J. Neurochem. (2010) 112, 92–102.
Some forms of seizure activity can be stopped by gap junctional (GJ) blockade. Here, we found that GJ blockers attenuate hippocampal seizure activity induced by a novel seizuregenic protocol using Co2+. We hypothesized that this activity may occur because of the altered expression of connexin (Cx) and/or pannexin (Panx) mRNAs and protein. We found a 1.5-, 1.4-, and 2-fold increase in Panx1, Panx2, and Cx43 mRNAs, respectively. Significant post-translational modifications of the proteins Cx43 and Panx1 were also observed after the Co2+ treatment. No changes were observed in the presence of tetrodotoxin, indicating that seizure activity is required for these alterations in expression, rather than the Co2+ treatment itself. Further analysis of the QPCR data showed that the Cx and Panx transcriptome becomes remarkably re-organized. Pannexin (Panxs 1 and 2) and glial connexin mRNA became highly correlated to one another; suggesting that these genes formed a transcriptomic network of coordinated gene expression, perhaps facilitating seizure induction. These data show that seizure activity up-regulates the expression of both glial and neuronal GJ mRNAs and protein while inducing a high degree of coordinate expression of the GJ transcriptome.
artificial cerebrospinal fluid
The activity mediated by gap junctions has been increasingly implicated in brain pathologies (Condorelli et al. 2002; Rouach et al. 2002; Salameh and Dhein 2005), particularly in epileptic seizures (Carlen et al. 2000; Perez-Velazquez and Carlen 2000; Traub et al. 2004). It has been shown that gap junctional (GJ) communication contributes to neuronal synchrony and rhythmicity, and GJ blockers are effective in reducing seizures in several in vitro (Jahromi et al. 2002; Samoilova et al. 2003; Gigout et al. 2006) and in vivo (Gajda et al. 2005, 2006; Bostanci and Bağirici 2006, 2007) models. Alterations in GJ expression may be involved in epileptogenic processes, since seizure activity has been shown to alter or increase the mRNA expression and protein of Connexin (Cx) GJs (Naus et al. 1991; Li et al. 2001; Samoilova et al. 2003; Collignon et al. 2006; Gajda et al. 2006; Zappalàet al. 2006). Recently, it was shown that Cx43 mimetic peptides inhibit spontaneous epileptiform activity in organotypic hippocampal slice cultures (Samoilova et al. 2008). Although studies have shown, using semi-quantitative methods, that alterations in Cx expressions are associated with seizure activity, to date, there is no published literature showing alteration in pannexin (Panx) expressions in any seizure models.
We have shown previously that mechanisms involving activity-dependent facilitation of GJ communication may play a major role in Co2+-induced epileptiform discharges (He et al. 2009). To further understand the underlying cellular mechanism of the Co2+-induced seizure activity, the expression levels of Cx and Panx mRNA and protein were analyzed using quantitative RT-PCR (QRT-PCR) and western blotting in an in vitro mouse hippocampal model.
Materials and methods
All experimentation conducted in this study has been reviewed and approved by the animal care committee of our institution.
Preparation of mouse hippocampal tissues
C57BL/6 male mice (postnatal 15 days; Charles River Laboratories, Montreal, QC, Canada) were decapitated under isoflurane anesthesia, and their brains were quickly dissected out and maintained in an ice-cold, oxygenated artificial cerebrospinal fluid (ACSF in mM: 125 NaCl, 3.5 KCl, 1.25 NaH2PO4, 25 NaHCO3, 2 CaCl2, 1.3 MgSO4, and 10 glucose) at pH ∼7.4 when aerated with 95% O2-5% CO2 for a few minutes before further dissection. For preparing whole hippocampal isolates (Wu et al. 2002), we used a fine glass probe and gently separated the dentate gyrus from the CA1 area, and then removed the dentate gyrus area while preserving the CA3-CA1 tissues. The whole hippocampal isolates were maintained in warmed ACSF (31–32°C) for 1 h before further experimental manipulations.
