|Home | About | Journals | Submit | Contact Us | Français|
The A3 adenosine receptor is emerging as an important regulator of neuronal signaling, and in some situations receptor stimulation can limit excitability. As the NMDA receptor frequently contributes to neuronal excitability, this study examined whether A3 receptor activation could alter the calcium rise accompanying NMDA receptor stimulation. Calcium levels were determined from fura-2 imaging of isolated rat retinal ganglion cells as these neurons possess both receptor types. Brief application of glutamate or NMDA led to repeatable and reversible elevations of intracellular calcium. The A3 agonist Cl-IB-MECA reduced the response to both glutamate and NMDA. While adenosine mimicked the effect of Cl-IB-MECA, the A3 receptor antagonist MRS 1191 impeded the block by adenosine, implicating a role for the A3 receptor in the response to the natural agonist. The A1 receptor antagonist DPCPX provided additional inhibition, implying a contribution from both A1 and A3 adenosine receptors. The novel A3 agonist MRS 3558 (1’S,2’R,3’S,4’R,5’S)-4-(2-chloro-6-(3-chlorobenzylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0] hexane-1-carboxamide and mixed A1/A3 agonist MRS 3630 (1’S,2’R,3’S,4’R,5’S)-4-(2-chloro-6-(cyclopentylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo [3.1.0]hexane-1-carboxamide also inhibited the calcium rise induced by NMDA. Low levels of MRS 3558 were particularly effective, with an IC50 of 400 pM. In all cases, A3 receptor stimulation inhibited only 30-50% of the calcium rise. In summary, stimulation of the A3 adenosine receptor by either endogenous or synthesized agonists can limit the calcium rise accompanying NMDA receptor activation. It remains to be determined if partial block of the calcium rise by A3 agonists can modify downstream responses to NMDA receptor stimulation.
Adenosine acts as a transmitter throughout the nervous system, stimulating each of four receptors, termed A1, A2a, A2b and A3 (Fredholm et al., 2001a). While roles for the A1 and A2a receptors are well established, and that of the A2b receptor is emerging, the contribution of A3 receptors to neural signaling remains only partially elucidated. Both the restricted distribution of A3 receptors and the delayed availability of specific agonists and antagonists have hindered investigation (Rivkees et al., 2000; Jacobson & Knutsen, 2001).
In spite of these restrictions, evidence is accumulationg that the A3 receptor exerts an inhibitory influence on at least some neuronal systems. Adenosine acting at cerebral A3 receptors is neuroprotective in stroke models (Shen et al., 2005), while A3 agonists can limit ischemic damage (Von Lubitz et al., 1999). Moreover, A3 receptor knockout mice are more susceptible to neurodegeneration after hypoxic challenge (Fedorova et al., 2003). Stimulation of the A3 receptor prevents apoptosis of cultured cortical neurons exposed to extended chemical hypoxia in vitro, and repeated injections of A3 agonists confer prolonged protection from the effects of middle cerebral artery ligation (Chen et al., 2006). A3 receptors contribute to the inhibition of cortical synaptic potentials during hypoxia (Hentschel et al., 2003) and inhibit synaptic transmission in the enteric nervous system (Wunderlich et al., 2008). A3 receptors have also been shown to protect retinal ganglion cells following the Ca2+ elevation and cell death accompanying stimulation of the P2X7 receptor (Zhang et al., 2006c). The neuronal actions of the A3 receptor are varied, and receptor stimulation can also damage neurons under some circumstances (Pugliese et al., 2003), with the timing and concentration of drugs likely to determine the overall effect (Abbracchio & Cattabeni, 1999; Von Lubitz et al., 2001; Pugliese et al., 2007). However, the evidence above suggests that A3 receptors limit neuronal excitation in at least some situations.
Glutamate is the primary excitatory transmitter in the nervous system, mediating its effects through ionotropic and metabotropic receptors. The NMDA, AMPA, and kainate receptors comprise the ionotropic group, with activation leading to channel opening and the influx of sodium and calcium (Dingledine et al., 1999). The corresponding depolarization can open voltage-dependent ca2+ channels, leading to a secondary pathway of Ca2+ influx. While the kainate and AMPA receptors inactivate rapidly, NMDA receptors inactivate much more slowly; the sustained elevation in Ca2+ levels accompanying NMDA receptor stimulation ensures that the receptor makes a major contribution to neuronal excitability. Modulating the rise in Ca2+ that accompanies NMDA receptor activation is a key pathway to control neuronal excitability and thus a likely target for the inhibitory actions of A3 receptors. NMDA-dependent synaptic transmission is depressed by A3 receptor stimulation (Rubaj et al., 2003), suggesting the receptors can indeed modulate excitatory glutaminergic signaling. However, it is not presently known whether the mechanism underlying this inhibition involves attenuation of the Ca2+ rise.
