Address correspondence and reprint requests to Dr Stefano Ferroni, Department of Human and General Physiology, University of Bologna, Via S. Donato 19/2, 40127 Bologna, Italy. E-mail: Stefano.firstname.lastname@example.org
Guanosine (Guo) is an endogenous neuroprotective molecule of the CNS, which has various acute and long-term effects on both neurones and astroglial cells. Whether Guo also modulates the activity/expression of ion channels involved in homeostatic control of extracellular potassium by the astrocytic syncytium is still unknown. Here we provide electrophysiological evidence that chronic exposure (48 h) to Guo (500 μm) promotes the functional expression of an inward rectifier K+ (Kir) conductance in primary cultured rat cortical astrocytes. Molecular screening indicated that Guo promotes the up-regulation of the Kir4.1 channel, the major component of the Kir current in astroglia in vivo. Furthermore, the properties of astrocytic Kir current overlapped those of the recombinant Kir4.1 channel expressed in a heterologous system, strongly suggesting that the Guo-induced Kir conductance is mainly gated by Kir4.1. In contrast, the expression levels of two other Kir channel proteins were either unchanged (Kir2.1) or decreased (Kir5.1). Finally, we showed that inhibition of translational process, but not depression of transcription, prevents the Guo-induced up-regulation of Kir4.1, indicating that this nucleoside acts through de novo protein synthesis. Because accumulating data indicate that down-regulation of astroglial Kir current contributes to the pathogenesis of neurodegenerative diseases associated with dysregulation of extracellular K+ homeostasis, these results support the notion that Guo might be a molecule of therapeutic interest for counteracting the detrimental effect of K+-buffering impairment of the astroglial syncytium that occurs in pathological conditions.
sodium dodecyl sulphate–polyacrylamide gel electrophoresis
Extracellular nucleotides and nucleosides are a class of signalling molecules that play critical roles both in physiological conditions and in the pathophysiology of several neurodegenerative disorders (for review see Neary et al. 1996). The cellular and molecular mechanisms underlying the activity of adenine-based nucleotides and nucleosides on both neuronal and glial cells have been well characterized (Schubert et al. 1997; Abbracchio and Burnstock 1998), whereas little is known about the effects of their guanine-based counterparts. Recently, a major role of the guanine nucleoside guanosine (Guo) in neuroprotection has been hypothesized (Rathbone et al. 1999). Guo has been reported to have various cellular effects on astroglial cells that may contribute to its protective action, including the synthesis and secretion of neurotrophic and pleiotrophic factors (Middlemiss et al. 1995; Rathbone et al. 1998; Ciccarelli et al. 1999), and stimulation of glutamate uptake in cultured astrocytes (Frizzo et al. 2001). It has also been demonstrated that Guo preserves glial cell viability in experimental conditions that normally cause astrocytic death (Litsky et al. 1999; Di Iorio et al. 2004; Yoo et al. 2005). The importance of Guo as a protective molecule is further supported by in vivo findings indicating that chronically administered Guo prevented the development of seizures and cell death in a model of glutamate excitotoxicity (Lara et al. 2001; Vinadèet al. 2003).
Because there is clear evidence that extracellular levels of Guo remain raised for up to 1 week following insult and injury to the CNS (Uemura et al. 1991), and in view of the possibility that Guo may also exert its neuroprotective action by reinforcing the ability of astroglial cells to control [K+]o homeostasis, in this study we have addressed the question whether chronic exposure to Guo could also modify the activity/expression pattern of K+ channels in cultured rat cortical astrocytes. Consistent with our previous observation, we found that primary cultured astrocytes do not possess functional Kir channels (Ferroni et al. 1995). However, after a prolonged (48 h) treatment with Guo, astrocytes acquired a large Kir conductance. Immunoblot (IB) analysis and comparative functional studies with recombinant Kir channels indicated that the Guo-induced K+ current is largely mediated by the Kir4.1 subtype. Finally, we showed that this effect of Guo necessitates de novo synthesis of Kir4.1 protein. Collectively, these data suggest that the neuroprotective effect exerted by Guo may also rely on its ability to reinforce the astrocyte-mediated regulation of [K+]o homeostasis.