Whole hippocampal isolates (left and right) from C57BL/6 male mice were prepared and pre-incubated separately in warmed ACSF (31–32°C) for 1 h before further experimental manipulations. Then, 100 μM CoCl2 (Sigma-Aldrich, Oakville, ON, Canada) was added to the ACSF of one of the isolates. Both control (without cobalt) and treated hippocampal isolates were continuously incubated for 1 h at the same temperature. The Co2+-treated hippocampal isolates were then transferred to the standard ACSF and allowed to recover for at least 1 h before further experiments. In some experiments, tetrodotoxin (TTX, 1 μM) was added in the Co2+-containing ACSF to block seizure-like activities. Here, the optimum incubation time for Co2+ treatment was chosen, based on previous studies (He et al. 2009).
Hippocampal isolates were recorded in a submerged chamber at 32°C (Wu et al. 2005). Extracellular field potentials were recorded by NaCl-filled glass electrodes (∼2 MΩ) using an Axoclamp-2B amplifier (Axon Instruments/Molecular devices, Sunnyvale, CA, USA). Data acquisition, storage, and analyses were done using a digital-analog interface board (Digidata 1200 or 1300A; Axon Instruments/Molecular devices) and PCLAMP software (version 8 or 9; Axon Instruments).
Total RNA preparation, first strand synthesis
Whole hippocampal isolates obtained from each mouse was divided into two groups: control and treated (Co2+ or Co2+/TTX). The total RNA was isolated using PureLinkTM Micro-to Midi Total RNA Purification System kit (Invitrogen, Burlington, ON, Canada) according to the manufacturer’s protocol. The isolate was treated with DNase I to remove contaminating genomic DNA and the RNA was eluted with 30 μL of nuclease-free water. The absorbance readings were taken using Nanodrop Spectrophotometer ND-1000 (Nanodrop Technologies Inc., Wilmington, DE, USA) to calculate the concentrations of each RNA preparation. The OD260/80 ratio of all samples was greater than 1.8. For reverse transcription, 1 μg total RNA was used to make cDNA according to the manufacturer’s protocol (Invitrogen). This reaction was then diluted to give the final concentration of 20 ng total RNA equivalent/μL of the template. For each QRT-PCR reaction, 2 μL of this solution was used.
PCR primer design and selection
Primers were designed for ‘Universal reaction conditions’ using the program Primer ExpressTM (Applied Biosystems, Foster City, CA, USA) and Primer 3. Each primer was designed to produce 90–110 bp amplicon. The gene-specific primers used for QRT-PCR were synthesized by Sigma-Genosys (Oakville, ON, Canada) and listed in (Table 1). To identify the optimum annealing temperature for each primer pair gradient PCR was conducted using Platinum® Taq DNA Polymerase (Invitrogen). The annealing temperature range was set between 53°C and 65°C. PCR products from each reaction were then run on a 2% agarose gel. A visual inspection then determined what annealing temperature was best. For nearly all reactions, 58°C appeared to be optimum. This temperature was used to assess primer efficiency in the next step.
|Genes||Gene-specific primer sequences||Slope||r-Value||Efficiency (%)|
Cycling was performed on the ABI 7900 HT (Applied Biosystems) using the following protocol: 3 min at 95°C followed by 50 cycles of 15 s at 94°C and 30 s at 58°C. Finally, a dissociation profile of the PCR product(s) was obtained by a temperature gradient running from 60°C to 94°C. The Ct (cycle threshold) is defined as the number of cycles required for the fluorescent signal to cross the threshold (exceeds background level). Ct levels are inversely proportional to the amount of target nucleic acid in the sample (the lower the Ct level the greater the amount of target nucleic acid in the sample). Cts < 29 are strong positive reactions indicative of abundant target nucleic acid in the sample. The Ct values for the reference gene, Synaptophysin (Syn) were used to take into account different mRNA levels of the sample for normalization. Synaptophysin was used, as it has been shown to be stably expressed under a number of instances (Chen et al. 2001).
Real time PCR was performed using SYBR-Green RT-PCR kit (Bio-Rad Laboratories Ltd., Mississauga, ON, Canada), according to the manufacturer’s protocol in a 20 μL volume. The real time PCR conditions were used as follows: 3 min at 95°C followed by 50 cycles of 15 s at 94°C and 30 s at 58°C. Finally, a dissociation profile of the PCR product(s) was obtained by a temperature gradient running from 60°C to 94°C.
