Single channel properties of pannexin-1 and connexin-43 hemichannels and P2X7 receptors in astrocytes cultured from rodent spinal cords

Abstract

Astrocytes express surface channels involved in purinergic signaling. Among these channels, pannexin-1 (Px1) and connexin-43 (Cx43) hemichannels (HCs) release ATP that acts directly, or through its derivatives, on neurons and glia via purinergic receptors. Although HCs are functional, i.e., open and close under physiological and pathological conditions, single channel properties of Px1 HCs in astrocytes has not been defined. Here, we developed a dual voltage clamp technique in HeLa cells expressing human Px1-YFP, and then applied this system to rodent spinal astrocytes to compare their single channel properties with other surface channels, i.e., Cx43 HCs and P2X7 receptors (P2X7Rs). Channels were recorded in cell attached patches and evoked with ramp cycles applied through another pipette in whole cell voltage clamp. The mean unitary conductances of Px1 HCs were comparable in HeLa Px1-YFP cells and spinal astrocytes, ~42 pS and ~48 pS, respectively. Based on their unitary conductance, voltage-dependence, and unitary activity after pharmacological and gene silencing, Px1 HCs in astrocytes could be distinguished from Cx43 HCs and P2X7Rs. Channel activity of Px1 HCs and P2X7Rs was greater than that of Cx43 HCs in control astrocytes during ramps. Unitary activity of Px1 HCs was decreased and that of Cx43 HCs and P2X7Rs increased in astrocytes treated with fibroblast growth factor 1 (FGF-1). In summary, we resolved single channel properties of three different surface channels involved in purinergic signaling in spinal astrocytes, which were differentially modulated by FGF-1, a growth factor involved in neurodevelopment, inflammation and repair.

Graphical Abstract

Introduction

Pannexins (Pxs) and connexins (Cxs) are members of two different gene families that can form gap junctions (GJs) from hemichannels (HCs) in apposed membranes. In the central nervous system (CNS) these proteins are detected in neurons, glial cells, pericytes, and endothelial and vascular smooth muscle cells (Giaume et al., 2021; Mai-Morente et al., 2021; Nagy et al., 2018; Nagy & Rash, 2000; Vanlandewijck et al., 2018). Pxs and Cxs are assembled into oligomers in the ER-Golgi compartments to form HCs, which are inserted into the plasma membrane. When two compatible surface HCs dock in series, one HC from each opposed cell, they form an intercellular channel that may be incorporated into a cluster of hundreds of channels, a so-called GJ plaque. Chordate Pxs are orthologs of invertebrate innexins, the GJ forming proteins in protostomes and deuterostomes, except for echinoderms and hemichordates which have no identified GJs (Baranova et al., 2004; Bruzzone et al., 2003; Garré & Bennett, 2009; Panchin et al., 2000). Although glycosylation at HC docking sites limits Px GJ formation in mammalian tissues (Ruan et al., 2020), Px HCs (or pannexons) may dock in series to form GJ channels in Xenopus oocytes (Bruzzone et al., 2003) and mammalian cell lines transfected with exogenous Px1 cDNA (Lai et al., 2007; Sahu et al., 2014; Vanden Abeele et al., 2006).

Both Pxs and Cxs have been shown to form unopposed functional HCs in glial cells and modulate a wide range of physiological and pathophysiological processes including cell death, inflammation, neural development, neurotransmission, behavior, and neurodegeneration (Avendano et al., 2015; Bennett et al., 2012; Contreras et al., 2002; Garré et al., 2010; Garré et al., 2016; Giaume et al., 2021; Iglesias et al., 2009; Naus & Giaume, 2016; Orellana et al., 2011; Prieto-Villalobos et al., 2021; Retamal et al., 2007; Santiago et al., 2011; Scemes et al., 2019; Seo et al., 2021; Stehberg et al., 2012; Velasquez et al., 2016).

Distinguishing between Px HC and Cx HC pathways at both the cellular and system levels is still a challenge, since genetic manipulations induce compensations and specific pharmacological tools are not currently available (Battulin et al., 2021; Dermietzel et al., 2000; Naus et al., 1997). Techniques that improve measurements of unitary conductance in combination with pharmacology and gene manipulation would provide very informative tools to distinguish between Px HC and Cx HC pathways in cell culture models of rodent glial cells and human-derived glial cells obtained from patients with neurological disorders.

GJ channels are nominally twice the length of one HC, and an intercellular GJ channel has about half the conductance of an unopposed HC. For example, the unitary conductance of Cx43 HCs is about 220 pS, twice that of the gap junction channels, ~ 110 pS (Bukauskas et al., 2001; Bukauskas & Peracchia, 1997; Contreras et al., 2003). In contrast to Cx43 GJ, Px1 GJ channels may open with two main conductances, a prevalent voltage-sensitive state of ~35 pS, and an infrequent and voltage-insensitive state of ~170 pS (Palacios-Prado et al., 2022). While early studies reported that when expressed exogenously in oocytes, human Px1 forms unopposed HCs that open with a maximum unitary conductance of ~500 pS (Bao et al., 2004; Qiu & Dahl, 2009), others have found that Px1 HCs have unitary conductances no larger than ~ 100 pS in excised outside and inside out patches (Chiu et al., 2017; Ma et al., 2012), and there is still no general agreement on the single channel conductance of vertebrate unopposed Px1 HCs in the whole cell. Furthermore, there is as yet no study describing the single channel properties of Px1 HCs in glial cells including astrocytes.

To study the single channel properties of unopposed Px1 HCs and compare with those of other functionally (but not structurally) related ion channels expressed by astrocytes, i.e., Cx43 HCs and P2X receptors (P2XRs), we used a dual voltage clamp technique in which current through a cell attached patch was measured during administration of a voltage applied with a whole cell voltage clamp in single (isolated) cells. This strategy allows recording unitary events from a few channels in a small surface area, while maintaining strict control of V_m in the whole cell, minimizing the generation of confounding results due to overlapping of channel openings from the whole cell activity. After validation of the technique in HeLa cells overexpressing exogenous human Px1 tagged with yellow fluorescent protein (HeLa Px1-YFP) and rat Cx43 tagged with cyan fluorescent protein (HeLa Cx43-CFP), we recorded HCs from astrocytes cultured from both embryonic and neonatal rodent spinal cords, which express endogenous Px1 and Cx43 (Garré et al., 2010). Px1 HCs were activated by voltage ramp cycles of 2 s duration and up to - 100 mV to + 100 mV amplitude. The mean unitary conductance of Px1 HCs in dual voltage clamp experiments was comparable between HeLa Px1-YFP cells and astrocytes, 42.1± 4.1 pS and 48.8 ± 5.0 pS, respectively. We found no evidence of single channel openings of larger unitary conductance, i.e., 300–500 pS, using single clamp whole cell and cell attached patch and dual voltage clamp techniques in HeLa Px1-YFP cells and spinal astrocytes. Unitary conductance and voltage sensitivity of Px1 HCs differed from those of Cx43 HCs and P2X7Rs in spinal astrocytes. Furthermore, we evaluated other single channel properties of Px1 HCs, Cx43 HCs and P2X7Rs, including total open time, closed time between openings, and number of openings per ramp, approaching the overall activity of these channels in spinal astrocytes. Single channel activity of Px1 HCs, Cx43 HCs, and P2X7Rs in spinal astrocytes was differentially modulated by FGF-1, a growth factor implicated in neurodevelopment, inflammation and repair.

