Adenosine inhibits calcium channel currents via A1 receptors on ...

Journal of Neurochemistry, 2002, 81, 550–556

Adenosine inhibits calcium channel currents via A1 receptors on salamander retinal ganglion cells in a mini-slice preparation Xiaolu Sun,*, ,à,1 Steven Barnes ,à and William H. Baldridge*,à Departments of *Anatomy and Neurobiology, Physiology and Biophysics and àOphthalmology, Dalhousie University, Halifax, Nova Scotia, Canada

Abstract The effects of adenosine on high-voltage-activated calcium channel currents in tiger salamander retinal ganglion cells were investigated in a mini-slice preparation. Adenosine produced a concentration-dependent decrease in the amplitude of calcium channel current with a maximum inhibition of 26%. The effects of adenosine on calcium channel current were both time- and voltage-dependent. In cells dialyzed with GTPc-s, adenosine caused a sustained and irreversible inhibition of calcium channel current, suggesting involvement of a GTPbinding protein. The inhibitory effect of adenosine on calcium channel current was blocked by the A1 antagonist 8-cyclopentyltheophylline (DPCPX, 1–10 lM), but not by the A2 antagonist 3-7-dimethyl-1-propargylxanthine (DMPX, 10 lM), and was mimicked by the A1 agonist N 6-cyclohexyladenosine

(CHA, 1 lM) but not by the A2 agonist 5¢-(N-cyclopropyl) carbox-amidoadenosine (CPCA, 1 lM). Adenosine’s inhibition of calcium channel current was not affected by the L-type calcium channel blocker nifedipine (5 lM). However, adenosine’s inhibition of calcium channel current was reduced to approximately 10% after application of x-conotoxin GVIA (1 lM), suggesting that adenosine inhibits N-type calcium channels. These results show that adenosine acts on an A1 adenosine receptor subtype via a G protein-coupled pathway to inhibit the component of calcium channel current carried in N-type calcium channels. Keywords: A1 receptors, adenosine, calcium, ganglion cells, retina, salamander. J. Neurochem. (2002) 81, 550–556.

Adenosine has been identified as an important neuromodulator in various regions of the CNS where, acting in most cases via A1 receptors, it has been shown to inhibit the release of a broad range of classical neurotransmitters (for review see Dunwiddie and Masino 2001). The principal mechanism of A1 receptor-mediated inhibition of neurotransmitter release is G protein-coupled inhibition of calcium (Ca) channels (Fredholm and Dunwiddie 1988; Yawo and Chuhma 1993; Wu and Saggau 1997). While a great deal is known about the role of adenosine in the brain, less is known about the actions of adenosine in the retina. In mammalian retina, A1 receptors are localized predominantly in the optic nerve fiber layer, ganglion cell layer and inner plexiform layer and, in addition, are found on the terminals of retinal ganglion cells in the brain (Goodman et al. 1983; Braas et al. 1987; Blazynski et al. 1989). This suggests the potential for adenosine-dependent regulation of ganglion cell Ca channel activity and therefore control of neurotransmitter release from ganglion cell terminals within the brain. Indeed, Zhang and Schmidt (1998, 1999) have shown that retinotectal synaptic transmission can be modu-

lated by adenosine in the goldfish and that this is likely due to A1 receptor-mediated modulation of N-type Ca channels on ganglion cells. In the present study we investigated the effect of adenosine on high voltage-activated (HVA) Ca channel currents of ganglion cells in the well-characterized tiger salamander retina. Retinal ganglion cell Ca channel currents were recorded using whole-cell patch techniques in a Ômini-sliceÕ

550

Received December 18, 2001; revised manuscript received January 21, 2002; accepted January 22, 2002. Address correspondence and reprint requests to William H. Baldridge, Department of Anatomy & Neurobiology, Dalhousie University, Halifax, Nova Scotia B3H 4H7, Canada. E-mail: [email protected] 1 The current address of Xiaolu Sun is Department of Pediatrics, Section of Respiratory Medicine, Yale University, School of Medicine, 333 Cedar Street, New Haven, CT 06520, USA Abbreviations used: Ca, calcium; CHA, N6-cyclohexyladenosine; CPCA, 5¢-(N-cyclopropyl) carbox-amidoadenosine; DMPX, 3-7-dimethyl-1-propargylxanthine; DMSO, dimethyl sulfoxide; DPCPX, 8-cyclopentyl-1,3-dipropylxanthine; HVA, high voltage activated; TTX, tetrodotoxin.

