Gap26, TAT-L2 and ¹⁰panx1 peptides were obtained from Genscript (New Jersey, USA). HEPES, DMEM, DNAse I, poly-L-lysine, CPP, A74003, MRS2179, brilliant blue G (BBG), oATP, ethidium (Etd) bromide, and probenecid (Prob) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Fetal calf serum (FCS) was obtained from Hyclone (Logan, UT, USA). Penicillin and streptomycin were obtained from Invitrogen (Carlsbad, CA, USA). Normal goat serum (NGS) was purchased from Zymed (San Francisco, CA, USA). Cx43^E2, a Cx43 hemichannel antibody to the second extracellular loop was kindly provided by Dr. Jean Jiang, Department of Biochemistry, University of San Antonio, USA (Siller-Jackson et al., 2008).

All animal experimentation were conducted in accordance with the guidelines for care and use of experimental animals of the National Institute of Health (NIH) and local guidance documents generated by the ad hoc committee of the Chilean Research Organization (CONICYT). The studies were performed according to protocols approved by the bioethical committee of Universidad Andrés Bello, Chile.

Wild-type C57BL/6 or Panx1^−/− male mice weighting between 25 and 35 g were used. The generation of Panx1^−/− (KO) mice has been described previously (Anselmi et al., 2008). Mice were housed individually in plastic homecages in a temperature controlled room at 24°C, under a 12 h:12 h illumination cycle (lights on at 8:00 AM). All animals were kept in individual cages throughout the study and had ad libitum access to standard rodent food pellets and tap water. Animals were maintained under standard laboratory conditions for at least 2 weeks before starting the stress protocol. To stress animals, we used a modified version of the restraint protocol described by Mozhui et al. (Mozhui et al., 2010). Animals were segregated in three groups: acute stress, chronic stress and control. For acute stress, animals were placed in ventilated 50 ml Falcon tubes only once for 2 h, prior to behavioral tests. For chronic stress, each mouse was placed into a tube for 2 h per day (14:00 P.M. to 16:00 P.M.) for 10 consecutive days before behavioral evaluations. Non-restrained mice (control group) remained in the home cage until behavioral evaluations.

Mice were decapitated and brains were dissected and placed in ice-cold artificial cerebral spinal fluid (ACSF) containing (in mM): 125 NaCl, 2.5 KCl, 25 glucose, 25 NaHCO₃, 1.25 NaH₂PO₄, 2 CaCl₂, and 1 MgCl₂, bubbled with 95% O₂/5% CO₂, pH 7.4. Hippocampal coronal slices (400 µm) were cut using a vibratome (Leica, VT 1000GS; Leica, Wetzlar, Germany) filled with ice-cold ACSF. The slices were transferred at room temperature (20–22°C) to a holding chamber and immersed in oxygenated ACSF, pH 7.4, for a stabilization period of 30 min before using them.

For “snapshot” experiments, acute slices were incubated with 100 µM Etd for 15 min in a chamber oxygenated by bubbling gas mixture (95% O₂ and 5% CO₂) into ACSF, pH 7.4. Slices were then washed five times with ACSF, fixed at room temperature with 4% paraformaldehyde for 30 min, rinsed extensively in PBS and stored overnight at 4°C in a cryoprotectant solution (30% sucrose). Next day, slices were frozen and dissected in 12–16 µm-thick sections with a cryostat. The sections were then mounted in Fluoromount and incubated in 0.1% PBS-Triton X-100 containing 10% NGS for 30 min. Afterwards, sections were incubated overnight at 4°C with either anti-Iba-1 polyclonal antibody (1:300, Wako), anti-GFAP monoclonal antibody (1:300, Sigma) or anti-NeuN monoclonal antibody (1:400, Chemicon) to detect microglia, astrocytes or neurons, respectively. All antibodies were diluted in 0.1% PBS-Triton X-100 with 2% NGS. After five rinses in 0.1% PBS-Triton X-100, sections were then incubated for 1 h at room temperature with goat anti-rabbit Alexa Fluor 488 (1:1,500), goat anti-mouse Alexa Fluor 488 (1:1,500) or goat anti-mouse Alexa Fluor 647 (1:1,500) antibody. After several washes, slices were mounted in Fluoromount, coverslipped and examined in a confocal laser-scanning microscope (Olympus Fluoview FV1000, Tokio, Japan). Stacks of consecutive confocal images taken with a 63 X objective at 500 nm intervals were acquired sequentially with two lasers (argon 488 nm and helium/neon 543 nm), and Z projections were reconstructed using Fluoview software. Dye uptake ratio was calculated as the subtraction (F−F0) between the fluorescence (F) from respective cell and the background fluorescence (F0) measured where no labeled cells were detected. At least six cells per field were selected from at least three fields in each hippocampal slice.