This study specifically asked whether stimulation of the A3 adenosine receptor could limit the Ca2+ rise that accompanies NMDA stimulation in neurons. Isolated rat retinal ganglion cells were chosen as the assay system because they have been shown to express A3 receptors on a molecular (Zhang et al., 2006a) and functional (Zhang et al., 2006c) level, in addition to possessing well characterized NMDA receptors (Sucher et al., 1991; Sucher et al., 2003; Hartwick et al., 2004; Goebel & Winkler, 2006; Gustafson et al., 2007). Freshly dissociated ganglion cells can be isolated from other retinal cells using a two-step immunopanning procedure, ensuring that neurochemical interactions are not confounded by interference from glial cells (Barres et al., 1988; Hartwick et al., 2004). This preparation is also amendable to Ca2+ measurements using the ratiometric dye fura-2, providing a quantifiable measurement of intracellular levels. Isolated ganglion cells were thus used to examine the effect of A3 adenosine receptors on the Ca2+ rise triggered by NMDA receptors.
Preliminary versions of some of the data have been presented in abstract form (Zhang et al., 2006b).
Purification of ganglion cells using the immunopanning procedure was performed as described previously (Zhang et al., 2005) based upon standard protocols (Barres et al., 1988; Hartwick et al., 2004). Retinas were removed from rat pups on post-natal day (PD) 5-11 from untimed pregnant Long-Evan rats (Jackson Laboratory Inc., Bar Harbor, ME). All animals were treated in accordance with University of Pennsylvania IACUC approved protocols and the National Institutes of Health guide for the care and use of laboratory animals. All efforts were made to minimize animal suffering and to reduce the number of animals used.
Retinas were enzymatically dissociated for 30 min with 15 U/mL papain, 0.2 mg/mL DL-cysteine and 0.004% DNAse I (Worthington/Cooper, Lakewood, NJ). The tissue was washed, centrifuged, resuspended and filtered. Cells were incubated with rabbit antimacrophage antibody (1:75, Accurate Chemical, Westbury, NY), then incubated in a 100 mm dish coated with goat anti-rabbit IgG antibody (1:400, Jackson ImmunoResearch Laboratories Inc, West Grove, PA). Non-adherent cells were removed to a second petri-dish coated with goat anti-mouse IgM (1:300, Jackson ImmunoResearch Laboratories Inc) and anti-Thy 1.1 antibody (from hybridoma T11D7e2; American Type Culture Collection, Rockville, MD). After 30 min, non-adherent cells were washed off and adherent ganglion cells were incubated with 0.125% trypsin for 8 min at 37°C. Digestion was stopped with fetal bovine serum (30%) in Neurobasal medium and cells were centrifuged and plated on coverslips coated with poly-L-lysine and laminin. The basic growth medium contained Neurobasal medium with 2 mM glutamine, 100 μg/ml gentamicin, 0.025ml/ml B27 supplement (all Invitrogen Inc., Carlsbad, CA), 0.7% methylcellulose (Stemcell Technologies Inc., Vancouver, BC, Canada) and 2.5% rat serum (Cocalico Biologicals Inc., Reamstown, PA). Retinal ganglion cells were incubated at 37°C with 5% CO2. Quantification with labeled cells indicated that the immunopurified preparation contained at least 98% ganglion cells.
Unlabeled RGCs from isolated ganglion cell preparations attached to coverslips for 24 hrs were loaded with 10 μM fura-2 and 2% Pluronic F-127 (Invitrogen, Inc.) for 60-90 min at room temperature, rinsed and maintained in fura-2-free solution for 30 min before data acquisition began. The coverslips were mounted on a Nikon Diaphot inverted microscope and visualized with a 40x objective. To obtain Ca2+ measurements, the field was alternatively excited at 340nm and 380nm with a scanning monochromator and the light emitted at >520 nm from a region of interest surrounding individual retinal ganglion cells was imaged with a CCD camera and analyzed (all Photon Technologies International, Inc., Lawrenceville, N.J.). Cells were perfused with a control solution at the start of Ca2+ imaging experiments containing (in mM) 105 NaCl, 5 KCl, 4 NaHepes, 6 Hepes acid, 5 NaHCO3, 1.3 CaCl2, 5 glucose, 60 mannitol, pH 7.4. Drugs were dissolved in the control solution. Calibration was performed separately on each cell after the experiment by perfusing cells in the presence of 5 μM ionomycin and control solution (with 1.3 mM Ca2+) followed by ionomycin in the base solution without Ca2+ and with the addition of 5 mM EGTA (pH 8.0). The 340/380 ratio was converted to Ca2+ concentration as previously described (Zhang et al., 2005). In some experiments, data are expressed as this ratio as the large number of solutions precluded conversion. Ca2+ measurements were performed at room temperature.