Materials and Methods

Ethics

Animal experimentation. Experimental procedures performed in making rat and mouse cell cultures were approved by the Institutional Animal Care and Use Committee of the Albert Einstein College of Medicine.

Cell cultures

HeLa cells expressing Cx43-CFP or Px1-YFP were used as previously reported (Bukauskas et al., 2001; Garré et al., 2016). For one set of experiments, HeLa cells stably expressing Px1-YFP or Cx43-CFP and parental cells were seeded 1 day before experiments at low density. Single channel recordings were conducted in isolated single cells.

For astrocyte cell cultures, rat spinal cords were removed from pups at postnatal day 1 (P1). Spinal cord tissue was mechanically dissociated and treated for 25 min with 0.001% trypsin-0.0002% EDTA solution in PBS (stock: 0.25% trypsin / 0.1% EDTA). Trypsin activity was stopped by adding 1 mM EDTA for 5 min, followed by further mechanical dissociation in maintenance medium composed of DMEM supplemented with 10% FBS, penicillin (100 UI/ml), and streptomycin (100 mg/ml). Cells were plated (at 2.0 × 10⁴ cells/cm²) on 12 mm glass coverslips or in 60 mm plastic culture dishes and kept in a maintenance medium. Cultures were grown at 37ºC in a humidified atmosphere of 5% CO2 and 95% O2 for ~10 days until reaching confluence.

For patch clamp experiments, only single isolated spinal astrocytes were patched in sub-confluent monolayers 1–3 days after cell dissociation of the confluent astrocyte monolayer. Astrocyte cultures were > 90% pure as determined by their GFAP immunoreactivity, and there were no cells of myeloid linage (CD11b+ and CD11c+).

Pannexin-1 and Cx43 deficient rat and mouse astrocytes

Subconfluent astrocyte cultures were treated with a Stealth siRNA duplex oligo ribonucleotide (siRNA Px1, 0.2 μM, ThermoFisher Scientific, catalog number 1330001) targeting rat Px1 mRNA and preventing Px1-mediated dye uptake (Garré et al., 2016). For control, we used a non-targeting siRNA, siRNA(−) (ThermoFisher Scientific). For transfection, subconfluent cultures were treated overnight with siRNA Px1 mixed with 10 μl/ml Lipofectamine reagent (ThermoFisher Scientific). Transfection reagents were removed and cells incubated for 4–6 h in DMEM medium (2% FBS) before 7 h treatment with FGF-1: Heparin (10 ng/ml: 5 IU/ml), as previously reported (Garré et al., 2016). Astrocytes were treated 7 h with heparin (50 IU/ml) as a control.

Gja1tm1Kdr mice (Cx43 KO) (Reaume et al., 1995) were obtained from Dr. Eliana Scemes (Albert Einstein College of Medicine, Bronx, NY). Using the same protocol as that for rat astrocytes, Cx43−/− astrocyte cultures were prepared from mouse embryos (E18), because pups die at birth of cardiac failure. The presence of targeted mutation was confirmed by PCR.

Electrophysiology: dual voltage clamp

HeLa cells and astrocytes were grown on 22-mm number 0 coverslips, which were transferred to an experimental chamber mounted on the stage of an inverted microscope equipped with phase-contrast optics (IX70 microscope, Olympus). A fluorescence imaging system was used with HeLa cells to visualize Cx43 or Px1 with enhanced cyan fluorescent protein (Cx43-CFP) or yellow fluorescent protein (Px1-YFP) attached to the C-terminal.

Patch clamp recordings were conducted using the experimental chamber and either one or two EPC-8 patch clamp amplifiers (HEKA). The chamber was perfused with a modified Krebs-Ringer external solution containing (in mM): 140 NaCl, 4 KCl, 2 CaCl₂, 1 MgCl₂, 2 CsCl₂, 1 BaCl₂, 5 HEPES, 5 glucose, 2 pyruvate (pH 7.4). Patch pipettes for whole cell recording were filled with a solution containing (in mM): 10 NaAsp, 140 KCl, 0.2 CaCl₂, 1 MgCl₂, 2 MgATP, 5 HEPES (pH 7.3), 2 EGTA, 5 mM tetraethylammonium Cl-. Patch pipettes for cell attached recording were filled with the same solution as used for whole cell recording except that ATP was omitted to minimize activation of purinergic receptors. Since we generally opened the patch at the end of dual voltage clamp experiments, having both electrodes with very similar solutions should minimize changes associated with changing the internal solution. Pipette resistance was in the range of 5 to 10 MΏ. As a control for dual voltage clamp experiments, we performed experiments with external solution in the cell attached patch pipette, which revealed minor differences from single channel conductances observed in symmetrical conditions. We also used external solution in all single clamp, cell attached patch experiments (see Figure S8).

During a dual voltage clamp experiment, one pipette was in cell attached mode clamped to V₁ = 0 to record channel activity (with a sensitivity of 50 mV/pA), and the other pipette (V₂) was in whole cell mode to clamp V_m = V₂ - V₁ and stimulate (recording sensitivity of 1mV/pA). In some experiments to validate the method, the gain in the amplifier connected to the whole cell pipette was increased to 50 mV/pA to simultaneously record channel activity at the same gain with both pipettes. To present inward currents in the same direction in both whole cell and cell attached patch recordings, currents recorded through the V₁ electrode were multiplied by −1.

Voltage ramps used for stimulation started with a fast hyperpolarization from 0 to −100 mV, where V_m (V₂ - V₁) remained at – 100 mV for 0.1 s. This was followed by a linear rise in V_m from - to + 100 mV over 1.5 s, V_m remained stationary again for another 0.1 s. Ramp cycle ended with a fast return to 0 mV, where V_m remained for 0.3 s (total cycle time: 2 s, Figure 1a). Voltages and currents were recorded and subsequently digitized using a MIO-16X A/D converter and acquisition software (Bukauskas et al., 2001). Records were digitized at a rate of 5 kHz and then passed through a digital Gaussian filter with pass frequency between 0.2– 0.5 kHz.

Figure 1.

