• Gap junction channels;
  • Synchrony;
  • Connexins;
  • Development;
  • Epileptogenicity;
  • 4-Aminopyridine


  1. Top of page
  2. Abstract
  7. Acknowledgments

Summary: Purpose: The functional significance of gap-junction (GJ) channels in seizure susceptibility and induction and maintenance of seizures in the developing rat brain was investigated on the 4-aminopyridine (4-AP) in vivo epilepsy model.

Methods: In electrophysiological experiments, GJs were manipulated with a blocker or opener before induction or at the active epileptic foci between postnatal days 9 and 28 (P9–28). Semiquantitative reverse transcriptase–polymerase chain reaction (RT-PCR) amplification was used to measure the levels of connexin (Cx) 26, 32, 36, and 43 mRNAs at the untreated cortex or epileptic foci.

Results: The basic electrocorticogram (ECoG) and Cx messenger RNA (mRNA) expression patterns exhibited characteristic maturation; the 4-AP–induced epileptiform activity correlated well with these changes. Cx mRNA expressions were significantly upregulated around P16 (except for Cx26). The Cx26, 36, and 43 gene inducibility was highest around P16 and then declined significantly. In the youngest animals, the GJ opener induced rhythmic synchronous cortical activity. On maturation, the seizures became focalized and periodic; the discharges accelerated their amplitude and frequency increase. A transient decrease (P13–14) and then increase (P15–16) in seizure susceptibility were followed by a tendency to periodicity and focalization.

Conclusions: The study suggests that GJ communication is involved in rhythm genesis and synchronization of cortical activity and may enhance the epileptogenicity of the developing brain.

Clinical experience and various experimental data indicate that the developing nervous system is more sensitive than the mature one to different convulsive effects (1–5). Although the physiological factors underlying this differential epileptogenicity have not been fully clarified, the higher susceptibility of the immature brain can be explained by certain characteristic neurobiologic features. The developing brain exhibits a high metabolic rate, abundant neuronal and synaptic networks, the overexpression of receptors and enzymes, the depolarizing effect of γ-amino-acid, the hypersynchrony of neuronal circuits, and enhanced synaptic plasticity (6,7). In addition, the immature cerebral cortex and hippocampus have higher densities of excitatory amino acid receptors and gap junction (GJ) channels as compared with the adult organs (4,8).

Intercellular communication via GJ channels is an important form of cell-to-cell communication in early brain development (8–14). Electrical coupling via GJ channels has been reported both between pairs of inhibitory neurons and among inhibitory and excitatory neurons during the early postnatal days in the rat cortex (10). Moreover, the incidence of coupling between neurons and glia has been observed at this early age in rats (8,12,15,16). It is believed that a possible correlation exists between the high seizure susceptibility of the immature brain and the elevated communication through the GJ channels (4). However, the role of GJ coupling in epilepsy in the adult and the developing nervous system is still not fully understood.

Accordingly, the purpose of the present study was to investigate the functional significance of GJ channels in the epileptogenicity and seizure susceptibility of the immature mammalian brain. With this aim, we made use of the K+ channel blocker 4-aminopyridine (4-AP)-induced epileptiform activity in rats between postnatal days 9 and 28 (P9–28). This neocortical seizure model is appropriate for studying epileptiform activity in the developing nervous system, because infants and children often have seizures of neocortical origin (17).

We investigated whether the intensity of GJ communication influences the process of development of epilepsy in normal brain tissue (which is called epileptogenesis) and the possible contribution of GJ communication in the induction, manifestation, and propagation of seizures at already established epileptic foci (ictogenesis) at different developmental points. By combining electrophysiology and pharmacology, we examined the effects of the functional state of the GJ channels modified by the local application of either a blocker or opener on the basic electrocorticogram (ECoG) and on the seizure activity of the neocortex of the developing brain. Further, the developmental expression levels and the plastic changes induced in the connexin (Cx) 26, 32, 36, and 43 messenger RNA (mRNA) levels by epileptiform activity were examined in the rat neocortex by me means of semiquantitative reverse transcriptase–polymerase chain reaction (RT-PCR) analysis. This work is an extension of our earlier studies carried out on adult animals (18–20).


  1. Top of page
  2. Abstract
  7. Acknowledgments


For electrophysiologic and RT-PCR studies, P9–28 and adult Wistar rats of both genders were used. The animals were bred in our laboratory and housed under standard laboratory conditions, with food and water available ad libitum. The pups were housed together with their mothers in individual cages until chosen for study or until weaning on day P24 (the day of birth was taken as day P0). For the assessment of seizure susceptibility, animals were selected randomly from a litter of the appropriate age. To avoid interlitter variations in susceptibility to seizures, the animals for tests of the same kind of treatment were chosen from different litters (five animals for each experimental paradigm, in the five developmental stages). Data were pooled for statistical analysis.

Under general anesthesia (sodium pentobarbital, 50 mg/kg, i.p.), the heads of the animals were secured in a stereotaxic instrument. For the recording of ECoGs, four holes (2–3 mm wide) were drilled in the skull, and the dura mater was carefully removed at the prospective site of the primary focus (Pf). The experimental paradigm is illustrated in the schematic Fig. 1.


Figure 1. Representative samples showing the basic ECoG (left column) and the effects of trimethylamine (TMA) (right column), recorded from the same animal at different ages (A–E). The scheme illustrates the arrangement of the ECoG recording electrodes and the sites of Pf and Mf. P, here and in the following figures means postnatal day.

Download figure to PowerPoint

The Pf was induced by the local application of crystalline 4-AP (on a piece of filter paper soaked with saline) to the somatosensory cortical surface. Four silver-ball electrodes were used to record the ECoG from the Pf, from the secondarily induced mirror focus (Mf), and from two other points, to detect the propagation of epileptiform discharges. The electrophysiologic data were recorded continuously by means of an eight-channel electroencephalograph (with a low-frequency filter of 0.1 Hz and a high-frequency filter of 70 Hz) and stored in a computer memory for off-line analyses.