Materials and methods
Cell culturing and Guo treatment
Primary cultures of pure cortical rat astrocytes were prepared as described previously (Ferroni et al. 1995). Briefly, cerebral cortices devoid of meninges were triturated and placed in cell culture flasks containing Dulbecco's modified Eagle's medium (DMEM)–glutamax medium with 15% fetal bovine serum (FBS) and penicillin–streptomycin (100 U/mL and 100 µg/mL respectively) (all products purchased from Gibco-Invitrogen, Milan, Italy). Culture flasks were maintained in a humidified incubator with 5% CO2 for 2–5 weeks. Immunostaining for glial fibrillary acidic protein (GFAP) and the flat, polygonal morphological phenotype of the cultured cells indicated that more than 95% were type 1 cortical astrocytes (Ferroni et al. 1995). At confluence, astroglial cells were enzymatically dispersed using trypsin–EDTA in 33-mm diameter petri dishes at a density of 1 × 104 per dish and maintained in culture medium containing 15% FBS. Six hours after plating, the medium was replaced with DMEM containing 1% FBS. After 24 h, 500 µm Guo was added and astrocytes were investigated 48 h later. The use of such relatively high concentration of Guo was due to the fact that it has been demonstrated that during long-term incubation (> 3 h) of cultured astrocytes, the amount of extracellular Guo decreases to about 30% of the nominal level because of extracellular metabolism and activity of nucleoside transporters (Di Iorio et al. 2002; Peng et al. 2005). As a control, astrocytes plated in petri dishes were maintained for the same time period in medium containing 1% FBS. In experiments analysing the effect of the protein synthesis inhibitors cycloheximide (Sigma, St Louis, MO, USA) and actinomycin D (Sigma), the inhibitors were added (2.5 µg/mL and 100 ng/mL respectively) 1 h before Guo treatment and left throughout the whole period of incubation (cycloheximide) or the first 8 h of Guo treatment (actinomycin D). In experiments exploring the time course of Guo action, 500 µm Guo was added to cells maintained in medium containing 1% FBS for 1 or 12 h before performing the electrophysiological recordings. In some experiments the effect of 6 h of Guo treatment was investigated after 48 h; following the 6-h incubation, the Guo-containing medium was replaced with fresh medium containing 1% FBS for an additional 42 h. To minimize variability, results of each treatment were determined by comparing the treatment effects with results obtained on the same day in control astrocytes subjected to the same conditions but without Guo.
Current recordings were obtained with the whole-cell configuration of the patch-clamp technique (Hamill et al. 1981). Patch pipettes were prepared from thin-walled borosilicate glass capillaries to have a tip resistance of 2–4 MΩ when filled with standard internal solution. Membrane currents were amplified (List EPC-7 amplifier, Darmstadt, Germany), filtered at 2 kHz (− 3dB) and acquired at a sample rate of 5 kHz on a microcomputer for off-line analysis (pClamp 6, Axon Instruments, Foster City, CA, USA; Origin 6.0, MicroCal, Northampton, MA, USA). Because of the large amplitude of the currents often measured, the access resistance (below 10 MΩ) was corrected for to 70–90%. Cell capacitance was estimated by on-line electronic correction of capacitive currents of the recorded cell. Values of current density were calculated by dividing the whole-cell currents by cell capacitance. Experiments were carried out at room temperature (20–24°C).
Solutions and chemicals
Salts and other chemicals were of the highest purity grade (Sigma). For electrophysiological experiments the standard bath saline contained (in mm) 140 NaCl, 4 KCl, 2 MgCl2, 2 CaCl2, 10 HEPES, 5 glucose, pH 7.4, with NaOH and osmolarity adjusted to ∼ 315 mOsm with mannitol. The intracellular (pipette) solution was composed of (in mm): 144 KCl, 2 MgCl2, 5 EGTA, 10 HEPES, pH 7.2 with KOH and osmolarity ∼ 300 mOsm. When using external solutions with different ionic compositions, salts were replaced equimolarly. The different saline solutions containing the pharmacological agents were applied with a gravity-driven, local perfusion system at a flow rate of ∼ 200 µL/min positioned within ∼ 100 µm of the recorded cell. Guo was dissolved in 1 N NaOH and the final concentration of NaOH in culture dishes was 0.01%.
Total RNA extraction, cDNA synthesis and RT–PCR analysis
Total RNA was extracted from cell cultures with TRI-reagent following the manufacturer's instructions (Sigma). First-strand cDNA was synthesized from total RNA using Moloney Murine Leukemia Virus (M-MLV) reverse transcriptase according to the manufacturer's instructions (Gibco Invitrogen). First-strand cDNA was used as template for PCR amplification (Taq DNA polymerase; Takara, Otsu, Shiga, Japan) of specific Kir fragments. The reaction was carried out in a final volume of 50 µL containing 1 ×Taq polymerase PCR buffer, 0.5 µm primers, 0.2 mm dNTPs, 2.5 units Taq polymerase (5 U/µL), and specific Kir, β-actin or GFAP primers (Table 1). The following PCR cycles were then carried out: 94°C for 2 min, 35 cycles of 94°C for 40 s, 58°C for 1 min and 68°C for 2 min, and finally 1 min at 68°C. Some 10 µL of the amplification reaction was run in parallel with a known molecular weight marker (Fermentas, Hanover, MA, USA) on a 2% agarose gel and stained with ethidium bromide.
Table 1. RT–PCR primers for Kir channels
Size of PCR product (bp)
Sequences of oligonucleotides used for RT–PCR reactions and relative expected sizes (bp) of PCR products used to amplify different members of the Kir channel family.