To examine GJ mRNA expression, hippocampi isolated from 21 mice were divided into three groups: controls, Co2+- treated, and Co2+/TTX-treated (see Materials and methods). Within each group, Ct values were normalized with reference gene, Syn, to compare changes in GJ gene expressions. The expression levels of the mRNA analyzed are presented as normalized Ct values (Ctn) rather than △△Ct that is more commonly used. This is calculated as the difference in Ct reference gene and the Ct gene of interest, giving a normalized Ct (Ctn). As the threshold for the reference gene is usually smaller than the cycle threshold for the genes of interest, the Ctn is usually negative. Numbers that are less negative indicate higher expression. The reason for this presentation style is so that we can better illustrate relative transcript abundance. Also this presentation style shows the range where the alterations in abundance occur. In the text and figures legends, where relevant, we give the average fold change in the expression, as well. Statistical analyses were performed on the Ctn values as described below.
Statistical analyses were done with Stat-View software (Abacus Concepts, Berkeley, CA, USA) using anova with Fischer’s PLSD post hoc test. The normalized Ct values were used in pair-wise comparisons among group levels using anova. This determines the significance of the effects or factors (cobalt treatment) by calculating how much of the variability in the dependent variable (Normalized Ct values of the control vs. treatment) can be explained by the different treatments. Using Fisher’s multiple comparisons test; pair-wise comparisons were made to determine which populations are significantly different from others by calculating the mean difference. The level of significance was set to p < 0.05. Pearson product correlations were conducted independently for glial and neuronal GJ mRNAs, to determine the relations between specific family members. A Z-test was used to determine if the calculated r-values for these relations were significant. In the tables, significant (p < 0.05) correlations are indicated in bold.
Western blot analysis
For protein extraction the whole hippocampal isolates from 10 animals were homogenized in ristocetin-induced platelet agglutination buffer (RIPA) containing 50 mM Tris (pH 8.0), 150 mM NaCl, 1% sodium dodecyl sulfate (SDS), 1% Igepal CA-630, 0.5% sodium deoxycholate, 1 mM NaF, 10 μM/mL phenylmethylsulfonyl fluoride (PMSF), 1 mM Na3VO4, and protease inhibitor cocktail (Hoffmann-La Roche Limited, Mississauga, ON, Canada) on ice. Homogenates were incubated on ice for 30 min and sonicated twice for 20 s on ice. Protein concentrations were determined with the bicinchoninic acid protein assay reagent (Pierce/Thermo Fisher Scientific, Nepean, ON, Canada). Proteins were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE). Thirty micrograms of each homogenate was loaded into each lane. For Western blotting, the gel was transferred to polyvinylidene difluoride membrane (PVDF; Bio-Rad) and blocked in 5% (wt/vol) non-fat dry milk and 0.05% Tween 20 in Tris-buffered saline (pH 7.6) for 1 h and then incubated with Cx43 (1 : 500 dilution, Millipore, MAB3068; Lot# LV1519578), Panx1 (1 : 1000 dilution, Millipore AB9886; Lot# LV140345), Panx2 [1 : 250 dilution (AVIVA SYSTEMS BIOLOGY, San Diego, CA, USA), ARP42778_T100; Lot# NP_443071], Syn (1 : 500 dilution, Sigma, S5768; Lot# 045K4828, Canada) and β-actin (1 : 1000 dilution,Stressgen,905-733-100) antibodies in Tris-buffered saline with 5% non-fat dry milk and 0.05% Tween 20 overnight. Membranes were washed and incubated with anti-rabbit (for Panx1, Panx2 and β-actin assays, Cedarlane, Canada) or anti-mouse [for Cx43 and Syn (Sigma)], secondary antibody conjugated with horseradish peroxidase. The immunoreaction was detected with a chemiluminescence kit (Amersham Biosciences Inc./GE Healthcare, Baie d'Urfe, Quebec, Canada). To confirm a consistent protein loading for each lane, membranes stained for Syn β-actin. After the X-ray films were scanned, the signal intensities of the bands were analyzed with Image J software (NIH Image, Bethesda, MD, USA). Similar assays have been performed using hippocampal isolates treated with Co2+ and TTX.
