High Activity of Hemichannels Permeable to Calcium Ions Leads to ROS Generation and Reduced Cell Viability

TL;DR

High activity of certain cell channels lets calcium in, causing cell damage and death through increased oxidative stress.

Contribution

Identifies Cx45 as a new calcium-permeable hemichannel and links its activity to oxidative stress and cell death.

Findings

Cx45 hemichannels allow calcium influx, leading to increased cytosolic calcium levels.

Elevated hemichannel activity under alkaline conditions causes lipid peroxidation and reduced cell viability.

Calcium influx through Cx43 and Cx45 hemichannels triggers oxidative stress and cell death.

Abstract

Connexins (Cxs) and pannexin1 (Panx1) form hemichannels (HCs) that enable the exchange of ions and small molecules between the intracellular and extracellular compartments. Since an elevated cytoplasmic Ca2+ concentration promotes cell death and elevated HC activity has been implicated in pathological conditions, we investigated whether high HC activity contributes to Ca2+ influx and cell death. HeLa parental cells and HeLa cells expressing Cx39, Cx43, Cx45, or Panx1 were exposed to an alkaline extracellular solution (pH 8.5) to increase HC activity. Under these conditions, dye uptake assays revealed high HC activity in all transfected cells but not in parental control cells. Previous studies have shown that Cx43 HCs, but not Cx39 and Panx1 HCs, allow the influx of extracellular Ca2+. Here, we also found that exposure of Cx45 transfectants to pH 8.5 activated HCs and allowed the influx of extracellular Ca2+. Only in cells expressing functional HCs permeable to Ca2+ did the elevated HC activity heighten the cytosolic Ca2+ concentration, which promoted lipid peroxidation and reduced cell viability. The effects were also abolished by the removal of extracellular divalent cations, suggesting that a Ca2+ influx that triggers downstream deleterious effects is required. Our findings identify Cx45 as a novel Ca2+-permeable HC, and they reveal that alkaline stress promotes Ca2+ entry via Cx43 and Cx45 HCs, which in turn leads to oxidative stress and cell death.

Figures

Full text

1. Introduction

In mammalian cells, hemichannels (halves of gap junction channels; HCs) are constituted of six protein subunits in the case of connexins (Cxs) or seven subunits in the case of pannexins (Panxs) [1,2]. At the cell surface, these large-pore channels can facilitate communication between intracellular and extracellular compartments. These HCs are permeable to ions and small molecules, including metabolites as well as autocrine and paracrine signals (e.g., ATP, cyclic nucleotides, prostaglandin E2, NAD^+^, glutathione and glutamate) [3].

During physiological responses, Cx and Panx1 HCs may open briefly without compromising cell survival [4,5]. However, diverse pathological conditions or gain-of-function mutations can induce sustained HC opening, inflammation and cell death [6,7,8]. Several of the HC permeants can contribute to cell death. In this respect, one of the most important are calcium ions (Ca^2+^), because they can activate Ca^2+^-dependent proteases, lipases and nucleases, which participate in numerous intracellular metabolic pathways [9]. Cx26 and Cx43 HCs have been shown to be permeable to Ca^2+^ [10,11]. In contrast, Cx39 and Panx1 HCs are impermeable to Ca^2+^ [12,13]. However, recent work has demonstrated that Panx1 HCs become permeable to Ca^2+^ and induce cell death only after C-terminal truncation of the protein [13]. Also, HeLa cells transfected with Cx45 and treated for 7 h with FGF-1 show an elevated cytoplasmic Ca^2+^ signal, suggesting that Cx45 HCs might be permeable to this divalent ion [14]. Nevertheless, direct demonstration of Cx45 HC permeability to Ca^2+^ remains unresolved.

