Angiotensin II-Induced Mesangial Cell Damage Is Preceded by Cell Membrane Permeabilization Due to Upregulation of Non-Selective Channels

Connexin43 (Cx43), pannexin1 (Panx1) and P2X7 receptor (P2X7R) are expressed in kidneys and are known to constitute a feedforward mechanism leading to inflammation in other tissues. However, the possible functional relationship between these membrane channels and their role in damaged renal cells remain unknown. In the present work, we found that MES-13 cells, from a cell line derived from mesangial cells, stimulated with angiotensin II (AngII) developed oxidative stress (OS, thiobarbituric acid reactive species (TBARS) and generated pro-inflammatory cytokines (ELISA; IL-1β and TNF-α). The membrane permeability increased progressively several hours before the latter outcome, which was a response prevented by Losartan, indicating the involvement of AT1 receptors. Western blot analysis showed that the amount of phosphorylated MYPT (a substrate of RhoA/ROCK) and Cx43 increased progressively and in parallel in cells treated with AngII, a response followed by an increase in the amount in Panx1 and P2X7R. Greater membrane permeability was partially explained by opening of Cx43 hemichannels (Cx43 HCs) and Panx1 channels (Panx1 Chs), as well as P2X7Rs activation by extracellular ATP, which was presumably released via Cx HCs and Panx1 Chs. Additionally, inhibition of RhoA/ROCK blocked the progressive increase in membrane permeability, and the remaining response was explained by the other non-selective channels. The rise of activity in the RhoA/ROCK-dependent pathway, as well as in Cx HCs, P2X7R, and to a minor extent in Panx1 Chs led to higher amounts of TBARS and pro-inflammatory cytokines. We propose that AngII-induced mesangial cell damage could be effectively inhibited by concomitantly inhibiting the RhoA/ROCK-dependent pathway and one or more non-selective channel(s) activated through this pathway.


Introduction
In chronic kidney disease, hypertension is one of the most common complications that generate predispositions to other health problems, affecting several organs. Hypertensive nephropathy begins in the glomerulus by increasing intraglomerular pressure. These early events activate and damage mesangial cells, epithelial cells and podocytes in the glomerulus. In turn, these cells produce vasoactive and pro-inflammatory agents, which increase cell damage and promote fibrosis, reducing renal blood flow, permeability and, eventually, glomerular filtration [1].
Cx37, Cx40, Cx45 and Cx46) and Panx1 have been found in the kidney [21][22][23], and changes in their amounts have beendetected in several models of renal damage [10,24,25]. P2X 7 Rs have also been detected in the kidney [26]. However, it remains unknown whether Cx-and Panx-based channels as well as P2X 7 Rs interact, and whether they are regulated by AngII. To our knowledge only a few studies have described the participation of Cx HCs in renal damage [27], and no signaling pathway has been clearly associated with these changes. In the present study, we found that AngII increases membrane permeability in MES13-cells via AT1 receptors, as well as the activation of a RhoA/ROCK-dependent intracellular signaling pathway, followed by the up-regulation of three non-selective channels, and the generation of OS and pro-inflammatory cytokines.

AngII Increases Cell Membrane Permeability of MES-13 Cells
Since Etd + is a membrane-impermeant cationic dye, which becomes fluorescent when intercalated in nucleic acid strands normally found in the intracellular space [28], we used it to evaluate changes in membrane permeability in MES-13 cells under control conditions and after treatment with AngII.