Co2+ induced seizure activity in vitro
In previous studies, we have developed a novel in vitro model of Co2+-induced seizures in thick mouse hippocampal slices (He et al. 2009). Since Co2+-induced seizures originate from the CA3 area, for this study, whole hippocampal isolates were prepared after removal of the dentate gyrus, to focus on the GJ expression in the CA3-CA1 areas. When monitored via extracellular recordings from the CA3 or CA1 area, the Co2+-treated hippocampi display spontaneous ictal-like discharges that manifested as repetitive spikes with peak amplitudes of up to 2 mV and durations of up to 15 s (Fig. 1a, top panel). These ictal-like discharges were reversibly abolished by carbamazepine, an anti-convulsive agent known to suppress sodium channel activities. In the presence of carbamazepine (applied at 50 μM for 8–9 min), intermittent interictal-like sharp waves (200–300 ms) were recognizable (Fig. 1, middle panel). The gap junction blocker, octanol, applied at 1 mM for 8–10 min (Fig. 1b, bottom panel) abolished both the ictal-like and interictal discharges. The blockade of ictal-like discharges by carbamazepine or octanol was observed in seven of seven hippocampi examined. Prior to the application of carbamazepine or octanol, the mean duration of ictal-like discharges was 8.84 ± 0.74 s (n = 3 tissues) or 9.0 ± 1.1 s (n = 4 tissues) (Fig. 1c). Collectively, these observations indicated that GJs play a pivotal role in the generation of ictal-like discharges following the cobalt treatments.
Changes in GJ mRNA expression as a result of Co2+-induced seizures
The efficiency of each primer used in this study was determined from the slope of the plot between cycle thresholds and total brain RNA (in grams) that had previously been reverse transcribed. The linearity was verified by regression analysis; r > 0.90 was set as a cut off for acceptable linearity. The efficiency of the amplification was determined from the slope of this linear relationship (see Table 1 for calculated slopes r-values and derived efficiency).
In control and in both Co2+- and Co2+ + TTX-treated hippocampal tissues, the level of expression showed the order of Cx43 ∼ Cx30 ∼ Panx2 > Panx1 ≥ Cx32 ∼ Cx47 > Cx26 > Cx36 ∼ Cx45 > Cx40. Although, seizure activity after Co2+ treatment significantly increased the levels of highly expressed GJ mRNAs, i.e., Cx43, Panx2, and Panx1, the relative abundance of all GJ mRNA remained the same.
In our experiments, mRNAs of Panx1, Panx2, and Cx43 were considered significantly elevated because the change between the control and treated brain was ≥ 1.5-fold with a p < 0.05. Panx1, Panx2, and Cx43 mRNAs increased by 1.5-, 1.4- and 2-folds, respectively (Fig. 2a). Although the Cx47 was apparently increased by nearly twofold, this effect did not reach statistical significance (Fig. 2a). The up-regulation of Cx43, Panx1, and Panx2 mRNAs was directly associated with seizure-like activity, as this up-regulation was absent when the seizures were prevented by TTX incubation which blocks neuronal activity (Fig. 3a).
To determine whether the change in mRNA levels causes a change in protein content, western blot analysis was done. In Fig. 2(b), we showed that after Co2+-treatment, the level of the native form of Panx1 (∼43 kDa) was decreased but the glycosylated form (∼48 kDa) was increased (p < 0.05). Another band at ∼46 kDa disappeared after Co2+-treatment. Similarly, after treatment, one of the phospho-Cx43 protein (∼44 kDa) was increased whereas another phospho-Cx43 protein (∼46 kDa) and the native form of Cx43 (∼41 kDa) showed a statistically insignificant slight increase. Although Panx2 proteins at ∼62 kDa (monomer) and ∼120 kDa (dimer) did not show any significant changes, a small band at ∼70 kDa appeared in the treated tissue, which could represent its post-translational modification form (phosphorylated or glycosylated). The same samples measured for Syn or β-actin showed no changes at the protein level. Similar analysis using Co + TTX treatment of the tissues showed no change in Cx43, Panx1, and Panx2 proteins (Fig. 3b).