Although numerous reports have suggested that HCs are involved in cell death mediated by necrosis or apoptosis [15,16,17], the relative importance of HCs formed by Cxs and Panx1 remains elusive. The issue may be even more complex when considering that pathological conditions may alter the relative level of expression of these proteins as occurs in mammalian skeletal muscles. Differentiated cells in this tissue normally express Panx1 but do not express Cxs. Yet, under several pathological conditions (i.e., denervation, muscle dystrophies, sepsis, diabetes and long-term glucocorticoid treatment) there is de novo expression of three Cxs (i.e., Cx39, Cx43 and Cx45) and up-regulation of Panx1 levels [18]. Interestingly, under different pathological conditions, skeletal muscles show increased generation of reactive oxygen species (ROS) and muscle fiber death [19,20,21,22]. ROS levels also increase in skeletal muscle after dexamethasone treatment, a phenomenon that does not occur in mice lacking Cx43 and Cx45 [23]. However, the individual contribution of each Cx HC to cell death remains unknown.

The present work was undertaken to study whether elevated HC activity leads to an increase in Ca^2+^ influx, the generation of ROS and cell death. We found a direct correlation of these changes only in cells expressing Cx43 or Cx45 HCs.

2. Results

2.1. Extracellular Alkaline Solution Increases Cell Membrane Permeability Through Cx45 Hemichannels

It has previously been reported that an extracellular alkaline solution increases the activity of Cx43 and Cx39 HCs, leading to an increase in Etd^+^ uptake [11,12]. In contrast, Panx1 HCs show minimal Etd^+^ uptake under basal or alkaline conditions; their activity is more reliably detected using DAPI [13,24]. Among the Cx HCs expressed in pathological conditions in skeletal muscle, Cx45 HCs have not yet been characterized for their behavior under alkaline conditions.

To elucidate the effect of extracellular alkaline solution on Cx45 HCs, we first evaluated DAPI uptake. Representative recordings of DAPI uptake over time in HeLa parental cells and HeLa-Cx45 cells are shown in Figure 1A. HeLa parental cells did not exhibit significant changes in uptake independent of the experimental conditions (Figure 1A,B). In contrast, HeLa-Cx45 displayed a marked increase in DAPI uptake at pH 8.5 compared with pH 7.4 (a ~4-fold increase, p < 0.0001). This response was inhibited by 200 μM La^3+^, indicating that the effect was mediated by Cx45 HCs (Figure 1A,B).

We next assessed Etd^+^ uptake under the same conditions. Representative time-lapse recordings of Etd^+^ uptake are shown in Figure 1C. Consistent with the results obtained with DAPI, HeLa parental cells showed no significant changes in Etd^+^ uptake, whereas HeLa-Cx45 exhibited a significant increase at pH 8.5 compared with pH 7.4 (a ~4-fold increase, p < 0.0001) (Figure 1C,D). This effect was also fully blocked by 200 μM La^3+^, supporting the involvement of elevated Cx45 HC activity.

Taken together, these results demonstrate that extracellular alkaline solution enhanced the activity of Cx45 HCs, promoting permeability to both DAPI and Etd^+^, whereas parental cells did not show significant changes.

2.2. Extracellular Alkaline Solution Induces Ca2+ Influx in HeLa-Cx45 Cells

Increased permeability to Etd^+^ and/or DAPI in Cx43, Cx39 and Panx1 HCs does not necessarily correlate with increased permeability to Ca^2+^ [12,13]. This prompted us to test whether Cx45 HCs are permeable to Ca^2+^. For this purpose, we evaluated intracellular Ca^2+^ dynamics using the ratiometric indicator Fura-2 AM in HeLa-Cx45 cells. The exposure of these cells to extracellular pH 8.5 induced a rapid and sustained increase in the 340/380 ratio of Fura-2-emitted fluorescence, consistent with Ca^2+^ influx (Figure 2A). Quantitative analysis of the area under the curve (AUC) demonstrated a significant increase in intracellular free Ca^2+^ concentration at pH 8.5 compared with pH 7.4 (p < 0.05); this effect was prevented by La^3+^ (Figure 2B). In the representative recording shown in Figure 2A, the magnitude of the Ca^2+^ response (AUC) correlated with the fluorescence intensity of the GFP–Cx45 reporter (R^2^ = 0.92), supporting a direct relationship between Cx45 expression levels and Ca^2+^ influx. These findings indicate that the extracellular alkalinization-induced opening of Cx45 HCs allows Ca^2+^ to permeate and enter the cells; the slight increase in Ca^2+^ signal detected over time (more evident between 500 and 1000 s of recording) in parental cells might have resulted from photodynamic damage.

