Localization of Cannabinoid CB1 Receptors in the Human Anterior Eye and Retina

Abstract

purpose. To determine the presence and distribution of CB1 cannabinoid receptors within the human eye.

methods. A subtype-specific affinity-purified polyclonal antibody to the cannabinoid CB1 receptor was used to determine CB1 localization. Postmortem human eyes were fixed in methacarn and embedded in paraffin. Sagittal sections were mounted on slides and immunostained using antibodies to the CB1 receptor. Antibody binding was detected either by using peroxidase conjugated secondary antibodies and developing with diaminobenzidine or by using fluorescent secondary antibodies.

results. Strong CB1 receptor labeling was detected in the ciliary epithelium, the corneal epithelium, and endothelium of the anterior human eye. Strong-to-moderate levels of CB1 staining were found in the trabecular meshwork and Schlemm’s canal. Moderate labeling was detected in the ciliary muscle and in the blood vessels of the ciliary body. Moderate-to-light labeling also was detected in the sphincter papillae of the anterior human eye. Staining for CB1 receptors also was detected in human retina. The two synaptic layers of the retina and the inner and outer plexiform layers, were both moderately stained for CB1. In addition, moderate labeling was detected in the inner nuclear layer, and the ganglion cell layer. Strong labeling was detected in the outer segments of photoreceptors. No staining was observed in the corneal stroma or in the choroid.

conclusions. The wide distribution of cannabinoid CB1 receptors in both the anterior eye and the retina of humans suggests that cannabinoids influence several different physiological functions in the human eye.

Ingestion or topical application of cannabinoids present in marijuana and hashish lowers intraocular pressure (IOP), suggesting that they may be useful for treating glaucoma. ^{1 2 3 4 5} Potential sites of action lie either in the ciliary epithelium, where aqueous humor is formed, or in the aqueous humor outflow pathways. Cannabinoid consumption also has been linked to corneal opacification, accommodative changes, photophobia, and alterations of vision. ^{6 7 8} Though some groups have sought to explain cannabinoid action as non–cannabinoid receptor mediated, ⁹ little hope existed for settling this question before the cloning of the first cannabinoid receptor. ¹⁰ Since then, two cannabinoid, or CB, receptors have been identified. CB1 is enriched in the brain. ¹¹ Another receptor, known as CB2, ¹² is thought to be limited to the periphery, with functions relating to the immune system, ¹³ though some work suggests that CB2 may be present in the CNS, including the retina. ^{14 15}  

Numerous CB1 receptor–mediated effects have been observed, ranging from modulation of nociception and glutamate transmission to inhibition of long-term potentiation. ^{16 17 18} In addition, candidate endogenous ligands have been identified: arachidonylethanolamide (anandamide or AEA), ¹⁹ 2-arachidonylglycerol (2-AG), ¹⁷ and palmitylethanolamide (PEA). ²⁰ Of these, 2-AG and anandamide both lower IOP when applied topically. ^{9 21 22} PEA, the sole putative ligand for the CB2 receptor, ²⁰ has no effect on IOP. ²¹ Interestingly, the selective CB1 antagonist SR141716A increases IOP on its own and opposes the effects of a synthetic CB1 ligand CP-55,940 but not that of anandamide. ²³ This suggests that anandamide itself may influence IOP by a non–CB1 receptor–mediated pathway. The presence of anandamide amidohydrolase, the enzyme thought to break down anandamide, has been found to be active in the retina. ²⁴ Our own studies detected 2-AG and PEA in the retina but not anandamide (Straiker AJ, Stella N, Piomelli D, Mackie K, Karten HJ, Maguire G, unpublished observations). This does not exclude possible circadian or activity-dependent release of anandamide, nor does it exclude the presence of anandamide elsewhere in the eye. 

Recently, RT-PCR has been used to indicate the presence of CB1 mRNA in the retina and anterior eye. ²⁵ Unfortunately, these results give little insight into the actual presence or precise localization of CB1 receptors. The availability of antibodies against the CB1 receptor has made it possible to determine its presence and distribution in the eye. We recently used a CB1 receptor antibody to localize CB1 immunoreactivity in the retinas of monkey, mouse, chick, goldfish, and salamander. ²⁶ Here we report the distribution of CB1 labeling in the anterior segment of the human eye, as well as in human retina. 

