INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Adenosine is a potent neuromodulatory substance that is widely distributed throughout the CNS. Adenosine (or P1) receptors are divided into four known molecular and pharmacological subtypes: A₁, A_2A, A_2B, and A₃ receptors (Ralevic and Burnstock 1998). Adenosine is well characterized as an inhibitory transmitter in the CNS (Ribeiro 1995), but excitatory actions of adenosine have also been observed (Sebastiao and Ribeiro 1996). Inhibitory actions of adenosine are commonly mediated by the A₁ receptor that couples negatively to adenylyl cyclase via a G_i protein to reduce cAMP levels in the cell (Dolphin et al. 1986; Ralevic and Burnstock 1998; van Calker et al. 1979). A₂ receptors have been considered to mediate excitatory effects in the brain because they couple positively to adenylyl cyclase through a G_s protein to increase cAMP levels (Ralevic and Burnstock 1998; Sebastiao and Ribeiro 1996; van Calker et al. 1979). However, synaptic activation of A₂ receptors is not always associated with an increase in I_Ca or transmitter release. A₂ receptor activation enhances P-type but not other Ca²⁺ channel subtypes in neurons (Satoh et al. 1997; Umemiya and Berger 1994). Adenosine acting at A₂ receptors has been shown to inhibit both N- and L-type Ca²⁺ channels in PC12 cells via stimulation of cAMP and protein kinase A (PKA) (Kobayashi et al. 1998; Park et al. 1998). A₂ receptor activation also inhibits I_Ca, and thus GABA release in suprachiasmatic and arcuate nuclei neurons (Chen and van den Pol 1997). It has thus been suggested that the increased excitability associated with A₂ receptor activation arises from inhibition of I_Ca and the resulting decrease in release of inhibitory transmitter (reviewed in Edwards and Robertson 1999).

Adenosine is released in darkness from the retina (Blazynski and Perez 1991; Paes de Carvalho et al. 1990; Perez et al. 1986). Adenosine is present in human photoreceptors (Braas et al. 1987), and [³H]adenosine is taken up into photoreceptors and horizontal cells via transporters for adenosine (Paes de Carvalho et al. 1990; Studholme and Yazulla 1997), suggesting that adenosine is removed from the extracellular space following release. The presence of A₂ receptors in the outer retina and particularly photoreceptors has been demonstrated by autoradiography and in situ hybridization (Blazynski 1990; Kvanta et al. 1997). Adenosine acting on A₂ receptors stimulates melatonin synthesis (Valenciano et al. 1998) and myoid elongation in cone photoreceptors (Rey and Burnside 1999), providing evidence for a physiological role for A₂ receptors on photoreceptors.

Rod photoreceptors use dihydropyridine (DHP)-sensitive L-type Ca²⁺ channels to control release of the neurotransmitter, L-glutamate (Ayoub and Copenhagen 1991; Copenhagen and Jahr 1989; Corey et al. 1984; Kourennyi and Barnes 2000; Schmitz and Witkovsky 1997). Photoreceptor I_Ca can be modulated by various ions (H⁺, cations, anions) (Barnes et al. 1993; Piccolino et al. 1996; Thoreson et al. 1997, 2000) and neurotransmitters (dopamine, somatostatin, and nitric oxide) (Akopian et al. 2000; Kurenny et al. 1994; Stella and Thoreson 2000). Barnes and Hille (1989) found that adenosine inhibits the Ca²⁺-activated chloride current in cones, but the actions of adenosine on photoreceptor I_Ca have not previously been examined.

The purpose of the present study was to examine the effect of adenosine on I_Ca and Ca²⁺ influx in rod photoreceptors. Driven by evidence that adenosine inhibits I_Ca in many other preparations, adenosine is released from retinal neurons in the dark, A₂ receptors are located on photoreceptors, and an uptake system for adenosine is present at the photoreceptor synapse, we hypothesized that adenosine acting on A₂ receptors may inhibit I_Ca and thus decrease Ca²⁺ influx into rods, which would in turn reduce transmitter release. Whole cell perforated patch-clamp recordings and Ca²⁺-imaging experiments with fura-2 were performed on rods in retinal slice and isolated cell preparations. Our results indicate that adenosine inhibits both I_Ca and the depolarization-evoked Ca²⁺ influx in a dose-dependent manner and that this inhibition is mediated by A₂ receptor activation coupled to a PKA pathway. To our knowledge, this is the first report of inhibition of presynaptic I_Ca by A₂ receptors at a glutamatergic synapse. Abstracts describing some of these results have previously been presented (Stella and Thoreson 1999; Stella et al. 1999)

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Tissue preparation

Larval tiger salamanders (Ambystoma tigrinum, Kons, Germantown, WI or Charles Sullivan, TN; 7-10 in) were cared for according to institutional guidelines. Retinal slices were prepared according to the methods of Werblin (1978) and Wu (1987); exact procedures are described in further detail by Thoreson et al. (1997). Briefly, salamanders were killed by decapitation and pithed, the eyes were enucleated, and the anterior portion of an eye including the lens was removed. The resulting eyecup was cut into sections and placed vitreal side down onto a piece of filter paper (Millipore 2 × 5 mm, Type GS, 0.2 µM pores). After the retina adhered to the filter paper, the sclera, choroid, and retinal pigment epithelium were removed under chilled amphibian superfusate. The isolated retina was then cut into 100- to 150-µM slices using a razor-blade tissue chopper (Stoelting). Retinal slices were rotated 90° for viewing of the retinal layers when placed under a water immersion objective (×40, 0.7 NA or ×60, 1.0 NA) and viewed on an upright fixed stage microscope (Olympus BHWI or Nikon E600FN). All procedures were performed under dim light or under infrared illumination using Gen III image intensifiers (Nitemate NavIII, Litton Industries).