After each treatment, cultured astrocytes were washed twice with ice-cold phosphate-buffered saline (PBS) and then incubated with biotinamidohexanoic acid 3-sulfo-N-hydroxysuccinimide ester sodium salt (a reagent that preferentially labels primary amines, but also secondary amines) to a final concentration of 25 µg/mL (Sigma) for 30 min at 4°C in ice-cold biotinylation buffer (50 mm Na borate, pH 8, 150 mm NaCl). To stop the reaction, cells were incubated with cold Tris-HCl, pH 8.8, for 15 min. After two washes with ice-cold PBS, astrocytes were harvested for 30 min with lysis buffer (50 mm HEPES pH 7.4, 150 mm NaCl, 10% glycerol, 1% Triton X-100), and centrifuged at 14 000 g for 30 min at 4°C. Protein concentration was determined in the supernatant before and after precipitation using a Bio-Rad protein assay (Bio-Rad Laboratories, Hercules, CA, USA) and equal amounts of protein were used for the purification with streptavidin–agarose (Sigma), which specifically binds biotin-conjugated cell-surface proteins. Affinity-purified protein complexes were washed three times with lysis buffer, denatured with Laemmli sample buffer (80°C for 10 min). Total lysate containing 15 µg total protein (pre-P) and post-precipitation samples (post-P), and 15 µL affinity-purified protein complexes (membrane fraction, P) were separated by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE), transferred on to nitrocellulose and probed by western blot analysis.
Protein immunoprecipitation (IP)
At the end of each treatment, astrocytes were washed twice with ice-cold PBS, harvested in lysis buffer (50 mm Tris-HCl, pH 7.4, 100 mm NaCl, 1 mm EGTA pH 7.4, 0.5% sodium deoxycholate, 1% Triton X-100) and scraped off. The lysate was centrifuged at 14000 g for 30 min at 4°C. Protein concentration was determined in the supernatant. The supernatant was used as a whole-cell extract or immunoprecipitated with anti-Kir4.1 antibody (10 µg antibody per 0.500 mg whole-cell extract protein) at 4°C overnight. Immunocomplexes were captured with agarose-conjugated protein G (75 µL; ImmunoPure® immobilized protein G; Pierce), washed five times with 500 µL lysis buffer at 4°C, dissociated with 75 µL SDS–PAGE sample buffer, boiled for 5 min at 80°C, and analysed by SDS–PAGE and immunoblotting. Some 15 µg whole-cell extract (pre-IP), immunoprecipitated samples (post-IP), 40 µL of affinity-purified protein immunocomplexes (membrane fraction, IP) and 25 µL pre-immune sample (pre-immune) were separated by SDS–PAGE, transferred on to nitrocellulose and probed by western blot analysis with anti-Kir5.1 antibody (1 : 100; Santa Cruz Biotechnology, Santa Cruz, CA, USA). Immunoreactive bands were visualized using an enhancing chemiluminescence detection system (ECL-Plus; Amersham Biosciences Europe, Milan, Italy).
Western IB analysis
At the end of each treatment, astrocytes were washed twice with ice-cold PBS, harvested in lysis buffer (50 mm HEPES, pH 7.4, 150 mm NaCl, 10% glycerol, 1% Triton X-100) and scraped. The lysate was centrifuged at 14000 g for 30 min at 4°C. Protein concentration of each supernatant, collected as total protein extract, was quantified with the Bio-Rad protein assay. Samples containing 25 µg total protein were separated on SDS–polyacrylamide gels (12%) and electrotransferred on to nitrocellulose membranes (Bio-Rad Laboratories), blocked in 5% fat-free milk in PBS containing 0.05% Tween 20 and probed overnight with primary antibodies against Kir4.1 (1 : 400; Alomone Laboratories, Jerusalem, Israel), Kv1.1 (1 : 400; Alomone Laboratories), Kir 2.1 (1 : 100; Santa Cruz Biotechnology), Kir 5.1 (1 : 100; Santa Cruz Biotechnology), catalytic subunit of protein kinase A (cat PKA) (1 : 1000; BD Transduction Laboratories, San Jose, CA, USA), GFAP (1 : 500; Sigma) and actin (1 : 400; Sigma). After washing with phosphate-buffered saline plus Tween 0.05% v/v (PBST) membranes were incubated with horseradish peroxidase-conjugated IgG secondary antibodies (Sigma) and developed with ECL-Plus.
The Kir4.1/pBF clone was a gift from Professor Mauro Pessia (University of Perugia, Perugia, Italy). For expression in mammalian COS-7 cells, Kir4.1 clone was removed from the pBF vector and subcloned into the mammalian expression vector pCDNA3.1 (Invitrogen, Carlbad, CA, USA). The sequences of the transferred segments were verified by restriction analysis and automated DNA sequencing.
COS-7 cell culture and transfection
COS-7 cells were cultured as described previously (Caprini et al. 2001). Briefly, the day before transfection, COS-7 cells were re-plated in 35-mm petri dishes at a density of 2–5 × 104 per dish and maintained in supplemented DMEM. COS-7 cells were co-transfected with the Kir4.1 construct and the reporter gene encoding enhanced green fluorescent protein by the DEAE-dextran method (Sigma). Electrophysiological measurements were performed 48–72 h after transfection (Caprini et al. 2001).
Currents elicited with families of voltage steps or by voltage ramps were analysed with Clampfit (Axon Instruments) The current traces were plotted with Origin software (Microcal). Data are expressed as mean ± SEM of several cells (n) for each condition. Because of possible differences in cell size, membrane currents were normalized and are shown as current densities. Statistical evaluation was performed with two-tailed Student's t-test or anova followed by Bonferroni's post hoc test as appropriate; p < 0.05 was taken as statistically significant.