2.3. Exposure to Extracellular Alkaline Solution Induces Oxidative Stress Only in HeLa-Cx43 and HeLa-Cx45 Cells

Since increases in the intracellular free Ca^2+^ and/or Mg^2+^ concentration are directly associated with increases in cellular oxidative state [25,26], we evaluated whether exposure of HeLa-Cx43 or HeLa-Cx45 cells to divalent cation-containing extracellular alkaline solutions for 1 h altered the amount of thiobarbituric acid-reactive substances (TBARSs), which are widely used as biomarkers of lipid peroxidation and reflect the degree of oxidative stress-associated cellular injury. Levels of TBARS increased from 0.53 μM/10^6^ cells at pH 7.4 to 15.10 μM/10^6^ cells at pH 8.5 in HeLa-Cx43 cells (Figure 3). Similarly, the amount of TBARS in HeLa-Cx45 cells increased from 0.98 μM/10^6^ cells at pH 7.4 to 14.90 μM/10^6^ cells at pH 8.5 (Figure 3). These increases in TBARS levels were abrogated by 5 min pretreatment of HeLa-Cx43 or HeLa-Cx45 cells with 200 μM CBX, and they did not occur in DCFS. In contrast, the amounts of TBARS in HeLa parental, HeLa-Cx39 and HeLa-Panx1 cells were not significantly affected by changing extracellular pH values from 7.4 to 8.5, or by pretreatment with CBX (200 μM for HeLa parental and HeLa-Cx39 cells or 10 μM for Panx1 HeLa cells) (Figure 3).

2.4. The Extracellular Alkaline Solution Decreased Cell Viability in Cx43 and Cx45 but Not Cx39 and Panx1 HeLa Transfectants

Increases in intracellular free Ca^2+^ concentration and/or oxidative stress induce cell death [27], and Ca^2+^ or Mg^2+^ overload leads to the generation of ROS by activating several intracellular pathways [26,28]. Because some of the HeLa transfectants showed an increase in both the intracellular Ca^2+^ signal and TBARS when exposed to extracellular alkaline solution, we evaluated whether these changes were associated with cell death. For this purpose, we evaluated the loss of plasma membrane integrity by propidium iodide staining in cells exposed to alkaline (pH 8.5) or physiological (pH 7.4) Krebs–Ringer solution in the presence or absence of extracellular divalent cations for 3 h. We found that 98.9% of HeLa-Cx43 cells stained with propidium iodide at pH 8.5, compared to 5.8% at an extracellular pH of 7.4; these cells were scored as dead (Figure 4). The same treatment caused 52.1% mortality in HeLa-Cx45 cells vs. 1.6% at pH 7.4 (Figure 4). In contrast, much lower levels of mortality were scored at pH 8.5 in HeLa parental cells (1.4%), HeLa-Cx39 cells (7.9%) and HeLa-Panx1 cells (0.9%); these values were not significantly different than those obtained at pH 7.4 (Figure 4). In addition, no significant cell death was scored in any of the Cx transfectants after 3 h exposure to DCFS regardless of whether the pH was 7.4 or 8.5 (Figure 4). Indeed, mortality in DCFS pH 8.5 in HeLa parental, HeLa-Cx39, HeLa-Cx43 and HeLa-Cx45 cells was 3.0%, 10.9%, 1.6% and 0.8%, respectively (Figure 4).

The presence of 200 μM CBX prevented the increase in PI staining induced in HeLa-Cx43 and HeLa-Cx45 cells by exposure to Krebs–Ringer solution at pH 8.5. Under these conditions, mortality was only 1.7% in Hela-Cx43 and 1.4% in HeLa-Cx45 cells (Figure 4). Cell death also decreased in CBX-treated HeLa-Cx39 cells, reaching a value of 1.6% (Figure 4). No significant changes were detected in the mortality of Hela parental and HeLa-Panx1 cells exposed to 200 µM CBX (Figure 4).