Pharmacologic manipulation of GJ communication


To investigate the possible contribution of GJ communication to the seizure induction in normal brain tissue, in one group of animals, carbenoxolone, a broad-spectrum GJ blocker (21) (10 mM, dissolved in saline), or trimethylamine (TMA), an intracellular alkalinizing agent that opens GJ channels (22) (100 mM dissolved in saline), was used. The application of the drugs started 5 min or 10 min, respectively, before the induction of the primary epileptiform activity (the filter paper containing the crystalline 4-AP was applied on top of the filter paper soaked with the drug). The control values for these experiments were collected from animals of the same litter submitted to an identical experimental paradigm, but with a piece of filter paper soaked with saline.


In another group of animals, the already active Pf was treated locally with carbenoxolone or TMA. In both cases, treatment started 60 min after the appearance of epileptiform activity induced by 4-AP. A piece of filter paper was soaked with a solution of the appropriate drug, applied on top of the filter paper containing crystalline 4-AP, and left in place until the end of the experiment to diffuse into the cortex. The data collected from the same animals before the application of the drugs served as the control values.

During the experiments, the general state of the animal (level of anesthesia and pupil size) was regularly checked, the body temperature was maintained at 37–39°C by a heating lamp, and the exposed cortical surface was kept wet with 39°C saline. Adequate measures were taken to minimize unnecessary pain and discomfort to the animal and to minimize animal use. All experimental procedures were conducted in accordance with the United States Public Health Service's Guidelines for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee at the University of Szeged. At the end of the experiments, the animals were given a lethal dose of sodium pentobarbital.

The effects of the drugs were assessed by measuring the number and duration of seizures, and the summated ictal activity (determined by multiplying the number of individual ictal periods by their durations measured during a 20-min period), and by analyzing the pattern (frequency and amplitude) of the seizure discharges. Measurements in treatment experiments started 10 min (in the case of carbenoxolone) or 40 min (in the case of TMA) after the application of the drug.

Data were stored in a computer memory with the aid of Digidata 1200B (BD, BNC; Axon Instruments, Inc., Sunnyvale, CA, U.S.A.) in parallel with the EEG and analyzed after the experiments. Results are given as means ±SD. Student's t test was used to assess significant differences between the control and experimental groups of data. The level of statistical significance was set at p ≤ 0.05.

Carbenoxolone, trimethylamine, and 4-AP were purchased from Sigma (St. Louis, Mo, U.S.A).

Tissue isolation and RT-PCR

The cortical tissues of the Pf area were isolated after 1 h of repeated ictal discharges (four animals from each developmental time point) followed by a 20-min recovery period, frozen immediately in liquid nitrogen, and stored at −80°C. Identical cortical areas of animals without induced epileptic activity (four animals from the same litter as the 4-AP–treated ones, at each developmental time point) were used as controls.

An RT-PCR–based strategy was used to quantify the expression levels of the Cx 43, 36, 32, and 26 genes and their inducibility by cortical seizure discharges during postnatal development. About 3 μg total RNA, prepared from the Pf and the identical area of the control animals, was used as template for first-strand cDNA synthesis, as described earlier (18). PCR amplification was performed on 3 μl RT product, by using 33 cycles of 95°C for 40 s, 55°C for 40 s, and 72°C for 60 s for the Cxs, and 26 cycles for β-actin. For the normalization of Cx mRNAs, the level of β-actin mRNA was used as an internal standard. The relative levels of Cx mRNAs are expressed as ratios (Cx/β-actin × 100). Images of ethidium bromide–stained agarose gels were digitized with a GDS 7500 Gel Documentation System and analyzed with GelBase/GelBlot Pro Gel Analysis Software (Ultra-Violet Products, Cambridge, United Kingdom).

RT-PCR reactions for each sample were performed in triplicate to increase the reliability of the measurements. Data were analyzed by averaging across animals and expressed as means ± SD. The statistical analysis of the data was done by a two-way analysis of variance (ANOVA; Jandel SigmaStat Statistical Software, Version 2, Biosoft, Cambridge, United Kingdom). When significant differences were found with the overall analysis, the Tukey HSD test was used for post hoc comparisons between groups (significance criterion, p ± 0.05).

These primers were used.












  1. Top of page
  2. Abstract
  7. Acknowledgments

During the development, both the basic ECoG and the Cx mRNA expression patterns exhibited a characteristic maturation; the 4-AP–induced epileptiform activity correlated well with these changes.

Basic electrocortical activity

The structure of the basic ECoG progressively changed with age, with the appearance of new, faster components with higher frequencies (Figs. 1 and 2 first column). The basic ECoG on P9–P12 was rather simple, characterized by slow sinusoid-like waves of amplitude 0.1–0.2 mV and frequency 1–3 Hz. By the age of P28, the amplitudes and frequency configurations were almost identical to those for the adult form.


Figure 2. Representative samples showing the basic electrocorticogram (left column) and the effects of carbenoxolone (right column), recorded from the same animal at different ages (A–E).

Download figure to PowerPoint

4-AP–induced epileptiform activity

On the basis of the pattern of 4-AP–induced ECoG activity, the animals from P9 to P28 could be divided roughly into five age groups (five animals in each group), without a sharp border between them (Figs. 3 and 4, first column).


Figure 3. Graphs illustrate the effects of carbenoxolone (A) and trimethylamine (TMA) (B) on the spike amplitude, spike frequency, seizure numbers and summated ictal activity of cortical seizure activity induced by 4-AP (considered as control, Co) at several developmental time points. The drugs were applied at the already active primary focus 60 min after repeated seizures. Data were collected for 20 min between the 10th and 30th min of the presence of the drug. Each bar represents the mean values ± S.D of five animals per group. Asterisk (*) indicates statistically significant differences in comparison with the control. Significance criterion: p ≤ 0.05.

Download figure to PowerPoint


Figure 4. Representative samples showing the cortical activity induced by 4-aminopyridine (left column) and 10 min after the application of carbenoxolone (right column), recorded from the same animal at different ages (A–E).