Long-term exposure to Guo up-regulates a plasma membrane Kir conductance in cultured cortical astrocytes
Whole-cell membrane currents were recorded in cultured cortical astrocytes under control conditions and after exposure to 500 µm Guo for 48 h. Low-density cortical astrocytes had the typical polygonal, flat morphological phenotype (Ferroni et al. 1995). Astrocytes were voltage clamped at a holding potential (Vh) of − 60 mV and, after stepping to − 120 mV for 400 ms, a slow ramp (180 mV/600 ms) from − 120 to 60 mV (inset in Fig. 1a) was applied to evoke whole-cell currents. With control intracellular and extracellular saline, the ramp current displayed a strong outward rectification (Fig. 1a). Rapidly activating, non-inactivating voltage-dependent whole-cell currents were elicited with a voltage-step protocol at potentials positive to − 40 mV (Figs 1c and e). These data were consistent with our previous observation (Ferroni et al. 1995) that primary cultured cortical astrocytes display only voltage-gated K+ channels that are activated at membrane potentials more positive than − 40 mV. Astrocytes treated for 48 h with 500 µm Guo showed the same morphological phenotype, but displayed significant changes in passive membrane properties. The resting membrane potentials shifted to more negative values (− 41 ± 1 mV, n = 64, in untreated astroglia; − 69 ± 2 mV, n = 71, in Guo-treated cells; p < 0.01) an effect that was accompanied by a significant decrease in input resistance at − 60 mV (782 ± 42 MΩ in untreated astrocytes; 255 ± 24 MΩ in Guo-treated cells; p < 0.01) and a slight but significant increase in membrane capacitance (49 ± 3 pF in control astrocytes; 65 ± 2 pF in Guo-treated cells; p < 0.05). The typical current profile of an astrocyte after treatment with Guo for 48 h treated is shown in Fig. 1(b). The current reversal potential (Erev) was significantly more negative (∼ − 80 mV) and the ramp current showed a double rectification profile with large inward currents also activated at membrane potentials more negative than − 40 mV (Fig. 1f). The voltage-step protocol illustrates that, at negative membrane potentials, currents activated fully within 50 ms and did not show any time-dependent inactivation (Fig. 1d).
Because of the negative zero-current membrane potential close to the equilibrium potential for K+ (EK) under our experimental conditions (∼ − 90 mV) and the temporal kinetics of the current activated upon hyperpolarization, we hypothesized that Guo-treated astrocytes bear Kir current, and so the next set of experiments was designed to verify this possibility. To this end, [K+]o was raised from 4 to 40 mm, and ramp currents were measured at each [K+]o in both untreated and Guo-treated astrocytes (Fig. 2a). Increasing the [K+]o in control astrocytes did not cause a significant shift in Erev and current kinetics (Fig. 2a), whereas in Guo-treated cells the same procedure caused a ∼ 58-mV positive shift in Erev and a large increase in ramp currents at membrane potentials below Erev (Fig. 2b). In Guo-treated astroglial cells, the input resistance measured at − 60 mV was four-fold lower (Fig. 2c), and the specific conductance, calculated by normalizing the cell conductance to the membrane capacitance to avoid variability in cell size, was significantly increased upon increasing [K+]o (Fig. 2d). Collectively, these data support the notion that Guo-treated astrocytes express a Kir conductance, which is absent in untreated control astrocytes.
To gain further insight into the type of Kir conductance induced by Guo, we performed experiments at various [K+]o. In addition to displaying voltage sensitivity of gating that depends on [K+]o, Kir currents possess a high K+ selectivity over cations and they are multi-ion pores, a property associated with a non-linear membrane conductance (G)/[K+]o relationship (Nichols and Lopatin 1997). In agreement with previous results, raising [K+]o caused an increment in inward currents that was accompanied by a shift in zero-current potentials (Fig. 2e). The plot of the changes in Erev as function of [K+]o indicated that this conductance was highly K+ selective (Fig. 2f). Moreover, the fit of the chord conductance of the linear portion of the current–voltage (I-V) curves at negative potentials in various [K+]o indicated that conductance changes occurred proportionally to the square root of [K+]o as expected for multi-ion channels such as Kir channels (Hille 1992) (Fig. 2g).
We next attempted to isolate the Kir current underlying this Guo effect by the definition of its pharmacological properties. Because it is not possible to inhibit the astroglial delayed rectifier K+ current using tetraethylammonium and 4-aminopyridine without also affecting the Kir conductance (Ransom and Sontheimer 1995; Bordey and Sontheimer 1999), we used the alternative approach of inhibiting Kir channels with micromolar concentrations of Ba2+ and Cs+, which have previously been shown selectively to block Kir channels in spinal cord astrocytes (Ransom and Sontheimer 1995). Figure 3 shows the currents elicited in the absence and presence of 200 µm Ba2+ (Figs 3a and b) and 1 mm Cs+ (Figs 3c and d) in untreated and Guo-treated astrocytes. As expected, at these concentrations these two ions did not significantly alter the ramp currents of control astrocytes, but caused a depression of the inward K+ conductance in Guo-treated astrocytes. At − 120 mV the percentage blockade was 82 ± 13% for Ba2+ (n = 20) and 44 ± 8% for Cs+ (n = 13). Notably, in Guo-treated cells Cs+ did not affect the outward current, whereas Ba2+ superfusion caused a significant reduction. A similar behaviour was previously reported in spinal cord astrocytes expressing Kir channels (Ransom and Sontheimer 1995). Such pharmacological properties were interpreted as evidence that Cs+ blockage was steeply voltage dependent, being larger at more negative potentials and virtually absent at potentials positive to Erev. In contrast, Ba2+-induced reduction of the outward current was ascribed to its weak voltage-dependent blocking behaviour probably causing a significant inhibition of the Kir current elicited at potentials above Erev. Our data support this hypothesis because in Guo-treated astrocytes voltage-dependent kinetics of residual outward currents after Ba2+ superfusion overlapped those of untreated cells bearing only delayed rectifier K+ current (cf. Figs 3a and b). As a result, the Ba2+-sensitive currents showed a weakly inward rectifying profile (Fig. 3f). Thus, the large outward conductance observed in Guo-treated astroglia compared with control cells at potentials positive to − 80 mV was probably due to the contemporary opening of delayed and inward rectifier K+ channels. Taken together, these data are in line with those obtained for Kir currents in spinal cord astrocytes, thereby supporting the tenet that chronic exposure of cultured cortical astrocytes to Guo promotes the up-regulation of Kir channels.
