• Neocortex;
  • Dysplasia;
  • Epilepsy;
  • γ-Aminobutyric acid


  1. Top of page
  2. Abstract

Summary:  Purpose: Cortical dysplasia (CD) is associated with epilepsy in both the pediatric and adult populations. The mechanism underlying seizures with cortical malformations is still poorly understood. To study the physiology of dysplastic cortex, we developed an experimental model of CD.

Methods: Pregnant rats were given intraperitoneal injections of carmustine (1-3-bis-chloroethyl-nitrosourea; BCNU) on embryonic day 15 (E15). Cortical histology was examined in the resulting pups at P0, P28, and P60. In addition, evoked and spontaneous field potential recordings were obtained in cortical slices from adult control and BCNU-exposed rats. Finally, we used whole-cell recordings to compare physiologic properties of pyramidal neurons and γ-aminobutyric acid (GABA) responses in control and BCNU-treated animals.

Results: Features characteristic of CD were found in the offspring, including laminar disorganization, cytomegalic neurons, and neuronal heterotopias. Dysplastic cortex also contained abnormal clusters of Cajal–Retzius (CR) cells and disruption of radial glial fibers, as demonstrated with immunohistochemistry. Under conditions of partial GABAA-receptor blockade with 10 μM bicuculline methiodide (BMI), slices of dysplastic cortex demonstrated a significant increase in the number of spontaneous and evoked epileptiform discharges. Individual pyramidal neurons in dysplastic cortex were less sensitive to application of GABA compared with controls.

Conclusions: BCNU exposure in utero produces histologic alterations suggestive of CD in rat offspring. Dysplastic cortex from this model demonstrates features of hyperexcitability and decreased neuronal sensitivity to GABA. Such physiologic alterations may underlie the increased epileptogenicity of dysplastic cortex.

Cortical dysplasia (CD) is a disruption in cortical architecture that is associated with epilepsy in both children and adults (1).

Electroencephalography frequently demonstrates foci of seizure activity within or near regions of CD (2). The cellular mechanisms underlying epileptogenesis in the setting of CD are still poorly understood (3,4). To investigate these mechanisms, several experimental models of CD have been developed. For example, CD can be produced by freeze lesioning (5), in utero exposure to γ-irradiation (6), or in utero exposure to methylazoxymethanol (MAM), which produces multiple cortical and subcortical abnormalities depending on the embryonic day of exposure (7). Various cellular and electrophysiologic abnormalities have been noted in these models, including “excessively bursting” neurons (8), epileptiform synaptic responses (9,10), and decreased numbers of inhibitory neurons (11,12).

We wished to investigate further the physiologic characteristics of neurons in the setting of CD. For this study, we developed a simple and reliable method for generating dysplasia in mammalian cortex by using carmustine (BCNU) delivered in utero. This DNA-alkylating agent has high lipophilicity and high central nervous system (CNS) penetration, suggesting that it would be effective when delivered to pregnant dams (13). In this article, we first show that BCNU exposure in utero produces CD in the offspring. Many morphologic features of the CD found in this model are shared with human pathological specimens including disruption of lamination, cytomegalic neurons, and heterotopias.

Next, we wished to confirm the physiologic correlates of CD. With field-recording techniques, under conditions of partial γ-aminobutyric acid (GABA) blockade, we demonstrate that, in our model, dysplastic cortex exhibits a higher number of spontaneous epileptiform events and increased excitability after white matter (WM) stimulation compared with control cortex. Roper et al. (12) previously demonstrated hyperexcitability for irradiated cortex under similar conditions.

Finally, we investigated the physiologic properties of pyramidal cells in the setting of CD by using whole-cell recording. We found that despite minimal differences in membrane properties, the pyramidal cells in dysplastic cortex demonstrate a decreased sensitivity to GABA, manifested by a reduction in the median effective concentration (EC50), suggesting that a decrease in the postsynaptic efficacy of GABA may partly underlie the increased excitability in CD.


  1. Top of page
  2. Abstract

Animal preparation

All procedures involving animals were performed according to protocols approved by the Institutional Animal Care and Use Committee of Columbia University. Pregnant Sprague–Dawley rats (240–340 g) were injected intraperitoneally (i.p.) with 20 mg/kg of BCNU on embryonic day 15 (E15). E15 was chosen because previous work with MAM in rats showed that CD and neuronal migration disorders were most common if that agent was delivered on E15 (7,14). Pilot experiments with BCNU resulted in similar findings (data not shown). BCNU was dissolved in sterile 5% dextrose in water (D5W) at 4 mg/ml (15). Control animals were injected with a corresponding volume of D5W at E15. Care was taken to avoid injecting the drug either intravascularly or into the uterus by aspirating before injecting and noting absence of blood or amniotic fluid. Pregnant animals were housed with 12-h light cycles and were allowed to take food and water ad lib. The day of birth was noted as P0, and pups were allowed to remain with their mother until P21, at which time they were weaned.