Solitary retinal neurons were prepared by isolating retina from the eyecup and finely mincing it using half of a double-edged razor blade. The minced retina was then gently triturated in amphibian superfusate with a large-bore fire-polished Pasteur pipette. Isolated cells were plated on slides coated with a salamander-specific antibody, Sal-1 (kindly provided by Peter MacLeish). Prior to beginning an experiment, cells were allowed to adhere to the slides for 15 min at 4°C.

Solutions and perfusion

Solutions were applied by a single-pass, gravity-fed perfusion system that delivered medium to the slice chamber (chamber volume: ~0.5 ml) at a rate of 1.0 ml/min. The normal amphibian superfusate that bathed the slices contained (in mM) 111 NaCl, 2.5 KCl, 1.8 CaCl₂, 0.5 MgCl₂, 10 N-2-hydroxyethylpiperazine-N' 2-ethanesulfonic acid (HEPES), and 5 glucose. After obtaining a photoreceptor recording, the superfusate was switched to a Ba²⁺ solution to enhance Ca²⁺ currents. The Ba²⁺ solution contained (in mM) 99 NaCl, 2.5 KCl, 10 BaCl₂, 0.5 MgCl₂, 10 HEPES, 5 glucose, 0.1 picrotoxin, and 0.1 niflumic acid. The pH of all solutions was adjusted to 7.8 with NaOH. The osmolarity measured with a vapor pressure osmometer (Wescor) was 242 ± 5 mOsm. Solutions were continuously bubbled with 100% O₂.

Electrophysiology

Patch pipettes were pulled on a Narashige PB-7 vertical puller from borosilicate glass pipettes (1.2 mm OD, 0.95 mm ID, omega dot) and had tips of ~1 µm OD with tip resistances of 10-15 M.

Pipettes were filled with a solution containing (in mM) 54 CsCl, 61.5 CsCH₃SO₄, 3.5 NaCH₃SO₄, and 10 HEPES. The pH was adjusted to 7.2 with CsOH and the osmolarity was adjusted, if necessary, to 242 ± 5 mOsm. To maintain endogenous second-messenger signaling pathways and avoid the rundown that accompanies conventional whole cell recording of I_Ca, we used the perforated patch method of whole cell recording with the pore forming antibiotic, nystatin (Rae et al. 1991). Nystatin was dissolved in dimethylsulfoxide (DMSO) at a concentration of 120 mg/ml and then diluted into the pipette electrolyte solution to achieve a final concentration of 480 µg/ml. The final working solution was vortexed vigorously for 20-30 s and stored in the refrigerator. Fresh antibiotic solutions were made every 3 h. In successful recordings, seals >1 G were obtained in 30 s and cells were usually fully perforated within 5 min of sealing.

Rods were identified by their long, rod-shaped outer segments. For electrophysiology, electrode placement was performed under infrared illumination using either Gen III image intensifiers mounted over the microscope eyepieces or with an infrared-sensitive video camera mounted on the trinocular head of the microscope.

The input resistance (R_in) of rods in the slice averaged 659.3 ± 48.9 (SE) M (n = 22). R_in of isolated rods averaged 2.4 ± 0.3 G (n = 5). Access resistance, clamp speed, and membrane capacitance were measured by analyzing capacitive transients evoked by hyperpolarizing steps from a holding potential of 70 mV. The time constant of the capacitative transient indicated a voltage-clamp speed of 0.97 ± 0.07 ms and a membrane capacitance for rods in the slice of 36.5 ± 3.7 pF (n = 22). For isolated rods, the clamp speed and capacitance determined from the capacitative transient averaged 2.4 ± 0.42 ms and 13.5 ± 1.8 pF (n = 5), respectively. Access resistance was typically between 22 and 30 M; recordings were considered acceptable only when the access resistance was <40 M. Estimates of the voltage errors introduced by the access resistance are included in the legends of figures showing examples of I_Ca.

Once whole cell access was achieved, light responses were obtained to determine the spectral sensitivity of the cell. Light-evoked responses in rods were generated by light from a tungsten light source that passed through a filter wheel containing interference filters of four different wavelengths (380, 480, 580, and 680 nm). The light stimulus was reflected into the microscope condenser using a beam splitter. Light intensity was controlled by neutral density filters (Wratten gel). The intensity of unattenuated light measurement with a laser power meter (Metrologic) was ~1.1 × 10⁶ photon s¹ µm² at 680 nm, 1.3 × 10⁶ photon s¹ µm² at 580 nm, and 2.1 × 10⁵ photon s¹ µm² at 480 nm (the power meter was not accurate at 380 nm). Generally, a flash of 1-s duration was used. The protocol for identifying the spectral sensitivity of a single photoreceptor involved successive flashes of increasing intensity (usually encompassing 2 log units of intensity) at the four wavelengths.