Acute hippocampal slices were immersed in oxygenated ACSF (as above), pH 7.4, at room temperature (20–22°C) for 30 min under control conditions or exposed to different agents. Then, extracellular ATP was measured using a luciferin/luciferase bioluminescence assay kit (Sigma-Aldrich), while extracellular levels of glutamate were determined using an enzyme-linked fluorimetric assay (Sigma-Aldrich). The amount of glutamate and ATP in each sample was inferred from standard curves as described previously (Orellana et al., 2011a,b). Briefly, after the experiments, the slices were washed twice with ACSF solution and sonicated in ice-cold PBS containing 5 µM EDTA, Halt (78440) and T-PER protein extraction cocktail (78510) according to manufacturer instructions (Pierce, Rockford, IL). Total proteins from tissue homogenates were measured using the Bio-Rad protein assay.

For each data group, results were expressed as mean ± standard error (SEM); n refers to the number of independent experiments. For statistical analysis, each treatment was compared with its corresponding control, and significance was determined using a one-way ANOVA followed, in case of significance, by a Tukey post hoc test.

Restrain stress impairs hippocampus-dependent spatial memory and hippocampal synaptic plasticity, inducing LTP deficits (Luine et al., 1994; Pavlides et al., 2002; Alfarez et al., 2003; Kleen et al., 2006), whereas hemichannel opening has been linked to glial and neuronal dysfunction (Takeuchi et al., 2006; Orellana et al., 2011a,b, 2013; Shestopalov and Slepak, 2014). Therefore, we investigated whether restrain stress affects the functional activity of hemichannels in hippocampal microglia, astrocytes and neurons. Anxiety-like symptoms increased as result of restraint stress using different models. A non-significant tendency for increased thigmotaxis was found in animals that underwent acute restraint stress, increment that became significant in mice subjected to chronic restraint stress, when compared to control mice (from 268.3 ± 4.3 s to 280.6 ± 2.7 s and 289.7 ± 2.7 s, respectively, n = 7, p < 0.05) (Figure 1A). Moreover, in the open field test, the time spent in the center was significantly reduced in mice subjected to acute restraint stress compared to control mice, a decrease that was even larger after chronic restraint stress (from 30.8 ± 4.6 s to 19.5 ± 3.8 s and 10.3 ± 2.7 s, respectively, n = 7, p < 0.05) (Figure 1B). In addition, in the dark and light exploration test, both acute and chronic restraint stress induced a significant reduction in the time spent in the lit compartment compared to control mice (from 95.5 ± 15.5 s to 48.5 ± 7 s and 43.4 ± 11.2 s, respectively, n = 7, p < 0.05) (Figure 1C). These results are indicative of anxiety-like symptoms in mice subjected to acute and chronic restrain confirming that they were stressed and suggesting that chronic stress induces more anxiety-like symptoms than acute stress.

