|Home | About | Journals | Submit | Contact Us | Français|
Interneurons, GABAA receptor density and subunit composition determine inhibitory function in pyramidal neurons and control excitability in cortex. Abnormalities in GABAergic cells or GABAA receptors could contribute to seizures in malformations of cortical development. Here we review data obtained in resected cortex from pediatric epilepsy surgery patients with Type I and Type II cortical dysplasia (CD) and non-CD pathologies. Our studies found fewer interneurons immunolabeled for glutamic acid decarboxylase (GAD) in Type II CD while there were no changes in tissue from Type I CD. GAD-labeled neurons had larger somata, and GABA transporter (VGAT and GAT1) staining showed a dense plexus surrounding cytomegalic neurons in Type II CD. Functionally, neurons from Type I CD tissue showed GABA currents with increased EC50 compared to cells from the other groups. In Type II CD, cytomegalic pyramidal neurons showed alterations in GABA currents, decreased sensitivity to zolpidem and zinc and increased sensitivity to bretazenil. In addition, pyramidal neurons from Type II CD displayed higher frequency of spontaneous inhibitory post synaptic currents. The GABAergic system is therefore altered differently in cortex from Type I and Type II CD patients. Alterations in zolpidem, zinc and bretazenil sensitivity and spontaneous IPSCs suggest that Type II CD neurons have altered GABAA receptor subunit composition and receive dense GABA inputs. These findings support the hypothesis that patients with Type I and Type II CD will respond differently to GABA receptor-mediated antiepileptic drugs and that cytomegalic neurons have features similar to immature neurons.
Cortical dysplasia (CD) is a malformation of cortical development responsible for many cases of intractable seizures in children (Chugani et al., 1990; Lerner et al., 2009). CD has been classified into sub-types depending on the severity of abnormalities (Taylor et al., 1971; Mischel et al., 1995). Palmini Type I CD is defined as cortical dyslamination and cellular misorientation usually associated with excessive numbers of subcortical white matter neurons. In Type II CD, the cortex often contains abnormal cells such as dysmorphic, immature, cytomegalic neurons, and balloon cells (Palmini et al., 2004). These histopathological features suggest that the etiology and pathogenesis of Type I and Type II CD may be different. Studies in tissue resected from adult patients have reported decreases in interneurons more pronounced in Type II than in Type I CD (Spreafico et al., 2000; Zamecnik et al., 2006). Inhibitory postsynaptic currents are also decreased in adult dysplastic cortex (Calcagnotto et al., 2005). In contrast, in pediatric CD patients, our group found that GABA synaptic activity was increased in areas of severe CD (Cepeda et al., 2005). The purpose of the present study was to determine whether interneurons were affected in pediatric CD, and if alterations in interneurons correlated with alterations in GABA currents. In this report we review evidence supporting the idea that GABA function is differentially affected in Type I and Type II CD. Some of the data described in this report are a synopsis of work previously published (Andre et al., 2008).
Data were collected form neocortical samples, resected from children with intractable seizures and classified into sub-groups: non-CD, Type I and Type II CD pathologies. Tissue from Type II CD was further classified into samples with or without cytomegalic neurons. Pre-surgery, all patients received antiepileptic drugs (AEDs). The number of patients taking any of the 18 different AEDs did not differ between non-CD, Type I and Type II CD patients. Thus, differences between cell types from CD tissue are not likely due to pre-surgery AEDs (Andre et al., 2008).
Double immunofluorescence stainings were performed with antibodies against NeuN, GAD and the GABA vesicular transporter, VGAT. Cell counts for NeuN and GAD, and soma measurements of GAD-positive neurons were performed and showed that the percent of neurons co-expressing NeuN and GAD was decreased in Type II CD tissue with cytomegalic neurons (Fig. 1A). There was also a decrease in the percent of GAD-labeled neurons in Type II tissue without cytomegalic neurons in layer III, although the decrease was less than in the tissue with cytomegalic neurons. In Type II CD tissue containing cytomegalic pyramidal neurons, abnormally large interneurons were found in 50% of the patients and GAD-positive neurons had significantly larger cell somatic area (Fig. 1A, 1C). In Type II CD tissue, GAD and VGAT immunoreactivities were mostly seen in terminals densely surrounding cytomegalic pyramidal neurons (Fig. 1B).
Solutions used for electrophysiology on acutely dissociated neurons and slices were described previously (Cepeda et al., 2005; Andre et al., 2008). Dissociated neurons were obtained after slices were enzymatically treated for 25–30 min with papain and mechanically dissociated. Drugs were applied through capillary tubes using a pressure-driven perfusion system synchronized by pClamp. An interval of 60 sec between each application was adequate to avoid GABA current run-down. Spontaneous inhibitory post synaptic currents (IPSCs) were recorded in slices with a cesium methanesulfonate based internal solution while holding the membrane potential at +10 mV.