After superfusion with the 10 mM BaCl₂ solution was started, voltage ramps (0.5 mV/ms) from 90 to +60 mV were used to assess I_Ca every 30 s. Voltage ramps yield a similar current-voltage profile as a step protocol but cause less rundown or inactivation of the current and provide more data points for fitting for analysis (Stella and Thoreson 2000). Drug solutions were applied after I_Ca appeared stable for 90 s.

Cells were voltage clamped at 70 mV using an Axopatch 200B patch-clamp amplifier (Axon Instruments, Foster City, CA). Currents were acquired and analyzed using Clampex 7.0 software (Axon Instruments). The leak conductance was assumed to be ohmic and equal to the minimum conductance between 75 and 55 mV and then digitally subtracted. Leak subtraction by blocking I_Ca with 0.1 mM Cd²⁺ yields almost identical voltage profiles.

Measurement of [Ca²+]_i transients

Intracellular Ca²⁺ changes were assessed using the ratiometric dye fura-2 (Grynkiewicz et al. 1985). Retinal slices were incubated for 45 min in the dark with 0.5 ml of 10 µM fura-2/AM +0.02% pluronic F-127 (Molecular Probes, Eugene, OR) in a slice chamber at 4°C. This was followed by an additional incubation in fura-2/AM alone for 1.5 h.

Digital fluorescent images were recorded with a cooled CCD camera (SensiCam, Cooke) on an upright fixed stage microscope (Nikon E600FN) equipped with a ×60 (1.0 NA) water-immersion objective. A 150-W Xe bulb (Opti-Quip) was mounted on a Sutter Lambda 10-2 filter wheel with 340- and 380-nm interference filters and coupled to the microscope by a liquid light guide (Sutter). The 380-nm intensity was attenuated with a 0.5-neutral density filter to balance the intensity of emissions evoked by 340 and 380 nm. The fluorescence emitted by the cells on stimulation with 340- or 380-nm light was filtered through a 510 ± 20-nm band-pass emission filter. Axon Imaging Workbench (AIW 2.2) was used to control the camera, filter wheel, and image acquisition. Pixel binning (2 × 2) of the images was used to decrease acquisition rate (acquisition time: 0.5-1 s). Images were subtracted for background camera noise but no averaging or masking was performed.

To activate voltage-dependent Ca²⁺ channels, cells were depolarized by increasing [K⁺]_o from 2.5 to 50 mM for 1 min. Elevated KCl applications were performed at 15 min intervals to reduce any Ca²⁺-dependent inactivation. Images were acquired at 5- to 10-s intervals during elevated [K⁺]_o applications. All experiments were performed at room temperature. For analysis, a region of interest was drawn over the rod inner segment and a change in the fluorescence ratio (340 nm/380 nm) was interpreted as reflecting changes in [Ca²⁺]_i. The ratio change produced by the elevated [K⁺]_o application in the test solution was compared with the ratio changed produced in control conditions. Ratio values for control conditions were determined by averaging the prior control and subsequent wash responses for each drug application. Multiple cells could be analyzed in each preparation; each experiment was performed on at least three different slice preparations.

Statistical analysis was performed using paired and unpaired Student's t-test (GraphPad Prism 3.0). Significance was chosen as P < 0.05, and variance is reported as ±SE.

Drug solutions

Adenosine, ATP, (R)-N⁶-(1-methyl-2-phenylethyl)adenosine (R-PIA), N⁶-[2-(3,5-dimethoxyphenyl)-2-(2-methylphenyl)-ethyl]adenosine (DPMA), N⁶-2-(4-aminophenyl)ethyladenosine (APNEA), and Rp-adenosine 3',5'-cyclic monophosphothioate triethylamine (Rp-cAMPS) were obtained from Sigma Chemical (St. Louis, MO). Solutions containing ()Bay K 8644 (Research Biochemicals International) or nisoldipine (Zeneca Pharmaceuticals) were prepared by diluting 10,000× DMSO stock solutions into the superfusate. R-PIA, DPMA, and APNEA stocks were dissolved in DMSO as 50 mM stock solutions. All other stock solutions (1,000×) were prepared in distilled water or superfusate. Superfusion with 0.1% DMSO alone did not affect any of the properties of I_Ca- or Ca²⁺-imaging responses that we studied.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Characterization of L-type calcium currents in rod photoreceptors

We will refer to the inward current evoked by a depolarizing voltage ramp in rods in the presence of 10 mM Ba²⁺ as I_Ca. In the presence of 10 mM Ba²⁺, a slowly developing U-shaped current-voltage relationship appeared above 50 mV that peaked between 25 and 10 mV and then diminished at more positive potentials. The current typically reversed between +50 and +60 mV. The peak amplitude of I_Ca in rods averaged 371.7 ± 31.7 pA (n = 23) and the voltage at which the current was half-maximal (V₅₀) averaged 31.0 ± 1.6 mV (n = 23).