Px1 HCs are activated by low frequency voltage ramps of either polarity. (a, b) Single channels were recorded using dual voltage clamp in HeLa Px1-YFP cells. Voltage ramps of 1.7 s duration and −/+ 100 mV amplitude applied through the whole cell pipette (V₂) to HeLa Px1-YFP cells (a), evoked hemichannel openings recorded in the cell attached patch (V₁) (b). (b1,b2) Same cell attached patch recording shown in b, but at higher current amplification. Two distinct channel behaviors are illustrated in b1 and b2. A single channel activity of ~ 40 pS was observed during the negative phase of the voltage ramp in both b1 and b2. In b1, a single channel remained mostly open during the positive phase of the ramp. In b2, at the most positive phase of the voltage ramp (+50 to +100 mV), a channel opening of slow gating was resolved. (c) Quantification of unitary conductances from various experiments like in (a,b). Unitary conductances were calculated by identifying single channel transitions during the linearly changing phase of the voltage ramps. These conductances were confirmed from the slopes of the relationship between current at the open (red dotted lines) and close states (blue dotted lines) and V_m (n = 40 ramps, 5 ramps/patch, from 8 different patches and cells). The mean unitary conductance were 50.8 ± 4.2 pS, V_m from −100 to 0 mV; and 48.3 ± 3.8 pS, from 0 to 100 mV. (d) Percentage of ramps with active Px1 HCs in cell attached patches (30 consecutive ramps as in panel a for each cell) in HeLa parental or Px1-YFP cells (n = 8, 13, patches, one patch per cell). (e) Total open and closed time during ramps (total open time or total closed time between two openings divided by total period of recording). Ramps with only one active channel were quantified ( n = 10, 6, patches, one patch per cell). (f) Mean number of openings per ramp recorded in HeLa Px1-YFP cells ( n = 7 patches, one patch per cell, 10 ramps per patch). Both pipettes contained internal recording solution because the cell attached patch was opened at the end of the experiment to calculate the voltage drop in the V₁ electrode (see Methods). MgATP was not added to the cell attached pipette solution to minimize activation of endogenous purinergic receptors. Currents recorded in V₁ were multiply by −1. Data are presented as dot plot distributions in (c, e) and mean ± SEM in (d, f). ns, not significant, *P < 0.05, ** P < 0.01, Wilcoxon matched-pairs signed rank test was used in (c, e), and Mann Whitney test was used in (d).

Single channel analysis

Channel gating events were identified as rapid transitions, opening and closing, between periods at constant voltage (steps) or during the linearly changing phase of voltage ramps. Repeated transitions of fixed amplitude (or conductance) from the basal level indicated activity of a single channel. Single channel events were characterized by three parameters, membrane potential, current change between open and closed states, and duration. One cell attached patch per cell was established with a seal resistance of 1–5 GΏ from at least 3 different cells for each experimental condition. For dual voltage clamp experiments, single-channel conductance was calculated from single channel currents in the cell attached pipette divided by the difference between the voltage in the whole cell and cell attached recordings (V₂-V₁). A correction factor, V₁/V₂, was calculated at the end of each experiment by opening the attached patch, shifting amplifier mode to current clamp in both pipettes and passing currents through V₁ comparable to those used during the experiment. V₁/V₂ was generally 0.8–1 at the end of the experiment, indicating that the bridge circuit that removes the voltage drop in the whole cell pipette was properly balanced. To present currents in the same direction in both whole cell and cell attached patch recordings, actual currents recorded through the V₁ electrode were multiplied by −1.

Statistics.

Data are presented as dot plot distributions and mean ± SEM; n expresses the number of ramps, trials, or patches per cell from a minimum of 3 independent experiments (i.e., HeLa cells subcultured on different days; astrocyte cell cultures prepared from ~10 pup spinal cords from at least 3 different litters). Means for each group were compared using a non-parametric test for continuous and non-paired variables, the Mann-Whitney test or paired Wilcoxon test, and one-way analysis of variance (ANOVA). Sidak’s test was applied to adjust the significance level of multiple comparisons. Statistics were performed using Graph Pad Prism 9 (2021/22) software. Differences were considered significant at P < 0.05. Graphics were prepared using Adobe Illustrator (2021), Sigma Plot (2000) and Graph Pad Prism 9 (2021/22).

Results

Unitary events recorded from Px1-YFP expressing HeLa cells

HeLa cells expressing human Px1-YFP (HeLa Px1-YFP) or rat Cx43-CFP (HeLa Cx43-CFP) (Figure S1a–c) offer simple models to record unitary events of Px1 HCs and Cx43 HCs, with less interference from activity of other ion channels. We improved signal to noise ratio and calculation of unitary conductances in single channel recordings by using a dual patch clamp technique. In this approach, one pipette was in cell attached mode to record activity of single channels in the limited area of the patch, and the other was in whole cell mode to clamp V_m in the whole cell (see diagram in Figure S2a). The whole cell pipette also recorded whole cell activity, including that of channels in the cell attached patch (Figure S2b–d).

HC openings were evoked with ramps of 1.7 s duration and –100 to +100 mV amplitude delivered every 2 s (Figure 1a, b). The signal to noise ratio of single channel recordings generally was greater in the cell attached patches than in the whole cell recordings (Figure S2). We computed ramp-evoked HC openings over a broad range of unitary conductances (25–110 pS). The means calculated from these distributions were 50.8 ± 4.2 pS and 48.3 ± 3.8 pS, at negative and positive V_m, respectively (P = 0.657) (Figure 1c).

HeLa cells express, at very low levels, endogenous Px1 (Imamura et al., 2020; Penuela et al., 2013). Accordingly, voltage ramps applied through the whole cell pipette produced Px1 HC openings in a larger percentage of cell attached patches of HeLa Px1-YFP cells (~80%) than of parental cells (~20%, Figure 1d, P < 0.05). Ramp polarity affected the total open time and total closed time between openings, suggesting a voltage-dependent gating. Open time of Px1 HC openings was greater at positive V_m than at negative V_m, whereas closed time between openings was lower at positive than negative V_m (Figure 1e, P < 0.05).

On average, the first Px1 HC opening in HeLa Px1-YFP cells occurred after 20.5 ± 6.5 s of starting a train of 0.5 Hz voltage ramps (Figure S3a–d, n = 26 patches, one patch per cell). After the first opening occurred, ramp-evoked HC openings were of shorter latency (< 2 s), and the number of Px1 HC openings, quantified from 10 consecutive ramps per patch, on average, was not affected by the polarity of the ramps (Figure 1f, P > 0.05).

Single channel properties were also determined at constant V_m in HeLa Px1-YFP cells after activation of HCs with a train of voltage ramps (Figure 2a–f). Reminiscent of the voltage-dependent activity recorded during ramps in Fig.1e, the total open time of Px1 HCs was greater at positive than negative V_m, whereas the closed time between openings were shorter at positive than negative V_m (Figure 2g, P < 0.05, positive vs. negative voltages). We confirmed voltage-dependent changes in open probability (Po) in the cell attached patches from the product NPo, which was estimated from recordings of long duration ( > 20 s) at V_m held constant at −/+ 80 mV. Po of Px1 HCs was significantly higher at positive than negative V_m (Figure 2h).

Figure 2.

Single channel activity of Px1 HCs is sensitive to membrane voltage. (a-f) After evoking Px1 HC openings with a 0.5-Hz train of voltage ramps, V_m was clamped through a whole cell pipette (V₂) for longer duration (~50 s) than the voltage ramp cycles shown in Fig. 1. (a) Single channel activity recorded in the cell attached pipette (V₁) with V_m clamped at – 80 mV through a whole cell pipette. (b, c) Magnified traces of single channel currents (b) and unitary conductances (c) calculated from experiments like those in a. (d-f) Single channel activity recorded as in a, but with V_m clamped at +80 mV through V₂. (e, f) Magnified traces of single channel currents (e) and unitary conductances (f) calculated from experiments like those in d. (g) Total open time and total closed time between openings at V_m held constant at −/+80 mV (n = 6, 6, patches, one patch per cell). (h) NPo estimation, i.e., total time one channel is open/ time of recording. V_m held constant to −/+ 80 mV through V₂. Only recordings longer than 20 s and with one active channel in the patch were quantified. (i) Single channel I-V curves from recordings as those shown in a,d, but of shorter duration (5–10 s, see also Fig. S4). On average, the slope of I-V curves from various HeLa Px1-YFP cells was 42.1 ± 4.1 pS (n = 13 patches, one patch per cell), which was comparable to the values directly calculated in (c,f) from the current traces in (b,e). Currents recorded in V₁ were multiply by −1. * P < 0.05, Wilcoxon matched-pairs signed rank test was used in (g). Mann Whitney test was used in (h).

