Potential roles of (endo)cannabinoids in the treatment of glaucoma: from intraocular pressure control to neuroprotection

Introduction

Glaucoma is an optic neuropathy characterized by apoptotic death of the retinal ganglion cells (RGCs) and loss of the axons that make up the

Corresponding author. Tel.: 39 06 7259 6147;

Fax: 39 06 2026232; E-mail: nucci@med.uniroma2.it aEqually senior authors

optic nerve (Weinreb and Khaw, 2004). The changes provoked by glaucoma are progressive, and if left untreated can produce severe visual-field deficits. The pathogenesis of glaucoma is complex and multifactorial, but elevated intraocular pressure (IOP) is one of the risk factors most closely associated with both onset and progression of the disease. Increases in the IOP are believed to be responsible for blood-flow alterations that ultimately lead to hypoxia and ischemia of the retina

452

and the optic nerve. Treatment — medical or surgical — is aimed at reducing IOP to prevent structural and functional damage, but this approach is not always sufficient to block progression of the disease and to prevent blindness (Heijl et al., 2002; Kass et al., 2002). Progressive optic nerve damage has in fact been documented in many glaucoma patients, whose IOP has been pharmacologically or surgically normalized (Heijl et al., 2002; Kass et al., 2002), and in others whose untreated pressure is well within the normal range (normal tension glaucoma) (Collaborative Normal Tension Glaucoma Study Group, 1998). Therefore, it seems clear that the damage caused by glaucoma is related in part to other factors (vascular, genetic, etc.) that have yet to be identified.

Experimental data suggest that the excitotoxic cascade triggered by glutamate plays a fundamental role in RGC damage associated with glaucoma (Nucci et al., 2005a). Elevated intravitreal levels of this neurotransmitter have been documented in experimental models of glaucoma (LouzadaJu´nior et al., 1992; Adachi et al., 1998) and in glaucoma patients (Dreyer et al., 1996), and while other authors have not confirmed these findings (Carter-Dawson et al., 2002; Honkanen et al., 2003; Kwon et al., 2005), agents that antagonize N-methyl-D-aspartate (NMDA) and non-NMDA glutamate receptors have produced neuroprotective effects in experimentally induced RGC death (Sucher et al., 1997; Adachi et al., 1998; Nucci et al., 2005b) and in a monkey model of glaucoma (Hare et al., 2004). So it may be possible that compromised RGC are more susceptible to damage by glutamate, even at normal level (Lipton, 2003) or it may reflect a technical difficulty in measuring only the extracellular glutamate content in tissue samples.

Despite the rational basis for NMDA receptor blockade as a means to prevent excitotoxic RGC death, trials to treat glaucoma have recently failed to provide evidence based support and this is reminiscent of failure of similar clinical trials to treat stroke. To reconcile this apparent discrepancy it can be suggested that abnormal activation of glutamate receptors might occur in a time and space locked fashion affecting downstream cellular mechanisms eventually responsible for

RGC death. Accordingly, long-term inhibition of excitatory neurotransmission does not afford neuroprotection if derangement of downstream mechanisms are defective. A great deal of information concerned with the interplay between the glutamatergic and endocannabinoid systems have been recently accumulated in the normal and pathological eye and these open new lanes of investigation.

Cannabinoids (CBs) are a class of molecules that seem to produce therapeutic effects in various types of central nervous system (CNS) pathologies, including Parkinson’s disease (Lastres-Becker et al., 2005), Alzheimer’s disease (Ramirez et al., 2005), head trauma (Panikashvili et al., 2001), multiple sclerosis (Centonze et al., 2007), and Huntington’s disease (Maccarrone et al., 2007). CBs include the active principles of Cannabis sativa (extracts like marijuana), as well as a number of endogenous and synthetic compounds that like the plant-derived counterparts, interact with two G-protein-coupled receptors referred to as CB1 and CB2 (Howlett et al., 2002). Some endocannabinoids like anandamide (N-arachidonoylethanolamine, AEA) bind to and also activate the type-1 vanilloid receptor (now called transient receptor potential vanilloid type 1, TRPV1), a ligand-gated, nonselective cation channel, and are therefore considered true ‘‘endovanilloids’’ (Starowicz et al., 2007). The discovery by Straiker et al. (1999) of CB receptors and their agonists at the ocular level sparked numerous attempts to determine whether and how this system might be involved in physiologic andor pathologic processes of the eye. There is a large body of experimental data showing that certain endocannabinoids and synthetic CBs can reduce IOP, and these findings have led to the suggestion that CBs might be used in the treatment of glaucoma, together with or instead of traditional antiglaucoma drugs (a2-agonists, carbonic anhydrase inhibitors, prostaglandin analogs, b-blockers) (Tomida et al., 2006). Additional support for the use of this approach has emerged from more recent studies, which indicate that the (endo)cannabinoids are also capable of producing specific neuroprotective effects at the level of the retina.