Figure 1 illustrates an example of the effects of the DHP agonist, Bay K 8644 (1 µM), and the DHP antagonist, nisoldipine (5 µM), on I_Ca recorded in a rod. In agreement with other studies on photoreceptors, I_Ca was enhanced by Bay K 8644 and incompletely blocked by a high concentration of nisoldipine (Kourennyi and Barnes 2000; Wilkinson and Barnes 1996). In addition to suppressing the peak current, nisoldipine produced a positive shift in the current/voltage relationship for I_Ca similar to that shown by Wilkinson and Barnes for cone I_Ca (Wilkinson and Barnes 1996). Three millivolts of the rightward shift shown in Fig. 1 can be accounted for by a decrease in I_Ca flowing across the access resistance. The remainder of this positive shift may arise from the more potent block of L-type I_Ca by nisoldipine at more negative potentials (Albillos et al. 1994). The residual I_Ca in photoreceptors recorded in the presence of nisoldipine has been shown to arise largely from unblocked L-type Ca²⁺ channels and not from a second channel type (Kourennyi and Barnes 2000; Wilkinson and Barnes 1996).

View larger version (17K): [in this window] [in a new window]   Fig. 1. Rod photoreceptors contain L-type ICa. A rod photoreceptor in a retinal slice was held at 70 mV and ramped from 90 to +60 mV at 0.5 mV/ms. The graph shows overlaid current-voltage relationships of voltage ramps in control superfusate (thin trace) followed by perfusion with Bay K 8644 (1 µM, thick trace) and then nisoldipine (5 µM, dashed trace). Bay 8644 (1 µM) enhanced the peak ICa without shifting the I-V relationship along the voltage axis. Nisoldipine (5 µM) suppressed ICa and shifted the I-V relationship to more positive potentials. Voltage error at the peak of the current due to uncompensated access resistance was estimated to be 5.4 mV for control, 8 mV for Bay K 8644, and 2.4 mV for nisoldipine.

Adenosine inhibits rod I_Ca

The effect of adenosine on rod I_Ca is shown in Fig. 2. Figure 2A illustrates current-voltage profiles of I_Ca in control conditions (Fig. 2B, 1) and in the presence of adenosine (Fig. 2B, 2). Figure 2B plots the amplitude of I_Ca measured with voltage ramps every 30 s. As illustrated in Fig. 2A, adenosine (50 µM) inhibited I_Ca in rods without shifting the current-voltage relationship along the voltage axis. The mean reduction in I_Ca by 50 µM adenosine was 22.8 ± 3.3% (n = 9, P = 0.0001). All of the rods tested at this concentration displayed a similar degree of inhibition in the presence of adenosine. Typically, inhibition of the rod I_Ca began within 30 to 60 s, the time required for complete solution exchange in the recording chamber. Full inhibition was observed after 2 min, and currents recovered after ~5 min washout of adenosine (Fig. 2B). As the concentration of adenosine was increased from 1 to 50 µM, a significant inhibition of rod I_Ca was seen at concentrations 10 µM, implying that there is a concentration-dependent effect of adenosine on rod I_Ca (Fig. 3).

View larger version (18K): [in this window] [in a new window]   Fig. 2. Adenosine (50 µM) inhibited rod ICa. A: the graph shows I-V relationships of voltage ramps obtained in control superfusate (thin trace) and in the presence of adenosine (50 µM, thick trace). Voltage error at the peak of the current due to uncompensated access resistance was estimated to be 14 mV in control and 11.5 mV with adenosine. B: time course of adenosine-induced changes in the amplitude of ICa in a rod. Voltage ramps in A were obtained at time points 1 (control) and 2 (adenosine). The voltage was ramped from 90 to +60 mV at 0.5 mV/ms.

View larger version (11K): [in this window] [in a new window]   Fig. 3. Effect of adenosine on rod ICa as a function of concentration. Data values were obtained from peak amplitude measurements of ICa from voltage ramps. Amplitudes of the currents obtained in test solutions were normalized to currents obtained in the control solution (Itest/Icontrol). Asterisks denote significant changes with respect to control (P < 0.05). Each point represents the mean ± SE. Number of experiments is shown in parentheses. *, P < 0.05 compared with control. Adenosine concentrations: 1 µM: 11.3 ± 3.8%, n = 3, P = 0.098; 10 µM: 13.5 ± 5.4%, n = 4, P = 0.089; 25 µM: 20.6 ± 3.7%, n = 5, P = 0.005; 50 µM: 22.8 ± 3.3%, n = 9, P = 0.001.

To test whether an intermediary cell type might be involved in adenosine modulation of rod I_Ca, solitary rods were mechanically isolated by gentle trituration and tested with 50 µM adenosine. As in retinal slices, adenosine significantly inhibited I_Ca in isolated rods by 19.0 ± 5.5% (n = 5, P = 0.026), suggesting that adenosine acts directly on rod photoreceptors to modulate I_Ca.