Active channels in HeLa Px1-YFP cells showed a linear single channel I-V relationship in which the mean slope was 42.1 ± 4.1 pS ( n = 13 slopes computed from 13 patches, one patch per cell, Figure 2i, and Figure S4). The slope calculated from these I-V curves was not significantly different from the values calculated from voltage ramps (P > 0.05). Although most of the data points (87%) computed from voltage ramps distributed from 25–90 pS, ~13% of openings scored were of high conductance (90–110 pS) (Figure 1e), broadening the range of unitary events recorded during ramps, as compared to voltage steps. We could not dissect whether these high conductance outputs corresponded to the openings of single channels, or in contrast, to the opening of duplets or triplets during ramp stimulation, i.e., two or three overlapping channel openings that cannot be distinguished as single channels during the ascending phase of the voltage ramp. For this reason, we opted to use the broadest range of conductances recorded during ramps, i.e., 25–110 pS, as a reference to identify Px1 HCs in astrocytes, rather than the average conductances of ~42 pS and ~48 pS.

Mechanical stress on the cell surface might alter voltage sensitivity of channels in the patch, and cell dialysis caused by the whole cell pipette could affect Po. As a control, in separate experiments, we monitored single channel activity in whole cell recordings, but not using a second, cell attached pipette. Although we used symmetrical conditions in dual voltage clamp and asymmetrical in single clamp whole cell recordings, conductances were comparable, although not identical (Figure S5).

Pharmacological block of Px1 HCs in HeLa Px1-YFP cells

To confirm that Px1 HCs mediate unitary currents recorded from cell attached patches in HeLa Px1-YFP cells, we inhibited channel activity pharmacologically. Agents in the cell attached pipette rapidly reach the outside of the cell membrane under the pipette. We added hemichannel blockers to the cell attached pipette solution and measured the channel activity and conductance in the patch 1 min after gigaseal formation. Channels were recorded as in Fig. 1b. The electrical conductance in the cell attached patch was calculated by measuring the slope of the I-V curve, where I in this case, was the current flowing through the cell attached pipette, and V_m was the voltage clamped through the whole cell pipette (V₂ - V₁). The conductances of channels in the patches of HeLa Px1-YFP cells exposed to control solution were significantly greater than those of cell attached patches of HeLa Px1-YFP cells to which CBX (0.05 mM) or probenecid (1 mM), two Px1 HC blockers, were applied (Figure 3a–c). CBX blocks gap junctions and Cx HCs as well as Px HCs (Garré et al., 2010), whereas probenecid blocks Px HCs with little effect on Cx HCs (Silverman et al., 2008). CBX and probenecid blocked more completely (i.e., traces show no single channel transitions) during the negative phase of the ramps than during the positive phase (Figure 3b, c) (negative phase, P < 0.001 probenecid and CBX vs. control; positive phase, P < 0.001 and P < 0.01, probenecid and CBX vs. control, respectively). Presumably, additional kinds of channels (or Px1 HCs that were not blocked completely) contributed to the positive phase. Furthermore, the number of active channels was reduced in patches exposed to CBX and probenecid, as compared to untreated cells (Figure 3d) ( P < 0.05). Together, these data indicate that the major single channel activity recorded in the cell attached patches in HeLa Px1-YFP cells was mediated by Px1 HCs.

Figure 3.

Px1 HCs are sensitive to carbenoxolone (CBX) and probenecid. (a) Pharmacological block of Px1 HCs in cell attached patches. CBX also blocks Cx HCs. Probenecid is a Px1 HC blocker with no effect on Cx HCs. In these I-V curves like those in Fig. 2, I was the current in the cell attached pipette, and V_m was the cell membrane voltage which was clamped through the whole cell pipette (V_m = V₂-V₁). Single channel openings (~ 40 pS) at negative voltages are indicated with black arrows in HeLa Px1-YFP control. Two channels (~ 80 pS) were open at positive V_m in HeLa Px1-YFP control. HeLa Px1-YFP treated with probenecid (1mM) or CBX (50 μM) showed no obvious single channel activity. (b) Conductance (pS) in the patch during the more negative phase of the voltage ramp (−50 to −100 mV) in HeLa Px1-YFP control (90.5 ± 13.0 pS) and treated with probenecid (1.1 ± 0.4 pS) or CBX (4.5 ± 1.1 pS) (n = 20 ramps, from 4 different patches and cells, probenecid and CBX vs. control). (c) Conductance in the patch during the positive phase of the voltage ramp (+50 to 100 mV) in HeLa Px1-YFP control (120.8 ± 21.8 pS) and treated with probenecid (39.4 ± 6.8 pS) or CBX (62.0 ± 5.6 pS) (n = 20 ramps, from 4 different patches and cells, probenecid and CBX vs. control, respectively). (d) Maximum number of channel openings observed per patch ( n = 5 patches, 10 ramps per patch, one patch per cell). Drugs were added to the cell attached pipette solution. Currents recorded in V₁ were multiply by −1. Data are presented as dot plots and the mean ± SEM. *P < 0.05, ***P < 0.001, one-way ANOVA followed by Sidak’s multiple comparisons test was used in (b, c, d).

Single channel properties of Px1 HCs and Cx43 HCs in rodent spinal astrocytes

In contrast to HeLa cells (Figure S5), our whole cell recordings in astrocytes did not show a signal to noise ratio adequate to discriminate small transitions < 50 pS, which would include unitary conductances for Px1 HCs and other channels of lower conductance such us P2X7Rs. Having used a dual voltage clamp technique in HeLa cells to record HCs, we asked whether the same technique would allow activity of Px1 HCs to be distinguished from that of Cx43 HCs and P2XRs in rat and mouse astrocytes.

We applied ramp stimulation (2 s ramp cycle, stimulation 1.7s, −/+ 80 mV) through a whole cell clamp and recorded channel activity in cell attached patches in spinal astrocyte cultures (Figure 4a–c). We first assigned channel activity to Px1 HCs, Cx43 HCs or P2X7Rs based on the expected unitary conductances from our recordings in HeLa cells and reported in the literature. Events ranging 25–110 pS were assigned to Px1 HCs (Figure 1c), and 160–250 pS events were assigned to Cx43 HCs (Figure S6). Events ranging 10–22 pS were assigned to rodent P2X7Rs, based on the unitary conductance of human P2X7Rs (Riedel et al., 2007). We confirmed the molecular identity of these channels based on single channel activity during pharmacological treatment and after gene depletion of Px1 or Cx43.

