Two studies for WhosCarl?

Ending Marijuana Prohibition Would Save $10-14 Billion Annually ... Report Endorsed by Milton Friedman and More Than 500 Economists

In "The Budgetary Implications of Marijuana Prohibition" (released June 2, 2005), Dr. Jeffrey Miron, visiting professor of economics at Harvard University, estimates that replacing marijuana prohibition with a system of taxation and regulation similar to that used for alcoholic beverages would produce combined savings and tax revenues of between $10 billion and $14 billion per year.


----

Cannabinoid Receptor Activation Differentially Modulates Ion Channels in Photoreceptors of the Tiger Salamander

Cannabinoid CB1 receptors have been detected in retinas of numerous species, with prominent labeling in photoreceptor terminals of the chick and monkey. CB1 labeling is well-conserved across species, suggesting that CB1 receptors might also be present in photoreceptors of the tiger salamander. Synaptic transmission in vertebrate photoreceptors is mediated by L-type calcium currents---currents that are modulated by CB1 receptors in bipolar cells of the tiger salamander. Presence of CB1 receptors in photoreceptor terminals would therefore be consistent with presynaptic modulation of synaptic transmission, a role seen for cannabinoids in other parts of the brain. Here we report immunohistochemical and electrophysiological evidence for the presence of functional CB1 receptors in rod and cone photoreceptors of the tiger salamander. The cannabinoid receptor agonist WIN 55212-2 enhances calcium currents of rod photoreceptors by 39% but decreases calcium currents of large single cones by 50%. In addition, WIN 55212-2 suppresses potassium currents of rods and large single cones by 44 and 48%, respectively. Thus functional CB1 receptors, present in the terminals of rod and cone photoreceptors, differentially modulate calcium and potassium currents in rods and large single cones. CB1 receptors are therefore well positioned to modulate neurotransmitter release at the first synapse of the visual system.

Cannabinoid CB1 receptors have been detected in the retinas of numerous species, including salamander, goldfish, chick, rat, mouse, monkey, and human (Straiker et al. 1999a,b; Yazulla et al. 1999) with a well-conserved pattern of staining that is particularly intense in both synaptic layers. In species for which double labeling has been performed---chick and monkey---outer plexiform layer staining is found to localize to photoreceptor terminals. CB1 is a member of the Gi/o-type G-protein-coupled receptor family and mediates a wide range of effects on ion channels, including voltage-dependent calcium and potassium channels (Deadwyler et al. 1995; Mackie and Hille 1992; Mackie et al. 1995; Mu et al. 1999). Retinal cannabinoid receptor activation has been shown to inhibit dopamine release in the guinea pig (Schlicker et al. 1996), to inhibit a delayed rectifier K+ current in ON cone bipolar cells of the goldfish (Yazulla et al. 2000), and to inhibit L-type calcium currents in bipolar cells of the tiger salamander (Straiker et al. 1999a).

Three recent reports have described G-protein-coupled receptor-dependent modulation of ion channel function in photoreceptors of the tiger salamander by dopamine, adenosine, and somatostatin (Akopian et al. 2000; Stella and Thoreson 2000; Stella et al. 2002). These studies find that the calcium currents (ICa) of rods and large single cones respond differentially to activation of the same receptor type.

To study the possible presence and action of cannabinoid receptors in rod and cone photoreceptors of the tiger salamander, we employed immunohistochemistry and electrophysiology. CB1 receptors were found to be present in both rod and cone photoreceptors, including the terminals of both cell types. We recorded from the tiger salamander retinal slice preparation using whole cell voltage-clamp techniques to examine the effect of cannabinoid receptor activation on voltage-dependent ion currents in rod and cone photoreceptors. Studying both ICa and potassium currents (IK) in rods and large single cones, we found that the cannabinoid receptor agonist WIN 55212-2 (WIN) differentially modulates ICa in rods and cones, enhancing the former and suppressing the latter. In addition, WIN was found to inhibit IK of both rod and large single cone photoreceptors. These actions were blocked by the cannabinoid receptor antagonist SR141716A, indicating a CB1 receptor-dependence.