We have previously shown that GABAA receptor transcript abundance may be correlated between experimental replicates and that these correlations may be perturbed or re-organized under some conditions (Merali et al. 2004). Highly correlated transcript abundance implies that the levels of expression are controlled relative to one another to provide proper functionality of the oligomeric proteins. Thus, as gap junctions are constructed from a wide range of proteins analogous to the GABAA receptor, we investigated whether similar relations are evident between GJ mRNAs and whether they may change because of seizures. Table 2 summarizes the correlation coefficients for each pair of mRNA species (control vs. Co2+-treatment). Panxs and Cx45 mRNAs, reported to be present in neurons (Söhl et al. 2005) were not highly correlated to one another or with other Cxs (only 3 of 38 correlated) in the control, however, after the Co2+-treatment causing recurrent seizures, Cx45, Panx1 and Panx2 showed strong correlations among themselves and with all of the glial GJ mRNAs. Neuronal Cx36, however, was not correlated to any other GJ mRNAs before or after the treatment. In control tissues, we found that Cxs 26, 30, 32, 40, 43, and 47, as a group, were not highly correlated, only 3 of 15 showing significant regression values (not shown). However a remarkable re-organization of expression was evident after treatment. All transcripts became highly correlated to one another (Table 2a). Correlation coefficient studies for the mRNAs from tissues treated with Co2+ + TTX showed no correlation before or after the treatment (Table 2b), implying that the levels of GJ expression became highly co-regulated as a result of seizure activity.
Direct intra-cortical or intra-hippocampal Co2+ application induces seizures in several animal species including rats, mice, cats, and monkeys. These cobalt-induced seizures share some common features with human epilepsies including intermittent occurrence, paroxysmal nature, and response to anti-convulsants (Craig and Colasanti 1992). We have chosen the in vitro cobalt model for this study because in this model, ictal-like discharges occur spontaneously that last up to 45 s without overall depolarization of hippocampal neurons (He et al. 2009). Pharmacologically, the cobalt-induced seizures in vitro are similar to in vivo ictal discharges as they are blocked by phenytoin and carbamazepine. It is unlikely cause neuronal death because the acute action of Co2+ is to suppress transmission and hence, reduce glutamate toxicity. In the absence of seizure-like events, neuronal activities are largely ‘normal’. Seizure-like activity occurs after washout of Co2+, and in the absence of pharmacological alteration of major neurotransmitter receptors or ionic channels.
We hypothesize that during the epileptiform discharges, the intensive firing of hippocampal pyramidal neurons, interneurons, and the associated synaptic activities, causes a large Ca2+ influx through voltage- and ligand-gated channels. The Ca2+ influx leads to a substantial decrease in extracellular Ca2+ because of fast Ca2+ entry and relatively slow redistribution of extracellular Ca2+(Heinemann et al. 1986; Rusakov and Fine 2003). The decrease in extracellular Ca2+ acts as a positive feedback factor that facilitates GJ communication (Perez-Velazquez et al., 1994; Perez-Velazquez and Carlen 2000; Carlen et al. 2000) and perhaps also increases the open probability of GJ hemi-channels (Goodenough and Paul 2003;Spray et al. 2006; Thompson et al. 2006, 2008; Scemes et al. 2008). Although the exact mechanisms by which Co2+ induces ictal-like discharges remain to be elucidated, data from our previous experiments suggested that an activity-dependent facilitation of GJs play a major role in the generation of these ictal-like discharges (He et al. 2009).
We show here that a brief Co2+-treatment leading to seizure activity causes discrete, quantifiable changes in GJ expression, i.e., increased expression of Panx1, Panx2, and Cx43. Such increased expression is a result of seizure-like activity because it is absent in the hippocampal tissues treated with Co2+ and TTX. Our present observations are in agreement with the previous studies that epileptic or seizure activity is accompanied by specific alterations in different GJ mRNA and protein levels (Naus et al. 1991; Li et al. 2001; Fonseca et al. 2002; Samoilova et al. 2003; Collignon et al. 2006; Gajda et al. 2006; Zappalàet al. 2006).
One of the unique characteristics of most Cx family members is that their physiological regulation is not only at the level of channel gating but also via processes leading to synthesis, assembly, internalization and turnover of the GJ proteins (Laird 2006). Previous studies have shown that phosphorylation of Cx43 enhances GJ intercellular communication (Vikhamar et al. 1998) whereas conversely, protein kinase C activation inhibits Cx43 hemi-channel conductance (Hawat and Baroudi 2008). Enhanced glial Cx43 GJ communication also could be related to wider and quicker transfer of potassium, calcium, epileptogenic substances, or electrical field effects. We have shown that blocking Cx43-based GJ communication by a Cx-specific mimetic peptide prevented spontaneous seizures in organotypic hippocampal slice cultures (Samoilova et al. 2008).