Figure 4.

FGF-1 differentially modulates single channel activity of Px1 HCs, Cx43 HCs, and P2X7Rs. (a-c) Channel transitions were recorded in the cell attached patch (V₁) during ramps of 1.7 s duration and −/+ 80 mV amplitude applied through a whole cell pipette (V₂). Representative single channel recordings of Px1 HCs (a), Cx43 HCs (b), and P2X7Rs (c) in rat spinal astrocytes after 7 h FGF-1 treatment. Red and blue dashed lines indicate the main open (O) and closed states (C), respectively. Insets show single channel transitions (pS) from the current traces in a-c, at negative (blue arrow, blue inset) and positive (red arrow, red inset) V_m. C, close state, O_{1 ,} one channel open, O_2, two channels open. (d) Distributions of active P2X7Rs, Px1 HCs and Cx43 HCs in astrocytes maintained in control solution and after treatment for 7 h with FGF-1. Events were grouped in 3 statistically different categories: single channel conductances of 10–22 pS (P2XRs including P2X7Rs), 25–110 pS (Px1 HCs), 160–250 pS (Cx43 HCs) ( n = 18, 36, n = 62, 36, n = 36, 18, n = 62, 18, patches, one patch per cell, control and FGF-1, respectively). Unitary events < 10 pS and ranging from 120 to 150 pS were undetermined. (e) Percentage of ramps with active P2X7Rs, Px1 HCs, and Cx43 HCs in control and 7 h FGF-1-treated astrocytes (n = 14, 18, n = 11, 20, n = 12, 12 , patches, one patch per cell, control and FGF-1, respectively). (f and g) distributions showing total open time and total closed time between openings during ramps of negative and positive polarity for Px1 HCs (f) and P2X7Rs (g) in astrocytes maintained in control medium for 7 h (n = 6, 6 patches in f; n = 5, 4 patches in g, open and closed times, respectively, one patch per cell). (h-j) Total open time and total closed time between openings of Px1 HCs (h), Cx43 HCs (i), P2X7Rs (j) in astrocytes control and treated with FGF-1 for 7h. Cx43 HCs were very rare in control astrocytes and not quantified in (i). Currents recorded in V₁ were multiply by −1. Data are presented as dot plots in (d,f,g) and mean ± SEM in (e,h,i, j). *P < 0.05, **P < 0.01, ***P < 0.001. One-way ANOVA followed by Sidak’s multiple comparisons test was separately used for control and FGF-1 in (d). Mann Whitney test was used in (e,h,i,j). Wilcoxon matched-pairs signed rank test was used in (f, g).

Single channels in the range of conductances expected for Px1 HCs and P2X7Rs were recorded in astrocytes maintained in control medium for 7 h during stimulation with short duration voltage ramps. Cx43 HC openings were very rare. Accordingly, in HeLa cells transfected with rat Cx43-CFP, we failed to see the predicted single channel activity of Cx43 HCs (~220 pS) with short ramp stimulation (Figure S6a, S6b, and S6f). Cx43 HC openings were recorded during 8 s ramp cycles of the same amplitude, mostly at positive voltages (Figure S6c–f), indicating Cx43 HC Po is very low at negative V_m. However, in spinal astrocytes treated with 7h FGF-1, short duration ramps evoked openings of conductance expected of Cx43 HC, indicating increased single channel activity of Cx43 HCs. Representative recordings of these three surface channels expressed by spinal astrocytes after 7 h treatment with FGF-1 are shown in Figure 4a–c. The spectrum of unitary events changed amplitude and frequency after treating astrocytes for 7 h with FGF-1, as compared to control astrocytes (Figure. S7a). The mean unitary conductances of the single channel events assigned to Px1 HCs, Cx43 HCs and P2X7Rs were significantly different from each other in both control and FGF-1-treated astrocytes (Figure 4d) (P < 0.001). The percentage of ramps with active Px1 HCs was reduced to ~30%, as compared to control astrocytes (~60%) (P < 0.05). In contrast, the percentage of ramps with active Cx43 HCs was increased to ~19%, as compared to astrocytes maintained for 7 h in control medium alone (< 1%) (P < 0.05). Both Px1 HC and Cx43 HC openings mediate ATP release, which in turn activates P2XRs. Accordingly, the percentage of ramps with active P2X7Rs was increased from ~ 16 % in control to 61 % after 7 h FGF-1 treatment (P < 0.01) (Figure 4e).

Similar to HeLa Px1-YFP, in astrocytes the fraction of time Px1 HCs remained in the open configuration was higher at positive than negative V_m, whereas the times in the closed states between openings decreased (Figure 4f, P < 0.05). Unlike Px1 HCs, total open times and total closed times between openings for P2X7Rs were voltage independent (Figure 4g, P > 0.05). There were no significant differences in Px1 HC open and closed times between control and FGF-1-treated spinal astrocytes (Figure 4h, P > 0.05, control vs. FGF-1). Cx43 HCs in astrocytes treated with FGF-1 opened for ~8 % and ~15 % of the ramp cycle duration (2 s), at negative and positive voltages, respectively (Figures 4i). In contrast to Px1 HCs, open time of P2X7Rs in astrocytes treated with FGF-1 was higher than in control, at both positive and negative V_m (Figure 4j, P < 0.05).

In ~10 consecutive ramps, the number of Px1 HC, Cx43 HC and P2X7R openings was not affected by the polarity of the ramps in control and FGF-1-treated astrocytes (Figure 5a–c). Nevertheless, 7 h treatment with FGF-1 increased the number of P2X7R openings per ramp, but had no effect on Px1 HC openings (Figure 5c, P < 0.05). Because in control astrocytes Cx43 HC openings were very rare (<1%), total open time, closed time between openings, and number of openings per ramp were not quantified (Figure 4e and 5b).

Figure 5.

Ramp polarity does not affect the number of Px1 HC, Cx43 HC and P2X7R openings. Single channel openings in spinal astrocytes recorded like those in Fig.4. (a) Mean number of Px1 HC openings per ramp in spinal astrocytes control and after 7 h FGF-1 ( n = 6 patches, one patch per cell, 10 ramps per patch). (b) Mean number of Cx43 HC openings per ramp in spinal astrocytes treated 7 h with FGF-1 (Cx43 HC, n = 4, one patch per cell, 10 ramps per patch). Cx43 HC openings in control astrocytes were not quantified, because these were very rare (< 1 %). (c) Mean number of P2X7Rs openings per ramp in spinal astrocytes control and after 7 h FGF-1 ( n = 6 patches, one patch per cell, 10 ramps per patch). FGF-1 increased the number of P2X7Rs openings in spinal astrocytes. Data are presented as mean ± SEM. ns, not significant, *P < 0.05, Mann Whitney test was used in (b), one way ANOVA was used in (a and c), followed by Sidak’s multiple comparisons test in (c) (control vs. FGF-1, −80 to 0 mV ramp; control vs. FGF-1, 0 to +80 mV ramp).

These data indicate that FGF-1 differentially modulates the single channel activity of Px1 HCs, Cx43 HCs and P2X7Rs, suggesting purinergic signaling in spinal astrocytes is adjusted during neurodevelopment, inflammation, and repair (see discussion).