Adenosine reduces depolarization-evoked Ca²+influx in rods

Ca²⁺-imaging experiments were performed on retinal slices to test whether adenosine reduced depolarization-evoked [Ca²⁺]_i increases in rod photoreceptors. Slices were incubated with fura-2/AM. Rod were depolarized by elevating [K⁺]_o from 2.5 to 50 mM for 1 min. Figure 4A shows a brightfield Nomarski image of rod photoreceptors in the retinal slice. Corresponding paired pseudocolor images of fluorescence from the same cells are displayed in Fig. 4B (340, 380, and 340/380 nm), showing cells in both normal (2.5 mM) and elevated (50 mM) [K⁺]_o amphibian superfusate. Figure 4C shows a plot of the [Ca²⁺]_i response measured in a rod photoreceptor before, during, and after the application of adenosine (50 µM). Adenosine produced a reversible inhibition of the depolarization-evoked [Ca²⁺]_i increase. On average, adenosine (50 µM) caused a 25.9 ± 3.4% (n = 14, P < 0.0001) reduction in the K⁺-evoked 340/380 ratio change in rod photoreceptors in the slice. Significant inhibition of the K⁺-evoked [Ca²⁺]_i increase was seen at concentrations as low as 1 µM (4.9 ± 1.4%, n = 11, P = 0.005), and inhibition increased in a dose-dependent fashion with concentrations 50 µM (Fig. 5).

View larger version (41K): [in this window] [in a new window]   Fig. 4. A: a brightfield Nomarski image of rod photoreceptors in the retinal slice of a tiger salamander retinal slice. B: corresponding paired pseudocolor images of fluorescence from rods are shown in A to illustrate the quality of fura-2 dye loading and the resolution of images obtained in the slice preparation. Images of fura-2 fluorescence emissions are displayed in normal (2.5 mM) and elevated (50 mM) [K+]o amphibian superfusate for 340, 380, and the 340 nm/380 nm ratio. C: effects of adenosine on depolarization-evoked [Ca2+]i increases in photoreceptor inner segments produced by a 1-min application of elevated (50 mM) [K+]o.

View larger version (13K): [in this window] [in a new window]   Fig. 5. Concentration-dependent inhibition of the K+-evoked [Ca2+]i increase in rods from retinal slice preparations by adenosine. Fura-2-loaded slices were stimulated with elevated [K+]o in the presence of different concentrations of adenosine (1-50 µM). Inhibition of K+-evoked ratio change produced by adenosine were normalized to the average of the initial control and the subsequent recovery following the test solution. Each point represents the mean ± SE. Number of experiments is shown in parentheses. *, P < 0.05 compared with control. Adenosine concentrations: 1 µM: 4.9 ± 1.3%, n = 11, P = 0.005; 10 µM: 14.1 ± 3.2%, n = 8, P = 0.003; 25 µM: 12.3 ± 3.1%, n = 12, P = 0.002; 50 µM: 25.9 ± 3.4%, n = 14, P < 0.0001.

A₂ receptor agonists inhibit rod I_Ca and Ca²+ influx

Adenosine can interact with P2 (purinergic) receptors (Ralevic and Burnstock 1998) that have been shown to modulate Ca²⁺ channels in various neurons (Brown et al. 2000; Dave and Mogul 1996). To test whether a P2 receptor may be involved in the modulation of voltage-gated Ca²⁺ channels in rods, ATP was bath applied to rod photoreceptors in the retinal slice preparation and examined with perforated patch recording of rod I_Ca and [Ca²⁺]_i imaging using fura 2. ATP did not significantly inhibit I_Ca (50 µM ATP: +7.1 ± 11.3%, n = 7, P = 0.549) or the K⁺-evoked Ca²⁺ increase in rod photoreceptors (75 µM ATP: +11.9 ± 7.3%, n = 15, P = 0.1296), suggesting that P2 receptors do not mediate the observed inhibition by adenosine.

The absence of an ATP effect suggests that adenosine acts at an adenosine receptor. There are four major subtypes of adenosine receptors: A₁, A_2A, A_2B, and A₃. R-PIA is an A₁-selective receptor agonist, DPMA is an A₂-selective receptor agonist that interacts with both A_2A and A_2B receptors, and APNEA is an A₃-selective receptor agonist (Bridges et al. 1988; Ralevic and Burnstock 1998). To characterize the adenosine receptor subtype that mediates inhibition of I_Ca, we tested the effects of these three agonists on the rod I_Ca. As shown in Fig. 6B, only the A₂-selective agonist, DPMA (2 µM), significantly inhibited the rod I_Ca. Like adenosine, DPMA did not cause a significant shift in the current-voltage relationship along the voltage axis (Fig. 6B, mean shift in V₅₀ = +0.8 ± 0.63 mV, n = 6, P = 0.1563). The A₁-selective agonist, R-PIA (2 µM), and A₃-selective agonist, APNEA (2 µM), did not significantly inhibit rod I_Ca (Fig. 6, A and C). Adenosine receptor agonists were tested at 2 or 10 µM. Figure 6D illustrates the overall results with R-PIA (2 µM), DPMA (2 µM), and APNEA (10 µM) on rod I_Ca. DPMA at concentrations of 2 and 10 µM inhibited I_Ca in rods by 15.1 ± 4.7% (n = 11, P < 0.0001) and 25.0 ± 5.8% (n = 3, P = 0.05), respectively. However, neither R-PIA nor APNEA had any effect on the rod I_Ca at any of the concentrations tested (R-PIA, 2 µM: 0.5 ± 4.7%, n = 6, P = 0.92; APNEA, 2 µM: 0.0 ± 3.5%, n = 5, P = 1.0; APNEA, 10 µM: +2.4 ± 5.8%, n = 7, P = 0.6982). These results suggest that the effects of adenosine on rod I_Ca are mediated by A₂ receptors.