2002 International Society for Neurochemistry, Journal of Neurochemistry, 81, 550–556

Adenosine inhibits Ca2+ currents in RGCs 551

tiger salamander retina preparation, which preserves much of the ganglion cell’s dendritic arbour. Our goal was to pharmacologically characterize the types of adenosine receptors present and to determine the Ca channel subtypes that were modulated by adenosine. We hoped to build on and extend previous findings made in goldfish (Zhang and Schmidt 1999), which would indicate that A1 receptor modulation of ganglion cell N-type Ca channels may be a common feature of the vertebrate retina.

Materials and methods Preparation of retinal mini-slices In accordance with guidelines provided by the Canadian Council for Animal Care, larval tiger salamanders (Ambystoma tigrinum) were killed by decapitation and the eyes were enucleated. After the anterior portion of the eye was removed, the retina was isolated in amphibian superfusate containing: 90 mM NaCl, 2.5 mM KCl, 3 mM CaCl2, 15 mM HEPES, and 10 mM glucose, and with pH adjusted to 7.6. The retina was minced into small mini-slices under a microscope with half of a double-sided razor blade. The mini-slices were then transferred into the recording chamber via pipette. Prior to beginning an experiment, slices were allowed to adhere to the bottom of the plastic chamber for 5–10 min. Application of drugs and perfusion Solutions were applied by a single-pass, gravity-feed perfusion system that delivered medium to the slice chamber (chamber volume < 0.5 mL) at a rate of 3–5 mL/min. The normal bathing solution for slices was the same solution used for slice preparation. To record Ca channel currents, the chamber was superfused with a barium bathing solution containing: 90 mM NaCl, 2.5 mM KCl, 5 mM CsCl, 10 mM BaCl2, 15 mM TEACl, 15 mM HEPES, and 10 mM glucose, with pH adjusted to 7.6 with NaOH. Experiments were conducted at room temperature (22–23C). Whole-cell patch clamp recording from retinal ganglion cells Barium currents flowing through Ca channels (referred to in this paper as Ca channel currents) were recorded from ganglion cells using the whole-cell patch-clamp technique. Barium, used as the charge carrier because it permeates high voltage-activated Ca channels better than Ca2+ and reduces current flow in K channels, would also be expected to shift activation by up to + 10 mV (Hille 2001). Patch electrodes were pulled from hematocrit capillary glass using a two-stage pipette puller (Kopf model 730, David Kopf Instruments, Tujunga, CA, USA). The intracellular solution contained: 95 mM CsCl, 3 mM MgCl2, 10 mM HEPES, 1 mM EGTA, 1 mM ATP, and 0.2 mM GTP, adjusted to pH 7.4 with CsOH. To dialyze cells with a non-hydrolyzable analog of GTP, 0.2 mM GTPcs (Sigma, St Louis, MO, USA) was included in the patch solution instead of GTP. The morphology of cells recorded from in the ganglion cell layer was visualized by including Lucifer yellow in the internal solution at a concentration of 1%. Electrodes had 2.5–6 MW resistance when filled with internal solutions. Membrane potential and membrane currents were recorded with an Axopatch 1-D amplifier (Axon Instruments, Foster City, CA, USA). Signals were