Gene depletion and pharmacological block of HCs in spinal astrocytes

To confirm the identity of HCs, we conducted dual patch clamp experiments in Px1- deficient and Cx43^−/− astrocytes. The percentage of ramps with active 25–110 pS channels was reduced in spinal astrocytes transfected with stealth siRNA against Px1 (siRNA Px1), as compared to astrocytes transfected with a non-targeting siRNA (siRNA(−)), indicating that these unitary events were mostly mediated by Px1 HCs (Figure 6a–c). Of note, siRNA Px1 treatment did not prevent openings of < 25 pS channels, suggesting this activity was not mediated by Px1 HCs. Furthermore, in Px1-deficient spinal astrocytes we observed a compensatory increase on the single channel activity of other channels, including P2XRs and Cx43 HCs (Figure 6a).

Figure 6.

Px1 HCs are functional in spinal astrocytes prepared from neonate spinal cords. (a) Single channels were recorded in the cell attached pipette (V₁) whereas V_m was controlled through a whole cell clamp (V₂) after 24 h transfection of rat spinal astrocytes with siRNA (−) and siRNA Px1. Single channel transitions were evoked by voltage ramps of 1.7 s and −/+ 80 mV. (b) Mean unitary conductance of Px1 HCs was not altered after transfection with siRNA (−) and siRNA Px1 (n = 32, n = 5, ramps with 25–110 pS channels, P > 0.05). (c) Percentage of ramps with active Px1 HCs (with transitions ranging 25–110 pS) was reduced after 24 h transfection with siRNA Px1 ( n = 45 ramps, 5–10 ramps/patch, 6 patches, one patch per cell ). (d) After evoking Px1 HC openings with a 0.5-Hz train of voltage ramps, V_m was clamped through V₂ at −70 mV. Representative single Px1 HC transitions recorded from spinal astrocyte controls. (e) On average, single channel transitions like those shown in d were 48.8 ± 5.0 pS (n = 6 patches, one patch per cell). Quantification from various experiments in which V_m was clamped at negative voltages, between −60 to −80 mV. (f) Single channel currents before and during bath application of probenecid (1 mM) were recorded in the cell attached pipette. V_m clamped at - 70 mV. (g) Single channel activity of Px1 HCs (red traces) is blocked by probenecid (green traces). Currents recorded in V₁ were multiply by −1. C, close state, O₁, open state. Data are presented as dot plots in (a, b, e). Mean ± SEM is included in c. **P < 0.01, ns, not significant, Mann Whitney test was used in (b, c).

To record Px1 HC openings with less interference from the opening of Cx43 HCs expressed by spinal astrocytes, single channel conductance of Px1 HCs in spinal astrocytes was also determined at constant negative V_m (in astrocytes maintained in control medium, Cx43 HC Po at negative V_m is very low). The mean unitary conductance of Px1 HCs in spinal astrocytes was 48.8 ± 5.0 pS, which was not significantly different from that recorded in HeLa Px1-YFP, 42.1 ± 4.1 (P > 0.05, Mann Whitney test) (Figure 5d, e). This single channel activity was blocked after bath application of probenecid (1 mM), confirming that these channel openings were mediated by Px1 HCs (Figure 6f, g).

The distribution of channel conductances in mouse astrocytes maintained in control medium was similar to that in rat astrocytes (Figure S7a,b). In culture spinal astrocytes prepared from Cx43 KO mouse embryos, we also recorded Px1 HC activity at constant negative V_m that was reduced after bath application of CBX (200 μM), further indicating that these channel openings were primarily mediated by Px1 HCs (Figure 7). As expected, Cx43 HC openings were not observed in Cx43−/− spinal astrocytes (Figure 7b and Figure S7b). Astrocytes express other connexins that potentially may form HCs, e.g., Cx30. However, Cx30 is expected to form HCs of larger unitary conductance (~300 pS) (Vogel et al., 2006) and is not detected in the CNS until P10 (Ribot et al., 2021). As compared to Cx43 or Cx30, Cx45 expression is negligible in astrocytes, and although compensatory upregulation might occur in Cx43 KO (Dermietzel et al., 2000), Po of Cx45 HC is also very low at negative V_m.

Figure 7.

Px1 HCs are functional in spinal astrocytes prepared from Cx43 KO mouse embryos. (a) Single channel currents before and during bath application of carbenoxolone (CBX) were recorded in the cell attached pipette (V₁) while V_m was clamped with a second whole cell pipette (V₂) at - 40 mV. After putative Px1 HC activity was recorded (~ 40 pS), the Px1 HC and Cx43 HC blocker CBX (200 μM) was applied to the bath to confirm channel identity. (b) Single channel distribution 40 s before (black) and 40 s after (red) CBX application. CBX reduced single channel transitions in the range of Px1 HC (25–110 pS). Single channel transitions in the range of P2X7Rs were not affected by CBX. Single channel transitions in the range of Cx43 HCs were not detected in Cx43^−/− mouse spinal astrocytes. (c-f) Single channel currents (c, d) and unitary conductances (e, f) from the traces in (a) before (c, e) and after (d, f) application of CBX. Currents recorded in V₁ were multiply by −1. C, closed state. O1, O2, O3, one, two, and three Px1 hemichannels opened, respectively. Single channel transitions in the range of P2X7Rs (10–22 pS) are indicated with black arrows in (d, f).

In summary, using both genetic and pharmacological tools we confirmed that the mean single channel conductance of Px1 HCs is ~48 pS at near physiological V_m, which was comparable to the mean values recorded in HeLa Px1-YFP (~42 pS and ~48 pS, voltage steps and voltage ramps, respectively).

Of note, in rat and mouse spinal astrocytes, we found another type of high unitary conductance channel; these channels opened with unitary conductances ranging 120–150 pS in astrocytes in both control and after 7 h FGF-1 (Figure S7a,b). These transitions were unlikely to mediated by Px1 HCs and Cx43 HCs, because they were not observed in HeLa Px1-YFP cells (Figure 1e), but were often detected in patches from Px1-deficient and Cx43^−/− astrocytes (Figure 6a and S7). Further studies will be required to determine the molecular basis of these channel openings.

Single channel properties of P2X7Rs in rodent spinal astrocytes

Using dual voltage clamp of rodent spinal astrocytes, we recorded unitary transitions < 25 pS resembling the single channel responses of P2XRs including P2X7Rs (Benham & Tsien, 1987; Riedel et al., 2007). The mean unitary conductances of P2X7Rs calculated from dual voltage clamp experiments in rat spinal astrocytes in control and after 7 h FGF-1 treatment were not significantly different, 16.9 ± 3.5 pS and 17.3 ± 4.8 pS, respectively (P > 0.05). Where P2X7R single channel transitions were identified during ramps in dual voltage clamp recordings, the base line current was not linear, likely because of distinct active channels in the patch (Figure 4c). The non-linearity in the current during the ramp was more evident at larger voltages (−50 to −100 mV and + 50 to +100 mV, Figure 4c).