3. Discussion

In the present work, we demonstrated that extracellular alkaline solution promotes the activity of HCs formed by Panx1 and different Cxs, but that only cells expressing Cx43 or Cx45 exhibited a sustained increase in cytosolic Ca^2+^, elevated TBARS levels, and reduced cell viability. These responses were inhibited by CBX and were also prevented by the absence of extracellular divalent cations, indicating that the Ca^2+^ responsible for the rise in the intracellular Ca^2+^ concentration originated from the extracellular space. Thus, our findings identify that Cx45 forms HCs that allow Ca^2+^ influx under alkaline conditions, thereby leading to oxidative stress and cell death. Consistent with a previous report [29], we confirmed that Cx43 HCs are permeable to Ca^2+^ and that this response leads to oxidative stress and cell death.

3.1. Activation of Cx45 Hemichannels by Extracellular Alkaline Solution

It has previously been reported that exposure of HeLa-Cx43 cells to extracellular alkalinization (pH 8.5) enhances Etd^+^ uptake by increasing the activity of Cx43 HCs [11]. We found that the Cx45 HC activity was also rapidly increased by extracellular alkaline pH. This permeability profile is not shared by all HCs. For instance, Cx39 and Cx43 HCs are permeable to both DAPI and Etd^+^ [12,30], while full-length Panx1 HCs allow entry of DAPI but not Etd^+^, unless the protein is truncated at the C-terminal domain [13]. These observations indicate that Cx45 HCs share the dye permeability properties of other Cx HCs, but differ from those of full-length Panx1 HCs. In this context, the slight residual increase in DAPI uptake observed in representative traces after acute La^3+^ application may reflect the contribution of low endogenous Panx1 expression in HeLa cells, since Panx1 hemichannels are permeable to DAPI but not to Etd^+^ and are not fully inhibited by La^3+^ [13]. Importantly, quantitative analysis revealed that DAPI uptake after La^3+^ application was not significantly different from basal conditions.

Previous work has shown that Cx43 hemichannels exhibit a higher permeability to Etd^+^ than to DAPI [31], whereas in the present study Cx45 hemichannels displayed comparable uptake rates for both dyes. Such difference is consistent with recent evidence demonstrating marked isoform-dependent permselectivity among connexin hemichannels, whereby some isoforms are permeable to ATP, glutamate, and lactate (e.g., Cx43), whereas others restrict the passage of specific metabolites independently of molecular size (e.g., ATP-impermeable Cx50 or glutamate-impermeable Cx36) [32].

Previous studies have shown that both Cx43 and Cx45 gap junction channels exhibit pH-dependent modulation of gating, with alkalinization favoring channel opening [33,34]. Our results are consistent with these findings, extending the concept to Cx45 HCs at the plasma membrane. Interestingly, pH sensitivity also appears conserved across different channel families, since HCs from zebrafish Panx1 increase dye uptake in response to alkaline conditions [35]. Together, these observations reinforce the idea that extracellular alkalinization promotes HC activity through a conserved mechanism of pH-dependent gating.

The effect of alkaline pH on Cx45 HCs may involve changes in gating dynamics. Previous studies have shown that extracellular Ca^2+^ and Mg^2+^ critically modulate HC activity by stabilizing the closed conformation [11,36,37]. In the case of Cx43, reducing extracellular Ca^2+^ expands the estimated pore diameter from ~1.8 nm (closed state) to ~2.5 nm (open state) [36]. A similar regulation has been suggested for other Cxs. Our finding that the activity of Cx45 HCs increases under alkaline pH in the presence of physiological concentrations of Ca^2+^ and Mg^2+^ suggests that extracellular alkalinization acts by altering the open probability and/or open time of the HCs. Whether this involves conformational changes in the pore architecture and/or an increased number of active HCs at the cell surface remains to be determined.

3.2. Differential Permeability of Cx HCs to Ca2+

Differences in Ca^2+^ permeability have been reported among Cx HCs. For instance, although Cx39 HCs are permeable to DAPI and Etd^+^, they are not permeable to Ca^2+^ when activated by extracellular alkaline solution [12]. Similarly, full-length Panx1 HCs do not allow Ca^2+^ entry under extracellular alkalinization [13,24], although truncation of the C-terminal domain renders them Ca^2+^-permeable [13]. In contrast, Cx43 HCs are permeable to Ca^2+^ upon alkalinization [11].