filtered at 2 kHz () 3dB, 4-pole Bessel) and digitized at 4 kHz using an Indec Systems interface for storage on the hard disk of a personal computer. Stimulus generation and data acquisition were controlled by BASIC-FASTLAB program software (Indec Systems, Sunnyvale, CA, USA). Before seals were made on cells, offset potentials were nulled. Capacitance subtraction was used in all recordings. Drugs and solutions The adenosine receptor agonists and antagonists used were adenosine (RBI/Sigma); 8-cyclopentyl-1, 3-dipropylxanthine (DPCPX, RBI); 3-7-dimethyl-1-propargylxanthine (DMPX, RBI); N6-cyclohexyladenosine (CHA, RBI); 5¢-(N-cyclopropyl) carboxamidoadenosine (CPCA, Sigma). The Ca channel blockers used were x-conotoxin GVIA (Alomone Laboratories, Jerusalem, Israel) and nifedipine (Sigma). Adenosine and DMPX were first dissolved in water at concentrations of 10 mM and 5 mM, respectively; DPCPX and nifedipine were dissolved in dimethyl sulfoxide (DMSO) at 10 mM and 50 mM, respectively; CPCA and CHA were both dissolved in 0.1 N HCl at 10 mM. Tetrodotoxin (TTX, Sigma) stock solution concentration was 1 mM. Aliquots of stock solutions were kept at ) 30C and were later diluted with external solution before use. Data analysis To study adenosine effects on Ca channel current, currents were elicited from a holding potential of )70 mV. Single steps to )20 mV or a series of steps between )60 mV and + 50 mV were used. Mean values during the last 5 ms of each 50-ms voltage step were used to evaluate Ca channel current amplitude. Percentage block of these currents was defined as [1–(Itest)/(Icontrol)] · 100. Activation curves were constructed from steady-state I–V relations (after leak subtraction) by dividing by the driving force and were then fit with the Boltzmann function ( f ¼ {1 + exp [(V–V1/2)/ m]})1) to determine the half activation voltage (activation midpoint, V1/2) and slope factor (m). Raw data were compared for statistical significance using anova and Fisher’s PLSD post hoc comparison. Data are presented as mean ± SEM.

Results

Characterization of the inward currents recorded in ganglion cell layer cells of a mini-slice preparation A total of 48 retinal ganglion cell layer cells were recorded in mini-slices using the whole cell patch clamp technique. Retinal ganglion cells in the tiger salamander slice preparation labeled with Lucifer yellow (Fig. 1) typically do not reveal an axon but retrograde staining suggests that most of the neurons in the tiger salamander ganglion cell layer are ganglion cells (Lukasiewicz and Werblin 1988, 1990). Therefore, the neurons in the present work shall be referred to as retinal ganglion cells. The cells were round or slightly ovoid in shape and had capacitances ranging from 7 to 20 pf with a mean of 11.3 ± 0.4 pF (n ¼ 44). Occasionally much bigger cells with


552 X. Sun et al.

Fig. 2 Inward currents recorded in ganglion cells. (a) Transient sodium currents were blocked by 0.1 lM TTX. (b) Sustained inward Ca channel currents were blocked by 100 lM Cd2+. Inward currents were measured during a voltage step to ) 20 mV. Holding potential was ) 70 mV.

these cells with 10 mM Ba2+ as the charge carrier was between 50 and 200 pA (89.7 ± 10.0 pA, n ¼ 44).

Fig. 1 A retinal ganglion cell filled with Lucifer yellow in a mini-slice tiger salamander retina preparation. (a) Combined bright-field and fluorescence image showing the Lucifer yellow-filled patch pipette and the position of the Lucifer yellow-injected cell within the ganglion cell layer with primary dendrites projecting into the inner plexiform layer. (b) Under fluorescence excitation only the morphology of the Lucifer yellow-filled somata and the dendrites within the inner plexiform layer are more distinct.

capacitance around 50 pf were recorded. Calcium currents were activated at higher test potentials in these cells and are therefore not included in the data analysis. Since TTX was not routinely included in the superfusion solution, the inward current recorded included a transient and a sustained component (Fig. 2). The transient current was TTX-sensitive (Fig. 2a), consistent with current in sodium channels. The sustained current was completely blocked by Cd2+ (100 lM, Fig. 2b), suggesting that this was Ca channel current. To study Ca channel current, all measurements were made near the end of each test voltage step, well after the time when sodium channels are inactivated. The Ca channel current of