We used cell attached patch single clamp (i.e., without controlling V_m through a second pipette) to confirm that P2X7Rs are functional in spinal astrocytes. Small unitary events were more frequent in patches when 50 μM or 500 μM ATP was added in the recording solution, as compared to control patches without ATP (Figure S8a–c). In cell attached single voltage clamp experiments, the unitary events were ~2.5 pA at V_p = + 90 mV, corresponding to unitary conductances of 22 ± 4 pS (n = 6 patches, one patch per cell). Under experimental conditions similar to those used in dual voltage clamp experiments in HeLa cells, the mean resting potential of cultured astrocytes recorded in separate experiments using current clamp mode in whole cell recordings was - 21.1 ± 2.6 mV. For these recordings, we used an internal recording solution containing 5 mM TEA, a K⁺ channel blocker that improved signal-to-noise ratio in whole cell recordings, but also reduced the resting potential of astrocytes (Ransom & Sontheimer, 1995).

To study whether responses of P2X7Rs to ATP are affected after FGF treatment, we also used cell attached single voltage clamp but adding ATP to the recording solution for spinal astrocytes in control and after 7 h FGF-1 (Figure S8d). In control astrocytes, the percentage of trials (voltages pulses, + 90 mV relative V_m, ~10 s duration) with active P2X channels increased in experiments in which 50 μM (41.5 ± 6.6 %) or 500 μM ATP (33.8 ± 7.3 %) was added to the cell attached pipette (50 μM, P < 0.05, 500 μM, P = 0.05), as compared to results in the absence of ATP (9.2 ± 4.3%). Similarly, in astrocytes treated with FGF-1 for 7 h, the percentage of trials with active P2X7Rs was increased when ATP (50 μM) was present in the recording solution (33.8 ± 6.9 %), as compared to control astrocytes in which ATP was not added to the cell attached pipette (P = 0.05) (Figure S8e). These data indicate FGF-1 treatment did not affect the responses of P2X7Rs to ATP, at the concentrations used in this study.

To confirm P2X7R mediation, an antagonist of these receptors, Brilliant Blue G (BBG, 10 μM), was added together with ATP in the cell attached pipette. Events of < 25 pS were absent in both control astrocytes and those treated with FGF-1 for 7 h. Moreover, at this high concentration of BBG, Px1 HCs and Cx43 HCs were functional in both control and FGF-1 treated astrocytes (Figure S8f). These data indicate that unitary currents (~22 pS) induced by high ATP concentrations (50, 500 μM) are primarily mediated by P2X7Rs, although other purinergic channels expressed by astrocytes at lower levels, including P2X4Rs and P2X2Rs, may contribute (P2rx7:P2rx4:P2rx2 relative gene expression in mature astrocytes is ~50:36:1, respectively, source: www.brainrnaseq.org).

Discussion

We determined single channel properties of Px1 HCs in spinal astrocytes using a dual voltage clamp technique with improved resolution, i.e, recording a reduced number of channels in a small area of cell membrane while controlling V_m in the whole cell. Using this approach, we compared the single channel properties of Px1 HCs with other channels involved in purinergic signaling and expressed by astrocytes, i.e, Cx43 HCs and P2X7Rs. Px1 HCs were effectively activated by alternating V_m with a train of low frequency voltage ramps. Px1 HCs could be differentiated from Cx43 HCs and P2X7Rs based on unitary conductance, voltage dependence, and activity after pharmacological and gene silencing. On average, Px1 HC single channel conductance was ~48 pS in rodent spinal astrocytes, which was comparable to that recorded in HeLa Px1-YFP cells (~42 pS). Single channel activity of Px1 HCs, Cx43 HCs and P2X7Rs in spinal astrocytes was differentially modulated by FGF-1, suggesting that the activity of these channels may be fine-tuned during neurodevelopment, inflammation and repair.

There is no agreement as yet on the single channel conductance of Px1 HCs. Previous reports stated that the maximum unitary conductance of human Px1 HCs in oocytes was ~500 pS. These data led to the idea that vertebrate Px1 HCs may form “pores” of large electrical conductance in the membrane. These “pores” could well be big enough to accommodate large inflammatory molecules including ATP and IL1-β as well as large dyes such as YO-PRO (Bao et al., 2004; Pelegrin & Surprenant, 2006). If activated in neurons, opening of these non-selective channels would drive neuronal and circuit dysfunction (Thompson et al., 2008). A later study suggested that Px1 HCs form anion selective channels of smaller unitary conductance (~ 70 pS) in mammalian cell lines, as evaluated in outside out patches (Ma et al., 2012). To explain this disparity, it was suggested that Px1 may have two different open channel configurations depending on the concentration of extracellular potassium, [K]_o. It is likely that the two different conductances of Px1 HCs would have different permeabilities to ATP; relatively small channels of 50 pS would likely be ATP-impermeable, whereas larger channels of 300–500 pS (activated by high [K]_o) would permit ATP efflux (Wang et al., 2014). In disagreement with these ideas, a recent study showed that the single channel conductance of Px1 HCs is far below 300–500 pS, and that channel openings of lower unitary conductance may be ATP permeable (Chiu et al., 2017). This study, which was conducted on excised inside-out patches, describes a caspase-3-dependent mechanism of Px1 HC activation, with channel openings of 12 pS and 96 pS, at negative and positive V_m, respectively. Other study suggests that physiological elevation in [Ca⁺²]_i is sufficient to open Px1 HCs with unitary conductances of ~45 pS via C-terminus phosphorylation by CaMKII. Px1 HC openings recorded in excised outside-out patches also showed outward rectification (Lopez et al., 2021).

In dual voltage clamp recordings, Px1 HCs were activated by alternating V_m with a train of 0.5 Hz voltage ramps. However, on average, the latency to the first opening in HeLa Px1-YFP cells was ~10 times the duration of one ramp cycle, suggesting gating with unusual voltage dependence. Unitary conductances of Px1 HCs computed from voltage ramps stimuli ranged from 25 to 110 pS in HeLa Px1-YFP cells, with similar averaged values at positive and negative voltages, ~48 pS and ~50 pS, respectively. After activation with voltage ramps, unitary conductance of Px1 HCs was determined in recordings of longer duration than 2 s ramps ( > 10 s), but at constant V_m. On average, Px1 HC unitary conductances were ~42 pS and ~48 pS in HeLa Px1-YFP and astrocytes, respectively. We confirmed that these unitary currents were mediated by Px1 HCs in HeLa Px1-YFP cells by pharmacology, i.e., they were blocked by probenecid or carbenoxolone. The single Px1 HC I-V curves show no signs of outward rectification, which is in contrast to the strong rectification observed in excised inside-out or outside-out patches (Chiu et al., 2017; Lopez et al., 2021). Furthermore, single Px1 HC I-V curves with no outward rectification and single channel conductances of ~100 pS and ~190 pS were recently recorded after reconstitution of Xenopus Px1 in lipid bilayers in symmetric ionic conditions (Narahari et al., 2021). Future studies using dual voltage clamp to characterize Px1 HC activation in the whole cell using both symmetrical and various asymmetrical conditions would help to clarify these discrepancies.