MRS 3558 (1’S,2’R,3’S,4’R,5’S)-4-(2-chloro-6-(3-chlorobenzylamino)-9H-purin-9-yl)-2,3-dihydroxy-N-methylbicyclo[3.1.0]hexane-1-carboxamide and MRS 3630 (1’S,2’R,3’S,4’R,5’S)-4-(2-chloro-6-(cyclopentylamino)-9H-purin-9-yl)-2,3-dihydro xy-N-methylbicyclo[3.1.0]hexane-1-carboxamide were developed at the Laboratory of Bioorganic Chemistry NIDDK, National Institutes of Health (Tchilibon et al., 2005; Jacobson et al., 2005). All other materials were obtained from Sigma Chemical Corp, (St. Louis, MO) unless otherwise indicated. Data are presented as mean ± standard error of the mean. Significance was generally evaluated using an appropriate Student’s t test for comparison between two variables or a one-way ANOVA with post-hoc test for more than two variables using SigmaStat 3.5 (Systat Software, Inc., San Jose, CA), with p<0.05 taken as a significant difference. “n” is defined as the number of individual ganglion cells from which measurements were made. The Ca2+ response was defined as the difference between peak and baseline for each application of agonist. Block by MRS 3558 was characterized by fitting data to an exponential decay with three parameters using a least-squared analysis with Sigmaplot 9 (Systat Software, Inc).
Initial experiments tested whether the relatively specific adenosine A3 receptor agonist 2-chloro-N6-(3-iodobenzyl)-adenosine-5’-N-methyluronimide (Cl-IB-MECA) could reduce the calcium elevation induced by NMDA in rat retinal ganglion cells. Neurons loaded with fura-2 were exposed to recurring 15-second applications of 10 μM NMDA followed by a washout. As NMDA receptors require glycine as a co-agonist (Danysz & Parsons, 1998), 10 μM glycine was always included in the mixture. All experiments were done in the absence of Mg2+ to limit the Mg2+-dependent block of the NMDA receptor (Nowak et al., 1984). This brief exposure to 100 μM NMDA + 10 μM glycine led to clear elevations in calcium levels that promptly returned to baseline upon the removal of agonist (Fig. 1A). The absolute increase in Ca2+ triggered by this level of NMDA was 658 ± 124 nM (n=44). The elevations were recurring upon reapplication of NMDA, with the magnitude of the Ca2+ elevation relatively consistent throughout multiple applications. Application of agents to alternate applications thus provided an ideal way to test for inhibition within a single cell, consistent with Hartwick et al., (2004).
To determine whether A3 receptor stimulation could modulate the response to NMDA, the A3 agonist Cl-IB-MECA was added to the bath and the response to NMDA was determined. Cl-IB-MECA led to a clear reduction in the calcium rise triggered by NMDA (Fig. 1B). The block by Cl-IB-MECA was reversible and repeatable within a given cell and led to a significant reduction of the mean calcium response (Fig. 1C).
While glutamate is the endogenous agonist at NMDA receptors, it can also stimulate other metabotropic and ionotropic receptors. To determine if the A3 receptor also attenuated the response to glutamate, isolated ganglion cells were exposed to brief, recurrent applications of 10 μM glutamate + 10 μM glycine. The responses were very similar to those induced by NMDA; Ca2+ levels rose sharply, and then returned to baseline promptly after agonist removal (Fig. 2A). The mean rise in calcium induced by glutamate was 612 ± 105 nM (n=24). This increase was not significantly different from that induced by NMDA, consistent with reports of the relative potency of the two agonists at NMDA receptors (Moriyoshi et al., 1991).
Stimulation of the A3 receptor with Cl-IB-MECA also inhibited the Ca2+ rise evoked by glutamate (Fig. 2B.). This inhibition was readily reversible, with glutamate producing a robust response upon removal of Cl-IB-MECA. Overall, Cl-IB-MECA led to a significant reduction in the Ca2+ elevation induced by glutamate (Fig. 2C).