Materials and Methods

Immunohistochemistry

Human tissue consisted of paraffin-embedded sections obtained from eyes received from the San Diego Eyebank. Ten eyes from donors of various ages from 44 to 90 years were fixed in methacarn (60% methanol, 30% chloroform, 10% glacial acetic acid) for 3 hours, then dehydrated, and embedded in paraffin. One of the 10 eyes (eye 80A) was fixed in formalin. Another eye (eye 68) was fixed in paraformaldehyde. For our experiments, anterior eye sections from five donors (49, 80A, 80B, 81, 86) and retina sections from seven donors were used (44, 68, 71, 80A, 81, 87, 90). Sections were heated at 56°C for at least 20 minutes and then deparaffinized in xylenes, followed by rehydration in an ethanol series. After washing in PBS, slides were preincubated with 3% H₂O₂ as a peroxidase suppressor. Tissue was allowed to incubate overnight at 4°C with the affinity-purified rabbit polyclonal CB1 receptor antibodies (1:200 for eyes 80A and 81,1:400 for all others, made in PBS, with 0.3% Triton X-100, 5% normal goat serum, 0.5% BSA). These antibodies have been characterized previously. ²⁷  

Immunoperoxidase labeling was obtained by subsequently treating the tissue with the biotinylated anti-rabbit IgG antibodies and then avidin horseradish peroxidase. The sections then were developed using diaminobenzidine (Biogenix, San Ramon, CA). 

In one case, an eye was immersed in 4% paraformaldehyde made in 0.1 M sodium phosphate buffer at pH 7.4 overnight. After fixation, the eyecup was kept in a 30% sucrose solution in phosphate buffer for at least 48 hours before being frozen in embedding medium. Sections 10μ m thick were cut on a cryostat and thaw-mounted onto glass slides. Slide-mounted sliced retina was washed in phosphate-buffered saline (PBS), incubated overnight at 4°C with the affinity-purified rabbit polyclonal CB1 receptor antibodies (1:200 dilution made in PBS, with 0.3% Triton, 0.5% BSA). After the overnight incubation, the sections were washed with PBS and then incubated with lissamine rhodamine goat anti-rabbit antibodies (1:100; Jackson Immunoresearch Laboratories, Inc., West Grove, PA) for 90 minutes at room temperature. Finally, the tissue was washed with PBS and coverslipped with glycerine carbonate. 

In control experiments, primary antibodies were omitted to determine the level of background labeling, which typically was low. As a second control, the immunizing protein (1–4 μg/ml) was mixed with the CB1 antibodies. In all cases, CB1 labeling was blocked successfully or was diminished substantially by blocking with the immunizing protein. Premixing the CB1 antibodies with a similar quantity of the immunizing protein for CB2 did not diminish labeling, though it did, in some instances, reduce background labeling. 

Results

Anterior Eye

CB1 labeling was detected in several locations in the anterior segment of the human eye. In the cornea and angle, strong labeling was detected in the corneal epithelium and endothelium (Fig. 1) . Strong labeling also was detected in the trabecular meshwork in a pattern consistent with trabecular epithelium. In one case, moderate labeling also was detected in Schlemm’s canal (Fig. 2) . Moderate labeling also was detected in the lining of blood vessels in the angle. No labeling was detected in the corneal stroma. 

In the uveal tract, the sphincter papillae muscle of the iris was moderately stained for CB1 receptors (Fig. 1) . The anterior border layer, stroma, pigment epithelium, and the root of the iris were unstained (not shown). Labeling in the myoepithelium of the iris was difficult to assess because of the heavy pigmentation in these cells. The ciliary smooth muscle fibers were strongly stained for CB1, as was ciliary nonpigmented epithelium (Fig. 2) . Labeling in the ciliary pigmented epithelium was difficult to distinguish from inherent pigmentation. 

The pattern of CB1 staining generally was consistent in the five eyes examined (Table 1) . The notable exception to this was the ciliary nonpigmented epithelium, which in several cases had slight nonspecific staining in the absence of cannabinoid receptor antibodies. Blocking of the CB1 antibody by premixing with immunizing protein always diminished or eliminated nonspecific staining, except in the ciliary epithelium, where the staining was only diminished, but not eliminated. This suggests that there was some recognition of a ciliary epithelium antigen by the secondary antibody. General differences in intensity of labeling was observed between eyes 80A and 81 and the others, a difference that appeared to be explainable by the fact that eyes from donors 80A and 81 were treated with higher concentrations (1:200) of CB1 antibody than eyes from other donors. 