We believe that our study is the first to describe properties of Px1 HCs with single channel resolution in glial cells. Using dual voltage clamp, Px1 HC identity was confirmed by their absence in Px1-deficient spinal astrocytes and sensitivity to CBX and probenecid. Furthermore, we described a voltage-dependent activity in HeLa cells that facilitated the identification of Px1 HCs in astrocytes, i.e., the total open times and closed times between openings were voltage dependent in both HeLa Px1-YFP cells and astrocytes. Our data indicate that the single channel activities of ~42 pS and ~48 pS recorded in HeLa Px1-YFP and spinal astrocytes, respectively, comprised the main functional state of Px1 HCs. We found no evidence of pores of higher electrical conductance, i.e., 300–500 pS, in either HeLa Px1-YFP or spinal astrocytes using dual voltage clamp and single voltage clamp whole cell and cell attached patch recordings. Of note, recordings were conducted in symmetrical conditions, i.e, using internal recording solutions (with high K⁺) in both the cell attached patch and whole cell pipettes, arguing against an effect of high [K]_o on inducing Px1 HC openings of large conductance (300–500 pS). Our data call for caution regarding the extrapolation of various Px1 HC configurations from dual whole cell recordings from pairs of electrically coupled cells via Px1 GJ channels (Palacios-Prado et al., 2022). In these recordings, a whole cell pipette monitors whole cell channel activity including that of Px1 GJ channels and unopposed Px1 HCs, and analysis of both unitary activities may be challenging. This problem was not prominent in studies of Cx43 GJ channels, because Po of unopposed Cx43 HCs is extremely low (< 0.01). Our dual voltage clamp approach in cellular systems deprived of cell-cell coupling (single cells) would be a better choice to determine the functional state of unopposed Px1 HCs.

There is less debate about the single channel properties of Cx43 HCs. In dual voltage clamp experiments in HeLa Cx43-CFP and spinal astrocytes, Cx43 HCs open with distributions ranging from 160 to 250 pS, which is consistent with the conductance reported in a previous study using single clamp whole cell recordings (Contreras et al., 2003). Furthermore, these unitary currents were absent in Cx43^−/− spinal astrocytes.

Cx43 HCs and Px1 HCs mediate ATP release from astrocytes (Kang et al., 2008; Narahari et al., 2021), which in turn activates ATP-gated ion channels (P2XRs) and G-protein coupled receptors (P2YRs) expressed by neurons and various glial cell types (Khakh & North, 2012). Interestingly, our dual voltage clamp technique also allowed us to record ion channels of smaller conductances (~ 17 pS) in cultured spinal astrocytes. We confirmed that these small unitary currents were primarily mediated by P2X7Rs using cell attached patch clamp and pharmacology, i.e., activating single channels in the patch with ATP and blocking them with BBG, a P2X7R antagonist.

In this study we also show that FGF-1 differentially modulates the single channel activity of Px1 HCs, Cx43 HCs, and P2X7Rs in spinal astrocytes. Notably, we found that single channel activity of Px1 HCs was decreased by 7 h FGF-1 treatment in spinal astrocytes, as evidenced by a reduced number of ramps with active Px1 HCs. A fraction of Px1 HCs was still active during voltage ramps after this treatment, explaining the Px1-dependent mechanism of dye uptake observed after 7 h FGF-1 in a previous study (Garré et al., 2010). Although Po of Cx43 HCs is very low, single channel activity was detected in spinal rat and mouse astrocytes after 7 h FGF-1 treatment using 2 s voltage ramp stimuli. In contrast, Cx43 HCs openings were negligible in astrocytes maintained in control medium. These data indicate that the open probability of Cx43 HCs was increased after treatment of astrocytes with FGF-1; this increase was due to a greater number of active HCs per patch and mean open times, with no changes in surface Cx43 (Garré et al., 2010).

FGF signaling has multiple roles in normal brain and spinal cord development (Diez Del Corral & Morales, 2017; Ford-Perriss et al., 2001) and dysregulation in FGF systems has been associated with neurodevelopmental and psychiatric disorders (Turner et al., 2016). In the adult CNS, FGF-1 may contribute to repair by suppressing astrocyte activation after brain injury (Kang et al., 2014). Although FGF-1 and FGF-2 promote axonal growth and functional recovery in spinal cord injury (SCI) models, FGFs in Amyotrophic Lateral Sclerosis (ALS) favor motor neuron death and accelerate disease onset (Cassina et al., 2005; Klimaschewski & Claus, 2021), suggesting a bimodal role of FGF signaling in spinal cord diseases. Beyond neurodevelopment and repair, FGF-1 is involved in inflammation (Xie et al., 2020) and modulates neuro-glia-immune communication in the spinal cord, for example, by inducing ATP release from astrocytes through Px1 HCs and Cx43 HCs (Garré et al., 2010; Garré et al., 2016). Openings of Px1 HCs and Cx43 HCs induced by FGF-1 cause regenerative ATP release from spinal astrocytes that activates microglia to produce IL1-β in spinal cord slices (Garré et al., 2016). In vivo, a similar neuroimmune response may be amplified by leptomeningeal macrophages, which express higher surface level of P2X7Rs than microglia (Garré et al., 2020), and possibly sense tissue damage or neuronal dysfunction when exposed to lower extracellular ATP concentrations. Our data provide molecular evidence that FGF-1 increases the single channel activity (total open time and number of openings per patch) of P2X7Rs in cultured astrocyte monolayers, likely as a consequence of ATP release mediated by Px1 HCs and Cx43 HCs. Copious amounts of ATP are also released from astrocyte HCs at early stages after traumatic spinal cord injury (Huang et al., 2012), which may activate P2X7Rs in infiltrating myeloid cells (Di Virgilio et al., 2017). P2X7R activation in other types of cells than astrocytes causes neuronal dysfunction in neurodegeneration (Wang et al., 2004), activation of NLRP3 inflammasome in leptomeningeal macrophages, and social behavioral deficits in Rett syndrome, a rare neurodevelopmental disorder (Garré et al., 2020). In addition to modulate communication between neurons, astrocytes and brain’s myeloid cells, FGF-1 may be involved in other forms of neuroimmune interactions during inflammatory conditions, for example, providing a co-stimulatory signal for T cell expansion and activation of CD8 T cells (Byrd et al., 1999). In this regard, antigen-specific CD8 T cells infiltrate in the spinal cord and contribute to pathogenesis in ALS-4, a juvenile form of ALS (Campisi et al., 2022). Future studies should determine whether FGFs modulate single channel activities of HCs and P2X7Rs and immune responses in disease-associated T cells.

In summary, we determine single channel properties of Px1 HCs, Cx43 HCs, and P2X7Rs, and provide molecular insights into how FGF-1 modulates channel activities in spinal astrocytes. By using a dual voltage clamp technique with improved resolution, we were able to record a small number of channels on the cell surface under strict control of V_m in the whole cell, avoiding confounding “phenotypes” caused by overlapping of various channel openings from whole cell activity.

Main points.

• Single channel properties of astrocyte Px1 HCs, Cx43 HCs and P2X7Rs are determined with a patch clamp technique of improved resolution

• In spinal astrocytes, unitary activity of Px1 HCs, Cx43 HCs, and P2X7Rs is differentially modulated by FGF-1

Data availability statement

The data that support the findings of this study are available from the authors upon reasonable request.