Adenosinergic modulation of neuronal activity in the pond snail Lymnaea stagnalis

Adenosine acts at various parts of the central nervous system to modulate neuronal activity. Chemically classified as a nucleoside, it plays various different roles as an intercellular messenger in the brain, which expresses high concentrations of adenosine receptors and where its role has been implicated in both normal and pathological processes (for a review, see Buck, 2004; Downey et al., 2007; Pagonopoulou et al., 2006). Under conditions of increased demand and reduced supply of ATP, such as hypoxia and prolonged wakefulness, there is an increase in adenosine turnover and adenosine receptor stimulation (Lutz and Kabler, 1997; Lutz and Manuel, 1999; Roman et al., 2008). This is achieved by dramatically increasing adenosine concentrations by the breakdown of high-energy phosphates during episodes of increased nerve activity, hypoxia or ischemia (Schubert et al., 1997). Under these conditions, degradation or transport of adenosine is prevented by adenosinergic transmission-potentiating agents (adenosine deaminase and kinase inhibitors), and elevated adenosine levels offer protection against ischemic or excitotoxic neuronal damage (Wardas, 2002).

The important neuroprotective role of endogenous adenosine in the vertebrate CNS has been well characterized. Increases in extracellular adenosine concentration in response to anoxia have been observed in the striatum of the red-eared turtle, Trachemys scripta striatum, leading to the suggestion that the initial energetic shift during hypoxia (90–120min) is due to the effects of adenosine. Ultimately, the remarkable metabolic rate depression in this species that allows prolonged anoxic survival is achieved through release of inhibitory transmitters [e.g. GABA and glycine (Nilsson and Lutz, 1992)]. The defensive role of adenosine has also been implicated in the anoxia-induced increase in cerebral blood-flow that confers neuroprotection through increased delivery of glycolytic substrates and enhanced removal of acidic anaerobic end products from sensitive brain areas. For example, in the crucian carp, Carassius carassius, a 2.2-fold increase in brain blood-flow rate was observed during anoxia with a similar increase following perfusion of the brain with adenosine during normoxia (Nilsson et al., 1994). Although adenosine appears to have a definitive role in mediating anoxia tolerance through different pathways in turtles and the crucian carp, results are variable because it does not have a neuroprotective role during oxygen deprivation in the leopard frog, Rana pipens (Soderstrom-Laurtizen et al., 2001) or the shark Hemicyllium ocellatum (Soderstrom et al., 1999).

Despite the large number of anoxia-tolerant invertebrate species few studies have utilized invertebrate models to examine a possible neuromodulatory role of adenosine (Table 1). Adenosine is metabolized by many invertebrates (bivalves, gastropods, echinoderms, crustaceans, cephalopods and polychaetes) as indicated by the presence of adenylate-metabolizing enzymes in their tissues (Lazou, 1988). Adenosine was shown to inhibit the acetylcholine-induced depolarization of a specific neuron in the mollusk Helix aspersa (Cox and Walker, 1987). Furthermore, the marine invertebrate Sipunculus nudus depressed its metabolic rate by 70% during anoxia and neuronal tissue levels of adenosine increased during hypoxia, hypercapnia, and to an even greater degree during anoxic hypercapnia (Reipschlager et al., 1997). Of greater importance, adenosine applied during normoxic conditions depressed oxygen consumption for more than 90 min, whereas application of an adenosine antagonist increased oxygen consumption rate after 30 min of perfusion during hypercapnia (Reipschlager et al., 1997). The finding that adenosine had an excitatory effect on the spontaneous activity and responsiveness of certain interneurons to chemical and electrical stimuli in the spiny lobster Panulirus argus (Derby et al., 1987) suggests that invertebrates may have more variable responses.

Given the dearth of knowledge regarding CNS effects of adenosine in invertebrates, we sought to examine the neuromodulatory role of adenosine in the anoxia-tolerant freshwater mollusk Lymnaea stagnalis. L. stagnalis can tolerate anoxia for about 40 h in a N₂-bubbled environment at 20°C, with a mortality rate of 10% on day 7 of normoxic recovery (Wijsman et al., 1985). Recent studies have utilized this ability of L. stagnalis to examine hypoxia-induced modulation of different protein profiles (Fei and Feng, 2008; Fei et al., 2007). Extensive data on its anaerobic metabolism is also available (Wijsman et al., 1985) along with a detailed morphological and electrophysiological map of the identifiable neurons and neuronal clusters of its paired pedal ganglia (Kyriakides et al., 1989). We used whole-cell patch-clamp techniques and fluorescence microscopy to measure changes in neuronal electrical properties and intracellular calcium levels in response to adenosine receptor modulation (vertebrate A₁R subclass-like) using neurons from cluster F of the left or right pedal dorsal ganglion (LPeDG or RPeDG) within the central ring ganglia of L. stagnalis.