Kir4.1 channel contributes to the Kir conductance up-regulated by Guo
There is substantial evidence that astroglial cells in vivo possess different subtypes of Kir channels, which are differentially expressed according to brain region and developmental stage (Kressin et al. 1995; Poopalasundaram et al. 2000; Higashi et al. 2001). Thus, we next sought to analyse, at the mRNA level, the expression profiles of the channels that most likely gate the Kir current in Guo-stimulated astrocytes. RT–PCR analysis demonstrated that mRNAs for Kir2.1, Kir 4.1 and Kir5.1 were present in Guo-treated astrocytic cultures, whereas Kir1.1, Kir3.1 and Kir6.1 were not detected (Fig. 4a); such a messenger profile also was seen in untreated astrocytes (data not shown).
Under our experimental conditions, RT–PCR analysis provided only qualitative information about the presence of a given RNA. Moreover, the RNA transcript does not necessarily correlate with expression of the corresponding protein. Therefore, we next evaluated by IB analysis the molecular relationship between the signals for Kir2.1, Kir 4.1 and Kir5.1 transcripts, and expression of their respective proteins. Whole-cell extracts from astrocyte cultures were probed with specific antibodies against the Kir channels previously detected at the mRNA level to assess whether the difference in K+ current phenotype was associated with variations in the level of expression of specific Kir channels between untreated and Guo-treated astrocytes (Fig. 4b). Among the Kir transcripts identified in astrocytes bearing Kir currents, Kir5.1 showed the strongest IB signal and Kir2.1 the weakest. Compared with control astroglia, Guo treatment caused a down-regulation of Kir5.1 and left the expression level of Kir2.1 unchanged. In contrast, monomeric Kir4.1, identified as a band at ∼ 50 kDa, was strongly up-regulated upon exposure to Guo. The molecular weight of this band did not correspond to that of thr Kir4.1 previously identified in total brain lysate of mouse (Hibino et al. 2004), but was identical to that of Kir4.1 previously shown to be expressed in cultured cortical astrocytes (Olsen and Sontheimer 2004). As an additional internal control, we used the voltage-gated channel Kv1.1, which has been described in astroglia both in situ and in vitro (Allen et al. 1998; Hallows and Tempel 1998). Our data indicated that Kv1.1 was not affected by Guo treatment. Collectively, these findings identify Kir4.1 as the likely molecular entity that mediates the Kir current.
This notion was further supported by the results of a cell-surface biotinylation assay to specifically label Kir4.1 channels at the plasma membrane. The result of comparative studies of membrane-biotinylated Kir4.1 protein (denoted as P) in Guo-treated and untreated cells is shown in Fig. 4(c). The strongly up-regulated band at ∼ 50 kDa in Guo-treated astroglia was the only one found in the membrane fraction; no signal was present in the streptavidin precipitate when biotin was omitted (Fig. 4d). The lack of biotin labelling of a cytosolic protein, cat PKA (Fig. 4e), shows that the biotinylation assay was carried out in non-permeabilized cells under our experimental conditions. These data confirmed that the increase of biotinylated Kir4.1 in Guo-treated astroglia was due to an increase in the amount of Kir4.1 at the plasma membrane.
As further evidence that Kir4.1 underlies the Kir current in Guo-treated astrocytes, we compared some biophysical and pharmacological properties of the astrocytic Kir conductance with those of the recombinant Kir4.1 channel expressed in COS-7 cells. The ramp current profile of the Kir-bearing astrocytes and that of Kir4.1 expressed in COS-7 cells elicited in 4 and 40 mm[K+]o were nearly identical (Figs 5a and b). Furthermore, the current families evoked with voltage pulses had comparable time- and voltage-dependent kinetics. Notably, although these channels had inwardly rectifying behaviours, they allowed detectable outward current at potentials positive to Erev. These features are fully compatible with Kir4.1 properties that several studies have indicated to be less strongly inward rectifier compared with other subtypes of the same channel family (for review see Nichols and Lopatin 1997). Furthermore, the Ba2+ sensitivity was virtually identical in the two cell preparations (insets to Figs 5a–d). Taken together, these data strongly suggested that Kir4.1 mediates the majority of the Kir conductance expressed in cultured cortical astrocytes upon treatment for 48 h with Guo.