Rat retinal ganglion cells express multiple adenosine receptors (Kvanta et al., 1997), and the A1 receptor is known to attenuate the response to NMDA (Hartwick et al., 2004). As adenosine is the endogenous agonist at all 4 receptor subtypes, it was important to determine whether the A3 receptor contributed to the block by adenosine itself.
Adenosine (10 μm) attenuated the calcium elevation induced by glutamate. This concentration of adenosine is high enough to stimulate all four adenosine receptors, although the actual concentration reaching the membrane may have been reduced by adenosine deaminase or transporters (Fredholm et al., 2001b). To evaluate the contribution of the A3 receptor to the block by adenosine, the A3 antagonist MRS 1191 was added. The Ca2+ response to glutamate was substantially inhibited by adenosine. However 1 μM MRS 1191 significantly reduced the block produced by adenosine (Fig. 3A and B). This demonstrates that the A3 receptor contributes to the attenuation by adenosine.
To determine whether the residual block remaining in the presence of MRS 1191 and adenosine was due to adenosine acting at the A1 receptor, the effect of A1 antagonist 1,3-dipropyl-8-cyclopentylxanthine (DPCPX) was examined (Fig 4A). The mean Ca2+ rise induced by NMDA in the presence of both adenosine and MRS 1191 was 15% less than that produced by NMDA alone (Fig. 4B), not significantly different than the reduction of the glutamate rise (p=0.36). DPCPX eliminated this residual, consistent with the theory that both A1 and A3 receptors contribute to the inhibitory effects of adenosine on these neurons.
The ability of the A3 receptor to reduce the NMDA signal was further examined with the A3 agonist MRS 3558. MRS 3558 is 100-fold more selective for the A3 receptor over other adenosine receptors and is effective in the low nanomolar range (Tchilibon et al., 2005). While the compound has been shown to prevent ischemic damage in lung tissue (Matot et al., 2006) the effect of MRS 3558 in neurons has not been determined. At 1 nM, MRS 3558 produced a substantial inhibition of the response to NMDA (Fig. 5A-B). MRS 3558 also reduced the response to glutamate (not shown). Given this low effective level, the concentration-dependence of block by MRS 3558 was examined. Exposure of cells to 1 nM, 10 nM or 100 nM MRS 3558 produced a significant reduction in Ca2+ response. Even 100 pM MRS 3558 produced a small, albeit insignificant, reduction. The response to NMDA was blocked by MRS 3558 with an IC50 of 400 pM (Fig. 5C), although only 30% of the Ca2+ elevation was inhibited by saturating levels.
As the results above, and previous findings by others (Hartwick et al., 2004), suggest the A1 adenosine receptor can also limit the Ca2+ rise from NMDA in these cells, initial experiments were performed with the joint A1/A3 agonist MRS 3630. MRS 3630 has a radioligand binding Ki of 17.4 nM against rat A1 and 5.8 nM against rat A3 receptors respectively (Jacobson et al., 2005). MRS 3630 significantly inhibited the Ca2+ rise (Fig. 5D-E). Of note, MRS 3630 blocked only half the Ca2+ elevation even at 1 μM, suggesting the Ca2+ rise had both adenosine sensitive and adenosine insensitive components.
This study has demonstrated that stimulation of the A3 adenosine receptor limits the rise in Ca2+ that accompanies stimulation of the NMDA receptor in rat retinal ganglion cells. Several points support the identification of the A3 receptor in this role. The calcium elevation triggered by NMDA was inhibited by the agonist Cl-IB-MECA. Cl-IB-MECA is relatively specific for the A3 receptor, with a potency of 115/2,100/10,000/11 nM at A1/A2A/A2B and A3 receptors respectively (Freedholm et al., 2001a). While the block produced by 1 μM Cl-IB-MECA could have theoretically involved the A1 receptor, antagonism of the adenosine effect with MRS 1191 confirmed action of the A3 receptor, as MRS 1191 blocks the human A1/A2A and A3 receptors with a potency of 40,000/100,000 and 31 nM respectively (Jacobson et al., 1997). Although the compound is somewhat less effective at rat than at human A3 receptors, 1 μM still acts selectively at rat A3 receptors (Dunwiddie et al., 1997). As such, a block by 1 μM MRS 1191 strongly implicates the A3 receptor. Finally the specific A3 agonist MRS 3558 blocked the Ca2+ rise with an IC50 of 400 nM. MRS 3558 has an binding affinity for A1/A2A/A2B and A3 receptors of 260/2300/>10,000/0.29 nM respectively (Tchilibon et al., 2005), so an inhibition by 1, 10 and 100 nM confirms actions at the A3 over the A1 receptor. The similarity between our effective IC50 of 400 pM and the binding affinity support the correlation. Together, these findings imply that stimulation of the A3 adenosine receptor can reduce the Ca2+ influx accompanying NMDA receptor activation.