Effects of adenosine on voltage-dependent Ca2+ currents The effects of adenosine on Ca channel currents were determined using the same single-pulse protocol as in Fig. 2 (from )70 to )20 mV). Figure 3 shows a representative Ca channel current (Fig. 3a) and the time course of Ca channel current inhibition by 10 lM adenosine (Fig. 3b). Application of adenosine reversibly inhibited evoked Ca channel current with a rapid onset and a rapid recovery. After exposure to adenosine, Ca channel current amplitude decreased in less than 30 s. The current returned to control levels within 1 min after withdrawal of the drug. As shown in Fig. 3(b), the response of Ca channel current returned to the control level completely, often following a brief rebound. Application of adenosine at concentrations of 0.1–100 lM caused a dose-dependent inhibition of Ca channel current, with significant inhibition saturating at 10 lM (26.4 ± 2.8%, n ¼ 6, p < 0.01; Fig. 3c). Adenosine (1–100 lM) produced modulation of Ca channel current in all ganglion cells tested, suggesting that adenosine receptors are common to most, if not all, tiger salamander ganglion cells. Voltage-dependent inhibition by adenosine To study the voltage-dependence of adenosine effects, Ca channel current was elicited from a holding potential of ) 70 mV to depolarized test potentials ranging from ) 60 to + 50 mV for 50 ms. Figure 4(a) shows the averaged current– voltage relationships obtained from five neurons. The inhibition produced by adenosine was associated with a small positive shift of the Ca channel activation curve (Fig. 4b). Fitting of the activation curves, which were calculated from the averaged current–voltage relations in Fig. 4(a), with the Boltzmann function, showed that the voltage for half activation shifted to more positive potentials in the presence of adenosine by 2.9 mV (from ) 21.0 to ) 18.1 mV). Figure 4(c) shows summary data from the voltage dependence of the inhibition of Ca channel current by



Fig. 3 Adenosine reduces Ba2+ currents through Ca channels in retinal ganglion cells. (a) Superimposed Ca channel current traces in response to voltage steps to ) 20 mV from a holding potential of )70 mV before (control), during and after (wash) application of 10 lM

adenosine (AD). (b) The time course of 10 lM adenosine (AD)-induced inhibition of Ca channel current. (c) A dose–response curve of Ca channel current inhibition induced by adenosine (AD). Data are averaged from three to six experiments.

Fig. 4 Voltage-dependence of adenosine inhibition on Ca channel current. (a) Summary of the current-voltage relationship of Ca channel current before and after application of 10 lM adenosine (AD). Calcium channel current was activated by voltage steps between ) 60 mV and

+ 50 mV from a holding potential of ) 70 mV; (b) activation curves constructed from the mean currents shown in (a). (c) Voltagedependence of 10 lM adenosine-induced inhibition on Ca channel current. Data are averaged from five experiments. d, Control; s, AD.

adenosine (10 lM). Adenosine inhibited Ca channel current at voltages between ) 30 and + 20 mV with a peak inhibition (22.8 ± 4.9%, n ¼ 5) at ) 20 mV.

reduce adenosine-induced inhibition (19.7 ± 2.7% of control, n ¼ 3; p < 0.05; Figs 6b and c). We further compared the ability of CHA, an A1 receptor agonist, and CPCA, an A2 receptor agonist, to block Ca channel current in ganglion cells. Consistent with the data obtained using adenosine antagonists, CHA (1 lM) mimicked the effects of adenosine on Ca channel current with a significant inhibition of Ca channel current (23.8 ± 4.1%, n ¼ 4; p < 0.01; Fig. 6c), whereas CPCA at 1 lM failed to significantly inhibit Ca channel current (n ¼ 5; Fig. 6c). These results suggest that adenosine induces inhibition of Ca channel current through an A1 receptor. In two of the five cells studied, CPCA was found to enhance Ca channel current.