To address whether restrain stress affects hemichannel activity of brain cells, Etd uptake was measured in hippocampal slices of mice that underwent each experimental condition. Etd crosses the plasma membrane of healthy cells passing through poorly selective channels, including connexin and pannexin hemichannels (Schalper et al., 2008). Upon binding to intracellular nucleic acids, Etd becomes fluorescent, and inhibition of this signal with specific blockers is indicative of dye uptake through hemichannels (Schalper et al., 2008; Sáez and Leybaert, 2014). Etd uptake was evaluated in “snapshot” experiments in Iba-1-positive microglia, GFAP-positive astrocytes and NeuN-positive neurons in hippocampal slices. All three cell types from control mice showed a low Etd uptake ratio (Figures 2A–C, 3A–C, 4A–C) as demonstrated previously (Karpuk et al., 2011; Orellana et al., 2011b). However, acute restraint stress increased drastically the amount of Etd uptake in microglia and astrocytes (401 ± 78.6% and 204.4 ± 8.9%; respectively, compared to control, n = 3, p < 0.05) (Figures 2J, 3J), but not in pyramidal neurons (Figure 4J). Microglia have been shown to express functional unopposed pannexin1 (Panx1) and connexin43 (Cx43) hemichannels (Orellana et al., 2011b, 2013; Sáez et al., 2013). The possible role of Panx1 hemichannels in acute stress-evoked Etd uptake was studied using probenecid and the mimetic peptide ¹⁰panx1 with an amino acid sequence homologous to the second loop of Panx1 (Pelegrin and Surprenant, 2006; Silverman et al., 2008). Probenecid (500 µM) and ¹⁰panx1 (200 µM) nearly abolished the increase in microglial Etd uptake triggered by acute restraint stress (from 401 ± 78.6% to 110.3 ± 5.9% and 98.3 ± 15.9%, respectively, n = 3, p < 0.005) (Figures 2G–J). In contrast, mimetic peptides homologous to the cytoplasmic (TAT-L2), first (Gap26) or second (Gap27) extracellular loop of Cx43 (Wang et al., 2013a) and a Cx43 hemichannel antibody (Cx43^E2) (Siller-Jackson et al., 2008), did not affect acute stress-induced Etd uptake by microglia (Figure 2J). Astrocytes express functional unopposed hemichannels formed by Cx43 (Contreras et al., 2002) and Panx1 (Iglesias et al., 2009), thereby we used TAT-L2, Cx43^E2, Gap26, probenecid and ¹⁰panx1 to determine the contribution of each channel type in acute stress-induced Etd uptake in astrocytes. TAT-L2, Cx43^E2, Gap26 and Gap27 fully reduced astroglial cell Etd uptake evoked by acute restraint stress (from 204.4 ± 8.9% to 103 ± 3.8%; 108.4 ± 4.5%, 101.4 ± 5.9% and 101.8 ± 0.3%, respectively, n = 3, p < 0.005) (Figures 3G–J). In contrast, ¹⁰panx1 and probenecid did not affect the stress-induced Etd uptake (Figure 3J).