Larval tiger salamanders (Ambystoma tigrinum, Charles Sullivan, TN, 7-10 in) used in these experiments were killed by decapitation followed by pithing of the brain, after which the eye was promptly dissected out. All procedures used in this study were approved by the IACUC of the Salk Institute, La Jolla, CA, and conform to the guidelines of the National Institutes of Health on the Care and Use of Laboratory Animals.

Immunohistochemistry

For preparation of frozen slide-mounted tissue, the anterior eyepoles and vitreous were cut away. The eyecups were bathed overnight in 4% paraformaldehyde made in 0.1 M sodium phosphate buffer at pH 7.4. After fixation, the eyecups were kept in a 30% sucrose solution in phosphate buffer for >= 48 h before being frozen in embedding medium. Sections 10-µm thick were cut on a cryostat and thaw-mounted onto glass slides (Fisherbrand Superfrost/Plus, Fisher, Pittsburgh, PA).

Slide-mounted sliced retinas were washed in phosphate-buffered saline (PBS), incubated overnight at 4°C with the affinity-purified rabbit polyclonal CB1 receptor antibodies developed against the amino terminus of the rat CB1 receptor (1:500 dilution, made in PBS, with 0.3% Triton, 10% nonfat milk; the generous gift of Dr. Ken Mackie, University of Washington). After the overnight incubation, the sections were washed with PBS and then incubated with Alexa 568 goat anti-rabbit antibodies (1:500, Molecular Probes, Eugene, OR) for 90 min at room temperature. Finally, the tissues were washed with PBS and coverslipped with glycerine carbonate.

In separate double-labeling experiments the antibodies to calbindin (1:4,000, Sigma, St. Louis, MO)/CB1(1:500) were applied simultaneously to slide-mounted tissue. The calbindin antibodies, made in mouse, were then each detected with Alexa 488 goat anti-mouse antibody (1:500 dilution, Molecular Probes) while the CB1 was detected as before by Alexa 568 anti-rabbit antibodies. In control experiments, CB1 and calbindin antibodies were omitted to determine the level of background labeling, which was typically low. As a further control, the immunizing protein (a fusion protein consisting of the first 77 amino acids of the rat CB1 receptor and glutathione-S-transferase, 1 µg/ml) was pre- and co-incubated with the CB1 antibodies. In all cases, CB1 labeling was successfully blocked by inclusion of the immunizing protein. CB1 receptor immunolabeling was performed on retinas from two different salamanders with identical results.

Imaging was performed using a Leica confocal laser-scanning microscope with TCS-NT (Leica Microsystems) acquisition/imaging software. Images were subsequently prepared with Adobe Photoshop. Images were not subjected to any processing beyond ordinary adjustment of brightness and contrast.

Electrophysiology

Retinal slices were prepared by removal of anterior eye region, forming an eyecup, which was submerged in cold extracellular solution. The retina was removed from of the eyecup, detached from retinal pigment epithelium, and then flattened against a nitrocellulose membrane filter, ganglion cell layer down. The preparation was then sliced to yield transverse sections. Slices were incubated for 1 h at 4-7°C. Recordings were not made >6 h after slicing. Although the cells remained superficially healthy after this time, even sealing more readily onto patch pipettes, we found that cells exhibited more rapid rundown and unreliable responses to WIN. For this reason, experiments requiring longer-term incubation (i.e., pertussis toxin treatment) were not attempted.