Effects of dialysis with GTP-c-s on adenosine-induced inhibition of Ca channel current In order to determine if the effect of adenosine on Ca channel current was also mediated through a G protein, the nonhydrolyzable analog of GTP, GTP-c-s, was included in the patch pipette. With 0.2 mM GTP-c-s in the patch-pipette solution, adenosine elicited a significant inhibition of Ca channel current (30.2 ± 10.5%, n ¼ 3, p < 0.05; Fig. 5). In addition, GTP-c-s rendered the effects of adenosine irreversible (40.7 ± 15.2% of control, n ¼ 3, p < 0.05; Fig. 5), suggesting that a G protein was involved in the effects of adenosine. Pharmacological analysis of adenosine receptor subtypes The selective adenosine receptor antagonists DPCPX (A1 receptors) and DMPX (A2 receptors) were used to classify the adenosine receptor involved in the inhibition of Ca channel current in tiger salamander ganglion cells. As shown in Fig. 6, DPCPX (1 lM) completely abolished the adenosine-induced inhibition of Ca channel current with no significant difference from control () 4.7 ± 8.3% of control, n ¼ 8; Figs 6a and c). In contrast, DMPX (10 lM) failed to

Effects of blockers of Ca channels on inhibition by adenosine Tiger salamander retinal ganglion cells have been previously shown to express pharmacologically distinct types of Ca channels, including N-type and L-type channels (Zhang et al. 1997; Shen and Slaughter 1998; Hirooka et al. 2000a,b). To detect which types of voltage-dependent Ca channels contributed to the component of Ca channel current that was inhibited by adenosine, we used the selective Ca channel blockers x-conotoxin-GVIA (N-type) and nifedipine (L-type). As shown in Fig. 7, nifedipine alone (5 lM) blocked Ca channel current by about one-third


554 X. Sun et al.

Fig. 5 Effect of dialysis with GTP-c-s on adenosine inhibition of Ca channel current. (a) Superimposed Ca channel current traces before (control), during addition of 10 lM adenosine (AD), and after (wash) in a cell dialyzed with the non-hydrolysable GTP analog GTP-c-s (200 lM). (b) The time course of 10 lM adenosine (AD)-induced inhibition of Ca channel current in the same cell. (c) Group data of

adenosine (AD)-induced inhibition of Ca channel current with and without GTP-c-s. Ca channel current was measured during a voltage step to ) 20 mV from a holding potential of ) 70 mV. Data are averaged from three to six experiments. *p < 0.05, **p < 0.01 compared to control current.

Fig. 6 Pharmacological characterization of the receptor subtype involved in adenosine-induced Ca channel current inhibition. (a,b) Representative experiments illustrating the effects of adenosine antagonists on Ca channel current. The selective A1 antagonist DPCPX (1 lM) but not the A2 antagonist DMPX (10 lM) blocked the adenosine (10 lM, AD)-induced inhibition of Ca channel current.

(c) Summary of the effects of adenosine alone and in combination with the adenosine receptor antagonists DPCPX and DMPX, and the agonists CHA and CPCA on Ca channel current. Calcium channel current was measured as described in Fig. 5. Average data are shown from three to eight experiments. *p < 0.05, **p < 0.01 compared to control current.

Fig. 7 Analysis of Ca channel subtypes contributing to adenosineinduced Ca channel current inhibition. (A) Representative experiment illustrating the effect of 10 lM adenosine (AD) on Ca channel current after application of the L-type Ca channel blocker nifedipine (5 lM). (b) Superimposed traces showing the effect of 10 lM adenosine (AD) on

Ca channel current after addition of the N-type Ca channel blocker x-conotoxin GVIA (1 lM). (c) Summary of effects of Ca channel blockers and 10 lM adenosine (AD) on Ca channel current. Calcium channel current was measured as in Fig. 5. Data are averaged from three to four experiments. *p < 0.05, **p < 0.01 compared to control current.