View larger version (15K): [in this window] [in a new window]   Fig. 6. Effect of adenosine receptor agonists on rod photoreceptor ICa. A: the A1-selective adenosine receptor agonist, (R)-N6-(1-methyl-2-phenylethyl)adenosine (R-PIA, 2 µM), did not inhibit rod ICa. Voltage error at the peak of the current due to uncompensated access resistance was estimated to be 9 mV in both records. B: the A2-selective agonist, N6-[2-(3,5-dimethoxyphenyl)-ethyl]adenosine) (DPMA, 2 µM), inhibited rod ICa. Voltage error at the peak of the current due to uncompensated access resistance was estimated to be 2.6 mV in control and 2.1 mV in DPMA. C: the A3-selective adenosine receptor agonist, N6-2-(4-aminophenyl)ethyladenosine (APNEA, 10 µM), did not inhibit rod ICa. Voltage error at the peak of the current due to uncompensated access resistance was estimated to be 4 mV in both records. Each graph shows I-V relationships of voltage ramps obtained in control superfusate () and test solution (- - -). D: bar graph comparing effects of R-PIA, DPMA, and APNEA on ICa in rods. Amplitude of the currents obtained in test solutions were normalized to currents obtained in the control solution (Itest/Icontrol). *, significant changes with respect to control (P < 0.05). R-PIA, 0.5 ± 4.7%, n = 6, P = 0.92; DPMA, 15.1 ± 4.7%, n = 11, P < 0.0001; APNEA, +2.4 ± 5.8%, n = 7, P = 0.69.

Using Ca²⁺ imaging techniques, we tested the effects of the same three adenosine receptor agonists on depolarization-evoked [Ca²⁺]_i increases. Figure 7A displays responses to a series of K⁺-evoked depolarizations measured in the inner segment of a rod photoreceptor in the presence of R-PIA (10 µM), DPMA (10 µM), and APNEA (10 µM). Only the A₂-selective agonist, DPMA, inhibited the depolarization-evoked [Ca²⁺]_i increase in rods. Figure 7B shows the overall results of adenosine receptor agonists on the depolarization-induced [Ca²⁺]_i increases in rods. In agreement with the effects of these agonists on rod I_Ca, DPMA (10 µM) reduced the K⁺-evoked [Ca²⁺]_i increase in rods by 30.9 ± 4.9% (n = 12, P < 0.0001), while R-PIA (10 µM) and APNEA (10 µM) did not significantly alter the K⁺-evoked [Ca²⁺]_i increase (R-PIA, +3.0 ± 3.3%, n = 12, P = 0.355; APNEA, 4.4 ± 2.7%, n = 12, P = 0.129).

View larger version (18K): [in this window] [in a new window]   Fig. 7. Effect of adenosine receptor agonists on [Ca2+]i increases evoked by elevated [K+]o in rods. A: a graphic representation of successive applications of adenosine receptor agonists. Depolarization-evoked [Ca2+]i changes in photoreceptor inner segments produced by 1-min applications of elevated (50 mM) [K+]o were measured in the presence of the A1 receptor agonist, R-PIA (10 µM), the A2 receptor agonist, DPMA (10 µM), and the A3 receptor agonist, APNEA (10 µM). B: bar graphs comparing the effects of R-PIA, DPMA, and APNEA on [Ca2+]i responses in rods. K+-evoked [Ca2+]i increases in test solutions were normalized to the average of the initial control and subsequent recovery. Each bar graph represents the mean ± SE. *, P < 0.05 compared with control. R-PIA: +3.0 ± 3.3%, n = 12, P = 0.36; DPMA: 30.9 ± 4.9%, n = 12, P < 0.0001; APNEA: 4.4 ± 2.7%, n = 12, P = 0.13.

A₂ receptors inhibit rod I_Ca and Ca²+ influx through a PKA-dependent pathway

The primary signaling mechanism by which A₂ receptors transduce their signals intracellularly is to couple positively to adenylyl cyclase and thus stimulate the production of cAMP, which in turn activates PKA to phosphorylate various proteins that can include Ca²⁺ channels subunits (Alexander et al. 1994; De Jongh et al. 1996; Gubitz et al. 1996; Mogul et al. 1993). Similar to the effects of the A₂ receptor agonist DPMA, stimulation of cAMP production with forskolin or activation of PKA with Sp-cAMPS inhibits I_Ca in rods (Stella and Thoreson 2000). We therefore tested whether Rp-cAMPS, a cell-permeant analogue of cAMP known to inhibit PKA activity (Chik et al. 1997; Dolphin 1995), could block the effects of DPMA. Rp-cAMPS (10 µM) was applied to retinal slices prior to application of DPMA (2 µM). As illustrated in Fig. 8A, DPMA did not inhibit rod I_Ca while in the presence of Rp-cAMPS. Figure 8B shows current-voltage profiles of rod I_Ca obtained in control conditions (Fig. 8A, 1), in the presence of Rp-cAMPS (Fig. 8A, 2), and in the presence of both Rp-cAMPS and DPMA (Fig. 8A, 3). The bar graph in Fig. 8C shows that I_Ca recorded in the presence of DPMA was significantly larger when Rp-cAMPS (10 µM) was also present. The increase in I_Ca above the baseline in the presence of Rp-cAMPS and DPMA may reflect ability of Rp-cAMPS to enhance I_Ca in rods (Stella and Thoreson 2000).