It is worth noting, however, that these results do not completely rule out the possibility that the increase in Kir conductance was due to an increase in formation of heterotetrameric proteins composed of Kir4.1 and Kir5.1, as has been postulated to occur in vivo in cortical astrocytes (Hibino et al. 2004). To test this possibility, experiments were performed using intracellular saline buffered at the mild acidic pH of 6.5, which does not affect hometetrameric Kir4.1 channels (for review see Ruppersberg 2000) but causes a large depression of heteromeric Kir4.1/Kir5.1 conductance (Tucker et al. 2000; Pessia et al. 2001). Intracellular dialysis for up to 7 min with saline at pH 6.5 did not elicit a significant diminution of the astrocytic Kir current, a result mirrored by that obtained in COS-7 cells transfected with recombinant Kir4.1 (Fig. 6a). This time interval was sufficient to clamp the intracellular pH to the set value because control experiments performed in COS cells expressing Kir4.1 demonstrated that such a period of intracellular dialysis with saline at pH 5.0, which was previously shown to depress this channel strongly (Tanemoto et al. 2000), was long enough to cause a complete blockade of the inward currents (data not shown). Finally, further evidence that the Guo-induced conductance was not formed by heterometric Kir4.1/Kir5.1 channels was obtained in co-immunoprecipitation experiments, in which there was no sign of Kir4.1 and Kir5.1 heteromeric assembly (Fig. 6b).
Effect of Guo on Kir current requires high concentrations and de novo protein synthesis
Finally, we asked whether shorter incubation periods and lower concentrations of Guo were able to induce the functional appearance of the Kir conductance. To this end, cultured astrocytes were exposed for a shorter time to 500 µm Guo and expression of the Kir current was analysed electrophysiologically. The results summarized in Fig. 7(a), showing the specific conductance at − 60 mV, indicate that a 1–12-h incubation period was not sufficient to induce the functional appearance of the Kir channel. Likewise, a 48-h exposure to a 10-fold lower concentration of Guo was ineffective. Taken together, these findings suggest that Guo exerts this action only at high micromolar concentrations and acts by promoting de novo protein synthesis of Kir channels. Furthermore, co-treatment with cycloheximide, but not with actinomycin D, prevented the effect of Guo, indicating that Guo was acting at a translational level (Fig. 7b). The relatively high concentration of Guo necessary to achieve the described effect may be due to its turnover, which may affect the bioavailability. Whether Guo acts directly, via its metabolites, or by promoting the synthesis and release of trophic factors that interact in an autocrine/paracrine fashion with the cultured astrocytes, remains to be determined.
We have provided functional and molecular evidence that the guanine-based nucleoside Guo induces the expression of Kir conductance in cultured rat cortical astrocytes. Functional, pharmacological and molecular analyses indicate that this effect is probably mediated by de novo protein synthesis of the Kir4.1 subtype of Kir channels. Because Kir4.1 has been demonstrated in vivo to be involved in the control of extracellular K+ homeostasis by the astroglial syncytium, this represents the first study clearly identifying one of the molecular targets that may be involved in the neuroprotective role of Guo.
Previous studies showed that a brief application of low micromolar concentrations of Guo caused an increase in glutamate uptake both in primary cultured astrocytes and brain slices (Frizzo et al. 2001, 2002, 2003), suggesting that one of the mechanisms mediating the action Guo may be to limit the toxic extracellular rise of glutamate under pathophysiological conditions (Vinadèet al. 2005). This conclusion is further supported by the observation that orally administered Guo prevented the development of kainate-induced seizures (Lara et al. 2001). Because the predominant glutamate uptake mechanism in astroglia is the electrogenic, sodium-dependent glutamate uptake, which is favoured by membrane hyperpolarization (Anderson and Swanson 2000), it may be envisaged that an increase in the density of K+ currents in astroglial cells such as that produced by Guo would increase the efficiency of glutamate uptake. However, the effect of Guo described here is unlikely to play such a role for at least two reasons. First, the time course of the effect of Guo on the induction of astrocytic Kir conductance was too long (more that 12 h) compared with its action on glutamate uptake (within few hours) (Frizzo et al. 2002). Second, low micromolar concentrations of Guo did not affect astrocytic K+ currents even upon prolonged exposure (up to 48 h).