AngII Promotes Phosphorylation of MYPT and Increases the Amount of Cx43, Panx1 and P2X 7 R in Mesangial Cells
AngII binding to AT1 receptor activates RhoA and Rho kinase (ROCK) [3], and Cx43 HCs can mediate changes in membrane permeability in different cells types [12,30], we decided to evaluate the activity of RhoA/ROCK and Cx43 HCs. To this end, we first measured the amount of phosphorylated MYPT-a downstream effector of the RhoA/ROCK pathway-and the relative amount of unphosphorylated Cx43 in MES-13 cells at different time periods after treatment with AngII (10 −7 M). Moreover, and since open Panx1 Chs and P2X 7 Rs could increase membrane permeability and both are co-expressed in several cell types undergoing inflammatory responses [11,31], we also evaluated the relative amount of Panx1 and P2X 7 R.
Following AngII treatment, the amount of phosphorylated MYPT detected in MES-13 cells was significantly increased at 24 h (From 0.15 ± 0.03 AU to 0.28 ± 0.09 AU), and increased even more at 48 h (0.65 ± 0.16 AU) and 72 h (1.2 ± 0.2 AU) ( Figure 2). Similarly, Cx43 was detected as a single band and its amount increased significantly and progressively at 24, 48 and 72 h of stimulation with AngII (From 0.16 ± 0.02 AU at 0 h to 0.30 ± 0.02 AU at 24 h, 0.49 ± 0.02 AU at 48 h and 0.70 ± 0.02 AU at 72 h) ( Figure 3A). Since mesangial cells also express Cx40 and Cx45 [32], we evaluated their presence in MES-13 cells. As expected, these two Cxs were detected, but their relative amounts were not affected after treatment with AngII ( Figure 3A). This suggests that the effect of AngII could be Cx43-specific. Similarly, the relative amount of Panx1 and P2X 7 R were not significantly different at 24 and 48 h, but were significantly increased at 72 h post-AngII treatment (Panx1 from 0.20 ± 0.03 AU at 0 h to 0.60 ± 0.06 at 72 h and P2X 7 R from 0.24 ± 0.04 AU at 0 h to 0.74 ± 0.10 AU at 72 h) ( Figure 3B,C).    To assess whether the AngII-induced increase in cell membrane permeability was mediated by a RhoA/ROCK-dependent pathway, we evaluated the effect of Fasudil and Y-27632-two selective inhibitors of ROCK [3,9]-on AngII-induced Etd + uptake in MES-13 cells. After 72 h, AngII-treated cells showed a significantly higher dye uptake rate compared to that of cells under control conditions (AngII: 34.7 ± 1.34 AU/min; control: 3.16 ± 0.34) ( Figure 4). However, the dye uptake rate induced by AngII (72 h) was drastically decreased (~50%) by the mimetic peptide Gap27 (100 µM, 16.9 ± 1.20 AU/min)-a selective Cx43/Cx37 HC blocker [33,34]-as well as apyrase (2 units/mL, 16.3 ± 1.04 AU/min) or probenecid (PBC, 500 µM; 3.5 ± 0.6 AU/min)-a blocker of Panx1 Ch and P2X 7 R [35,36] applied 24 h before completing 72 h of AngII treatment ( Figure 4). This suggests that such blockers prevent the progression, but do not reverse the effect on membrane permeability already regenerated by AngII. On the other hand, cells treated with AngII for 72 h and treated with Fasudil (15 µM) or Y-27632 (15 µM) 24 h before evaluating membrane permeability showed increase in Etd + uptake rate that was~50% lower than those cells treated only with AngII (AngII + Fasudil: 15.4 ± 1.10 AU/min; AngII + Y-27632: 15.8 ± 1.8 AU/min) ( Figure 4). This suggests the persistent involvement of membrane pathways found downstream of ROCK. Moreover, since greater Etd + uptake rates in cells treated with AngII for 48 h (11.4 ± 1.70 AU/min) ( Figure 1) were like those of cells treated with AngII for 72 h plus Fasudil or Y27632 during the last 24 h, it is possible that ROCK inhibition during the last 24 h of AngII treatment prevented the progression, but did not reverse the downstream-induced increase in membrane permeability already triggered by AngII.