Here we showed that Cx45 HCs also mediated Ca^2+^ influx in response to extracellular alkaline solution. This finding is consistent with previous work showing that stimulation of HeLa-Cx45 cells with FGF-1 for 7 h induced a rise in intracellular Ca^2+^ [14]. However, our results revealed that acute exposure to an alkaline extracellular solution containing physiological concentrations of Ca^2+^ and Mg^2+^ is sufficient to rapidly increase cytosolic Ca^2+^ through Cx45 HCs. In agreement with the interpretation that the observed signal arises from Ca^2+^ influx rather than mobilization from intracellular stores, the removal of extracellular divalent cations completely abolished the alkalinization-induced TBARSs and reduced cell suvival.

3.3. Production of Reactive Oxygen Species and Decreased Viability of HeLa-Cx43 and HeLa-Cx45 Cells Exposed to Extracellular Alkaline pH

The induction of oxidative stress by extracellular alkalinization was restricted to cells expressing Cx43 or Cx45 whose HCs (unlike those of Cx39 and Panx1) are permeable to Ca^2+^. These findings are consistent with the notion that intracellular Ca^2+^ overload triggers oxidative stress, since it activates enzymes such as lipoxygenases, cyclooxygenases, and NADPH oxidases, and promotes mitochondrial dysfunction, all of which amplify ROS generation [25,38,39]. When ROS production exceeds the antioxidant capacity of the cell, oxidative stress ensues, resulting in damage to lipids, proteins, and nucleic acids, processes that underlie numerous pathological conditions [40]. These mechanisms provide a strong explanation for our observations, linking Ca^2+^ entry through Cx43 and Cx45 HCs to ROS production and subsequent cell death. This conclusion is also consistent with the observation that in vivo treatment with dexamethasone induces Cx39, Cx43 and Cx45 expression and increases oxidative stress in skeletal muscle, effects that do not occur in mice with muscle-specific deletion of Cx43 and Cx45 [23]. Since Cx39 HCs are not permeable to Ca^2+^, these results suggest that HC-dependent induction of oxidative stress requires HC-mediated Ca^2+^ permeation.

The role of Cx43 in oxidative stress and induction of apoptosis is supported by several independent studies. In endothelial cells, LPS stimulation increases Cx43 protein levels at the plasma membrane, which correlates with enhanced ROS generation and apoptosis. Selective inhibition of Cx43 HCs with Gap19 abolished these responses, demonstrating that Cx43 HC activity at the membrane is required for oxidative damage [41]. Likewise, Cx43 HCs mediate the propagation of Ca^2+^, ATP, ROS and NO signaling in radiation-induced bystander effects (which mimic the direct effects of ionizing radiation including apoptosis), where DNA damage in non-irradiated endothelial cells is dependent on HC activity [29]. Together, these studies and the data presented here establish Cx43 as a central mediator of Ca^2+^-dependent oxidative stress and cell death in different pathological conditions. Our findings extend this conclusion to Cx45 and provide the first evidence that Cx45 HCs are permeable to Ca^2+^ and that their elevated activity (resulting from extracellular alkalinization) triggers a Ca^2+^ influx, leading to increased lipid peroxidation and reduced cell viability. Both effects were prevented by carbenoxolone or by removal of extracellular Ca^2+^, suggesting that Ca^2+^ entry through Cx45 HCs is the initiating event.