Laboratory-raised stocks of the freshwater snail L. stagnalis L. were maintained in a natural photoperiod for Toronto, ON, Canada. All animals were raised in aquaria filled with well-aerated filtered tap water at 22°C. L. stagnalis were fed lettuce ad libitum and experiments were performed on snails with a shell length of 10–20 mm (~2 months old).

All experiments were conducted at room temperature of 22°C. L. stagnalis saline solution included (in mmol l⁻¹): 10 glucose, 51.3 NaCl, 1.7 KCl, 4 CaCl₂, 1.5 MgCl₂, 10 HEPES, set to pH 7.9 using 12moll⁻¹ HCl. Snails were anesthetized briefly in L. stagnalis saline solution containing 30% Listerine, then de-shelled with forceps. Snails were pinned dorsal surface up to the bottom of a Sylgard-filled dissection dish, covered with L. stagnalis saline solution, and a medial incision was made from the base of the mantle to the head. Central ring ganglia were removed with a pair of fine surgical scissors, and placed in a modified flow-through perfusion chamber (RC-26 with a P1 platform, Warner Instruments, CT, USA) before being pinned out. The bottom of an RC-26 chamber was modified by replacing the 22mm×40mm glass coverslip with 22mm×40 mm plexiglass plates and a 4mm layer of Sylgard (total thickness 5mm). The connective tissue sheath was removed from cluster F neurons within the (LPeDG and RPeDG) with a pair of fine forceps. After being transferred to the microscope stage, the chamber was perfused at a rate of 2–3 ml min⁻¹ by gravity flow.

Whole-cell recordings were performed using pipettes with resistance in the range of 2–3 MΩ. Pipettes were pulled from borosilicate glass capillaries (Fisher Scientific, Nepean, ON, Canada) and filled with a filtered intracellular solution (pH 7.4) composed of the following (in mmoll⁻¹): 8NaCl, 0.0001CaCl₂, 10NaHEPES, 110 potassium gluconate, 1 MgCl₂, 0.3 NaGTP and 2 NaATP. An Axopatch-1D amplifier (Axon Instruments, CA, USA) was used for recordings. The series resistance was monitored throughout each recording and if it varied by >10%, the recording was rejected. No electronic compensation for series resistance was used. After obtaining a gigaohm-seal onto neurons, whole-cell patches were obtained under voltage-clamp mode at a holding membrane potential of −60mV by applying a brief suction. Spontaneous neuronal activity was recorded after switching to current clamp mode. Recordings were low-pass filtered at 2 kHz using a CV-4 headstage, and a Digidata 1200 (Axon Instruments, CA, USA) and then digitized and stored on a computer using Clampex 7 software (Axon Instruments, CA, USA). Neurons were perfused with L. stagnalis saline solution until AP frequency and V_m stabilized (10–20 min). Following this stabilization period, baseline AP frequency and V_m were recorded for 20 min before exposing the preparations to treatment saline for up to 40 min. The preparation was then reperfused with control saline solution to wash out the treatment saline. Each preparation was exposed to only one treatment protocol.

The Ca²⁺-sensitive, membrane-impermeable pentapotassium salt of fluo-4 (Molecular Probes Inc.) was used to determine intracellular Ca²⁺ [Ca²⁺]_i levels. Fluo-4 was used because of its low background absorbance and increased brightness at lower concentrations (Paredes et al., 2008). Fluo-4 was dissolved in the intracellular recording solution at a final concentration of 100 μmol l⁻¹. Fluo-4 was excited at 488 nm for 0.1 s using a DeltaRam X high-speed random-access monochromator and a LPS220B light source (PTI, NJ, USA). Fluorescent emissions above 510nm were isolated using an Olympus DM510 dichroic mirror and fluorescence images were acquired (526 nm) at 10-s intervals to limit photobleaching, using an Olympus BX51W1 microscope and a QImaging Rolera MGi EMCCD camera (Roper Scientific Inc., IL, USA). The obtained images were quantitatively analyzed for changes in fluorescence intensities within cells using EasyRatioPro software (PTI). Values of fluo-4 measurements are reported as means ± standard error (s.e.m.), with N being the number of cell bodies tested from different animals. The amplitude of the fluo-4 responses was analyzed from the soma and the data expressed as relative fluorescence intensity.