Responses to acute stress are generally adaptive, but long lasting stress can cause persistent changes and even irreversible damage (Millán et al., 1996; Dhabhar and McEwen, 1997). In agreement with this notion, we found that Etd uptake (% compared to control conditions) induced by chronic stress in microglia and astrocytes was stronger than that found after acute stress (401 ± 78.6% vs. 811.1 ± 124.1%; respectively; and 204.4 ± 8.7% vs. 525.3 ± 4.5%; respectively; n = 3, p < 0.05) (Figures 2D–F,J, 3D–F,J). Probenecid and ¹⁰panx1 nearly abolished the increase in microglial cell Etd uptake triggered by chronic restraint stress (from 811.1 ± 124.1% to 100.9 ± 13.9% and 97.5 ± 7.7%, respectively, n = 3, p < 0.005) (Figure 2J), whereas TAT-L2, Cx43^E2, Gap26 and Gap27 failed to affect this response (Figure 2J). The above findings suggest that in microglia, Panx1 but not Cx43 hemichannels, mediate the restraint stress-induced Etd uptake. This interpretation was supported by the absence of chronic stress-induced microglia hemichannel activation in hippocampal slices from Panx1^−/− mice (Figure 2J). On the other hand, TAT-L2, Cx43^E2, Gap26 and Gap27 partially reduced astroglial Etd uptake evoked by chronic restraint stress (from 525.3 ± 4.6% to 287 ± 13.1%; 276.2 ± 13.1%, 286 ± 7.7% and 290.8 ± 0.6%, respectively, n = 3, p < 0.05) (Figure 3J). Moreover, contrary to the results observed in astrocytes from acute stress mice, ¹⁰panx1 and probenecid inhibited prominently the chronic stress-induced Etd uptake (from 525.3 ± 4.6% to 277.3 ± 2.5% and 288.6 ± 12.1% respectively, n = 3, p < 0.005) (Figure 3J). These data were in agreement with the fact that chronic stress triggered a partial increase of astroglial hemichannel activity in hippocampal slices from Panx1^−/− mice (Figure 3J). Moreover, consistent with the idea that both Cx43 and Panx1 hemichannels were the main contributors to chronic stress-induced Etd uptake in astrocytes, simultaneous blockade of these channels with TAT-L2 and ¹⁰panx1 fully reduced the response (from 525.3 ± 4.6% to 109.2 ± 9.8%, respectively, n = 3, p < 0.005) (Figure 3J). In contrast to the lack of effect of acute stress on neuronal hemichannel activity, chronic stress evoked a prominent increase on Etd uptake in pyramidal neurons (641.9 ± 61.7%, n = 4) (Figures 4A–F). Since most available evidence support the notion that neurons express hemichannels formed by Panx1 (Thompson et al., 2006), we used ¹⁰panx1 and probenecid to determine the contribution of these channels on chronic restraint stress-induced neuronal Etd uptake. ¹⁰panx1 and probenecid strongly reduced the stress-induced Etd uptake observed in pyramidal neurons (from 641.9 ± 61.7% to 81.8 ± 11% and 80.1 ± 17%, respectively, n = 3, p < 0.005) (Figures 4G–J), whereas TAT-L2, Gap19 and Gap26 caused a similar inhibition (from 641.9 ± 61.7% to 96.5 ± 14.9%; 105.1 ± 15.4% and 106 ± 41.5%, respectively, n = 3, p < 0.005) (Figure 4J). Accordingly, chronic restraint stress failed on evoke Etd uptake in hippocampal neurons from Panx1^−/− mice (Figure 4J). Moreover, basal levels of Etd uptake in microglia, astrocytes or neurons from Panx1^−/− mice were similar to that observed in wild type mice (data not shown). Overall, these data indicate that both acute and chronic restraint stress increase hemichannel opening of glial cells and neurons, being chronic restraint much more powerful than acute restraint stress in evoking this response.

Given that pathological conditions affect the expression of connexins and pannexins in the CNS (Rouach et al., 2002; Orellana et al., 2009), we examined whether chronic or acute restraint stress could modulate Cx43 and Panx1 levels in brain cells by confocal analysis. Interestingly, chronic but not acute restraint stress evoked a significant increase on Panx1 levels in astrocytes and neurons when compared to control conditions (Figures 5A–H,J,K). However, neither Cx43 nor Panx1 levels were affected in microglia in mice subjected to chronic restraint stress (Figure 5I). Similarly, for all tested conditions, Cx43 remained unchanged in astrocytes (Figure 5J).