Intracellular Ca^2+^ overload can lead to ROS generation by activating several cytoplasmic and mitochondrial oxygen-dependent reactions, which can further contribute and/or exacerbate the increase in intracellular Ca^2+^ and cause deleterious effects (including loss of plasma membrane integrity as assessed by propidium iodide uptake in our study) leading to cell demise. Because oxidative damage was assessed using a TBARS assay, these results are interpreted as evidence of increased lipid peroxidation, a widely used biomarker of oxidative stress-associated cellular injury; future studies employing complementary approaches will be necessary to further characterize the specific ROS involved. The resulting loss of cell viability was assessed by propidium iodide uptake, which reflects a loss of plasma membrane integrity and therefore does not allow for discrimination between necrotic and late apoptotic mechanisms of cell death. In this respect it is worth mentioning that aberrant HC activity has been associated with both cell death mechanisms depending on the Cx subtype and/or cell type. In other cellular contexts, it has been shown that Cx HCs promote apoptosis [42], that gain-of-function mutants of Cx31 cause excessive HC opening that is directly associated with ROS generation and necrotic cell death [7], and that a gain-of-function Cx26 mutant forms hyperactive heteromeric hemichannels with Cx30 leading to Ca^2+^ overload and apoptotic cell damage [43]. Within this framework, our results identify Cx45 HCs as novel contributors to Ca^2+^-dependent oxidative stress. Nonetheless, since increased cytoplasmic levels of Mg^2+^ or Zn^2+^ can also cause cell death [26], it remains to be determined whether HeLa-Cx45 cell death induced by alkaline pH is mediated exclusively by Ca^2+^ influx or whether additional divalent cations (such as Mg^2+^ or Zn^2+^) contribute to this process. In this respect it is worth mentioning that genetic ablation of Cx45 reduced retinal cell loss by about 50% under ischemic conditions, indicating that this Cx contributes to neuronal death in vivo [44], reinforcing the idea that Cx45 HCs can critically modulate cell fate under stress.

Taken together, and considering that aberrant expression of Cx43 and Cx45 occurs in skeletal muscle under diverse pathological conditions, including muscular diseases [18], our results suggest that Cx45 may represent a previously unrecognized contributor to oxidative damage in vivo and a potential pharmacological target for diseases associated with oxidative stress.

4. Materials and Methods

4.1. Reagents

Ethidium (Etd^+^) bromide and carbenoxolone (CBX) were purchased from Sigma-Aldrich (St. Louis, MO, USA). DAPI and Fura-2 AM were obtained from Invitrogen (Waltham, MA, USA). Stably transfected clones were selected by their resistance to 0.5 mg/mL puromycin and 0.3 mg/mL geneticin (Gibco, Carlsbad, CA, USA).

4.2. Cell Culture

Parental cells were obtained from ATCC (Manassas, VA, USA). HeLa cells transfected with mouse Cxs 39, 43, 45 (HeLa-Cx39, HeLa-Cx43, HeLa-Cx45, respectively) were kindly provided by Dr. Klaus Willecke, University of Bonn, Germany. HeLa cells stably transfected with mouse Panx1 (HeLa-Panx1) were obtained from Dr. Felixas Bukaukas, Albert Einstein College of Medicine, New York, USA. All cells were cultured on plastic tissue culture dishes (Nunc, Rochester, NY, USA) in DMEM medium (Gibco, Thermo Fisher Scientific, Waltham, MA, USA), supplemented with antibiotics and 10% fetal bovine serum (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) and were maintained in incubators at 37 °C with 5% CO_2_ and 98% relative humidity. Sub-confluent cultures (~70%) were used in all experiments.

4.3. Extracellular Solution

Cells were bathed with Krebs–Ringer solution (in mM: 145 NaCl, 5 KCl, 3 CaCl_2_, 1 MgCl_2_, 5.6 C_6_H_12_O_6_, 10 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), pH 7.4 or pH 8.5) or divalent cation (Ca^2+^ and Mg^2+^)-free solution (DCFS: Krebs–Ringer solution without divalent cations) (in mM: 154 NaCl, 5.4 KCl, 5 C_6_H_12_O_6_, 10 HEPES and 0.5 EGTA, pH 7.4 or pH 8.5). The final pH of each solution was adjusted to the desired value using Tris base.