View larger version (16K): [in this window] [in a new window]   Fig. 8. Rp-cAMPS (10 µM) prevented inhibition of ICa by DPMA (10 µM) in rods. A: time course of changes in ICa amplitude during application of Rp-cAMPS and Rp-cAMPS plus DPMA. B: current-voltage relationships of voltage ramps obtained at time points 1 (control), 2 (Rp-cAMPS), and 3 (Rp-cAMPS + DPMA) from A. Control: thick solid trace. Rp-cAMPS (10 µM): thin solid trace. DPMA (2 µM) + Rp-cAMPS (10 µM): thin trace. Voltage error at the peak of the current due to uncompensated access resistance was estimated to be 3 mV in all 3 records. C: bar graph summarizing relative changes in ICa amplitude induced by Rp-cAMPS or Rp-cAMPS plus DPMA. Each bar represents the mean ± SE. Asterisk, P < 0.05 compared with control. *DPMA significantly inhibited ICa (15.1 ± 4.7%, n = 11, P < 0.0001) but not in the presence of Rp-cAMPS (DPMA + Rp-cAMPS ICa: +12.2 ± 6.8%, n = 4, P = 0.14). DPMA produced significantly less inhibition in the presence of Rp-cAMPS (DPMA, n = 11 vs. DPMA + Rp-cAMP, unpaired t-test, P = 0.0001).

Figure 9 shows effects of Rp-cAMPS on [Ca²⁺]_i increases produced by application of elevated [K⁺]_o on rod photoreceptors in the slice. Similar to rod I_Ca, application of Rp-cAMPS (10 µM) prevented DPMA from inhibiting the K⁺-evoked Ca²⁺ increase (Fig. 9A, +3.6 ± 10.8% n = 8, P = 0.7499). To confirm the efficacy of DPMA, DPMA (2 µM) was reapplied after washout of Rp-cAMPS and found to inhibit the K⁺-evoked Ca²⁺ increase (Fig. 9A, 36.1 ± 3.9%, n = 8, P < 0.0001). In agreement with the I_Ca results of Fig. 8, Rp-cAMPS (10 µM), significantly reduced the ability of DPMA (2 µM) to inhibit the depolarization-evoked [Ca²⁺]_i increase (Fig. 9B, DPMA (2 µM) versus DPMA (2 µM) + Rp-cAMPS (10 µM): paired t-test, n = 8, P = 0.0212). The results of these experiments suggest that stimulation of PKA activity is primarily responsible for A₂ receptor modulation of voltage-dependent L-type Ca²⁺ channels in rods.

View larger version (19K): [in this window] [in a new window]   Fig. 9. Rp-cAMPS (10 µM) prevented the inhibition of depolarization-evoked [Ca2+]i increases by DPMA (2 µM) in rod photoreceptors. A: effects of successive elevated [K+]o applications in control superfusate and in the presence of Rp-cAMPS and DPMA. DPMA reversibly inhibited depolarization-evoked [Ca2+]i responses in the same cell following ~20 min washout of Rp-cAMPS + DPMA. B: bar graph summarizing relative changes in depolarization-evoked [Ca2+]i responses induced by Rp-cAMPS (10 µM) and Rp-cAMPS (10 µM) plus DPMA (2 µM) in experiments similar to those illustrated in A. *DPMA (2 µM) significantly inhibited the K+-evoked Ca increase in rods after washout of Rp-cAMPS + DPMA (36.1 ± 3.9%, n = 8, P < 0.0001). DPMA (2 µM) produced significantly less inhibition in the presence of Rp-cAMPS (10 µM; paired t-test, n = 8, P = 0.02).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study indicates that activation of A₂ receptors by adenosine stimulates PKA activity, which in turn inhibits voltage-gated L-type Ca²⁺ channels in rod photoreceptors. Similar concentrations of adenosine also produced a comparable inhibition of depolarization-evoked [Ca²⁺]_i increases. Adenosine released in the retina in darkness (Blazynski and Perez 1991; Paes de Carvalho et al. 1990; Perez et al. 1986) may therefore inhibit I_Ca and reduce Ca²⁺ entry into rods that would likely inhibit their release of glutamate.