Whole cell voltage-clamp recordings from photoreceptors were carried out using an Axopatch 200B amplifier (Axon Instruments, Burlingame, CA). The standard extracellular solution contained (in mM) 100 NaCl, 3 KCl, 2 CaCl2, 2 MgCl2, and 10 HEPES. pH and osmolarity of all external solutions were adjusted to 7.6 and 242 mosM, respectively. Calcium currents were assessed with the use of a high-barium external solution containing (in mM) 80 NaCl, 2.5 KCl, 10 BaCl2, 0.5 MgCl2, 10 HEPES, 20 tetraethylammonium, 0.1 picrotoxin (to block inhibitory GABAA-mediated currents), and 0.1 niflumic acid (to block calcium-dependent chloride currents). Barium was used in place of calcium because it yields a more prominent current through the L-type calcium channels found in rod and cone photoreceptors and because barium is less likely than calcium to activate ICl(Ca) and IK(Ca), to inactivate ICa, or to stimulate neurotransmitter release, all of which might confound our measurements.

The CB1 receptor agonist WIN 55,212-2 (WIN) was bath-applied for 30-60 s. Continuous flow of solution through the bath chamber (~1 ml/min) ensured rapid drug clearance. WIN was prepared as a stock solution in DMSO, then diluted on the day of experiments into extracellular solution at a final concentration of 1 µM. The concentration of DMSO did not exceed 0.04%. To block effects of WIN, the selective CB1 receptor antagonist SR141716A was both incubated with the slice and coapplied with WIN.

For isolation of calcium currents, recording pipettes of 4-8 MOmega were filled with (in mM): 100 CsGluconate, 2 MgCl2, 10 HEPES, 2 MgATP, and 0.1 LiGTP. For isolation of potassium currents, KGluconate was substituted for CsGluconate. Internal solutions were set to pH of 7.3 and osmolarity adjusted to 242 mosM. Solutions did not contain calcium chelators. Access resistance was monitored, and only cells with stable access resistance were included in the data analysis. The membrane potential was held at -60 mV. Current-voltage (I-V) curves were generated by ramping the membrane potential from -80 to +30 mV, at a rate of 0.5 mV/ms, every 20 s. Data were acquired at a rate of 5 kHz. Graphing and statistical analyses were performed using GraphPad Prism (GraphPad Software, San Diego, CA). Statistical analyses consisted of Student's t-test or paired t-test, as appropriate. Unless otherwise noted, all drugs and chemicals were obtained from Sigma.

Photoreceptor types

The main types of photoreceptors in tiger salamander retina are rods, large single cones, double cones, and small single cones. The rods are the most numerous (Mariani 1986), easily distinguished by their cylindrical outer segments, and may be divided into two subpopulations. The M-type rods are far more numerous (~65% of all photoreceptors) than the S-type (1%) and are the best characterized (Bader et al. 1982; Mariani 1986). We did not attempt to record from S-type rods. The largest population of cone photoreceptors is the large single cones (16% of all photoreceptors). Large single cones are morphologically distinct from the double cones and the smaller---and rarer---small single cones and are also better characterized (Barnes and Hille 1989). No effort was made to record from small single cones due to their small size, sparsity, and relative inaccessibility, but we did attempt recordings in the accessory cones of the double cone pairs. Accessory cones are relatively large, readily identifiable by their bulbous almond-shaped inner segment and sufficiently numerous to attempt recordings. Recordings were only attempted in intact pairs.

RESULTS


Cannabinoid CB1 receptors are expressed in both rod and cone photoreceptors

Cannabinoid CB1 receptor labeling has previously been observed in the outer and inner plexiform layers of the salamander and in the cone terminals, or pedicles, of goldfish, chick, and monkey (Straiker et al. 1999a; Yazulla et al. 2000). Calbindin is a cone marker in a variety of species (Hamano et al. 1990) and has been shown to label both large single cones and double cones in salamander (Deng et al. 2001). Double-labeling of CB1 receptors and calbindin was observed in the terminals of cone photoreceptors (Fig. 1C, arrows). Additional CB1 labeling in the outer plexiform layer was observed at the base of clearly identifiable rod photoreceptors (Fig. 1A), likely corresponding to CB1 receptor expression in rod terminals. This localization is consistent with the CB1 expression pattern observed in other species using light (Straiker et al. 1999a) and electron microscopy (X. Liu and D. S. Williams, personal communication).