AngII Reduces Intercellular Communication Mediated by GJs in MES-13 Cells
Since increases in membrane permeability via HCs induced by various conditions also reduces intercellular communication mediated by GJs [15,16,39], we decided to evaluate the effect of AngII on GJs expressed by MES-13 cells. The functional state of GJs can be demonstrated by means of an intercellular transfer of dyes microinjected in single cells of clusters or monolayers [12]. With this experimental approach, control MES-13 cells were found to be dye-coupled ( Figure 5A), and the coupling incidence decreased (44 ± 6%) after 72 h treatment with AngII (10 −7 M) compared to control cells (99 ± 1%) ( Figure 5A). In addition, the number of coupled cells (coupling index) also decreased from 5.0 ± 0.2 cells under control conditions to 2.0 ± 0.1 cells after 72 h treatment with AngII ( Figure 5B,C), hence indicates that AngII impacts membrane permeability and GJ channels in an opposite manner.
We also evaluated the effect of Fasudil (15 µM) on the functional state of GJs in AngII-treated MES-13 cells. After 72 h treatment with AngII (10 −7 M), cells showed a reduced coupling incidence (44 ± 6%) compared to control cells (control, 99.0 ± 0.8% and control + Fasudil, 100 ± 0%), but Fasudil added the last 24 h to cells treated with AngII reduced the inhibitory effect of AngII on intercellular communication via GJs (incidence of coupling: 93.0 ± 3.4%) ( Figure 6A). In addition, the coupling index was also decreased (1.0 ± 0.   Each treatment is denoted below each bar with a plus sign (+). Each bar represents the mean value ± SE of 4 independent experiments. In each experiment the dye was microinjected into at least 10 cells. Statistical significance ** p < 0.01 and *** p < 0.001 vs. Ctrl; ## p < 0.001 and ### p < 0.001 vs. AngII. (C) Etd + was microinjected into the brightest cell (arrow) and diffused to neighboring cells. Top panels are the phase contrast images and bottom panels show Etd + fluorescence in cells under different treatments. Scale bar = 20 µm.

Inhibition of RhoA/ROCK Reduces the Amount of Phosphorylated MYPT and Cx43, but Does Not Impact the Amount of Panx1 or P2X 7 Receptors in AngII-Treated MES-13 Cells
An increase in phosphorylation of MYPT was detected in MES-13 cells after 72 h of AngII stimulation (1.4 ± 0.1 AU) compared to control cells (control 0.45 ± 0.09 AU) or control cells treated with Fasudil (0.38 ± 0.10 AU) or Y-27632 (0.35 ± 0.06 AU) ( Figure 7A). This suggests that Rho/ROCK activity is very low under control conditions. Moreover, we found that Fasudil and Y-27632 added the last 24 h to cell cultures pretreated with AngII for 72 h reduced the amount of phosphorylated MYPT (AngII + Fasudil, 0.28 ± 0.08 AU and AngII + Y-27632, 0.33 ± 0.10 AU), reaching a phosphorylation state as low as that of control cells ( Figure 7A). But, Fasudil or Y-27632 added the last 24 h to MES-13 cells pretreated with AngII for 72 h drastically prevented a rise in Cx43 induced by AngII (AngII + Fasudil, 0.45 ± 0.10 AU and AngII + Y-27632, 0.4 ± 0.1 AU), reaching values close to those found in control cells ( Figure 7B). In contrast, Rho/ROCK inhibition during the last 24 h of AngII treatment did not significantly affect the amount of Panx1 or P2X 7 R, which were comparable to those of cells treated only with AngII ( Figure 7C,D). Therefore, it can be inferred that increases in phosphorylated MYPT and Cx43 are probably controlled by the same RhoA/ROCK-dependent pathway, and the amounts of Panx1 and P2X 7 R are controlled by a mechanism independent of RhoA/ROCK.