All chemicals were obtained from Sigma Chemical Co. (Oakville, ON, Canada). Adenosine was dissolved in L. stagnalis saline solution at a final concentration of 200 μmol l⁻¹. N⁶-cyclopentyl-adenosine (CPA) and 8-cyclopentyl-1,3-dipropylxanthine (DPCPX) were initially dissolved in 100% dimethylsulfonic acid (DMSO); and then diluted in L. stagnalis saline solution to a final concentration of about 0.3% v/v. Vehicle application alone did not affect the stability of patch-clamp recordings (data not shown). A final concentration of 100 μmol l⁻¹ was used for both CPA and DPCPX.

AP frequency and V_m were monitored for a 5 min interval in the last 5 min of control (immediately before the switch to treatment perfusion), last 5 min of treatment, and last 5 min of each wash-out perfusion at a sampling rate of 5 kHz. There was significant variation in the AP frequency between neurons (minimum number of APs: 0.02 spikes s⁻¹; maximum number of spikes: 3.2 spikes s⁻¹). Therefore, in order to compare AP frequency between neurons and show the extent of variance within the control values, 5 min of control recordings were used as baseline to normalize data during a single experiment. To assess if neuronal activity changes as a function of time under control conditions without any pharmacological intervention, AP frequency and V_m were determined by analyzing a continuous trace for a 5 min interval at the end of every 10 min, for up to 60 min. For concomitant AP frequency and fluorescence measurements, AP frequency and fluo-4 signals were analyzed for a 1min interval every 5min of treatment perfusion for 20 min. Fluo-4 signals were normalized to values obtained immediately prior to treatment perfusion. A dose–response curve was fitted to the Four-Parameter Logistic (4PL) equation. All results are reported as means ± s.e.m. Statistical analysis was performed using SigmaStat software (Point Richmond, CA, USA). Results were analyzed using repeated measures one-way ANOVA with a Tukey's post-hoc test or a paired t-test, and significance was determined at P<0.05 unless otherwise indicated. Values expressed as percentage change were calculated as measured value minus baseline value divided by baseline value multiplied by 100%.

Whole-cell recordings were obtained from visualized neurons in cluster F on the dorsal surface of the RPeDG or LPeDG, which are part of the central ring ganglia in L. stagnalis [see figure 1A,B in Cheung et al. (Cheung et al., 2006)]. Recordings were restricted to neurons in cluster F of the PeDG because they have similar electrophysiological properties and their action potential shape has been well characterized (Kyriakides et al., 1989). The majority of neurons tested were spontaneously active, showing APs along with excitatory and inhibitory postsynaptic potentials at their resting membrane potential. Modulation of neuronal activity was determined by changes in AP frequency and V_m. In cells obtained during the summer (May to August), adenosine-mediated decreases in AP frequency were minute and this depression was not statistically significant at any given concentration (Table 2). Significant changes in neuronal activity in response to adenosine treatment were observed during winter months (December to March), and the following results represent data collected during the wintertime.

Under control conditions, there was no significant change in AP frequency for up to 60 min of the recording period. AP frequency remained stable: 1.77±0.56 spikes s⁻¹ at 30 min and 1.72±0.39 spikes s⁻¹ at 60 min (Fig. 1A,B; N=8). Similarly, no significant change in V_m occurred during the first (−64.9±3.1 mV) and second (−64.2±3.1 mV) halves of the recording period (Table 3; N=8). Application of 200μmoll⁻¹ adenosine caused a significant decrease in AP frequency from 1.08±0.22 to 0.57±0.14 spikes s⁻¹, a 47% reduction (Fig. 2A and Fig. 3; N=14). AP frequency during the recovery period increased to 0.78±0.16 spikes s⁻¹ and was not statistically different from the control recording prior to the adenosine perfusion (Fig. 2B and Fig. 3; N=14). Adenosine application also hyperpolarized V_m from −63.8±3.6 to −70.5±4.6 (Table 3; N=14). An adenosine concentration of 200 μmol l⁻¹ was chosen because this was the lowest concentration that resulted in changes in V_m and AP frequency. Cells did not respond to 50 and 100 μmol l⁻¹ adenosine, but did respond to 200, 500 μmol l⁻¹ and 1 mmol l⁻¹ adenosine (Fig. 4).