Conversely, the results may be relevant in the context of the long-term trophic effects of Guo (for review see Neary et al. 1996). Guo has been reported to promote the proliferation of primary cultured astrocytes (Kim et al. 1991); it was also shown to enhance the release of nerve growth factor (NGF) from cultured astroglia (Middlemiss et al. 1995; Rathbone et al. 1998). Finally, there is evidence that Guo-mediated release of trophic factors also occurs in vivo following traumatic injuries (Rathbone et al. 1999). Because extracellular levels of Guo remain raised for several days following stroke (Uemura et al. 1991), and Guo release in vitro is enhanced by hypoxia/hypoglycemia (Ciccarelli et al. 1999), it has been postulated that the endogenous increase in Guo may contribute to the remodelling of brain circuitry and to the recovery of brain function following traumatic and ischaemic injuries (Rathbone et al. 1998). In this context, the up-regulation of the Kir current may be of particular relevance. It has been widely demonstrated that the equipment of K+ channels in astroglial cells in vivo is altered following various brain insults. Numerous studies have revealed that Kir channels are down-regulated in regions adjacent to the lesion site where reactive gliosis occurs (D'Ambrosio et al. 1999; Schroder et al. 1999; Koller et al. 2000; Bordey et al. 2001). It is worth noting, however, that increases in Kir conductance have also been reported (Bordey et al. 2000; Anderova et al. 2004). These observations suggest that functional expression of astroglial Kir in vivo is subject to reciprocal modulations that may depend on the type of brain injury, the site of lesion and/or the release of signalling molecules able to modify the expression/activity of Kir channels (for review see Leis et al. 2005). Notably, it has been reported that the pleiotropic cytokine tumour necrosis factor-α, which plays a crucial role in inflammatory disease occurring as a result of various brain insults, caused a reduction of Kir current in cultured cortical astrocytes (Koller et al. 1998).
Accumulating evidence suggests that astroglial Kir conductance plays a pivotal role in the control of [K+]o homeostasis by mediating the spatial buffering of K+ (for review see Kofuji and Newman 2004). In retinal Muller glial cells the efficiency of [K+]o clearance through K+ siphoning was significantly altered in mice in which the Kir current was genetically abrogated (Kofuji et al. 2000). It was also reported that a diminution of K+-buffering capability of the astrocytic syncytium due to an impairment of Kir activity led to the generation of seizure activity observed in some pathological conditions (D'Ambrosio et al. 1999; Hinterkeuser et al. 2000). Hence, the neuroprotective role of Guo may also reside in its ability to reinforce the process of astroglial K+ buffering. It has been shown recently that Guo administered chronically in vivo is anticonvulsant and protects rodents subjected to α-dendrotoxin-induced seizures (Vinadèet al. 2003). Although a major contribution of the up-regulation of glutamate uptake is likely to mediate this protective action of Guo, a Guo-induced potentiation of the ability of the astroglial syncytium to avoid seizure-induced extracellular K+ accumulation through an increase in Kir activity cannot be ruled out. Of note, in glial cells from humans suffering from pharmacoresistant temporal lobe epilepsy there is also a reduction in Kir density and inward rectification, which probably contributes to seizure generation owing to a defect in the regulation of [K+]o homeostasis (Hinterkeuser et al. 2000).
The specific up-regulation by Guo of the weak inwardly rectifying Kir4.1 is not surprising because in brain cortical astrocytes Kir4.1 is expressed at the endfeet that surround neuronal synapses and blood vessels (Poopalasundaram et al. 2000; Higashi et al. 2001). By contrast, in retinal glial cells (Muller cells) K+ buffering appears to be mediated by a co-operation of Kir4.1 and the strong inward rectifier Kir2.1. In these cells the two channels show a different pattern of localization (Kofuji et al. 2002), being more abundantly distributed at cell sites contacting blood vessels (Kir4.1) and neuronal cells (Kir2.1). Our finding that Kir2.1 immunoreactivity in cortical astocytes was very weak and was not up-regulated is in line with these in situ results. In primary culture, it has been reported that the level of Kir4.1 transcript is very low in proliferating astrocytes and is substantially higher in confluent cultures (Li et al. 2001). Our experiments were performed in subconfluent cultures, and so the small number of astrocytes bearing Kir current (∼ 20%) under control conditions is consistent with these observations and may explain the IB signal reported in untreated astroglia. The up-regulation of Kir4.1 in non-confluent Guo-treated cultures may signify that, under our experimental conditions, this molecule is able to promote a partial, functional differentiation of astrocytes in culture. It remains to be established whether the effect of Guo is partially mediated through a modification of the interaction of Kir channels with anchoring cytoskeletal proteins, as reported previously (Horio et al. 1997).
The biotinylation of Kir4.1 might appear surprising because it is well known that brain Kir4.1 protein, also called BIRK-1 and KAB-2, lacks lysine residues in the extracellular domains (Bredt et al. 1995; Takumi et al. 1995). However, it must be pointed out that under certain conditions sulpho-NHS-biotin can also react with secondary amines of arginine residues, resulting in a stable imide linkage. Notably, the Kir4.1 sequence contains an arginine residue at the extracellular vestibule of the pore. The specificity of the cell-surface immunoreaction of Kir4.1 was confirmed by the observation that the ∼ 50-kDa Kir4.1 band was not detectable in a precipitation experiment performed in non-biotinylated astroglia. It is unlikely that our results could be explained by biotinylation of intracellular Kir4.1 owing to cell permeabilization because a cytosolic protein (cat PKA) was not biotinylated. Moreover, the bands representing post-translationally modified forms of Kir4.1 in the total lysate were not seen in the streptavidin precipitate. An alternative explanation is that the biotin labelling occurred on a transmembrane protein strongly associated with Kir4.1 in a macromolecular complex (Connors and Kofuji 2006).