Inhibition of RhoA/ROCK, Cx HCs or P2X 7 Rs, but Not Panx1 Chs, Prevents Increases in Lipid Peroxidation and Inflammatory Responses Induced by AngII in MES-13 Cells
Since AngII is known to induce Ca 2+ influx as well as Ca 2+ release from intracellular stores [40,41] leading to oxygen radical generation and cell damage in several kidney diseases [5,42], and P2X 7 Rs as well as some Cx HCs are Ca 2+ conductive channels [43,44], we decided to evaluate whether AngII treatment causes OS and pro-inflammatory cytokine production in mesangial cells. Here, we show that the extracellular amount of TBARS increased (16.40 ± 0.70 µmol/L) in MES-13 cells treated with AngII (10 −7 M) for 72 h compared to control conditions (3.40 ± 0.40 µmol/L) ( Figure 9A). It was also found that this AngII treatment increased the extracellular amount of IL-1β and TNF-α (IL-1β, 2.30 ± 0.40 ng/mL; TNF-α, 1.50 ± 0.20 ng/mL) compared to control conditions (IL-1β, 0.08 ± 0.02 ng/mL; TNF-α, 0.31 ± 0.21 ng/mL) ( Figure 9B). Moreover, treatment with Fasudil at 24 h prior to the end of the 72 h incubation with AngII significantly reduced the extracellular amount of TBARS (3.31 ± 0.23 µmol/L) as well as IL-1β and TNF-α (IL-1β, 0.30 ± 0.06 ng/mL; TNF-α, 0.08 ± 0.01 ng/mL), as compared to what was found in the extracellular medium of cells treated only with AngII, and were similar to those found in the extracellular space of control cell cultures ( Figure 9).
Interestingly, treatment with Gap27 or A740003 24 h prior to the end of the 72 h post-treatment with AngII significantly reduced the amount of TBARS (Gap27: 3.04 ± 0.22 µmol/L and A740003: 3.70 ± 0.40 µmol/L). However, treatment with PBC caused only a partial reduction in the amount of TBARS (8.03 ± 0.82 µmol/L) ( Figure 9A). Moreover, the extracellular amount of IL-1β and TNF-α in MES-13 cell cultures treated with AngII for 72 h was more than 10 times higher than that of control cultures ( Figure 9B,C). However, the amount of pro-inflammatory cytokines in cultures treated with AngII for 72 h and then exposed to Gap27 (IL-1β, 0.08 ± 0.02 ng/mL and TNF-α, 0.06 ± 0.02 ng/mL) or A740003 (IL-1β, 0.13 ± 0.04 ng/mL and TNF-α, 0.09 ± 0.01 ng/mL) 24 h before harvesting the extracellular solution was comparable to that of cell cultures under control conditions ( Figure 9B,C). Unexpectedly, in cultures treated with AngII plus PBC the amount of pro-inflammatory cytokines was slightly, but not significantly, lower as compared to cultures treated only with AngII (IL-1β, 1.76 ± 0.25 ng/mL and TNF-α, 1.16 ± 0.19 ng/mL) ( Figure 9B,C).