Download figure to PowerPoint

P9–12: After 4-AP application, the rats of this age exhibited a sustained, generalized rhythmic activity with relatively slow waves of amplitudes between 0.2 and 0.6 mV and frequencies between 0.8 and 3 Hz. In some cases, either the frequency or the amplitude of these potentials fluctuated (Figs. 3 and 4A). The alternation of ictal and interictal-like periods, characteristic of the epileptic activity of adults could not be recognized at this age. Paradoxically, in some cases, we observed a depressed activity at the area of 4-AP application, probably because of a possible transfer inhibition originating from the contralateral side, through the rostral corpus callosum and/or the commissura anterior.

P13–14: A transient reduction in seizure susceptibility occurred at the age of P13–14 (Figs. 3 and 4B). The 4-AP–induced activity was mostly restricted to the periodic appearance of groups of roughly synchronous ictal-like discharges. These potentials, however, were shorter in duration and often higher in amplitude (0.6–0.8 mV) and frequency (6–10 Hz) than in the younger animals.

P15-16: At this age, profound changes occurred in the manifestation of the 4-AP–induced activity (Fig. 4C). The elevated seizure susceptibility was characterized by the generalized occurrence of synchronized rhythmic seizure discharges in the ECoG (with amplitudes of 0.8–1.2 mV and frequencies of 4–6 Hz), superimposed with randomly repeated and generalized seizures. The amplitudes and the frequencies of the seizure discharges gradually became higher (0.8–1.2 mV and 8–12 Hz) in comparison with the younger animals, although the discharges of highest frequency (9–15 Hz) appearing on the initiation of the seizures in adult rats were missing (Figs. 3 and 4E). Although the number of the seizures basically did not change, the duration of the individual seizures increased significantly (from 5–8 s to 60–90 s) in comparison to the P13–14 animals, resulting in a remarkable increase of the summated ictal activity (Figs. 3 and 4C). In this age group, noticeable differences were found between the activity patterns of the two hemispheres in most animals. On the primary side (the side of 4-AP application), an obvious tendency for the separation of periodic ictal-like and interictal-like activity patterns was noted.

P17–22: In the P17–P22 animals, the seizure susceptibility decreased in parallel with a strong tendency to focalization and periodicity. The appearance of more typical ictal-like periods was followed by interictal phases resembling the basic ECoG activity (Figs. 3 and 4D). Typical synchronous epileptiform activity occurred only at the site of 4-AP application and was not manifested at the identical point on the contralateral side.

P23 and later: The 4-AP–induced epileptiform activity was focal (restricted to the Pf and the Mf) and periodic (15–25 seizures/h) with the highest-frequency (9–15 Hz) discharges at the initiation of the tonic phase and spike–wave complexes of 1–3 Hz (clonic phase) at the end (Figs. 3 and 4E). The ictal periods were 50–80 s long, followed by interictal periods. In most animals, after a few repetitions of the seizures, a secondary Mf developed at the identical point on the side contralateral to 4-AP application, with activity synchronous with that of the Pf, this being characteristic for 90% of the animals after P23. In ∼30% of the animals, the epileptiform activity became secondarily generalized.

Opening or closing of GJ channels by pharmacologic tools

We observed a dynamic fluctuation in the effects of carbenoxolone and TMA on the basic ECoG and the induction and maintenance as well as the propagation of seizures at the different age groups (between P9 and P28), both in the pretreatment and treatment experiments, supposedly depending on the actual composition and size of the GJ pool. The effect of pretreatment with both carbenoxolone and TMA was quite obvious at P9–12, whereas at P13–14, pretreatment was apparently not effective. From the age of P13, pretreatment with carbenoxolone did not noticeably modify either the basic ECoG or the induction of seizures by 4-AP. At P15–16, TMA pretreatment slightly enhanced both the ECoG and the epileptiform activity, whereas in older animals, only the propagation of seizures was facilitated. Conversely, either a proconvulsive or anticonvulsive effect of treatment at the already active epileptic focus with TMA or carbenoxolone, respectively, was much more obvious at any age studied. Treatment with TMA transformed the 4-AP–induced periodic seizure activity to a continuous and generalized rhythm (like status epilepticus). See details later.

Opening of GJ channels by TMA

P9–12: At this age, pretreatment with TMA, even in the absence of the convulsant 4-AP, noticeably enhanced the neuronal synchrony, resulting in the generalized appearance of rhythmic potentials (0.5–0.8 mV and 0.5–0.7 Hz) resembling seizure discharges (Fig. 1A, second column). These discharges were rather uniform in their pattern and often occurred in 25- to 30-s periods followed by a silent phase. The subsequent application of 4-AP did not noticeably modify the TMA-induced activity pattern (Fig. 5A, first column). Similarly, treatment with TMA did not obviously influence the 4-AP–induced activity pattern (Fig. 5A, second column).


Figure 5. Comparison of the effects of pretreatment with the gap-junction opener trimethylamine (TMA) on the manifestation of 4-aminopyridine–induced activity (left column) and the effect of TMA treatment at the already active epileptic focus (right column) at different developmental time points (A–E).

Download figure to PowerPoint

P13–14: In these animals, TMA failed to induce the synchronous, rhythmic activity in the basic ECoG described for the P9–12 animals (Fig. 1B, second column). The 4-AP–induced activity after TMA pretreatment did not differ considerably from that without TMA pretreatment (Fig. 5B, first column). Conversely, treatment with TMA 60–65 min after 4-AP application perceptibly altered the activity pattern (Figs. 3 and 5B, second column). Although the number of the seizures decreased significantly, the summated ictal activity increased considerably because of the long seizures (Figs. 3B and 5C, second column). In some animals, the amplitudes of the discharges at the site of the Mf were higher in comparison with those appearing at other cortical areas (not shown). This could indicate a transient desynchronizing role of GJ communication locally at an already active epileptic focus, and a facilitating effect on the propagation of seizure discharges to the identical point of the contralateral hemisphere.