The finding that Kir4.1 does not appear to form heteromeric channels with Kir5.1 is surprising because recent in situ immunolabelling data in mouse brain tissue demonstrated that Kir4.1 and Kir5.1 co-assemble in astroglia of the neocortex (Hibino et al. 2004). However, the different cellular context may partly explain this discrepancy. A contribution of the Kir2.3 (IRK3) channel to the Guo-induced Kir conductance is also unlikely because this channel is expressed only in adult reactive astrocytes (Perillan et al. 2000). Moreover, Kir2.3 is strongly inwardly rectifying and its activity is rapidly reduced upon cell dialysis in the presence of intracellular Mg2+ (Perillan et al. 2000). Under our experimental conditions Kir currents were stable for a period of recording up to 15 min after accessing the cell, even in the presence of 2 mm Mg2+ in the pipette saline (Valentina Benfenati and Stefano Ferroni, University of Bologna, Bologna, Italy, unpublished observation, but see Fig. 6a).
A central question that arises from this study is through which mechanism is Guo able to promote the functional expression of Kir current in cultured astrocytes. This is particularly relevant because there is no clear evidence for the existence of Guo receptors. Recent data have demonstrated the presence of Guo-binding sites in rat brain (Traversa et al. 2002, 2003), and there is evidence that in brain slices administration of Guo produces a dose-dependent increase in intracellular cyclic AMP (cAMP) accumulation. Surprisingly, these studies also showed that pertussis toxin (PTX), a specific suppressor of Gi/o-coupled receptor activity usually linked to lowering of cAMP levels, caused a robust reduction in Guo binding. Altogether, these results were interpreted as evidence that in brain Guo acts by binding to a specific G protein-coupled Guo receptor, whose activation induces an increase in cytosolic cAMP in a PTX-sensitive manner. Interestingly, we previously showed that long-term treatment of cultured astrocytes with a cell-permeant cAMP analogue, dibutyryl cAMP, promoted the expression of Kir-mediated currents (Ferroni et al. 1995), thereby suggesting that an increase in intracellular cAMP might be involved in the action of Guo. Further studies are warranted to clarify this issue. Because of the long incubation time necessary to measure significant changes in Kir conductance induced by Guo, it is possible that Guo is not the direct mediator of the effect on Kir conductance, and that the positive modulation can occur only after other signalling molecule(s) have been released into the culture medium. Notably, Guo induces astroglial proliferation through the release of adenine-based purines (Ciccarelli et al. 2000). In cultured rat microglia, stimulation of adenosine A2A receptors modulates the expression of various K+ channels (Kust et al. 1999), indicating that adenine nucleoside is also able to modify the expression pattern of ion channels in brain cells. Alternatively, it might be envisaged that the Guo-mediated release of NGF plays a crucial role by activating, through an autocrine/paracrine loop, its astrocytic receptors. However, this hypothesis contrasts with the general view that the expression levels of high- and low-affinity NGF receptors (TrkA and p75NTR respectively) in astroglia in vivo under physiological conditions are very low or non-existent (Dougherty and Milner 1999; Barker-Gibb et al. 2001; Oderfeld-Nowak et al. 2003). Interestingly, both types of receptors are markedly up-regulated at the site of astrogliosis, which develops as a result of numerous acute brain insults (Oderfeld-Nowak et al. 2003; Lee et al. 1998). Because cultured astrocytes recapitulate some features of reactive astrocytes (McMillian et al. 1994; Wu and Schwartz 1998), it cannot be excluded that astrocytes possess NGF receptors under our experimental conditions (Hutton et al. 1992; Condorelli et al. 1994). In this context, the long time taken for Guo to up-regulate astrocytic Kir might be explained in terms of the period necessary for Guo to increase the level of extracellular NGF up to a critical point. NGF, in turn, may modulate Kir expression through binding with its astroglial receptors. It also remains to be clarified why cultured astrocytes have to be exposed to such high Guo concentration to become responsive. Binding studies of Guo on brain membranes have yielded Kd values in the nanomolar range (Traversa et al. 2002), and most cellular effects that have been directly attributed to an action of Guo had a Kd below 10 µm (Frizzo et al. 2001; Di Iorio et al. 2002). Nevertheless, a Guo-induced increase in cAMP has been reported, with increases in response to exposure up to 300 µm (Traversa et al. 2003). It must be pointed out that Guo undergoes rapid degradation in culture, dropping to 35% of starting levels by 3 h of incubation, and remaining stable thereafter (Di Iorio et al. 2004). This means that the bioavailability of Guo during the 48-h treatment may be significantly lower. The development of stable and specific Guo agonists should enable us to address this issue in more detail.
In conclusion, the results reported in this study indicate that the nucleoside Guo promotes the functional expression of Kir4.1 channels that regulate [K+]o homeostasis in cultured cortical astrocytes. Because of the critical importance of [K+]o homeostasis for normal brain activity, this study adds further support to the tenet that Guo is a molecule of potential therapeutic interest for the treatment of brain diseases whose pathogenesis results from substantial alterations of the K+-buffering capacity of the astrocytic syncytium.
We thank Alessia Minardi for preparation and maintenance of astrocytic cultures, and Michael Pusch and Maria Pia Abbracchio for valuable comments on the manuscript. We are grateful to Antonio Ferrer Montiel for helpful suggestions. This work was supported by grant COFIN/PRIN from the Italian Ministry of University and Research (SF) and FIRB negoziale RBNE01ARR4-003 (SF). Some of the results presented from part of the PhD thesis of VB.