Discussion
In the present work, high levels of TBARs, IL-1β and TNF-α were detected in the extracellular milieu of AngII-treated MES-13 cells, which are characteristic features of an inflammatory response. This outcome was preceded by an increase in membrane permeability due to the activation of at least three non-selective membrane channels (Cx HCs, Panx1 Ch and P2X 7 R) mainly through a RhoA/ROCK-dependent intracellular signaling pathway promoted by the activation of AT1 receptors. The generation of TBARS and pro-inflammatory cytokines was strongly dependent on the upregulation of functional Cx HCs as well as P2X 7 Rs, and to a lesser but still significant extent on Panx1 Chs ( Figure 10). Thus, we propose that in addition to AT1 receptor and RhoA/ROCK inhibition, AngII-induced mesangial cell damage could be efficiently prevented by inhibiting Cx43 HCs or P2X 7 Rs. Figure 10. Scheme of possible signaling pathways involved in regulating the functional state of Cx43 HCs, Panx1 Chs, P2X 7 Rs and Cx GJs in mesangial cells stimulated with AngII. High AngII concentrations (continuous black arrows) could promote Ca 2+ release from the endoplasmic reticulum (ER) and activation of a RhoA/ROCK-dependent pathway through angiotensin type 1 receptors (AT1R). The latter is inhibited by losartan. Once the RhoA/ROCK-dependent pathway is activated, the expression and release of pro-inflammatory cytokines such as TNF-α and IL-1β as well as formation of reactive oxygen species (ROS) that generate thiobarbituric acid reactive substances (TBARS) upon reaction with lipid of cell membranes can occur. The activated RhoA/ROCK-dependent pathway could affect the function of Cx43 HCs by opening them, since Fasudil or Y-27632 (ROCK blockers) inhibit this response. The resulting increase in intracellular Ca 2+ can also activate Panx1 Chs and together with activated Cx HCs enable release of ATP to the extracellular medium. Then, the extracellular ATP activates P2X 7 receptors that together with Cx43 HCs permit a drastic increase in Ca 2+ influx. The resulting intracellular Ca 2+ concentration, as already seen in other systems, promote the expression and release of pro-inflammatory cytokines such as TNF-α, IL-1β, and the generation of ROS. The release of ATP and influx of Ca 2+ establish a positive feedback loop. This loop is inhibited by different compounds: Apyrase, ATP hydrolase; the mimetic peptide Gap27, a selective Cx43 HCs blocker; A740003, a selective P2X 7 R blocker or probenecid (PBC), an inhibitor of Panx1 Chs. This increase in the cellular activity caused by AngII, where the RhoA/ROCK pathway could be involved, also reduces cell-cell communication through GJs, further affecting the cellular integrity. Discontinuous red and black arrows indicate cell responses identified in other systems, whereas continuous black arrows denote responses identified in the present work.
The AT1 receptor activated by AngII is a key factor in the pathogenesis of renal damage, since it triggers several intracellular signaling cascades, such as a RhoA/ROCK-dependent pathway. The latter contributes to inflammatory and oxidative changes observed in renal diseases [3,6], but the mechanism is not completely understood. InAngII-treatedMES-13 cells, we found an AT1 receptor-dependent increase in RhoA/ROCK activity reflected by the progressive increase in the amount of phosphorylated MYPT. This conclusion is supported by our finding that Losartan-an inhibitor of AT1 receptors-prevented the AngII-induced increase in membrane permeability. Furthermore, ROCK inhibition with Fasudil or Y-27632 prevented the AngII-induced increase in the amount of phosphorylated MYPT (Figure 10).
The half-life of bioactive peptides such as AngII is likely to be about a few hours [45], and the effect of AngII on membrane permeability described here was significant 2-3 days after its application, suggesting that its effects on MES-13 cells resulted from slow metabolic responses rather than the direct activation of channels present at the cell membrane of control cells. However, soon after AngII binds to AT1 occurs the increase in cytoplasmic free Ca 2+ concentration [40,41] as well as activation of Akt [46], and these responses might be sufficient to increase the open probability of Cx HCs [47] and Panx1 CHs [48,49]. Interestingly, these two channels are permeable to ATP, and could enable the release of this purine to the extracellular milieu, thus favoring the activation of P2X 7 Rs also expressed in control MES-13 cells. Since P2X 7 Rs and Cx43 HCs are permeable to Ca 2+ [19,43], it is possible that the concentration of AngII used in the present work caused a persistent increase in intracellular free Ca 2+ concentration, contributing to maintaining the activity of Cx HCs and Panx1 Chs available at the cell membrane ( Figure 10). In favor of this interpretation, we recorded high membrane permeability to Etd + 48 and 72 h after AngII treatment in cells bathed in AngII-free saline solution. A possible explanation for this delayed response could be that a calcium code triggered by AngII either enhanced the expression and/or reduced the degradation of proteins such as Cx43, Panx1 and P2X 7 R. Interestingly, the de novo expression of Cx HCs and P2X 7 Rs in denervated muscles has been shown to increase the activity of proteolytic pathways, which lead to muscle atrophy [50]. Thus, our data, showing a progressive increase in the amount of Cx43 caused by AngII, may have resulted from an increase in gene expression ( Figure 10) rather than a