Known currents in rod and cone photoreceptors of the tiger salamander

Five distinct currents have been observed in the inner segments of rod and cone photoreceptors of the tiger salamander (Bader et al. 1982; Barnes and Hille 1989). One is a voltage-dependent, dihydropyridine-sensitive calcium current (ICa) that is generally thought to be carried by L-type calcium channels. Second is a voltage-dependent, TEA-sensitive potassium current (IK). Third is a hyperpolarization-activated cationic current (Ih). Last there are two calcium-sensitive currents, one carried by potassium [IK(Ca)], the other by chloride [ICl(Ca)]. Of these five currents, we studied ICa and IK and, to a lesser extent, IK(Ca). Ih was not activated by the depolarizing ramping stimulus used here, whereas our bath solutions routinely contained niflumic acid to block ICl(Ca), so these currents were not studied.

CB1 receptor activation increases calcium currents in rod photoreceptors in a PKA-dependent manner

Application of WIN (1 µM) enhanced ICa in rod photoreceptors (Fig. 2, A-C; relative amplitude: 1.39 ± 0.05, n = 17). This effect was blocked by the cannabinoid receptor antagonist SR141716A (1 µM; Fig. 2C; relative amplitude: 0.96 ± 0.05, n = 6), indicating CB1 dependence. SR 141716A alone had no effect on these currents (Fig. 2C; relative amplitude: 0.99 ± 0.07, n = 4). The enhancement was reversed by washout of the WIN and typically recovered over the course of several minutes (e.g., Fig. 2B; 4.3 ± 0.5 min, n = 8). Fan and Yazulla (2002) have reported that lower concentrations of WIN (100-500 nM as opposed to 1-2 µM) act in an opposite manner on photoreceptor ICa and IK currents in goldfish cones. To explore this phenomenon in salamander rods, we tested the effect of 200 nM WIN on rod ICa and IK. WIN (200 nM) also enhanced the ICa (relative amplitude: 1.31 ± 0.08, n = 5), suggesting that ICa in rods, at least, is not modulated in an opposite manner by submaximal cannabinoid receptor activation.


ICa in rods was blocked by cadmium (Fig. 3A, 100 µM; relative amplitude: 0.20 ± 0.02; n = 7) and the dihydropyridine nifedipine (10 µM; relative amplitude: 0.43 ± 0.03; n = 5). The action of WIN was occluded by both cadmium and nifedipine (Fig. 3A). The substitution of barium for calcium in external solutions and the use of cesium-based internal solutions are both expected to block IK(Ca) (Bader et al. 1982); however as a further control, we treated cells with charybdotoxin (20 nM), which had no effect on our measured ICa (relative ICa amplitude: 1.05, n = 6). Taken together, these results indicate that WIN enhancement of ICa in rods is due to modulation of L-type calcium currents.

Elevation of cAMP levels has been shown to enhance ICa in rod photoreceptors and to underlie dopamine D2 receptor-mediated enhancement of these currents (Stella and Thoreson 2000). To test whether the mechanism of WIN action involves the cAMP/PKA-signaling pathway, we treated cells with the PKA inhibitor Rp-cAMPS followed by WIN + Rp-cAMPS. Rp-cAMPS (10 µM) enhanced rod ICa (Fig. 3B; relative current amplitude: 1.35 ± 0.12mV, n = 4), whereas subsequent application of WIN + Rp-cAMPS produced only a modest further increase in ICa (Fig. 3B; current amplitude relative to pretreatment control: 1.45 ± 0.17mV, n = 4; not significant by paired t-test). Thus Rp-cAMPS occludes the action of WIN, suggesting that the action of WIN on rod ICa is PKA-dependent.