To determine whether the changes in neuronal activity caused by adenosine were A₁R-mediated, AP frequency and V_m were analyzed in the presence of the A₁R agonist CPA. Addition of 100 μmoll⁻¹ CPA to the bulk perfusion resulted in a 43% reduction of AP frequency, from 1.11±0.27 to 0.62±0.11 spikes s⁻¹ (Fig. 2C and Fig. 3; N=7). As with adenosine application, AP frequency recovered to 0.67±0.11 spikes s⁻¹ after CPA addition, and this was not significantly different from the control value prior to treatment onset (Fig. 2C and Fig. 3; N=7). However, an all pair-wise post-hoc test revealed that only recovery from the adenosine treatment was significantly different from adenosine treatment alone. This may be the result of tighter binding of the pharmacological modulators and a requirement for a longer washout period. CPA application also hyperpolarized V_m, from −72.4±2.6 to −76.9±2.1 (Table 3; N=7). The A₁R-mediated change was further tested by incubation of the preparation with the A₁R antagonist DPCPX prior to and during adenosine application. Addition of adenosine to DPCPX-treated neurons caused no significant change in AP frequency (from 1.09±0.06 to 0.81±0.03 spikes s⁻¹; Fig. 2D and Fig. 3; N=6) or V_m (from −56.4±2.8 to −60.8±3.9 mV; Table 3; N=6).

The effect of adenosine on Ca²⁺ homeostasis of cluster F neurons was examined using the single wavelength Ca²⁺ indicator fluo-4. Since neurons were not incubated with fluo-4 and the dye was applied through a patch pipette, at least 15min of baseline recording was required prior to any treatment application to allow maximal loading of cells with the dye and for the fluorescence trace to reach a plateau. The spontaneous activity (both AP frequency and V_m) of these neurons did not change under control conditions and was similar to that observed in previous recordings lacking fluo-4 (data not shown). Thus inclusion of the dye to the intracellular recording solution did not adversely affect neuronal viability. Application of adenosine provoked a significant 12.5±1.5% increase in the fluo-4 fluorescence; this response had a slow onset and it reached a maxima at ~20min (Fig. 5A,B; N=4). AP frequency and fluo-4 fluorescence were then analyzed at 5min intervals after treatment onset for 20 min to determine whether the two measurement variables covaried. In response to adenosine perfusion, correlation analysis revealed a linear correlation between the two variables; as Ca²⁺ concentration increased, AP frequency decreased. The correlation between [Ca²⁺]_i and AP frequency was significant (r²=0.91, P=0.0125) and a linear regression analysis of the data resulted in a slope of −0.11 (Fig. 5C; N=4). Conversely, application of adenosine to DPCPX-treated neurons caused no significant change in intracellular calcium levels (Fig. 5A,B; N=4).

In this study, we examined the inhibitory effects of adenosine in the central ring ganglia from the pond snail L. stagnalis. In response to application of adenosine, cluster F neurons exhibited an 11% hyperpolarization of the resting membrane potential along with a 47% decrease in action potential firing frequency. The reversibility of the depression in neuronal activity upon reperfusion with control saline suggests that this phenomenon may be receptor mediated.

To further explore this possibility, we used a pharmacological A₁R agonist and antagonist to determine changes in neuronal electrical properties. CPA and DPCPX were selected because of their relative A₁R specificity and because the latter has previously proved effective as an A₁R antagonist in invertebrate preparations (Magazanik and Federova, 2003). When A₁-receptor mediated neurotransmission was antagonized with DPCPX, adenosine failed to elicit a robust inhibitory response. We further demonstrated that the inhibitory effects of adenosine are mediated via the A₁ receptor, by showing that the response is CPA sensitive. CPA caused a reduction in AP frequency and hyperpolarization of V_m that were comparable to those elicited by adenosine. The inhibitory actions of adenosine are consistent with a previous investigation (Barraco and Stefano, 1990) which showed that adenosine inhibited the release of excitatory neurotransmitters including serotonin and dopamine from the pedal ganglia of the marine bivalve Mytilusedulis, whereas an adenosine antagonist blocked the inhibitory effects on neurotransmitter release.