(29.8 ± 9.1%, n ¼ 3, Figs 7a and c). Subsequent addition of 10 lM adenosine caused a further 30.1 ± 6.2% reduction in the remaining Ca channel which was significant compared to nifedipine alone (p < 0.05; Fig. 7). This inhibition was similar to the 26% inhibition induced by adenosine alone (Fig. 3), suggesting that adenosine did not affect L-type Ca channel currents. Application of x-conotoxin GVIA (1 lM)

reduced currents by an average of 46.5 ± 2.9% (n ¼ 4, Figs 7b and c). Subsequent application of 10 lM adenosine produced only a weak and insignificant inhibition of the remaining Ca channel current (10.1 ± 4.6% of control, Fig. 7). These findings suggest that the inhibitory effects of adenosine on Ca channel current are due mainly to an effect on N-type Ca channels.



Discussion

Our results have identified both the receptor subtype and the Ca channel subtype involved in the actions of adenosine on tiger salamander ganglion cells. Our pharmacological results suggest that the inhibitory actions of adenosine are mediated through A1 receptors because the effects of adenosine were completely blocked by the A1 antagonist DPCPX and the A1 agonist CHA mimicked the effects of adenosine. The A2 antagonist DMPX did not block the inhibition produced by adenosine, and the A2 agonist was either without effect or, in a few cases, enhanced Ca channel current. Consistent with the G protein-mediated action of adenosine via A1 receptors, the inhibitory effect of adenosine on Ca channel current became irreversible in the presence of GTP-c-s. The reduction of Ca channel currents showed voltage sensitivity, an effect apparently derived from a positive shift in the activation curve of ganglion cell Ca channels. The data also showed that much of the current modulated by adenosine was carried in N-type Ca channels. We did not identify the subtype of any other component of the modulated current, but the L-type Ca channel blocker nifedipine did not affect the inhibition produced by adenosine on Ca channel current. Thus adenosine does not modulate L-type Ca channel currents in tiger salamander ganglion cells. A mini-slice preparation for the study of retinal ganglion cells In vitro preparations are commonly used to study the effect of putative neurotransmitters/neuromodulators on ganglion cells. This requires enzymatic digestion and/or mechanical isolation and this has the potential to damage the cells and alter their environment. At the very least, normal synaptic connections are lost as a result of isolation. The classic tiger salamander retinal slice preparation, developed by Werblin (1978), solves some of these problems. However, the method requires a great deal of practice to achieve the fine positioning and rotation of 150-lm slices, sometimes under conditions of limited illumination if light-evoked responses are to be preserved. The filter to which such slices are attached may also interfere with pharmacological manipulations by, for example, limiting rapid drug washout. The minislice preparation we introduce in the present study has the same advantages of the classical slice preparation but is much simpler to handle; it may also be easier to apply to the retinas from other species than the classical technique. Adenosine-induced inhibition of N-type Ca channels via A1 receptors in ganglion cells Using selective agonists and antagonists, we have shown that adenosine acts on A1 receptors to inhibit primarily N-type Ca channels. These findings are in agreement with previous studies in several types of neurons (Gross et al. 1989; Zhu and Ikeda 1993; Mynlieff and Beam 1994; Umemiya and