CB1 receptor activation inhibits ICa in large single cones but not accessory cones

In contrast to rods, the application of WIN to large single cones produced a robust inhibition of ICa (Fig. 4; relative ICa amplitude: 0.50 ± 0.07, n = 5). This effect was blocked by coapplication of SR141716A (1 µM; Fig. 4C; relative ICa amplitude: 0.97 ± 0.07, n = 5), indicating that the effect was cannabinoid receptor-dependent. In some cells, treatment with SR141716A alone produced a moderate increase in ICa, but this effect for the group as a whole was not statistically significant (Fig. 4C, relative ICa amplitude: 1.15 ± 0.07, n = 4). ICa recovered from WIN-inhibition after washout, typically in a few minutes (3.0 ± 0.3 min, n = 5). Because of the relative difficulty of recording from cone photoreceptors, more extensive studies of ICa and IK were not performed in these cells.

Sealing onto accessory cone members of intact double cone pairs proved difficult and even after successful commencement of recordings, ICa often declined rapidly. However, in several successful recordings the ICa of these cones were insensitive to treatment with WIN (relative amplitude: 0.97 ± 0.07, n = 4).

CB1 receptor activation inhibits IK in rod photoreceptors

Although the focus of this study lay with modulation of neurotransmission via calcium currents, the finding by Akopian et al. (2000) that photoreceptor IK is inhibited by somatostatin as well as previous reports of IK modulation by CB1 (Deadwyler et al. 1993; Mu et al. 1999), prompted us to examine cannabinoid modulation of these currents in rod photoreceptors. We found that WIN 55,212-2 inhibited the potassium currents in rod photoreceptors (Fig. 5; relative IK amplitude measured at +30 mV = 0.56 ± 0.07, n = 8). This effect was blocked by coapplication of SR141716A (Fig. 5C, relative IK amplitude: 0.97 ± 0.04, n = 5). SR 141716A alone did not alter the potassium currents in rod photoreceptors (Fig. 5C, relative IK amplitude: 0.99 ± 0.02, n = 6). As with rod ICa, application of lower concentration of WIN (200 nM) did not produce an opposite effect on IK but instead inhibited IK to a lesser extent (relative amplitude 0.87 ± 0.08, n = 7). Note that any WIN-mediated enhancement of IK(Ca)---due to effects on ICa---would oppose the inhibition of IK that we observe, i.e., CB1 receptor activation may therefore inhibit IK to an even greater degree than indicated by our results.



WIN-induced inhibition often recovered only partly if at all. Akopian et al. (2000) noted a similar lack of recovery even 15 min after washing off somatostatin. In our hands, of five cells maintained for >= 10 min, three displayed recovery to >= 80% of original IK, whereas in cells maintained for <10 min, none showed the same recovery. Our results suggest that the IK recovers when given enough time. The difference in time course of IK recovery, relative to that for ICa, indicates that different mechanisms may underlie modulation of potassium and calcium currents and may have implications for the time course of cannabinoid effects on neurotransmitter release from photoreceptors.

CB1 receptor activation inhibits IK in large single cone photoreceptors

To ascertain whether IK also exhibits differential modulation by CB1 receptor activation, we obtained recordings of these currents in large single cones. We found that WIN 55,212-2 inhibited the potassium currents in large single cones (Fig. 6; relative IK amplitude measured at +20 mV = 0.52 ± 0.09, n = 4). Thus while CB1 acts differentially on calcium currents in rods versus cones, CB1 receptor activation reduces the potassium current in both rods and cones. In contrast to rods, cone IK recovered promptly (1.1 ± 0.2 min, n = 4). To rule out the possibility that the observed inhibition of cone IK is the indirect result of CB1 modulation of ICa, we examined the effect of the voltage-dependent Ca channel blocker cadmium (100 µM) on cone IK. We found that cadmium blocked only a small portion of the net IK (relative IK amplitude measured at +20 mV = 0.85 ± 0.04, n = 3) and so inhibition of ICa (and subsequent inhibition of IK(Ca)) could not account for the robust modulation of potassium currents observed by CB1 receptor activation.