In this review, we will examine some of the latest evidence of the role of the endocannabinoid system

in the control of IOP and the therapeutic potential of CBs in preventing RGC death induced by glaucoma.

The endocannabinoid system in the eye

Endocannabinoids (eCBs) are amides and esters of long-chain polyunsaturated fatty acids. AEA and 2-AG are the two most studied members of this new group of lipid mediators (Piomelli, 2003; Bari et al., 2006). AEA and 2-AG are synthesized and released ‘‘on demand’’ by neurons and peripheral cells by cleavage of lipid precursors (Hansen et al., 2000; Ligresti et al., 2005). AEA synthesis occurs by a two step pathway: the first to originate the N- arachidonoyl-phosphatidylethanolamine (NArPE) precursor and the second to generate AEA and phosphatidic acid, through a specific N-acylpho- sphatidylethanolamine (NAPE)-hydrolyzing phospholipase D (NAPE-PLD) (Okamoto et al., 2004). Recent studies demonstrated that AEA has a presynaptic origin (Cravatt et al., 2008) and that the key enzyme of its synthesis, NAPE-PLD, is associated with intracellular calcium stores in several types of excitatory axon terminals (Di Marzo et al., 1994; Okamoto et al., 2004). The biosynthetic pathway of 2-AG starts from multiple routes for the hydrolysis of its precursors, by phospholipase C or phosphatidic acid phosphohydrolase, to generate diacylglycerol (DAG), which is converted to 2-AG by a sn-1-DAG lipase (DAGL) (Bisogno et al., 2003).

Once released, AEA and 2-AG bind to the cannabinoid receptors (CBRs) (Howlett et al., 2002; Pertwee and Ross, 2002), mimicking the effects of D9-tetrahydrocannabinol (THC), the main psychoactive compound of hashish and marijuana (Howlett et al., 2004; Mechoulam et al., 2002). CBRs are seven trans-membrane spanning receptors that belong to the rhodopsin family of G protein-coupled receptors, particularly those of the Gio family (Howlett et al., 2002). Type-1 cannabinoid receptors (CB1R) are present mainly in the CNS, but are also expressed in peripheral tissues; type-2 cannabinoid receptors (CB2R) are expressed predominantly by astrocytes, spleen, and immune cells, but are also present in the brain stem. Earlier studies demonstrated that CB1Rs are located in the

453

eye (McIntosh et al., 2007) and that functional CB2Rs are also expressed in the retina and trabecular meshwork (He and Song, 2007).

Endocannabinoids can induce a biological activity also via other CB receptors, like a purported CB3 (GPR55) receptor (Sawzdargo et al., 1999; Baker et al., 2006; Lauckner et al., 2008), via non-CB1non-CB2 receptors, and via non-CBRs. In the latter group TRPV1, which is activated by capsaicin, the pungent ingredient in hot peppers (De Petrocellis et al., 2001), has emerged as an important target of AEA.

The AEA–TRPV1 interaction, which occurs at a cytosolic binding side (De Petrocellis et al., 2001; Jordt and Julius, 2002), results in a cascade of intracellular responses that make of AEA a true ‘‘endovanilloid’’ (Starowicz et al., 2007). Since TRPV1 is expressed in peripheral sensory fibers and also in several nuclei of the CNS (Marinelli et al., 2003; Maccarrone et al., 2008), the endovanilloid activity of AEA may play a physiological control of brain function.

The biological activity of AEA at CB receptors is terminated by its removal from the extracellular space through cellular uptake by a purported AEA membrane transporter (AMT). Once taken up by cell, AEA is substrate for fatty acid amide hydrolase (FAAH), which breaks the amide bond and releases arachidonic acid and ethanolamine (McKinney and Cravatt, 2005). Recently, two new enzymes for AEA hydrolysis have been characterized: FAAH-2, that exhibits a much greater hydrolytic activity than FAAH with monounsaturated acyl chains, and shows a lower selectivity for AEA (Wei et al., 2006); and a novel lysosomal hydrolase N-acylethanola- minehydrolyzing acid amidase (NAAA) (Tsuboi et al., 2005) with a unique role in the degradation of N-acylethanolamines.