Depression of neurotransmitter release (glutamate in particular) by activation of adenosine receptor is achieved through G-protein-coupled inhibition of Ca²⁺ influx in nerve endings (Fredholm and Dunwiddie, 1988; Wu and Saggau, 1997) along with enhancement of K⁺ and Cl⁻ conductance (Trussel and Jackson, 1985; Mager et al., 1990). Adenosine receptor activation may also cause an increase in cytosolic Ca²⁺ levels via an inositol 3-phosphate (Ins(3)P)-mediated pathway. All adenosine receptors except the A_2A subtype activate phospholipase C (PLC), an enzyme converting phospholipids into diacylglycerol and Ins(3)P (Bickler and Buck, 2007; Dunwiddie and Masino, 2001). Ins(3)P receptor activation is consistent with the elevation of [Ca²⁺]_i observed in cluster F neurons upon perfusion with adenosine. Furthermore, a strong correlation was observed between elevated [Ca²⁺]_i and decreasing AP frequency. The reversibility of this phenomenon could not be tested because of technical limitations (whole cell patches were difficult to maintain beyond 60 min). Despite this shortcoming, it is clear that the Ca²⁺ increase was adenosine dependent because incubation with DPCPX failed to elicit a response in the presence of adenosine.

Unlike the deleterious accumulation that occurs during excitotoxic cell death (ECD) in response to lack of oxygen in hypoxia-sensitive species (Choi, 1994; Choi, 1992), minute increases in cytosolic Ca²⁺ levels can be neuroprotective in certain anoxia-tolerant species (Bickler et al., 2000; Shin et al., 2005). One of the outcomes of elevated Ca²⁺ is inactivation of N-methyl-d-aspartate receptors (Ehlers et al., 1996) that are Ca²⁺ permeable glutamate-gated ion channels regulated by [Ca²⁺]_i and required for fast excitatory neurotransmission in the central nervous system (Albensi, 2007). Thus, the adenosine-mediated decrease in AP frequency and hyperpolarization of V_m, along with a simultaneous increase in Ca²⁺ suggest a possible synergistic mechanism resulting in global neuronal depression.

L. stagnalis is an anoxia-tolerant pulmonate that hibernates during the winter (Jones, 1961). Tolerance of variable oxygen concentrations is often associated with inactivity and hypometabolism induced by cold temperatures. There is a “hemispheric asymmetry” in cold tolerance amongst species that frequently undergo freeze–thaw events (Sinclair et al., 2003; Chown et al., 2004). Because these events occur year-round in high latitudes of the Southern Hemisphere, endemic freeze-tolerant species retain their ability to survive cold temperatures throughout the year without energetically demanding metabolic shifts. By contrast, temperatures in high-latitude areas of the Northern Hemisphere remain sub-zero for several months during winter and rise well above freezing for long periods during the summer. Consequently, native species lose their cold tolerance in the summer, but show exceptional freeze tolerance in winter that requires costly physiological adaptations. These differing strategies amongst freeze-tolerant species in the Southern and Northern hemispheres raise the possibility that anoxia tolerance of L. stagnalis may have a seasonal dependence. Furthermore, it might explain our observation of diminished responsiveness to adenosine modulation in neurons obtained during the summer months (May to August). Sensitivity of L. stagnalis neurons to seasonal changes has been reported previously. Zapara et al. (Zapara et al., 2004) found that neurons harvested during summertime exhibited a depolarization of the resting membrane potential (anoxic depolarization) in response to ischemia and/or hypoxia, whereas those obtained during the winter showed resistance to anoxic depolarization. More research is required to elucidate the mechanisms that render neurons more sensitive to adenosine regulation during winter (December to March) than during summer.

In conclusion, we demonstrate that adenosine depresses neuronal activity in cluster F neurons of L. stagnalis via the A₁R (type I subclass) and this inhibition correlates with changes in [Ca²⁺]_i. Additional experiments are required to ascertain if the adenosine response is mediated entirely through the A₁R, by testing the effects of adenosine type 2 receptor modulators. Furthermore, it remains to be determined whether adenosine confers protection during severe hypoxia and whether this is mediated via changes in cytosolic Ca²⁺ levels.

A.M. and L.T.B. would like to thank Dr Melanie Woodin for her technical advice and assistance with the snail recordings. We also thank Dr Zhong-Ping Feng for helpful comments and generous supply of snails occasionally. This work was supported by an NSERC Discovery Grant to L.T.B.

 
• AP frequency

  action potential firing frequency

 
• A₁R

  adenosine receptor type I subclass

 
• [Ca²⁺]_i

  intracellular calcium concentration

 
• CPA

  N⁶-cyclopentyladenosine

 
• DPCPX

  8-cyclopentyl-1,3-dipropylxanthine

 
• PeDG

  pedal dorsal ganglia

 
• V_m

  membrane potential