DISCUSSION

The combined immunohistochemical and electrophysiological results presented in this study indicate that functional cannabinoid receptors are expressed in both rod and cone photoreceptors of the tiger salamander. Activation of cannabinoid receptors enhances ICa in rods but suppresses ICa in large single cones. In addition, cannabinoid receptor activation inhibits potassium currents in both rod and large single cone photoreceptors. These results represent the first evidence for differential and PKA-dependent modulation of ICa by cannabinoids. The data also suggest a novel site of action for cannabinoids---at the first synapse of the visual system---via modulation of at least two ion channels in rod and cone photoreceptors.

Functional significance

Photoreceptor membranes rest at depolarized potentials (approximately -40 mV) in the dark and tonically release neurotransmitter. Light causes photoreceptors to hyperpolarize, decreasing ICa and reducing neurotransmitter release. A CB1-mediated increase in ICa of rod photoreceptors will increase glutamate release. A decrease in rod photoreceptor IK will depolarize the membrane, further increasing ICa, and enhancing neurotransmitter release. The net effect of cannabinoid receptor activation on calcium and potassium currents in rod photoreceptors will therefore be a decrease in the sensitivity to light. In contrast, the effect of cannabinoid receptor activation on cone photoreceptor activation will be more complex, with a decreased neurotransmitter release and increased light sensitivity produced by inhibition of ICa, but opposition of these actions due to changes in net IK.

Divergent effects of CB1 activation on L-type calcium channels

The differential effect of CB1 activation on calcium currents of rod and cone photoreceptors may be most simply explained by a difference in channel subunit composition rather than any differences in signaling pathways activated by cannabinoid receptors. Currents carried by L-type channels containing alpha 1C or alpha 1S subunits are enhanced by cAMP (McDonald et al. 1994), whereas currents carried by L-type channels containing alpha 1D subunits are inhibited by cAMP (Chik et al. 1997). The tiger salamander retina may express at least three different L-type alpha subunits, one similar to alpha 1C and two with similarity to alpha 1D or the retina-specific alpha 1F (Hamid et al. 1999; Morgans et al. 2001), although the distribution of these subunits within the retina remains to be demonstrated. Our results might predict that rod photoreceptors express alpha 1D- or alpha 1F-like subunits and that cone photoreceptors express alpha 1C-like subunits.

Mechanism of action

While most studies of cannabinoid receptor action on ICa have focused on modulation of N and P/Q type calcium channels---presumably via direct Gbeta gamma subunit interactions with these channels (Garcia et al. 1998; Herlitze et al. 1996; Ikeda 1996)---several reports indicate that CB1 activation can also modulate L-type calcium channels in some systems (Gebremedhin et al. 1999; Straiker et al. 1999a). There is reason to think that GPCR modulation of L-type calcium currents in photoreceptors is mediated by cAMP via PKA. Specifically, three recent studies have described the G-protein-coupled receptor effects on ion channel function in photoreceptors of the tiger salamander: two examined the actions of dopamine D2 and adenosine A2 receptors, respectively, in modulating ICa (Stella and Thoreson 2000; Stella et al. 2002), whereas the other examined the effect of somatostatin on both calcium and potassium currents (Akopian et al. 2000). These results have been summarized along with our own in Table 1. Those studies that investigated both rods and cones found that ICa was modulated in opposite directions in rods and cones and, interestingly, also in opposite directions by the respective GPCRs. This latter difference may be due to differential coupling to adenylate cyclase for the GPCRs studied and consequent downstream actions on PKA. For example, the D2 and CB1 receptors, which activate Gi/o-type G proteins, have been found to be negatively coupled to adenylyl cyclase (Howlett 1984; Howlett and Fleming 1984; Onali et al. 1981)---reducing cAMP levels---whereas the Gs-activating adenosine A2 receptor is positively coupled to adenylyl cyclase (van Calker et al. 1979). Consistent with these findings, both CB1 and D2 effects on ICa were mimicked by PKA inhibitor Rp-cAMPS, and the D2 effect occluded by PKA activator Sp-cAMPS (Stella and Thoreson 2000), whereas the A2 effect was occluded by Rp-cAMPS (Stella et al. 2002). The cAMP/PKA-dependence of somatostatin action in salamander retina has not been investigated (Akopian et al. 2000).