Berger 1994), including goldfish retinal ganglion cells (Zhang and Schmidt 1999), where it was demonstrated that A1 adenosine receptors inhibit N-type Ca channels. In goldfish ganglion cells, Ca channels are mainly composed of the N-subtype (Bindokas and Ishida 1996), whereas it has previously been shown in tiger salamander that both N- and L-type Ca channels each contribute about one-third of the total ganglion cell calcium currents (Zhang et al. 1997; Shen and Slaughter 1998; Hirooka et al. 2000a,b). In the present study, we have shown that adenosine inhibits N-type channels but has no effect on L-type channels in tiger salamander ganglion cells. A2 receptors have been reported to enhance Ca channel current in other preparations (Mogul et al. 1993). In our study, an A2 receptor agonist enhanced Ca channel currents on some cells tested, suggesting A2 receptors may exist in some ganglion cells. Our preliminary assessment of apparent A2 receptor-mediated activation of Ca channel current in tiger salamander ganglion cells needs to be investigated further. Physiological implication of the inhibition of Ca channels by adenosine in retina The synaptic terminals of ganglion cells contact neurons in the brain and release glutamate in a calcium-dependent manner. Therefore, the modulation of ganglion cell Ca channels by adenosine suggests the potential for modulation of synaptic transmission from retina to brain should adenosine modulate the same type of Ca channels in the retinal ganglion cell terminals as in the somata. Indeed, Zhang and Schmidt (1998, 1999) demonstrated A1 receptor-mediated pre-synaptic inhibition at retinotectal synapses in goldfish. This is consistent with the localization of A1 receptors on the terminals of ganglion cells in mammals (Goodman et al. 1983; Braas et al. 1987). Retinal ganglion cells may also be the source of adenosine that acts on these receptors (Braas et al. 1987; Blazynski et al. 1989), suggesting that adenosine may be a co-transmitter released from ganglion cell terminals and, by modulating calcium currents on the terminals of ganglion cell, control the release of glutamate. A1 adenosine receptors have been identified within the mammalian retina and were localized to the optic nerve fiber layer, the ganglion cell layer and the inner plexiform layer (Braas et al. 1987; Blazynski et al. 1989). This localization is consistent with the production of A1 receptors within ganglion cells and subsequent transport down ganglion cell axons to terminals within the brain. However, such localization, in particular the localization of A1 receptors within the inner plexiform layer, leaves open the possibility that adenosine could modulate ganglion cell calcium currents within the retina. Modulation of calcium currents on the dendrites of ganglion cells could affect a number of calciumdependent processes within ganglion cells such as the calcium-dependent potassium current and action potential


556 X. Sun et al.

generation (Lukasiewicz and Werblin 1988; Fohlmeister and Miller 1997). In the mammalian retina it has been suggested that ÔstarburstÕ amacrine cells contain and release ATP that, if converted to adenosine by ecto-nucleotidases, could act on ganglion cell A1 receptors (Perez et al. 1986). Tiger salamander retinas possess an amacrine cell type similar to the mammalian ÔstarburstÕ cells (Deng et al. 2001) but it is not known if these cells are purinergic. Neuroprotective role of adenosine in the retina The potential neuroprotective role of adenosine within the CNS has now been well established (for review see Sweeney 1997; Dunwiddle and Masino 2001). Levels of extracellular adenosine are elevated during a variety of conditions that can lead to tissue damage (i.e. hypoxia, ischemia, hypoglycemia, excessive activity, etc.) due either to intracellular or extracellular conversion of ATP to adenosine. The neuroprotective effects of adenosine are mediated primarily by A1 receptors. A1 receptor-mediated inhibition of calcium influx into ganglion cell could represent part of the neuroprotective mechanism given the established link between neuronal damage and excessive or prolonged increases of intracellular calcium ion concentration. Acknowledgements This work was supported by grants from the Canadian Institutes of Health Research and by a grant from the E. A. Baker Foundation.

References Bindokas V. P. and Ishida A. T. (1996) Conotoxin-sensitive and conotoxin-resistant Ca2+ currents in fish retinal ganglion cells. J. Neurobiol. 29, 429–444. Blazynski C., Mosinger J. L. and Cohen A. I. (1989) Comparison of adenosine uptake and endogenous adenosine-containing cells in mammalian retina. Vision Neurosci. 2, 109–116. Braas K. M., Zarbin M. A. and Snyder S. H. (1987) Endogenous adenosine and adenosine receptors localized to ganglion cells of the retina. Proc. Natl. Acad. Sci. USA 84, 3906–3910. Deng P., Cuenca N., Doerr T., Pow D. V., Miller R. and Kolb H. (2001) Localization of neurotransmitters and calcium binding proteins to neurons of salamander and mudpuppy retinas. Vision Res. 41, 1771–1783. Dunwiddie T. V. and Masino S. A. (2001) The role and regulation of adenosine in the central nervous system. Annu. Rev. Neurosci. 24, 31–55. Fohlmeister J. F. and Miller R. F. (1997) Impulse encoding mechanisms of ganglion cells in the tiger salamander retina. J. Neurophysiol. 78, 1935–1947. Fredholm B. B. and Dunwiddie T. V. (1988) How does adenosine inhibit transmitter release? Trends Pharmacol. Sci. 9, 130–134. Goodman R. R., Kuhar M. J., Hester L. and Snyder S. H. (1983) Adenosine receptors: autoradiographic evidence for their

