs produce, and anti-emetic treatments (including cannabinoids) block conditioned gaping (see Parker, 2014 for review). Therefore, acute nausea-induced conditioned gaping is a reliable model of acute nausea in rats.

Contextually Elicited Conditioned Gaping, A Preclinical Model of Anticipatory Nausea
Rats also display conditioned gaping upon re-exposure to a nausea-paired context; this model is similar to the development of anticipatory nausea in humans . Furthermore, much like with human anticipatory nausea, a 5-HT 3 receptor antagonist does not reduce contextually elicited conditioned gaping (Limebeer et al., 2006;Rock et al., 2014). Humans are treated with nonspecific benzodiazepines for anticipatory nausea, similarly, benzodiazepines reduce contextually elicited conditioned gaping in rats . Therefore, there is face validity for contextually elicited gaping as a preclinical model of anticipatory nausea.

The Role of the Interoceptive Insular Cortex in Conditioned Gaping
Because the specific brain region(s) critical for nausea are still unclear, we are investigating the role of the endogenous cannabinoid system in nausea using the conditioned gaping model. One region of interest is the interoceptive insular cortex (IC), an area involved in the sensation of nausea in humans (Penfield and Faulk, 1955), as stimulation of the interoceptive IC (Ostrowsky et al., 2000;Isnard et al., 2004;Catenoix et al., 2008) and functional neuroimaging studies in humans (Napadow et al., 2013;Sclocco et al., 2014), pinpoint the interoceptive IC as a region critical for nausea.

Anticipatory nausea
CBD (1, 5 mg/kg, i.p.) or CBDA (0.001, 0.01, 0.1 mg/kg, i.p.) suppressed contextually elicited gaping in the absence of any locomotor impairments (Rock et al., 2008Bolognini et al., 2013), these effects were all reversed by a 5-HT 1A receptor antagonist (Bolognini et al., 2013). These results suggest a 5-HT 1A receptor mediated effect for CBD and CBDA in acute and anticipatory nausea and also a synergistic potential when combined with other anti-emetic agents.
Effect of AEA and FAAH Inhibition on nausea Acute nausea FAAH inhibition (by PF3845, but not URB597) reduces acute nausea by a PPARα mechanism of action, not a CB 1 receptor mechanism (Rock et al., 2015b). Previous work suggested that URB597 in combination with AEA also reduced LiCl-induced aversive responding, but not gaping per se (Cross-Mellor et al., 2007). The potential of TRPV1 or CB 2 receptor antagonists to reverse the anti-nausea effects of FAAH inhibition has not yet been evaluated. It is interesting that elevated OEA and PEA occur in serum of patients when they are experiencing cyclical vomiting (Venkatesan et al., 2016), suggesting that they may be playing a homeostatic protective role. Current investigations are underway to determine if the anti-nausea effects of FAAH inhibition (possibly by a PPARα mechanism of action) are peripherally or centrally mediated. 2 | Effect of exogenous cannabinoids and manipulations of the endogenous cannabinoid system on models of acute and anticipatory nausea in rats.

Anticipatory nausea
In the preclinical model of anticipatory nausea, both URB597 (0.3, 10 but not 0.1 mg/kg, i.p.) and PF3845 (10 and 20 mg/kg, i.p.) suppressed the expression of previously established contextually elicited gaping, with both effects blocked by CB 1 receptor antagonism, but not PPARα antagonism (Rock et al., 2008(Rock et al., , 2015b. In addition, the selective FAAH inhibitor, AM4303 (20 mg/kg, i.p.), also reduced contextually-elicited conditioned gaping, with an increase in interoceptive IC AEA levels . These results suggest that FAAH inhibition may reduce anticipatory nausea through a CB 1 receptor mediated effect; however, the potential of TRPV1 receptor antagonists and CB 2 receptor agonists to reverse LiCl-induced anticipatory nausea expression has not yet been evaluated.

CONCLUSIONS
Animal models suggest that, in general, 9 -THC, THCA, CBD, and CBDA, and manipulations of the endogenous cannabinoid system, have anti-emetic and anti-nausea properties. However, 2-AG and AEA's role in emesis is inconsistent across species. Further investigation is needed regarding the potential role of TRPV1 receptors in the anti-nausea effects produced by treatments that elevate AEA. It is time to take some of the preclinical findings (in particular CBDA, FAAH, and MAGL inhibition) into clinical trials for the treatment of acute and anticipatory nausea.

AUTHOR CONTRIBUTIONS
ER wrote the article; LP edited the article.