Adenosine receptor pharmacology and signaling pathways

Purine and pyrimidine receptors are divided into two large families, P1 (or adenosine) receptors and P2 (or purinergic) receptors (for review, Ralevic and Burnstock 1998). ATP (50-75 µM) had no effect on the I_Ca or the depolarization-evoked [Ca²⁺]_i increase in rods, indicating that P2 receptors are unlikely to contribute to the inhibition produced by adenosine. Adenosine receptor subtypes were classified originally by their effects on adenylyl cyclase activity: A₁ receptors inhibit and A₂ receptors stimulate adenylyl cylase (Londos et al. 1980; van Calker et al. 1979). However, the classification has since been broadened to include a new A₃ receptor subtype (Zhou et al. 1992) and two subtypes of A₂ receptors, A_2A and A_2B (Furlong et al. 1992; Maenhaut et al. 1990; Pierce et al. 1992). To discriminate among A₁, A₂, and A₃ receptor subtypes, we used selective agonists for each receptor subtype: R-PIA for A₁ receptors, DPMA for A₂ receptors, and APNEA for A₃ receptors. The ability of the A₂-selective agonist, DPMA, but not the A₁ or A₃ receptor agonists to inhibit both the depolarization-evoked Ca²⁺ increase and I_Ca in rods (Figs. 6 and 7) is consistent with binding studies that have localized A₂ receptors to photoreceptor inner and outer segments (McIntosh and Blazynski 1994). We did not attempt to pharmacologically discriminate between A_2A and A_2B receptors. However, mRNA for the A_2A receptor has been shown to be expressed in the outer nuclear layer and ganglion cell layer, whereas A_2B receptor mRNA is absent from the retina (Kvanta et al. 1997), suggesting that A_2A receptors are likely to be responsible for the effects observed in the present study. In a study on cone motility in teleosts, the rank order of potency for adenosine agonists was found to be consistent with A₂-like receptors, but antagonist potencies did not correspond precisely to either A_2A or A_2B receptors (Rey and Burnside 1999).

A₂ receptor activation elevates cAMP levels in whole retina (Blazynski et al. 1986; Paes de Carvalho and de Mello 1982). A₂ receptor-mediated inhibition of Ca²⁺ influx and I_Ca in rods is similar to the inhibition of rod I_Ca produced by stimulation of cAMP production or activation of PKA (Stella and Thoreson 2000) and inhibition of PKA with Rp-cAMPS blocked the inhibitory effect of DPMA in rods (Figs. 8 and 9). Taken together, these results indicate that A₂ receptor-mediated changes in Ca²⁺ influx from depolarization of I_Ca in rods are likely established by stimulation of PKA.

Sources of retinal adenosine

Adenosine has been detected in photoreceptors of human, monkey, guinea pig (Braas et al. 1987), fish (Ehinger and Perez 1984), and chick retinas (Paes de Carvalho et al. 1992). Adenosine has also been localized to ganglion cells (Ehinger and Perez 1984) and cells in the inner nuclear layer (Ehinger and Perez 1984; Paes de Carvalho et al. 1992) including rod horizontal cells (Studholme and Yazulla 1997) and amacrine cells (Blazynski 1989). Adenosine is released tonically from retinas in the dark (Blazynski and Perez 1991; Paes de Carvalho et al. 1990; Perez et al. 1986) and likely results from intracellular turnover of ATP in photoreceptor inner segments. In the dark, photoreceptors are depolarized by cations (e.g., Na⁺, Ca²⁺) entering through cyclic nucleotide channels in the outer segment (Shimazaki and Oakley 1986; Torre 1982). Increased intracellular adenosine levels are generated by the turnover of ATP from a highly active Na⁺/K⁺-ATPase in the inner segment that counters the cation influx in the outer segment. The increased ATP turnover and breakdown of adenine nucleotides elevates intracellular adenosine levels. Increased adenosine levels stimulate the efflux of adenosine down its concentration gradient via symmetrical adenosine transporters (Thorn and Jarvis 1996) which appear to be present on photoreceptors (Ehinger and Perez 1984).

ATP can be stored and released from vesicles (e.g., Corcoran et al. 1986; Jo and Schlichter 1999; Santos et al. 1999; von Kugelgen et al. 1994) and extracellular breakdown of ATP can be an important source of extracellular adenosine (Cunha et al. 1998). It appears unlikely that a significant amount of extracellular ATP is converted to adenosine at the photoreceptor synapse in the retinal slice preparation because application of 50-75 µM ATP did not inhibit I_Ca or the depolarization-evoked Ca²⁺ response. However, the finding that soluble nucleotidases are co-released with ATP from sympathetic neurons suggests that they may be co-localized in synaptic vesicles (Todorov et al. 1997), and this raises the possibility that ATP might be broken down by nucleotidases in synaptic vesicles to ADP, AMP, or adenosine prior to release. Consistent with such a possibility, Blazynski and Perez (1991) have shown that K⁺-evoked depolarization stimulates the release of adenosine in the retina.

Physiological significance

Previous studies have shown Ca²⁺ channels can be inhibited by A₂ receptors resulting in inhibition of transmitter release. For example, activation of A₂ receptors from cultured suprachiasmatic and arcuate nuclei neurons inhibit presynaptic I_Ca and GABA release (Chen and van den Pol 1997). A_2A activation also reduces the frequency of spontaneous and miniature inhibitory postsynaptic currents by 30% in striatal medium spiny neurons as a result of reduced GABA release (Mori et al. 1996). The present results show that A₂ receptor activation can also inhibit presynaptic I_Ca in glutamatergic neurons. Such a mechanism might account for the decreased release of L-glutamate evoked by depolarization following activation of A_2A receptors in rat striatum and cultured chick retinal neurons (Golembiowska and Zylewska 1997; Rego et al. 2000).

The results of the present study show that adenosine acting on A₂ receptors in rods can regulate L-type I_Ca and Ca²⁺ influx in rods. Thus the changing levels of adenosine in the retina that accompany changing levels of illumination could serve as an autocrine or paracrine signal to regulate synaptic output from rods.