While the mechanisms underlying the reduction in Ca2+ have not been defined, some inferences can be made. Stimulation of the A3 receptor never blocked the Ca2+ rise completely. Typical inhibition was 30-50%, even with saturating concentrations of MRS 3558 and MRS 3630. Although there are numerous explanations for the partial block, precedent suggests the A3 receptor does not act directly on the NMDA channels. The depolarization that accompanies NMDA receptor activation can subsequently open voltage-dependent Ca2+ channels (Sucher et al., 1991; Alagarsamy & Johnson, 1995), enabling Ca2+ to enter through both NMDA channels and Ca2+ channels. Stimulation of the adenosine A1 receptor blocks voltage-gated Ca2+ channels by direct inhibition by the Gβγ subunit (Schulte & Fredholm, 2002; Dolphin, 2003), and stimulation of the A1 receptor in salamander retinal ganglion cells inhibited the Ca2+ currents (Sun et al., 2002). Stimulation of A3 receptors can also activate Gβγ subunit released from Gi/o proteins (Schulte and Fredholm, 2002), suggesting a block of voltage dependent Ca2+ channels is possible. However, A3 receptors were recently shown to stabilize GABAA receptors in the membrane (Roseti et al., 2008), and as ganglion cells express these receptors (Akopian et al., 1998), this alternative explanation cannot be ruled out.
The demonstration that A3 receptors can limit excitatory rises in Ca2+ does not negate the contribution of the A1 adenosine receptor in these or other neurons. A1 receptors have clearly been shown to block Ca2+ influx in retinal ganglion cells (Sun et al., 2002; Hartwick et al., 2004), and stimulation of A1 receptors by endogenously-produced adenosine hyperpolarizes ganglion cells (Newman, 2001, 2003). The ability of specific A3 antagonist MRS 1191 to limit the effect of adenosine, along with the additive limiting effect of A1 antagonist DPCPX in the present study, imply that both receptors contribute to the block by adenosine in retinal ganglion cells. Although the A3 was originally thought to be less sensitive to adenosine, more recent analysis suggests both A1 and A3 receptors are highly responsive, with EC50 values of 310 nM and 290 nM respectively (Fredholm et al., 2001b). The overall contribution of each receptor to ganglion cell function will of course depend upon their relative expression in the membrane and the location of this expression, with the latter impossible to estimate using this isolated cell preparation.
The primary conclusion of this study is that A3 adenosine receptors limit the Ca2+ rise accompanying NMDA receptor stimulation. As such, it provides a potential explanation for the growing list of modulatory effects of A3 receptors in the nervous system mentioned above. This block may be restricted to stimulation of NMDA receptors, as A3 receptor stimulation did not protect against kainate triggered death of cortical neurons (Rebola et al., 2005). The effects will also be regionally and temporally dependent, as suggested by the varied response to A3 agonists in different neuronal regions. Given the evidence for the neuroprotective actions of the A3 receptor (Von Lubitz et al., 1999) it is tempting to predict that A3 agonists may be of some benefit under conditions where neurons succumb to damage from extended stimulation of the NMDA receptor. Any advantage may depend of the ability to restrict these agonists to the appropriate targets. Whether stimulation of the A3 receptor can protect retinal ganglion cells from ischemic damage as flavonoids and nicotinamide do remains to be determined (Ji et al., 2008; Jung et al., 2008)
This work is supported by grants from the NIH EY015537 and EY013434 (CHM), Vision Research Core Grant EY001583 (CHM and AML), Research to Prevent Blindness (AML), the Paul and Evanina Bell Mackall Foundation Trust (AML), the Jody Sack Fund (HH, MZ and XZ), the National Natural Science Foundation of China 30872831(XZ), and by support from the Intramural Research Program of NIDDK, NIH, Bethesda, MD (KAJ). The authors would like to thank Mortimer Civan for useful discussions and William Baldridge for assistance with the ganglion cell panning.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.