P15–16: In these animals, pretreatment with TMA slightly enhanced the ECoG activity (Fig. 1C, second column) and increased the amplitudes (from 1.2 ± 0.11 mV to 1.4 ± 0.13 mV) of the seizure discharges induced by 4-AP after TMA application (Fig. 5C, first column). Treatment with TMA at the Pf 60 min after the appearance of the first ictal period induced by 4-AP somewhat increased the manifestation of epileptiform activity. Seizure discharges occurred continuously without the separation of ictal and interictal periods, but the discharges became slower, and their frequency decreased from 8–12 to 4–6 Hz (Figs. 3, 5C, second column).

P17–22: TMA pretreatment did not visibly modify the basic ECoG at this age (Fig. 1D, second column), although it influenced the seizure activity induced after TMA pretreatment (Fig. 5D, first column). High-frequency spikes (12–15 Hz) with small amplitudes (0.3–0.4 mV) were accompanied by discharges of high amplitude (1.0–1.5 mV) but lower frequency (3–8 Hz) (Fig. 5D, first column; compare with Fig. 4D, first column). This activity pattern seemed to be generalized and was interrupted only by short, interictal-like periods. When TMA was applied 60 min after the induction of seizure activity, almost permanent rhythmic spiking developed, with frequencies of 5–8 Hz, and usually higher amplitudes (0.8–1.6 mV) at the Pf than in other cortical areas (0.1–0.2 mV; Figs. 3 and 5D, second column). The increase of the duration of the seizures resulted in significant increase of the summated ictal activity (Fig. 3B).

P23 and later: Pretreatment with TMA in P23 and older animals did not qualitatively influence the basic ECoG, as in adults (19). The seizures induced by 4-AP after TMA pretreatment were generalized in most animals and separated by interictal periods (Fig. 5E, first column). However, when TMA was applied at the already active Pf after a 60-min period of repeated seizures, the amplitudes of the discharges increased significantly, the seizures gradually became longer, resulting in significant increase in the summated ictal activity, and finally no interictal periods could be detected. This kind of epileptiform activity appeared in a rather uniform manner throughout the cortex (like generalized status epilepticus) (Figs. 3 and 5E, second column).

It is worthwhile to mention here that, in adult animals, TMA application after a longer period of repeated seizures (200 min, ∼120 seizures) converted the cortical activity into a synchronous rhythmic pattern, lacking the highest frequencies (Fig. 6), resembling the cortical activity in younger animals.


Figure 6. The cortical activity pattern of an adult animal after a long period (200 min, ∼120 seizures) of repeated seizure activity in the presence of the gap-junction opener trimethylamine (TMA) became similar to that seen in the young animals. A: Control electrocorticogram. B: Representative samples showing the cortical activity induced by 4-aminopyridine (4-AP) applied alone. C, D: Effects of TMA on the 4-aminopyridine (4-AP)-induced cortical activity when applied 40 and 200 min, respectively, after 4-AP.

Download figure to PowerPoint

Closing of GJ channels by carbenoxolone

P9–12: In these young animals, pretreatment with carbenoxolone locally eliminated the rhythmic oscillations in the basic ECoG (Fig. 2A) and suppressed the induction of synchronous activity induced by 4-AP (not shown). Treatment with carbenoxolone strongly depressed the maintenance of epileptiform activity, indicated by a significantly smaller spike amplitudes and reduced summated ictal activity (Figs. 3 and 4A). However, this effect was restricted to the site of application of carbenoxolone.

P13–14: From the age of P13, pretreatment with carbenoxolone did not noticeably modify the basic ECoG (Fig. 2B) nor the induction of the first seizures after pretreatment with carbenoxolone (not shown). However, when the already active epileptic focus was treated with carbenoxolone, it noticeably depressed the manifestation and propagation of the 4-AP–induced synchronous activity (Figs. 3 and 4B).

P15–16: Carbenoxolone pretreatment did not apparently modify either the basic ECoG (Fig. 2C) or the induction of 4-AP epileptiform activity (not shown). Conversely, the seizures became shorter, the amplitudes of the discharges became smaller, and their frequency became less, with the obvious separation of ictal and interictal periods, when carbenoxolone was applied 60 min after the appearance of the first ictal periods (Figs. 3 and 4C). Although the numbers of the seizures slightly increased, the summated ictal activity considerably decreased because of the short seizures. The seizures occurred in a generalized manner in most of the animals of this age, in spite of the treatment with carbenoxolone.

P17 and later: Local pretreatment of the cortical surface with carbenoxolone did not noticeably influence the basic ECoG in the animals older than 18 days (Fig. 2D and E). When the electrical synaptic transmission was depressed relative to the initial baseline before the induction of an epileptic focus, only a mild influence was seen on the induction of seizure discharges (not shown). By contrast, in another series of experiments, when carbenoxolone was applied at the already active epileptic focus, a significant decrease in the intensity of seizure activity and a considerable increase in the duration of the silent “seizure-free” period were detected both at the Pf and at the Mf (Figs. 3 and 4D and E; see also Szente et al., 2002).

Basic expression patterns of Cx mRNAs and their alteration by repeated seizure activity at several developmental time points

The expressions of Cx26, 36, and 43 did not display significant alterations during the first 2 weeks under physiologic conditions. The abundances of Cx26 and 43 mRNA were about the same, whereas the relative level of Cx36 mRNA was somewhat lower (Fig. 7). However, after the first 2 weeks, the expression patterns of these three Cx genes exhibited marked differences. The level of Cx26 mRNA had declined by ∼50% at P16 and remained around this low level till P23. In contrast, the Cx36 expression was significantly upregulated (175%) by P16. At P23, a somewhat lower specific mRNA level was detected (140%), but still significantly higher than that measured at P10–P14. A significantly elevated level of Cx43 expression was detected only by P23, when it reached 230% of the P10–14 value.