location on axon terminals of excitatory neurons. Science 220, 967–969. Gross R. A., Macdonald R. L. and Ryan-Jastrow T. (1989) 2-Chloroadenosine reduces the N calcium current of cultured mouse sensory neurones in a pertussis toxin-sensitive manner. J. Physiol. 411, 585–595. Hille B. (2001) Ion Channels of Excitable Membranes. Sinauer, Sunderland, MA. Hirooka K., Kelly M. E., Baldridge W. H. and Barnes S. (2000a) Suppressive actions of betaxolol on ionic currents in retinal ganglion cells may explain its neuroprotective effects. Exp. Eye Res. 70, 611–621. Hirooka K., Kourennyi D. E. and Barnes S. (2000b) Calcium channel activation facilitated by nitric oxide in retinal ganglion cells. J. Neurophysiol. 83, 198–206. Lukasiewicz P. D. and Werblin F. S. (1988) A slowly inactivating potassium current truncates spike activity in ganglion cells of the tiger salamander retina. J. Neurosci. 8, 4470–4481. Lukasiewicz P. D. and Werblin F. S. (1990) The spatial distribution of excitatory and inhibitory inputs to ganglion cell dendrites in the tiger salamander retina. J. Neurosci. 10, 210–221. Mogul D. J., Adams M. E. and Fox A. P. (1993) Differential activation of adenosine receptors decreases N-type but potentiates P-type Ca2+ current in hippocampal CA3 neurons. Neuron 10, 327–334. Mynlieff M. and Beam K. G. (1994) Adenosine acting at an A1 receptor decreases N-type calcium current in mouse motoneurons. J. Neurosci. 14, 3628–3634. Perez M. T., Ehinger B. E., Lindstrom K. and Fredholm B. B. (1986) Release of endogenous and radioactive purines from the rabbit retina. Brain Res. 398, 106–112. Shen W. and Slaughter M. M. (1998) Metabotropic and ionotropic glutamate receptors regulate calcium channel currents in salamander retinal ganglion cells. J. Physiol. 510, 815–828. Sweeney M. I. (1997) Neuroprotective effects of adenosine in cerebral ischemia: window of opportunity. Neurosci. Biobehav. Rev. 21, 207–217. Umemiya M. and Berger A. J. (1994) Activation of adenosine A1 and A2 receptors differentially modulates calcium channels and glycinergic synaptic transmission in rat brainstem. Neuron 13, 1439– 1446. Werblin F. S. (1978) Transmission along and between rods in the tiger salamander retina. J. Physiol. 280, 449–470. Wu L. G. and Saggau P. (1997) Presynaptic inhibition of elicited neurotransmitter release. Trends Neurosci. 20, 204–212. Yawo H. and Chuhma N. (1993) Preferential inhibition of omegaconotoxin-sensitive presynaptic Ca2+ channels by adenosine autoreceptors. Nature 365, 256–258. Zhang C. and Schmidt J. T. (1998) Adenosine A1 receptors mediate retinotectal presynaptic inhibition: uncoupling by C-kinase and role in LTP during regeneration. J. Neurophysiol. 79, 501–510. Zhang C. and Schmidt J. T. (1999) Adenosine A1 and class II metabotropic glutamate receptors mediate shared presynaptic inhibition of retinotectal transmission. J. Neurophysiol. 82, 2947–2955. Zhang J., Shen W. and Slaughter M. M. (1997) Two metabotropic gamma-aminobutyric acid receptors differentially modulate calcium currents in retinal ganglion cells. J. Gen. Physiol. 110, 45–58. Zhu Y. and Ikeda S. R. (1993) Adenosine modulates voltage-gated Ca2+ channels in adult rat sympathetic neurons. J. Neurophysiol. 70, 610–620.