Under activated state, glial cells release relevant amounts of gliotransmitters including glutamate and ATP, which underlie glia-to-glia and glia-to-neuron communication via glutamatergic and purinergic receptors, respectively (Perea et al., 2009; Perea and Araque, 2010). Because opening of hemichannels has been asociated with purinergic and glutamatergic signaling (Locovei et al., 2006; Thompson et al., 2008; Orellana et al., 2011a,b), we examined if NMDA and P2X₇ receptors are involved in chronic restraint stress-induced Etd uptake. We found that CPP, a NMDA receptor blocker, strongly abolished the Etd uptake evoked by chronic restraint stress in astrocytes (from 100% of stress-induced effect to 28.7 ± 5.4%, n = 3, p < 0.05) (Figure 6), whereas in microglia and pyramidal neurons caused a small inhibition (from 100% of stress-induced effect to 69.7 ± 3.8% and 49.5 ± 0.4%, respectively, n = 3) (Figure 6). Moreover, blockade of P2X₇ receptors with BBG, oATP and A740003 induced a prominent reduction on chronic stress-induced Etd uptake in microglia (from 100% of stress-induced effect to 29.6 ± 2.2%, 30.3 ± 0.7% and 30.5 ± 5.6%, respectively, n = 3, p < 0.05) and in neurons (from 100% of stress-induced effect to 44.2 ± 4.7%, 47 ± 3.4% and 53.6 ± 2.2%, respectively, n = 3, p < 0.05), with a lesser but significant decrease in astrocytes (from 100% of stress-induced effect to 67.3 ± 10.5%, 63 ± 10.2% and 60.5 ± 4.9%, respectively, n = 3) (Figure 6). To elucidate if in addition to P2X₇ receptors, metabotropic purinergic receptors might also be involved in chronic restraint stress-induced hemichannel opening, we used MRS2179, a blocker of P2Y₁ receptors, which has been previously linked to hemichannel opening in the CNS (Orellana et al., 2012; Sáez et al., 2013). MRS2179 did not affect the Etd uptake induced by chronic restraint stress in all brain cells studied (Figure 6). In agreement with the idea that both NMDA and P2X₇ receptors are involved in hemichannel opening induced by chronic restraint stress, blockade of both receptors with CPP and A740003, respectively, fully reduced this response in microglia, astrocytes and neurons (from 100% of stress-induced response to 8.0 ± 1.0%, 12.1 ± 0.4% and 10.3 ± 0.1%, respectively, n = 3, p < 0.05) (Figure 6). Taken together these data indicate that Etd uptake induced by chronic restraint stress depends on NMDA/P2X₇ receptor signaling.

Recently, gliotransmitters were shown to elicit their own release in an autocrine manner, via Cx43 and Panx1 hemichannels (Orellana et al., 2012, 2013). Given that NMDA/P2X₇ receptors are involved in the Etd uptake observed in hippocampal cells of mice subjected to restraint stress, we evaluated whether this condition affects the glutamate and ATP release from hippocampal slices via Cx43 and/or Panx1 hemichannels. Acute and chronic stress strongly increased the release of glutamate and ATP (from 32.5 ± 4.4 pmol/mg to 56.4 ± 8.5 pmol/mg and 138.8 ± 25.9 pmol/mg, respectively and from 13.5 ± 4.2 pmol/mg to 30.7 ± 4.6 pmol/mg and 75.3 ± 9.6 pmol/mg, respectively, n = 3, p < 0.05) (Figure 7). Interestingly, TAT-L2, Gap26, ¹⁰panx1 and probenecid prominently reduced the release of glutamate (from 138.8 ± 25.9 pmol/mg to 20.4 ± 3.3 pmol/mg, 31.7 ± 7.8 pmol/mg, 35.7 ± 7.5 pmol/mg and 25.5 ± 2.3 pmol/mg, respectively, n = 3) and ATP (from 75.3 ± 9.6 pmol/mg to 11.6 ± 2.3 pmol/mg, 13.8 ± 3.6 pmol/mg, 16.5 ± 5.2 pmol/mg and 12.5 ± 1.8 pmol/mg, respectively, n = 3, p < 0.05) induced by chronic restraint stress (Figure 7). These findings indicate that chronic stress increases the release of glutamate and ATP via opening of Cx43 and Panx1 hemichannels.

In support for the notion that gliotransmitters can elicit their own release, we found that CPP, BBG, oATP, A740003, CPP plus A74003, but not MRS2179 abolished almost completely the release of glutamate (from 138.8 ± 25.9 to 38.7 ± 5.9 pmol/mg, 40.4 ± 7.9 pmol/mg, 36.3 ± 7.9 pmol/mg, 45.6 ± 3.5 pmol/mg, 25.6 ± 5.6 pmol/mg and 145.6 ± 10.5 pmol/mg, respectively, n = 3, p < 0.05) and ATP (from 75.3 ± 9.6 pmol/mg to 34.9 ± 6.5 pmol/mg, 33.7 ± 8.5 pmol/mg, 30.7 ± 11.9 pmol/mg, 48 ± 17.1 pmol/mg, 17.4 ± 3.2 pmol/mg and 83.1 ± 17.9 pmol/mg, respectively, n = 3, p < 0.005) induced by chronic restraint stress (Figure 7). This evidence suggests that both glutamate and ATP evoke their own release by an autocrine pathway possibly mediated by unopposed Cx43 and Panx1 hemichannels.