reduction in protein degradation. Increases in Cx43 and Panx1 detected by western blot suggest that progressively more channels were available at the cell surface. Therefore, greater membrane permeability to Etd + in MES-13 cells induced by AngII could be explained by two responses: (1) increases in amounts of Cx43, Panx1 and P2X 7 R; and (2) increases in the open probability of Cx43 HCs and Panx1 Chs. In favor of the role played by Cx HCs and Panx1 Chs in increasing AngII-induced membrane permeability is the fact that these two channels are permeable to Etd + [28]. Additionally, greater permeability was abrogated by selective blockers of these channels (Gap27 for Cx43 HCs, and PBC and low CBX concentrations for Panx1 Chs) ( Figure 10). Possible implications of these findings could be relevant to explaining increases in the amount of Cx43 protein, mainly in mesangial cells found by others in at least three models of renal injury [10,51]. Interestingly, we found similar changes in RhoA/ROCK activity and ROCK inhibitors, which prevented increases in amount of Cx43 induced by AngII, indicating that expression and activation of RhoA/ROCK and Cx43 are regulated by the same membrane transduction mechanism and intracellular signaling pathway activated by AngII ( Figure 10). This interpretation is also supported by the strong reduction of the AngII dependent increase in the amounts of phosphorylated MYPT and Cx43 caused by inhibitors of ROCK (Fasudil applied in the last 24 h of the AngII treatment) (Figure 10). The same metabolic pathway seems to be involved in the reduction of cell-cell communication mediated by GJs between MES-13 cells, since inhibition of ROCK significantly reduced the cell-cell uncoupling induced by AngII (Figure 10). Both the activity of a RhoA/ROCK-dependent pathway and the amount of Cx43 has been observed to increase in various models of renal damage [10,24,25,52]. Interestingly, a comparable response has been observed in corneal epithelial cells, where a RhoA/ROCK-dependent pathway is involved in the formation of Cx43 GJs. In fact, ROCK inhibition resulted in greater cell-cell communication mediated by Cx43 GJs [53]. A similar relationship between the RhoA/ROCK pathway and Cx43 has been observed in fibroblasts, where fibroblast expansion in response to tissue stretch involves extracellular ATP signaling through the RhoA/ROCK pathway and the activity of Cx HCs. In this system, Y-27632 or Cx HC blockers (Octanol and CBX) prevent increases in extracellular ATP concentration and in fibroblast expansion [54]. In agreement with this mechanism, Cx HCs and Panx1 Chs enable the release of ATP to the extracellular milieu [35,55]. Despite the direct relationship between the RhoA/ROCK pathway and Cx43 postulated in the previous work mentioned above, in our work some remaining membrane permeability was observed in mesangial cells stimulated with AngII and treated with the mimetic peptide Gap27 (Cx43 HC blocker) as well as ROCK inhibitors (Fasudil and Y-27632). These results could be explained by the involvement of other membrane channels such as Panx Chs.
The involvement of extracellular ATP in the AngII-induced effects on MES-13 cells was evident, since AngII-induced membrane permeability was drastically diminished upon exposure to: (1) apyrase that degrades ATP; (2) A740003, which is a selective inhibitor of P2X 7 Rs activated by extracellular ATP; and (3) PBC or low concentration of CBX-inhibitors of Panx1 Chs-and Gap27, inhibitor of Cx43 HCs, which are blockers of two membrane pathways through which ATP can be released to the extracellular milieu ( Figure 10).
The AngII-induced generation of OS, IL-1β and TNF-α could explain decreases in intercellular communication mediated by GJs, since reduced redox potential and pro-inflammatory cytokines have also been shown to cause such effect in cortical astrocytes and in endothelial cells of the blood-brain barrier [15,30,[56][57][58]. Accordingly, Fasudil or Y-27632 has been shown to decrease AngII-induced inflammation and OS in models of kidney diseases [1,42]. For example, ROCK inhibitors protect rats subjected to chronic AngII infusion from translocation of p65 to the nuclei of cells from the afferent arterioles, suggesting that a RhoA/ROCK-dependent pathway is involved in NF-κB activation, and the ROCK/NF-κB axis contributes to the AngII-induced upregulation of angiotensinogen [59]. Additionally, Fasudil has a renoprotective action in spontaneous hypertensive rats under deoxycorticosterone-acetate salt treatment [60].
The participation of Cx HCs, Panx1 Chs and P2X 7 Rs in different tissues undergoing inflammation has been demonstrated, and the inhibition or ablation of Cx HCs has been shown to confer significant tissue protection [16,37,50,[61][62][63]. In MES-13 and primary mesangial cells, AngII seems to promote a similar feedforward mechanism in which these three non-selective channels, without excluding others ( Figure 10) (i.e., TRPC6 channels; [64]), maintain or even amplify inflammatory and oxidative responses, causing damage to kidney cells. Therefore, we propose that blocking AngII-induced progression in mesangial cell damage could be accomplished by inhibiting the RhoA/ROCK as previously demonstrated. Moreover, the effective reduction of initial AngII-induced alterations in cell membrane permeability-leading to activation of several metabolic pathways that promote OS and generation of pro-inflammatory cytokines-can be accomplished with selective and potent inhibitors of non-selective channels ( Figure 10).