Up-regulation of Cxs has been reported in several seizure models using semi-quantitative methods before, however, this is the first demonstration using a quantitative analysis and also showing that seizure activities alter Panx gene expression. Physiological studies revealed that Panx channel properties are distinct from those of Cxs, as Panxs appear to be unable to form sizable amounts of cell-to-cell channels under most conditions (Spray et al. 2006; Huang et al. 2007; Scemes et al. 2007, 2008). This lack of formation of functional GJs is likely because of the glycosylation of the extracellular loops of Panx protein (Boassa et al. 2007, 2008; Penuela et al. 2007; Shestopalov and Panchin 2008). The degree of glycosylation observed in Panx1 was also highly variable among tissues, i.e., very high in the brain but weak in skin and cartilage, suggesting that the functional role of Panx1 in brain tissue might be differentially regulated by the state of glycosylation (Penuela et al. 2007). Panx1 ‘hemi-channels’ have been demonstrated to be activated by a diverse set of experimental conditions including mechanical stress, oxygen and glucose deprivation, strong depolarization, activation of purinergic receptors including P2Y1r, P2Y2r, and P2X7r, by ATP, and other agonists (Lacovei et al. 2007; Thompson et al. 2006; Scemes et al. 2008; Thompson et al. 2008). Panx1 has been proposed to provide a paracrine-signaling pathway by releasing ATP (Bao et al. 2004; Locovei et al. 2006; Iglesias et al. 2009) thus modulating the range of the intercellular Ca2+ wave transmission between astrocytes in culture.
In this study, the up-regulation of a glycosylated form of Panx1 protein could be an indication of ‘hemi-channel’ activation in response to seizure activity. GJ channels and ‘hemi-channels’ share several common features, including permeability characteristics and sensitivity to blocking drugs. However, which GJ protein species actually establishes the functional membrane conductance and permeability is currently unknown because Panxs and Cxs have largely overlapping distributions in vertebrates. It is plausible to consider that Panxs and Cxs may have different roles; Panxs having a paracrine role, releasing ATP (Locovei et al. 2006), whereas Cxs forming GJ channels which coordinate the electrical activities of groups of cells and the direct transmission of Ca2+, other ions, and small molecules between cells.
The coordinated gene expression as measured by our correlation coefficient studies indicate that, with the exception of Cx36, the mRNA expression of all the Cxs and the two Panxs are highly correlated after the Co2+-induced seizures (Table 1). Although connexin proteins are tissue specific, some cells contain more than one type of connexin and there may be a common mechanism or pathway that controls the expression of their genes. Unlike other GJ genes, Cx36 belongs to the γ class, is the only gene reported to be neuron-specific, and has never been found in any other tissues (Kumar and Gilula 1992; O’Brien et al. 1996; Nagy et al. 2004). The fact that its expression is not correlated to any of the GJ genes after Co2+ treatment could be an indication that Cx36 expression follows a different signaling pathway.
The finding that up-regulation of glycosylated Panx1 proteins, in addition to the phosphorylated form of Cx43 forming the classical cell-to-cell GJ channels in the Co2+-induced seizure model, has opened a new field in the area of GJ proteins. These data also provide insights into the regulation of several other cell functions, for which the occurrence of hemi-channel may provide a plausible mechanism. Recent studies using microarray analysis, including expression, variability, regulation, and coordination of numerous genes in the brains of Cx43, Cx32, and Cx36 knockout mice, suggest that the brain transcriptome is organized into Cx-dependent transcellular networks, so that altering expression of a single gene can produce downstream and parallel ‘ripples’ contributing to the phenotypic changes in the coupled cells (Iacobas et al. 2007). In conclusion, our correlation coefficient studies suggest a remarkable overlap between the regulation of all glial Cx mRNAs and Panx mRNAs in the seizure-induced hippocampal tissue.
This research was supported by grants from the Canadian Institutes for Health Research (MOP-36415) to P.L.C. and M.O.P., Hospital for Sick Children Research Foundation and Epilepsy Canada to P.L.C.; M.O.P. is a NARSAD Independent Researcher. Special thanks to Michelle Everest for her technical support.