Retina

Antibodies to CB1 distinctly labeled the two synaptic layers of the retina and the outer and inner plexiform layers (Fig. 3) . CB1 also heavily stained the photoreceptor outer (and to a lesser extent inner) segments and portions of the ganglion cell layer. 

Light CB1 labeling was detected in the inner plexiform layer, with no evidence of stratification. Labeling in the inner nuclear layer was diffuse and difficult to assign to a particular cell type, though the labeling was consistent with the presence of CB1 in some amacrine cells (Figs. 3d 3e , paired left-pointing arrows and arrowheads). CB1 staining was detected in the ganglion cell layer and in the ganglion cell axon layer. In those eyes that had suffered the greatest postmortem retinal degradation (eyes 87, 90, 44, 81), labeling in the IPL was somewhat diminished, whereas labeling in the ganglion cell axon layer was increased (not shown). Additionally, in retinas from these donors, we detected a greater incidence of somatic labeling in a subpopulation of cells in the inner nuclear layer. No labeling was detected in the choroid. Labeling in the retinal pigment epithelium was difficult to assess, given the pigmented nature of the tissue. Negligible staining was observed in the control retina sections. This included controls in which CB1 antibody was incubated with immunizing protein and also in controls in which incubation with the primary antibody was omitted (Figs. 3c 3f) . 

Discussion

The presence of cannabinoid receptors in many different parts of the anterior eye is consistent with the many reported physiological effects of cannabinoid consumption. First, labeling in the ciliary pigment epithelium suggests that cannabinoids may have an effect on aqueous humor production. Second, staining in the trabecular meshwork and Schlemm’s canal suggests that cannabinoids may influence conventional outflow. Third, the presence of immunolabeling in the ciliary muscle suggests that cannabinoids may influence uveoscleral outflow. These observations suggest that IOP lowering by cannabinoids may reflect direct effects on ocular tissues. However, because CB1 receptors are distributed throughout much of the brain, ¹¹ IOP lowering by cannabinoids may reflect central regulation as well as local control. 

The initial discovery in 1971 that cannabinoids decrease IOP generated considerable interest. However, enthusiasm waned when it became clear that the undesired psychoactive properties of cannabinoids made them an imperfect treatment for elevated IOP. Also, the cannabinoid-induced lowering of IOP usually only lasts 4 to 6 hours, necessitating relatively frequent treatment. Tolerance to cannabinoids develops in humans, ²⁸ though the only controlled long-term animal study showed no tolerance in rabbits after 1 year of twice daily application of synthetic cannabinoids. ²⁹ Heavy users who abruptly stop treatment experience a rebound in IOP that temporarily increases it above pretreatment levels. ³⁰ Despite these drawbacks, cannabinoids still hold promise as a therapeutic agent to lower IOP. Thus, the possibility of a CB1-mediated influence on IOP warrants further investigation. Pate et al. ^{22 23 31 32} have begun to examine endogenous, exogenous, and synthetic cannabinoids as part of a search for more effective means of reducing IOP, with some encouraging results. Establishment of a role for CB1 in the modulation of IOP, along with the ongoing dissection of the pathways activated by cannabinoid receptors, may allow a detailed characterization of the metabolic machinery underlying the maintenance of ocular tension. 

The corneal endothelium, located on the posterior surface of the cornea, consists of a thin layer of simple squamous epithelial cells set in a honeycomb array. These cells play a vital role in maintaining corneal hydration, serving effectively to pump aqueous humor out of the cornea to maintain corneal clarity. ³³ When corneal hydration rises beyond a certain level, some precipitation occurs, resulting in corneal opacification. Intriguingly, corneal opacification has been seen in primates treated with high doses of tetrahydrocannabinol. ³⁴ It is possible, then, that CB1 receptor activation inhibits corneal endothelial mechanisms for removing aqueous humor from the cornea. 

CB1 was also detected in the corneal epithelium, the outermost cellular layer serving primarily as a barrier to protect the eye. Any potential role of cannabinoids in the corneal epithelium remains to be investigated but it is interesting to note that during corneal healing, cell migration occurs in a process mediated by cAMP and accompanied by the development of adhesion complexes. ³³ CB1 acts in part by altering levels of cAMP and has been shown to activate focal adhesion kinase, implying a potential role for CB1 in cell migration. These observations suggest CB1 agonists and antagonists may influence corneal wound healing. 