In the hippocampus, where neurotransmitter release is principally dependent on N- and P/Q-type calcium channels, cannabinoid modulation of these currents is thought to occur via direct G-protein beta gamma subunit interaction with the target channel. In the hippocampus, cannabinoid receptors have not been found to modulate L-type calcium channels (Twitchell et al. 1997). This stands in contrast to photoreceptors in the vertebrate retina, where neurotransmitter release relies on the action of L-type calcium channels. Here cannabinoids modulate L-type calcium currents in a PKA-dependent manner, an action that would be expected to occur via G-protein alpha or beta gamma subunit modulation of adenylyl cyclase (Simonds 1999). These results indicate a conservation of function---modulation of neurotransmitter release---by CB1 receptors across disparate neuronal settings, via distinct intracellular pathways and calcium channel effectors.

Although several groups have now studied the mechanism by which GPCRs modulate ICa in tiger salamander retina, the pathway for modulation of IK in these cells has received little attention. There is reason to think that the signaling pathway for modulation of IK involves second messengers in addition to cAMP/PKA---for example, PKC in the case of CB1 (Hampson et al. 2000). Indeed, given that CB1 modulates multiple channels in rods, separate pathways would very sensibly allow for differential intracellular modulation of CB1 action on each channel type. The nature of the pathway by which the several GPCRs studied in salamander retina modulate rod and cone potassium currents would be an excellent topic for future study.

Cannabinoid receptors at the first synapse of the visual system

Numerous visual effects have been reported for marijuana and hashish consumption, many of them involving an alteration---generally enhancement---of light sensitivity (Brown 1974; Dawson et al. 1977; Gregg et al. 1976a,b; Hepler and Frank 1971; Hollister 1971a,b; Hollister and Gillespie 1973; Laffi and Safran 1993; Shapiro 1974). Recent description of a cannabinoid signaling system in the vertebrate retina has raised the question of whether some of these visual effects may be due to a retinal site of action. CB1 receptors have now been shown to act at three sites in tiger salamander retina---bipolar cells and both rod and cone photoreceptors---in each case consistent with presynaptic modulation of glutamate release.

There is now evidence that---like dopamine (Djamgoz and Wagner 1992) and adenosine (Blazynski and Perez 1991)---endogenous cannabinoids are produced according to prevailing circadian and/or light conditions (Murillo-Rodriguez et al. 2001; A. Straiker, L. Walter, and N. Stella, personal communication). In view of these observations, differential effects of cannabinoid receptor activation on calcium currents in rods and cones are consistent with a model wherein cannabinoids are released from postsynaptic bipolar or horizontal cells onto presynaptic photoreceptors, differentially modulating rod and cone outputs to facilitate light or dark adaptation of the retina in response to circadian rhythms or ambient light conditions. Although there are critical differences between amphibian and mammalian retinae, the conservation of CB1 labeling across species may reflect a conservation of function and so extend to nonamphibian retinae.

In conclusion, CB1 differentially modulates ICa in rod and cone photoreceptors and also inhibits potassium currents in both rods and cones. CB1 action on photoreceptor ICa indicates that cannabinoid modulation of neurotransmitter release is conserved in disparate neuronal settings such as the retina and the hippocampus via distinct second messengers and calcium channel types. It would be of considerable interest to learn whether the cAMP/PKA dependence observed thus far extends to IK and/or to the small single cones, particularly because Stella and Thoreson (2000) found that dopamine D2 receptor activation enhanced ICa in small single cones. The questions of how endogenous cannabinoid signals originate and interact with other GPCRs to shape the net photoreceptor signal in the tiger salamander, and the extent to which such observations transfer to mammalian retinae, invite further study.

http://jn.physiology.org/cgi/content/full/89/5/2647