Figure 7. Expressions of Cx43, Cx36, Cx32, and CX26 genes in the cortical region of control and 4-aminopyridine–treated animals, at several developmental time points. Cx mRNA levels were normalized to that of β-actin mRNA. Data are expressed as mean ± SD from measurements on four animals at each time point. Each polymerase chain reaction was performed in triplicate to increase the reliability of the measurements. C, Control; Pf, primary focus. *Significant difference between control and Pf values at a given time point; o and #, values significantly different from that of at P10 within the basic and induced data set, respectively.

Download figure to PowerPoint

The Cx32-specific mRNA could not be detected on P10, but it then gradually increased to ∼15% of the β-actin mRNA level by P23, this being the least-expressed Cx gene examined in this study.

Expressions of Cx26, 32, 36, and 43 genes after repeated seizure activity

The seizure activity did not change the expression of Cx26 markedly on P10 and P14, but on P16 and P23 we observed a significant increase (50% and 37%, respectively) in the mRNA level as compared with the basic expression. This inducibility could be a consequence of the decreased basic expression at these ages, because the induced transcript levels are slightly but significantly lower than at P10.

The inducibility of the Cx36 gene after repeated seizures was not detected on P10–14, but the mRNA level was notably higher around P16, when the gene-specific transcript was ∼180% of the control value. By P23, the Cx36 expression was still affected in the epileptic foci; the mRNA level was ∼150% of the control.

We found the highest induced mRNA level for Cx43. This gene was not inducible with the convulsant 4-AP on P10, but its expression was highly upregulated in later stages (P14–23). The specific mRNA level reached 250% of the control level on P14–16 but was somewhat decreased by P23, when the transcript level was 185% of the control value.

Although Cx32 mRNA could not be detected on P10 in the controls, the seizure activity considerably induced its expression level at this age. The induced levels were about the same at the first three time points, differing significantly from the highest induced level measured on P23.


  1. Top of page
  2. Abstract
  7. Acknowledgments

With regard to recent findings that imply a role for GJ channels in the synchronization of neuronal activity, in the present study, we focused on possible roles of GJ channels in the epileptogenicity and seizure susceptibility of the immature mammalian brain. We carried out electrophysiologic experiments, combined with pharmacologic manipulations of the GJ channels, and measured developmental and 4-AP–induced synchronous activity-dependent plastic changes in the Cx26, 32, 36, and C43 mRNA levels in the rat neocortex in different postnatal developmental stages. The epileptic cortical activity induced by 4-AP is accompanied by specific quantitative alterations in the different Cx mRNA levels and suggests a cause–effect relation between the electrical communication pathway and epileptiform synchronization. Our RT-PCR results revealed a highly diverse basic expression pattern for the Cx26, 32, 36, and 43 genes and showed quantitative alterations in their expression during the epileptogenic process in an age-related manner. Each of the Cxs examined displayed a unique basic pattern of postnatal development with a common feature: a marked alteration in their expression levels was found at around P16. Expressions were significantly upregulated at around this age for all the followed genes except Cx26. The Cx26 mRNA level was characterized by a marked decrease after the first 2 weeks.

During the postnatal development, the epileptiform activity induced strong, subtype-specific changes in the levels of expression of all the examined Cx genes: the transcript level of Cx43 was gradually elevated up to P23, the mRNA level of Cx36 showed a transient increase with a marked peak expression around P16, whereas the mRNA levels of Cx32 and 26 demonstrated only very modest age-dependent changes after repeated seizures. The inducibility of the Cx26, 36, and 43 genes was highest at around P16 and had significantly declined by P23. Upregulation of Cx26 and Cx45 mRNA was detected in neuronal cells undergoing apoptotic cell death in vulnerable regions such as hippocampus, amygdala, and some thalamic nuclei, whereas Cx36 was downregulated after the kainate-induced in vivo model of status epilepticus (13). In our 4-AP–induced seizure model, characterized by isolated, repeating seizures, although some kind of neuronal and astroglial damage was detected (23) approximately at the time point of the sample collection for mRNA analysis, apoptotic cell death is not characteristic in the neocortex.

Pharmacologic manipulation of the GJ channels differentially affects the basic ECoG and the induction and maintenance of seizures in different postnatal developmental stages. In the basic ECoG, we found a characteristic inherent sinusoid-like, generalized pattern, restricted until P12. The ability of the cortex to express sustained wave activity around these postnatal days has already been reported (11). The widespread, synchronous rhythmic activity in the basic ECoG was markedly enhanced by 4-AP (Fig. 4) and strongly abolished by carbenoxolone (Figs. 2–4). GJ blockers were shown earlier to abolish horizontal wave propagation and to eliminate epileptiform waves completely in young animals (11). In contrast, the GJ opener TMA alone induced considerable rhythmic, highly synchronous seizure-like activity in the P9–13 animals, which was apparently not modified by 4-AP (Figs. 3 and 4). The failure of 4-AP to modify the TMA-induced cortical activity suggests that the targets of the two drugs may overlap each other and supports GJ involvement in the widespread, synchronous rhythmic activity at this age. 4-AP in P9–12 animals induced a rhythmic activity, characterized by the generalized occurrence of sustained seizure-like, relatively slow potentials, with variable amplitudes and low frequencies. On maturation, the seizure discharges became faster, with increasing amplitude and frequency.

The onset of spontaneous and the 4-AP– and/or TMA-induced generalized synchronous activity could be correlated in part with the extensive cell coupling by the GJ channels (12,24), overwhelming the poorly developed chemical synaptic systems (5) in the first 2 weeks in the rat neocortex. Under physiologic conditions, we measured relatively high levels of Cx26 and 43 mRNA at P10–14, and the presence of Cx 36 mRNA was also detected, although at a lower level (Fig. 7). In time, the level of Cx26 mRNA somewhat declined, whereas the Cx43 expression progressively increased up to P23, and the Cx36 transcript level increased transiently, with peak expression at around P16 (Fig. 7).