4.4. Ethidium (Etd+) and DAPI Uptake

Cells were plated on glass coverslips (25 mm in diameter; Marienfeld, Lauda-Königshofen, Germany) and maintained in tissue culture medium for 48 h. Then, they were bathed with Krebs–Ringer solution (pH 7.4 or pH 8.5) or DCFS (pH 7.4 or pH 8.5) containing 5 µM ethidium (Etd^+^) bromide or 5 µM DAPI. Etd^+^ and DAPI uptake was evaluated in HeLa-Cx45 cells. Subsequently, the HC activities of Cx HeLa transfectants were evaluated in solutions containing 200 µM lanthanum (La^3+^), a Cx hemichannel blocker [45]. The fluorescence intensity was recorded in regions of interest corresponding to the nuclei of at least 20 cells, and was viewed using an Olympus model BX41TF (Olympus Corporation, Center Valley, PA, USA). All images were captured every 15 sec with a QImaging model Retiga 13001 fast cooled 12-bit monochromatic digital camera (QImaging, Surrey, BC, Canada). The fluorescence emissions of Etd^+^ and DAPI bound to nucleic acids were recorded at 595 nm and 461 nm, respectively.

4.5. Intracellular Ca2+ Signal

HeLa cells transiently transfected with Cx45 cultured on coverslips for 48 h were loaded with 2 µM Fura-2 AM for 45 min. The cells were imaged every 3 s using a conventional Nikon Eclipse Ti fluorescent microscope (Nikon, Tokyo, Japan) to record the Fura-2 fluorescence intensities emitted at 510 nm following excitation at 340 and 380 nm. The Ca^2+^ signals were quantified as the 340/380 ratio of the emitted fluorescence intensities. Only reporter-positive HeLa-Cx45 cells were considered for analysis. Ca^2+^ signals from multiple cells within the same field were averaged to obtain a single value per independent experiment.

4.6. Quantification of Cellular TBARSs

Cells were exposed for 1 h at 37 °C ± 1 °C to Krebs–Ringer solution, pH 7.4 or pH 8.5, Ca^2+^- and Mg^2+^-free Krebs–Ringer solution, pH 8.5, or they were pretreated for 15 min with 200 μM CBX (in the case of Cx transfectants) or 10 μM CBX (in the case of Panx1 transfectants) and kept in the presence of the inhibitor during the 1 h exposure to Krebs–Ringer solution, pH 8.5. Then, the cells were harvested in ice-cold phosphate-buffered saline solution using a cell scraper and lysed by three 1 s sonication events. The content of thiobarbituric acid-reactive substances (TBARS) was measured in aliquots from the whole homogenates using a lipid peroxidation (MDA) Assay Kit (Abcam, Cambridge, MA, USA). In this kit, malondialdehyde (MDA) reacts with thiobarbituric acid to produce stable chromogenic adducts that can be quantified spectrophotometrically by measuring the absorbance at 532 nm. The amounts of TBARS in the samples were calculated by interpolation of the measured absorbance on a standard curve obtained with MDA [46].

4.7. Quantification of Cell Death

HeLa cells were exposed to either Krebs–Ringer solution, pH 7.4 or 8.5, for 3 h. In some experiments, the cells were exposed for 3 h to DCFS, pH 7.4 or pH 8.5, at 37 °C ± 1 °C, or they were pretreated and kept in the presence of 200 μM CBX during the 3 h exposure to one of the extracellular solutions. Then, the cells were incubated for 5 min with Krebs–Ringer solution containing 5 µM propidium iodide (plus added CBX only for cells that had been treated with this drug). Images of five different fields were acquired using an Olympus BX41TF microscope (model 389) equipped with epifluorescence illumination (λ_ex_ 535 nm and λ_em_ 617 nm) using a 20× objective (Olympus UPlanFL N, NA = 0.50). The total number of cells and the number of cells stained with propidium iodide were quantified in each picture.

4.8. Image Analysis and Statistical Analysis

Fluorescence images were analyzed using ImageJ software (version 1.64r) (NIH, Bethesda, MD, USA). Graphs and statistical analyses were performed using GraphPad Prism 5 software (San Diego, CA, USA). Statistical significance was determined using the Shapiro–Wilk test to assess data normality, followed by one-way ANOVA and Tukey’s post hoc test, or by the Mann–Whitney U test. Experiments were performed separately by three independent individuals working on different days with different batches of cells.