2-AG also could be degraded by FAAH releasing arachidonic acid and glycerol, but it is mainly hydrolyzed by a monoacylglycerol lipase (MAGL) (Dinh et al., 2002).

AEA and its congeners, along with the proteins that bind, transport, synthesize, and degrade these lipids, form the ‘‘endocannabinoid system’’ (Piomelli et al., 2003; Bari et al., 2006).

The presence of a functional endocannabinoid system in the eye has been widely documented. In

454

fact, porcine ocular tissues synthesize and degrade AEA (Matsuda et al., 1997), as do bovine (Bisogno et al., 1999) and rat retina (Glaser et al., 2005) and various human eye tissues (Chen et al., 2005). In addition, CB1 receptors, FAAH, and TRPV1 have been shown to be widespread in the retina of rats and other mammals by immunocytochemical methods (Yazulla and Studholme, 2004). Recently, it has been demonstrated that elevated hydrostatic pressure induces a TRPV1-dependent influx of extracellular Ca2+ in retinal microglia that produce and release IL-6 (Sappington and Calkins, 2008). The presence of a functional endocannabinoid system in the eye supports a role for eCBs in ocular physiology. Experimental evidence has accumulated in the last few years to support the role of AEA in the CNS, while it acts as neuroprotective or neurotoxic agent. In fact, eCBs have been shown to possess protective activity in an experimental model of allergic uveitis (Pryce et al., 2003), and to regulate photoreception and neurotransmission in the retina (Fan and Yazulla, 2005; Struik et al., 2006). Several studies have shown that topical administration of CB1 agonists lowers IOP in rabbits, nonhuman primates, and glaucomatous humans (Devane et al., 1992; Laine et al., 2001, 2002a, b; Chien et al., 2003), possibly due to an increase in aqueous humor outflow (Chien et al., 2003; Njie et al., 2006). More recently, plant-derived and synthetic CBs have been shown to exert neuroprotective actions in the eye, with potential implications for the treatment of glaucoma (Tomida et al., 2006).

The IOP-lowering effects of endocannabinoids

The first studies on the IOP-reducing effects of CBs date back to the 1970s. Hepler and Frank (1971) found that eating or smoking marijuana reduced IOP by 5–45% (mean, approximately 25%). However, since the effect lasted for only 3–4 h, patients would have to use cannabis 8–9 times a day to keep their IOP under control. In a review of the literature published some 20 years later, Green (1998) noted that marijuana produces detectable IOP-lowering effects in only 60–65% of healthy subjects or volunteers with glaucoma,

and the efficiency of treatment seemed to be dosedependent. Similar results were observed when CBs were administered intravenously (Perez-Reyez et al., 1976; Cooler and Gregg, 1977), although preliminary data on topical administration yielded negative results (Jay and Green, 1983). The beneficial effects of marijuana on ocular hypertension were also accompanied by several toxic effects, including orthostatic hypotension, increased heart rate, emphysema-like lung changes, diverse alterations of the mental state (euphoria, reduced attention, short-term memory deficits, and altered motor coordination), and at the ocular level they included conjunctival hyperemia associated with a 50% reduction in tear secretion (Brown et al., 1977). These findings prompted the American Academy of Ophthalmology’s Complementary Therapies Task Force to carry out a systematic review of the peer-reviewed literature on the IOP-lowering effects of CBs (see website www.aao.orgeyecaretreatmentalternative-therapies marijuana-glaucoma.cfm). The conclusions were that there was no scientific evidence to support the hypothesis that marijuana was more effective or safer for the treatment of glaucoma than the drugs and surgical procedures already widely used to reduce IOP.

At the same time, however, various studies revealed that there was a complex network of receptors for eCBs at the ocular level, and these findings led to a new wave of studies to determine how this endogenous system might be related to the control of IOP. These efforts identified a series of (endo)cannabinoids (cannabis derivatives and endogenous or synthetic molecules that interact specifically with CBRs), that were able to produce selective IOP-lowering effects without provoking systemic toxicity. Hosseini et al. (2006), for example, showed that topical administration (three times a day) of WIN-55-212-2, a synthetic CB that binds both CB1 and CB2 receptors, reduces the IOP by as much as 47% in rats with experimentally induced glaucoma, and this effect was maintained over a period of treatment of 4 weeks. The treatment was not associated with any psychotropic manifestations or symptoms of systemic or local toxicity. These findings were consistent with previous observations in primates