The coupling of combinations of pyramidal and nonpyramidal cells and between neurons and glia, mediated by Cx26, 36, and 43, has been reported at this early age in rats (8,10,12). Besides the dominating homotypic coupling, a heterotypic form of coupling is also involved in connecting pyramidal and nonpyramidal neurons, or neurons and astrocytes at P7 and P14 (12). In networks where large numbers of neurons transmit electrical signals directly through GJ channels, the temporal heterogeneity of the discharges decreases, the synchrony thereby being enhanced (25).

The generalized synchronized, rhythmic pattern of both the spontaneous and the epileptiform activity, and their sensitivity to pharmacologic manipulation of the GJ channels on P9–12, may indicate that GJ channels in the immature cortex link neurons and glia cells into extensive networks that may allow electrical activity to spread over long distances. The horizontal propagation of waves has been correlated with the presence of dendrodendritic GJ channels during the first 10–12 postnatal days in rats (11,24,26). Furthermore, GJ channels are also reported to be capable of producing large functional clusters of coupled neurons in vertical columns, serving to synchronize the activity of several cortical layers (24,27).

The complex role of GJ channels at this early age could additionally be related to the relatively low number of chemical synapses (5). Neurons with intrinsic bursting properties that are frequently presumed to be involved in synchronous, rhythmic activity are not present in the sensorimotor cortex of rats in the first 2 weeks (28).

In the first 2 weeks after birth, pyramidal cells have been shown to use predominantly Cx26, whereas the nonpyramidal cells may equally use both Cx26 and 43 for the formation of GJ channels (12). Cx26 has been reported both in the astrocytes and in the neurons of the developing brain and spinal cord (29,30), with the highest expression prenatally and during the first 3 weeks of postnatal life (8,12,24). In addition, more recently, Cx45 gene expression has also been reported in neurons with genetic approach during embryogenesis, with a peak of expression at P1, and declined subsequently during brain development (13,31). However, Cx45 protein is widely expressed in many developing and mature nonneural cell types beside selected subsets of neurons. Conversely, Cx36 may be of particular interest, as is expressed exclusively in neurons and is present both in the young neurons and also in adult cortex (31). Nevertheless, it is possible, that Cx45 might play certain role in some of the epileptiform events of the young animals described in this article, either participating in neuronal–neuronal or in neuronal–glia communication mediated by GJs, which should be the subject of further analysis.

The mRNA level of Cx36 underwent a transient increase, with a marked peak expression at ∼P16 in our study. The level of developmental change and its time course for Cx36 exhibits region specificity (32). In the cerebral cortex, a progressive increase was observed, with peaks between 7 and 16 days (9,32–36). In the early stages of postnatal development, Cx36 is detectable even in neuronal populations that are devoid of Cx36 mRNA in the adult stage (32). Upregulation of Cx36 has been linked to juvenile myoclonic epilepsy (37).

The astrocytic Cx43 is present in the cortex throughout the period of development. (16,29,32,38,39). Astrocytes in adults compose an astrocytic syncytium which gives physical and metabolic support to the neurons (40,41). In the first 2 postnatal weeks, astrocytes are more likely to couple to neurons than to other astrocytes (12) through GJ channels, and they may promote the synchrony of spontaneously active neural networks (30).

Recent findings have lent support to the concept that the Cxs, although redundant, may play a significant role in unstable, transient cell–cell contacts (42). With regard to the relatively high level and unspecific occurrence of the Cxs, it seems justified to presume that the extensive communication through the GJ channels in the first 2 weeks may play a key role in the enhanced seizure susceptibility and mediate not only the induction, but also the propagation of rhythmic synchronous activity. Consequently, the decline of this feature later could be explained by the decreased frequency of GJ coupling between the neurons as the cells mature and in parallel by the formation of more specific neuronal networks involving excitatory and inhibitory chemical synapses (5,12,24).

The spontaneous rhythmic pattern that was described at P9–12 was not manifested in the basic ECoG (Figs. 1 and 2), and it was not induced by TMA pretreatment in the animals after (Fig. 1). In addition, we observed a transient reduction in the epileptogenicity of the animals during a brief developmental window at around P13–14 (Fig. 4). The sudden disappearance of the rhythmic features of the basic electrocortical activity and the inability of carbenoxolone and TMA to influence the basic cortical activity (Figs. 3–5) suggest a weaker GJ involvement in this developmental stage. We detected a decreased expression of Cx26 by P15–23 (Fig. 7), whereas the level of expression of Cx43 gradually increased. The expressions of these Cxs probably became restricted to specific cell types after P16, contributing to more-specific neuronal networks. These observations confirm the findings of earlier experiments, in which the glutamate-independent coordinated activity decreased and a neurobiotin tracer indicated a low incidence of neuronal coupling at around P15 (11,12). Although the basal level of Cx43 mRNA was increased in the P17 and P23 animals, it was markedly elevated by the epileptiform activity (Fig. 7). This increment of Cx43 mRNA could be an indication of a more-specific astrocytic activity and adjustability to the elevated neuronal firing activity. Although the oligodendrocyte-specific Cx32 mRNA (43) was not detected at P10, it was dramatically induced by the convulsant 4-AP at this age (Fig. 7, P10). At later ages (P14–23), the basic level became detectable and gradually increased with age. As a consequence, the relative inducibility of this gene progressively decreased (Fig. 7).

The lack of manifestation of the fastest seizure discharges in animals before P13–14 could be related to the relative scarcity of chemical synaptic connections and the slow nature of synaptic potentials at this age (5). Accordingly, the increase in the frequency and the appearance of a faster discharge in older animals could be explained in terms of the observation that the number of excitatory synapses significantly increased during the second postnatal week, and these synapses are able to follow activation faithfully at high frequencies (5). In addition, the actual number and ratio of the open versus closed GJ channels can modulate the frequency of the discharges (20). The application of the GJ opener TMA at already active epileptic focus at most developmental points enhanced generalized synchronization and increased the duration of ictal events, resulting in the cortical activity pattern that is characteristic of young animals with abundant GJ communication (Figs. 3–5). These observations confirm our earlier findings in adult animals showing that epileptiform activity can upregulate the expression of Cxs 26, 32, 36, and 43 mRNAs, indicated also by the strong efficacy of pharmacologic manipulation of GJ communication by carbenoxolone and TMA (18–20). However, the inducibility of the Cx mRNAs examined here revealed some subtype specificity at the different developmental time points.