Dye Uptake
Cx HC activity was evaluated by using the dye uptake method previously published [55]. In brief, cells were plated onto glass coverslips and bathed with Locke's saline solution (in mM: 154 NaCl, 5.4 KCl, 2.3 CaCl 2 , 1.5 MgCl 2 , 5 HEPES, 5 glucose, and pH 7.4) containing 5 µM Etd + , which is a membrane-impermeant cationic dye. In time-lapse experiments images were recorded (at regions of interest in different cells) every 30 s during 13 min using a Nikon Eclipse Ti inverted microscope (Tokio, Japan) and NIS-Elements software. Basal fluorescence signal was recorded in cells only in the presence of Locke's saline solution that contained divalent cations.

Dye Coupling
MES-13 cells seeded on glass coverslips were bathed with Locke's saline solution, and cultures were observed using an inverted microscope equipped with a xenon arc lamp and a Nikon B filter (excitation wavelength: 450-490 nm, emission wavelength: above 520 nm). Etd + (25 mM) was microinjected through a glass microelectrode to one cell. Dye transfer to neighboring cells was evaluated two minutes after injection. We routinely performed all dye coupling experiments in the presence of La 3+ (150 µM) to block the Cx HCs and prevent Etd + leakage through open hemichannels that would reduce the intercellular diffusion among coupled cells [66]. The incidence of coupling corresponded to the percentage of cases in which the dye spread occurred to at least one neighboring cell. The coupling index was calculated as the average number of cells to which the dye had spread, divided by the number of positive cases. Intracellular coupling was tested in all experiments by injecting a minimum of 10 cells.

Western Blot Assays
Cell cultures were placed on ice, washed twice with ice-cold PBS (pH 7.4), harvested by scraping in 80 µL of a solution containing a protease and phosphatase inhibitor cocktail (Thermo Scientific, Pierce, Rockford, IL, USA; cat # 78430). Lysates were centrifuged (25,200× g, Eppendorf Centrifuge 5415C, Hamburg, Germany), and supernatants were collected for Western blot analysis. Protein concentration was determined by using the Lowry's method [67]. Samples of homogenized cell cultures (50 µg of proteins) under different conditions were resolved by electrophoresis in 10% SDS-polyacrylamide gel (SDS-PGE), and in one lane pre-stained molecular weight markers contained in an aliquot were resolved. Proteins were transferred to a PVDF membrane (pore size: 0.45 µm), which was blocked at room temperature with Tris pH 7.4, 5% skim milk (w/v) and 1% BSA (w/v). Then, the PVDF membrane was incubated overnight at 4 • C with anti-Cx43 (1:1000), anti-p-MYPT1 (1:500), anti-MYPT1 (1:1000), anti-Panx1 (1:1000) or anti-P2X 7 R (1:1000) antibody, followed by incubation with rabbit or mouse secondary antibody conjugated to peroxidase (1:2000 both of them) for 1 h at room temperature. Then, the PVDF membrane was stripped and reblotted with the anti-α-tubulin antibody (1:5000) used as loading control, following the same procedure described above. After repeated rinses, immunoreactive proteins were detected by using ECL reagents (Pierce Biotechnology, Rockford, IL, USA) according to the manufacturer's instructions. The bands detected were digitized and subjected to densitometry analysis using the software Image J (Version 1.50i, NIH, Washington, DC, USA).

Thiobarbituric Acid Reactive Substances (TBARS) Measurement
The amount of TBARS was estimated using the method described by Ramanathan and collaborators [68] with slight modifications. Culture medium was mixed with SDS (8% w/v), thiobarbituric acid (0.8% TBA w/v), and acetic acid (20% v/v), and heated for 60 min at 90 • C. Precipitated material was removed by centrifugation, and the absorbance of the supernatant was evaluated at 532 nm. The amount of TBARS was calculated using a calibration curve obtained with malondialdehyde (MDA) as standard. MDA was obtained from Merck (Darmstadt, Germany).

Enzyme-Linked Immunosorbent Assay
IL-1β and TNF-α ELISA assays were performed to determine the amount of IL-1β and TNF-α in the extracellular medium under different conditions, following the manufacturer's protocol (IL-1β and TNF-α EIA kit, Enzo Life Science, Farmingdale, NY, USA). Results were normalized by protein amount in ng/mL.

Conflicts of Interest:
The authors declare no conflict of interest.