Ciliary muscles serve to alter the accommodation of the lens, allowing us to focus on objects at various distances. Because ciliary muscles have an attachment to the trabecular meshwork, contraction of these muscles causes a significant change in the shape of the trabecular meshwork. This facilitates the escape of aqueous humor by conventional outflow via Schlemm’s canal, reducing IOP. ³⁵ If cannabinoids were to facilitate the contraction of the ciliary muscles, this might provide another explanation for cannabinoid effects on IOP. However, one would expect ciliary muscle contraction to produce a reduction in the range of accommodation in humans. Such an effect might serve to explain the difficulty reported by some people to read while under the influence of cannabis. Very little work has been done on this in humans, but two authors have observed just such a weakening of accommodation in patients known to smoke marijuana. ^{8 36} However, to our knowledge no controlled study of the effects of cannabinoids on accommodation has been undertaken. Cannabinoid action on ciliary muscle cells also might influence IOP by altering uveoscleral outflow, which passes through extracellular spaces in ciliary muscle. ³⁴  

The presence of CB1 receptors in the sphincter papillae muscle provides a possible site of action by cannabinoids on pupil dilation/contraction. A number of articles have either reported a constriction of the pupil (miosis) or no effect. ^{21 37 38} The labeling we observed suggests that a reinvestigation of the phenomenon is in order. 

In human retina, the overall pattern of CB1 labeling resembled that found in other vertebrates, particularly that of the primate retina. ^{26 39} Previous work in our laboratory using fluorescent labeling demonstrated the presence of CB1 receptors in the synaptic terminals of cone and rod photoreceptors, known as pedicles and spherules, respectively. ²⁶ Although CB1 is clearly present in the outer plexiform layer in a pattern similar to that of other vertebrates, including structures suggestive of photoreceptor pedicles and spherules, we were unable to identify with certainty spherule and pedicle structures with either the fluorescence or immunoperoxidase techniques. As in other species, CB1 labeling was detected in the outer and inner segments of the photoreceptors. In contrast to other species in which spherule and pedicle labeling constituted the strongest retinal labeling, we found the outer segments of human photoreceptors to be the most prominent. No labeling was detected in the somas of rod or cone photoreceptors. A wide range of visual effects have been ascribed to the use of marijuana and hashish, including an alteration of light sensitivity thresholds and glare recovery. ^{7 40 41} CB1 expression in photoreceptors may explain some of these effects and may thus represent a novel neuromodulatory system at the first level of visual processing. This is particularly so if CB1 is present in the synaptic terminals of human photoreceptors as we found in other vertebrates. ²⁶  

Labeling in the inner nuclear layer was suggestive of the presence of CB1 receptors in amacrine cells. This would be consistent with a recent report of CB1 receptor labeling in some amacrine cells. ³⁹ Activation of CB1 receptors has been shown to reduce cAMP levels by inhibiting adenylyl cyclase, activate inwardly rectifying K channels and I_A currents, and inhibiting P/Q- and N-type calcium channels. ^{42 43 44 45 46 47 48} Some of these channels are known to be present on amacrine, bipolar, and ganglion cells and may be influenced by cannabinoid receptor activation in the IPL. For example, we have shown that cannabinoid receptor agonists inhibit L-type calcium currents in bipolar cells of the tiger salamander (Straiker A, Stella N, Piomelli D, Mackie K, Karten HJ, Maguire G, unpublished observations). Müller cells possess several types of inwardly rectifying K channels in abundance. ⁴⁹ These channels are thought to play a role in the reuptake of potassium. Any effect of cannabinoids on these channels in Müller cells might serve to influence retinal pathology. ⁵⁰  

In conclusion, cannabinoid receptors represent part of a novel modulatory system both in the retina and in the anterior eye. Their ubiquity and distribution, combined with their known actions in other parts of the body, are suggestive of a role that extends well beyond the effects generally attributed to cannabinoids as drugs of abuse. Cannabinoids, acting via the CB1 receptor may substantively affect the maintenance of ocular tension, corneal hydration, corneal wound healing, and quite possibly vision itself. As such, further research into the mechanisms underlying these effects may provide a more thorough understanding of a wide range of interesting systems and open new therapeutic avenues in both the anterior eye and the retina. 