One possible explanation of the transiently elevated epileptogenicity of young animals at around P16–17 (Fig. 4) could be the appearance of the first intrinsic bursting neurons during the third postnatal week (5,44) and/or the transient hyperactivity of the N-methyl-d-aspartate–mediated neurotransmission and/or the immature stage of γ-aminobutyric acid (GABA)ergic inhibition (7). Because the percentage of neurons with intrinsic bursting capacity is high in early postnatal life (45,46), these cells can act as cellular pacemakers coupled by GJ channels (7).

In addition, although the basic level of neuron-specific Cx26 mRNA gradually decreased with age, the other neuron-specific Cx36 expression was significantly increased and approached the highest level at around P16 (Fig. 7). The elevated level of GJ communication in scattered subpopulations of cells express Cx36 mRNA can substantially contribute to the elevated seizure activity observed at this age. A recent in vitro study based on abrupt maturation at the end of the second postnatal week of synchronous activity among electrically coupled, low-threshold spiking inhibitory interneurons revealed that such activity was absent on earlier postnatal days (14). The strong synchronizing ability of this inhibitory cell network may contribute to the increased seizure susceptibility of rats after the second postnatal week.

All measured Cx mRNA levels exhibited an obvious increase after 60-min periods of seizure activity, indicating that, besides the increased efficacy of excitatory chemical neurotransmission, electrical coupling through the GJ channels may also contribute to the transiently elevated level of epileptogenicity of P15–16 animals. After this age, the progressive decline in seizure susceptibility could be an indication of the fine-tuning of local synaptic connectivity and the pruning back of some excess of the chemical and electrical synapses, in parallel with the development of the fully active inhibitory GABAA prune-back receptor system (5,7). The incidence of coupling between excitatory and inhibitory cells declines with age (10,47,48), and in the adult, most electrical coupling exists between homogeneous cell types, a condition that may diminish the susceptibility of the neural networks to synchronization (27).


  1. Top of page
  2. Abstract
  7. Acknowledgments

The findings of the present study suggest that GJ communication is not only involved in rhythm genesis and the synchronization of cortical activity, but also may be responsible for the elevated epileptogenicity of the developing brain. In addition, repeated seizures can induce changes in expression of the different Cx genes and the remodeling of GJ communication. If these plastic modifications after early epileptiform disorders are persistent, they may facilitate epileptogenesis and ictogenesis at a later developmental point without any discrete morphologic alterations.

Further studies are needed to analyze in detail the balance between excitatory and inhibitory functions and the establishment of chemical and electrical synaptic connections in the network and cellular levels.


  1. Top of page
  2. Abstract
  7. Acknowledgments

Acknowledgment:  This study was supported by OTKA grant T037505. We thank Peter Gahan for a critical reading of the manuscript