Four groups of cells were examined: normal-sized pyramidal neurons from non-CD, Type I and Type II CD tissue, and cytomegalic pyramidal neurons from Type II CD tissue. Cytomegalic pyramidal neurons had significantly larger membrane capacitance and lower input resistance than normal-sized pyramidal neurons (Andre et al., 2008). Postsynaptic GABA peak current amplitudes were larger in cytomegalic neurons compared to non-CD pyramidal neurons (Fig. 2A). In contrast, peak current densities (obtained after normalizing by cell capacitance) were smaller in cytomegalic neurons than in cells from non-CD and Type I (Fig. 2B). EC50 values were increased in Type I CD compared to non-CD and Type II CD neurons (Fig. 2C) while here were no differences in EC50 between non-CD pyramidal neurons, normal-sized and cytomegalic pyramidal neurons from Type II CD tissue. GABA current kinetics were different by cell type in CD tissue. Half-amplitude durations for GABA currents were significantly longer in cytomegalic pyramidal neurons compared to all other groups of pyramidal neurons (data not shown, see Andre et al., 2008). To determine if age at surgery or type of antiepileptic medication affected responses to GABA, we compared GABA currents among patient groups. There were no correlations between half-amplitude durations, EC50 values and age at surgery in any pyramidal cell group (Pearson correlation P>0.2). GABA currents were compared per patient groups and per medication. There was no significant effect of either BZD alone, barbiturates alone, or the combination of BZD/barbiturates on GABA peak currents, densities, EC50, or half-amplitude durations (ANOVA P>0.05).
Zolpidem is a type I benzodiazepine that binds more specifically to GABAA receptors containing α1 subunits. In acutely dissociated neurons, potentiation of GABA (10 µM) peak currents by zolpidem (100 nM) was significantly larger in non-CD neurons compared to normal-sized and cytomegalic pyramidal neurons from Type II CD (Fig. 2D). Modulation in Type I neurons was similar to non-CD neurons. Bretazenil is a partial agonist that enhances GABA currents in cells expressing α3, α4 and α5. Bretazenil (100 nM) increased GABA currents in non-CD and in Type II CD, but the modulation was significantly larger in Type II CD (Fig. 2E). We also testd the effects of zinc on GABA currents in CD because it blocks GABAA receptors devoid of γ2 subunits. Zinc (100 µM) block was significantly smaller in Type II CD cytomegalic neurons compared to normal-sized neurons from non-CD, Type I and Type II CD (Fig. 2F).
In order to examine synaptic GABA currents, we recorded IPSCs in slices. Cells from non-CD and Type I CD showed the same IPSC frequency and amplitude while cells from Type II CD showed increased IPSC frequency and amplitude compared to non-CD and Type I CD (Fig. 3A, B, C, D). IPSC frequency was the highest in cytomegalic neurons (Fig. 3A, C). Decay time and area under the curve were significantly higher in cytomegalic neurons compared to all other groups, indicating larger and slower synaptic GABA currents in cytomegalic neurons (Fig. 3E, F, G).
Our results indicate that GABAA receptors, GAD-positive cells and GABA terminals are differentially affected in Type I and Type II CD. In Type II CD, there were fewer GAD cells in tissue containing cytomegalic neurons, as reported by others in dysplastic regions (Spreafico et al., 2000; Thom et al., 2003; Deukmedjian et al., 2004). This suggests that tissue from Type II CD with cytomegalic cells, corresponding to regions with fewer interneurons could more easily induce or propagate seizures, secondary to decreased GABA neurotransmission. However, we found that regions with decreased numbers of GAD-positive cells had increased GAD cell soma size, consistent with findings from our laboratory showing cytomegalic interneurons (Andre et al., 2007). In addition, we and others have described a dense plexus of GABAergic fibers surrounding pyramidal cell somata in Type II CD (Spreafico et al., 1998; Alonso-Nanclares et al., 2005). These findings support the notion that, while there are fewer GABAergic cells in Type II CD tissue, the remaining GABAergic cells have the ability to support GABA function. However, the subsequent GABAergic network is likely to be abnormal, which could lead to weaker inhibition, synchronize populations of neurons in damaged areas and possibly be epileptogenic.
Although current densities were reduced in cytomegalic neurons, the overall GABA currents might still be increased, because of their large size, especially if they receive significant GABA innervation. The higher frequency of spontaneous IPSCs in both normal and cytomegalic neurons confirms that Type II CD neurons are connected by dense GABA inputs. In addition, slower GABA current kinetics suggest the channels stay open longer, which provides evidence for a stronger effect of GABA. In this case, and even if cytomegalic pyramidal neurons show hyperexcitable properties, an increase in the total number of GABAA receptors on the soma and slowly desensitizing GABA currents could produce a shunting of excitatory inputs if GABA is inhibitory. Changes in the effect of zolpidem and bretazenil support the concept that GABAA receptor subunit composition is probably altered in Type II CD, with responses consistent with decreases in the proportion of α1 relative to α3 α4 and α5. The presence of less α1- and more α3 α4 and α5-containing receptors corresponds to GABAA receptors inducing slow GABA currents described during embryonic development (Araki et al., 1992; Owens et al., 1999; Bosman et al., 2002). This would fit with the hypothesis that in Type II CD, some neurons show characteristics of immature brains. We also showed that cytomegalic neurons displayed less sensitivity to zinc, indicating more γ2 subunits. The presence of more γ2 subunits along with other subunits could render GABAA receptors more sensitive to some benzodiazepines in severe CD.
Our morphological and electrophysiological data show that GABA cells and GABA function are altered in Type II CD, with evidence showing increased GABA function. Ample GABA release, high GABA concentrations or prolonged GABA application could have a depolarizing effects on Type II CD neurons, especially if they show immature properties and altered intracellular chloride concentrations. In this case, AEDs increasing GABA function might not be beneficial in Type II CD and other options should be considered. In contrast, in Type I CD, we did not find any loss of GAD-labeled cells, any GABA terminal reorganization or any change in IPSC frequency but functionally, there was decreased GABA sensitivity compared to non-CD, which further suggests deficient inhibition in Type I CD. This indicates that AEDs increasing GABA function might be more efficient to control seizures in Type I CD patients.
This work was supported by NIH grant NS 38992.
Disclosure: None of the authors has any conflict of interest to disclose.
We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.