 Supported by National Institutes of Health Grants EY09133 (GM), The Glaucoma Foundation (GM), DA05908 (AS), and DA00286 and DA11322 (KM).

 Submitted for publication December 7, 1998; revised April 9, 1999; accepted April 21, 1999.

 Proprietary interest category: N.

 Corresponding author: Greg Maguire, Ophthalmology 0946, UCSD, La Jolla, CA 92093. E-mail: [email protected]

 

Figure 1.

 

CB1 cannabinoid receptor labeling in the cornea and iris. Low magnification photographs of cornea labeled with antibodies to the CB1 receptor (a) demonstrate labeling in the corneal epithelium (arrow) and corneal endothelium (arrowhead). Control sections with CB1 antibody omitted (b) and premixed with immunizing protein (c). Higher magnification image shows strong labeling of the corneal epithelium (d) and corneal endothelium (e). The sphincter papillae muscle of the iris is positively labeled (h, arrowhead), whereas tissue labeled with antibody premixed with immunizing protein is not (g). Magnified view of iris (i) shows detail of sphincter papillae muscle labeling. Scale bars, (a, b, c) 133 μm; (d, e, g, h, i) 66 μm. Original magnification, (a, b, c) ×25; (d, e) ×100; (g, h, i) ×25.

Figure 2.

 

CB1 cannabinoid receptor labeling in the ciliary body and angle. Lower magnification image (a) shows CB1 labeling in ciliary muscle, nonpigmented ciliary epithelium, trabecular meshwork, and in the blood vessels of the ciliary body. Control section (b) shows labeling from antibody premixed with immunizing protein. Section is unlabeled except for partially blocked staining in ciliary epithelium. Note that large patches of black pigment are due to presence of full depth of ciliary process in the sagittal section (a, arrowheads). Magnified image (c) shows greater detail of angle including CB1 labeling on epithelium of the trabecular meshwork (arrowheads) as well as on cells lining the Schlemm’s canals (arrow). Labeling in ciliary muscle is also evident (d, arrows) with labeled blood vessel (bv) visible. Image from same slice (e) shows strong labeling in nonpigmented ciliary epithelium (arrowheads). Scale bars, (a, b) 400 μm; (c) 10 μm; (d, e) 100 μm. Original magnification, (a, b) ×10; (c, d, e) ×100.

Table 1.

 

Summary of CB1 Receptor Labeling in Human Anterior Segment

CB1 Labeling	80 yo A	81 yo	80 yo B	49 yo	86 yo
Cornea
Epithelium	+++	+++	+	++	+
Stroma	−	−	−	−	−
Endothelium	+++	na	+	+	+
Iris
Anterior Border	−	na	−	−	−
Stroma	?	na	−	−	−
Sphincter	++	na	+	?	+
Pigment Epithelium	−	na	−	−	−
Root	?	na	−	−	−
Trabecular Meshwork	+++	+++	na	na	+
Schlemm’s Canal	++	?	na	na	+
Ciliary Body:
Nonpigment epithelium	+++	+++	+	na	+++
Ciliary Muscle Fibers	++	++	+	+	++
Blood Vessels	++	++	?	−	?

 (+++) Strong labeling; (++) moderate labeling; (+) light labeling; (−) background labeling

*  tissues from donors 80yo A and 81yo were treated with higher (1:200) concentration of antibody than others (1:400).

Figure 3.

 

CB1 receptors are present in human retina. Photographs of retinal cross sections labeled with CB1 antibodies (b, e) and control sections (c, f). Reference tissue sections are imaged with DIC optics (a, d). Prominent labeling is found in photoreceptor outer segments (b), outer plexiform layer, inner plexiform layer and ganglion cell layer (e). Labeling is also present in the inner nuclear layer in a pattern consistent with amacrine cell labeling. Outer plexiform layer staining may possess labeled structures corresponding to photoreceptor synaptic terminals (small arrowheads), a result that would be consistent with labeling in other species. Amacrine cells appear to possess CB1 receptors (d, e; paired left-pointing arrows and arrowheads). Labeling was detected with fluorescent (lissamine rhodamine) secondary antibodies and viewed with a confocal microscope. IS/OS, inner/outer segments of photoreceptors; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer. Scale bars, (a, b, c) 10μ m; (d, e) 20 μm. Original magnification,× 100.

The authors thank Anna M. Cervantes for technical assistance and the San Diego Eye Bank for providing the human eye tissue used in this study.