  1. Top of page
  2. Abstract
  7. Acknowledgments
  • 1
    Moshé SL, Albala BJ, Ackermann RF, et al. Increased seizure susceptibility of the immature brain. Dev Brain Res 1983;7:815.
  • 2
    Moshe SL. Epileptogenesis and the immature brain. Epilepsia 1987;281:315.
  • 3
    Holmes GL, Ben-Ari Y. Seizures of the developing brain: perhaps not so benign after all. Neuron 1998;21:12314.
  • 4
    Johnston MV. Developmental aspects of epileptogenesis. Epilepsia 1996;37:S29.
  • 5
    Swann JW, Hablitz JJ. Cellular abnormalities and synaptic plasticity in seizure disorders of the immature nervous system. Ment Retard Dev Disabil Res Rev 2000;64:25867.
  • 6
    Avoli M. Epileptiform discharges and a synchronous GABAergic potential induced by 4-aminopyridine in the rat immature hippocampus. Neurosci Lett 1990;11:938.
  • 7
    Ben-Ari Y, Khalilov I, Represa A, et al. Interneurons set the tune of developing networks. Neuroscience 2004;27:4227.
  • 8
    Nadarajah B, Jones AM, Evans WH, et al. Differential expression of connexins during neocortical development and neuronal circuit formation. J Neurosci 1997;17:20963111.
  • 9
    Prime G, Horn G, Sutor B. Time-related changes in connexin mRNA abundance in the rat neocortex during postnatal development. Dev Brain Res 2000;119:11125.
  • 10
    Venance L, Rozov A, Blatow M, et al. Connexin expression in electrically coupled postnatal rat brain neurons. Proc Natl Acad Sci U S A 2000;97:102605.
  • 11
    Peinado A. Immature neocortical neurons exist as extensive syncytia networks linked by dendrodendritic electrical connections. J Neurophysiol 2001;85:6209.
  • 12
    Bittman K, Becker DL, Cicirata F, et al. Connexin expression in homotypic and heterotypic cell coupling in the developing cerebral cortex. J Neurol 2002;443:20112.
  • 13
    Condorelli DF, Trovato-Salinaro A, Mudo G, et al. Cellular expression of connexins in the rat brain: neuronal localization, effects of kainate-induced seizures and expression in apoptotic neuronal cells. J Neurosci 2003;18:180727.
  • 14
    Long MA, Cruikshank SJ, Jutras MJ, et al. Abrupt maturation of a spike-synchronizing mechanism in neocortex. J Neurosci 2005;25:730916.
  • 15
    Nedergaard M. Direct signaling from astrocytes to neurons in cultures of mammalian brain cells. Science 1994;263:176871.
  • 16
    Nagy JI, Rash JE. Connexins and GJs of astrocytes and oligodendrocytes in the CNS. Brain Res Rev 2000;32:2944.
  • 17
    Wong M, Yamada KA. Developmental characteristics of epileptiform activity in immature rat neocortex: a comparison of four in vitro seizure models. Dev Brain Res 2001;128:11320.
  • 18
    Szente M, Gajda Z, Said Ali K, et al. Involvement of electrical coupling in the in vivo ictal epileptiform activity induced by 4-aminopyridine in the neocortex. Neuroscience 2002;115:106778.
  • 19
    Gajda Z, Gyengési E, Hermesz E, et al. Involvement of gap junctions in the manifestation and control of the duration of seizures in rats in vivo. Epilepsia 2003;44:1596600.
  • 20
    Gajda Z, Szupera Z, Blazsó G, et al. Quinine, a blocker of neuronal Cx36 channels, suppresses seizure activity in rat neocortex in vivo. Epilepsia 2005;46:158191.
  • 21
    Gareri P, Condorelli D, Belluardo N, et al. Anticonvulsant effects of carbenoxolone in genetically epilepsy prone rats (GEPRs). Neuropharmacology 2004;47:120516.
  • 22
    Köhling R, Gladwell SJ, Bracci E, et al. Prolonged epileptiform bursting induced by 0-Mg2+ in rat hippocampal slices depends on gap junctional coupling. Neuroscience 2001;105(3):57987.
  • 23
    Mihály A, Toth G, Szente M, et al. Neocortical cytopathology in focal aminopyridine seizures as related to the intracortical diffusion of 3H 4-aminopyridine. Acta Neuropathol Berl 1985;66:14554.
  • 24
    Peinado A, Yuste R, Katz LC. GJal communication and the development of local circuits in neocortex. Cereb Cortex 1993;3:48898.
  • 25
    White JA, Chow CC, Ritt J, et al. Synchronization and oscillatory dynamics in heterogeneous, mutually inhibited neurons. J Neurosci 1998;5:516.
  • 26
    Rorig B, Klausa G, Sutor B. Intracellular acidification reduced GJ coupling between immature rat neocortical pyramidal neurones. J Physiol 1996;490:3149.
  • 27
    Connors BW, Long MA. Electrical synapses in the mammalian brain. Ann Rev Neurosci 2004;27:393418.
  • 28
    Hoffman SN, Prince DA. Epileptogenesis in immature neocortical slices induced by 4-aminopyridine. Dev Brain Res 1995;85:6470.
  • 29
    Nagy JI, Li XB, Rempel J, et al. Connexin26 in adult rodent central nervous system: demonstration at astrocytic GJs and colocalization with connexin30 and connexin43. J Neurol 2001;441:30223.
  • 30
    Alvarez-Maubecin V, Garcia-Hernandez F, Williams JT, et al. Functional coupling between neurons and glia. Neuroscience 2000;20:40918.
  • 31
    Sohl G, Maxeiner S, Willecke K. Expression and functions of neuronal gap junctions. Nat Rev Neurosci 2005;6:191200.
  • 32
    Belluardo N, Mudo G, Trovato-Salinaro A, et al. Expression of connexin36 in the adult and developing rat brain. Brain Res 2000;865:12138.
  • 33
    Condorelli DF, Belluardo N, Trovato-Salinaro A, et al. Expression of Cx36 in mammalian neurons. Brain Res Rev 2000;32:7285.
  • 34
    Rash JE, Yasumura T, Dudek FE, et al. Cell-specific expression of connexins and evidence of restricted GJal coupling between glial cells and between neurons. J Neurosci 2001;216:19832000.
  • 35
    Srinivas M, Rozental R, Kojima T, et al. Functional properties of channels formed by the neuronal GJ protein connexin36. J Neurosci 1999;1922:984855.
  • 36
    Hormuzdi SG, Filippov MA, Mitropoulou G, et al. Electrical synapses: a dynamic signaling system that shapes the activity of neuronal networks. Biochim Biophys Acta Biomembranes 2004;1662:11337.
  • 37
    Mas C, Taske N, Deutsch S, et al. Association of the connexin36 gene with juvenile myoclonic epilepsy. J Med Genet 2004;41:e93.
  • 38
    Dermietzel R, Traub O, Hwang TK, et al. Differential expression of three GJ proteins in developing and mature brain tissues. Proc Natl Acad Sci U S A 1989;86:1014852.
  • 39
    Giaume C, Venance G. GJs in brain glial cells and development. Perspect Dev Neurobiol 1995;2:33545.
  • 40
    Walz W, Hertz L. Functional interactions between neurons and astrocytes: II. Potassium homeostasis at the cellular level. Prog Neurobiol 1983;20:13383.
  • 41
    Jefferys JGR. Nonsynaptic modulation of neuronal activity in the brain: electric currents and extracellular ions. Physiol Rev 1995;75:689723.
  • 42
    Segretain D, Falk MA. Regulation of connexin biosynthesis, assembly, GJ formation, and removal. Biochim Biophys Acta Biomembranes 2004;1662:321.
  • 43
    Belliveau DJ, Naus CC. Differential localization of GJ mRNAs in developing rat brain. Dev Neurosci 1995;17:8196.
  • 44
    Franceschetti S, Buzio S, Sancini G, et al. Expression of intrinsic bursting properties in neurons of maturing sensorimotor cortex. Neurosci Lett 1993;162:258.
  • 45
    Wellmer J, Su H, Beck H, et al. Long-lasting modification of intrinsic discharge properties in subicular neurons following status epilepticus. J Neurosci 2002;16:259266.
  • 46
    Yaari Y. Plasticity of intrinsic neuronal excitability after status epilepticus: time course and ionic mechanisms. Epilepsia 2005;46:206.
  • 47
    Meyer AH, Katona I, Blatow A, et al. In vivo labeling of parvalbumin-positive interneurons and analysis of electrical coupling in identified neurons. J Neurosci 2002;22:705564.
  • 48
    Naus CG, Bani-Yaghoub M. Gap junctional communication in the developing central nervous system. Cell Biol Int 1998;22:75163.