For the purposes of histologic examination, control and experimental animals (ages P0–P64) were anesthetized with ketamine (100 mg/kg) and xylazine (5 mg/kg) given i.p.. Sections were prepared from three control animals and three BCNU-exposed animals groups at the ages P0, P28, and P60. The animal was first transcardially perfused with phosphate-buffered saline (PBS; pH 7.4) and then with 4% paraformaldehyde. After removal from the skull, the brain was further fixed in 4% paraformaldehyde for 24 h, and then equilibrated in 30% sucrose, 4% paraformaldehyde for 72 h. Brains were then frozen at –30°C and stored at –80°C until cryostat sectioning. Cryostat sections were obtained at 20-μm thickness. Sections were stained with cresyl violet (0.5%) after mounting on slides with standard techniques (16). Sections were viewed and images were captured on a microscope (Axioskop; Zeiss, Oberkochen, Germany) fitted with a digital camera (Axiocam; Zeiss).


Cryostat sections, prepared as described earlier, were stained by using the standard peroxidase–antiperoxidase technique (Sternberger Monoclonal, Inc., Lutherville, MD, U.S.A.). The following antibodies were used at these dilutions: anti-parvalbumin (Swant, Inc., Bellinzona, Switzerland), 1:5,000; anti-calbindin (Swant, Inc.), 1:5,000; anti-calretinin (Sigma, St. Louis, MO, U.S.A.), 1:1,000; RT-401 (BD Pharmingen, San Diego, CA, U.S.A.), 1:5,000; and CR-50 (gift of Dr. Ogawa), 1:500. Cryostat sections were washed 3 times with a high-ionic-strength buffer (0.05 M Tris-HCl, pH 7.6, 1.5% NaCl) and treated with ice-cold methanol with 0.3% hydrogen peroxide for 20 min. Sections were then washed 3 times and blocked by using the same buffer with 10% normal goat serum (NGS). The primary antibody was applied in the same buffer with 1% NGS for 24–48 h at 4°C. Sections were then washed at room temperature (RT) with buffer 3 times. The secondary antibody (goat anti-mouse) was applied at 1:200 for 1 h at RT. The slices were again washed, and mouse peroxidase-antiperoxidase complex (Sternberger Monoclonal, Inc.) was applied (1:200) for 1 h at RT. After washing, the diaminobenzidine (DAB) substrate (Vector Laboratories, Burlingame, CA, U.S.A.) was applied for 8 min with agitation. Slices were again washed and mounted on gel-coated slides. Slides were air-dried, dehydrated with graded alcohols, and coverslipped with DPX.

Field recording

Coronal brain slices were obtained from adult (>P50) experimental (mean age, P65.5) and control rats (mean age, P67) for purposes of field recording. The details of slice preparation are given in Blanton et al. (17). In brief, experimental and control animals were anesthetized with ketamine and xylazine as described earlier. Rapid decapitation was followed by careful dissection of the brain from the skull. Coronal sections (400-μm thickness) were cut in ice-cold artificial CSF [aCSF; 125 mM NaCl; 2.5 mM KCl; 1 mM MgCl2; 2 mM CaCl2; 1.25 mM NaH2PO4; 25 mM NaHCO3, 25 mM glucose (310 mOsm, pH 7.4, when bubbled with 95% O2, 5% CO2)] or saCSF (220 mM sucrose, 2.5 mM KCl; 1 mM MgCl2; 2 mM CaCl2; 1.25 mM NaH2PO4; 25 mM NaHCO3, 25 mM glcuose). In later experiments, saCSF was used because of reports of increased neuron viability under these conditions (18). Before sectioning, the brain was allowed to recover in ice-cold oxygenated (95% O2, 5% CO2) aCSF or saCSF for 3 min. Brain slices (400-μm thickness) were made with a vibratome (Leica, Solms, Germany). Sections from the region of the anterior commissure were chosen for recording. Slices were allowed to recover in aCSF bubbled with 95% O2, 5% CO2 for ≥60 min at RT. Slices were then placed on a liquid/vapor interface chamber (Harvard Apparatus, Holliston, MA, U.S.A.) for recording. Bubbled aCSF was allowed to flow at a rate of 1–2 ml/min. The temperature was regulated at 34 ± 1°C. A bipolar electrode with 1-mm spacing (FHC, Bowdoinham, ME, U.S.A.) was placed at the cortical layer VI–WM interface for stimulation. Saline-filled (1 M NaCl) microelectrodes (impedance, 1–2 MΩ) were placed in layers II/III under microscopic guidance, and the response was recorded with a differential amplifier and filter (A-M Systems, Carlsborg, WA, U.S.A.). The amplitude of the electrical stimulus was adjusted to give the threshold of the maximal synaptic response. Typical electrical stimuli were in the range of 1–2 V, 0.1 ms. The recording electrode position was adjusted to a maximal recording of the response. Both control and dysplastic slices were subjected to the same experimental regimen, similar to that described by Roper (19). Initially, 20 responses to electrical stimulation were recorded and stored digitally for later analysis (pClamp 8.0; Axon, Union City, CA, U.S.A.) at a sampling rate of 10 kHz. Each stored episode lasted 4 s. Stimuli were given every 30 s.

Next, the perfusion solution was changed to aCSF with 10 μM BMI. Slices were perfused with this solution without stimulation for 15 min. After this, the same electrical stimulation was applied for 20 additional episodes. Under conditions of GABAA-receptor blockade with BMI, the normal response of the cortex to electrical stimulation consists of one or more negative field potentials (NFPs) (19–21). The number of NFPs per episode was quantified and compared for control and dysplastic cortex.

Next, spontaneous activity from cortical slices was recorded in 10 μM BMI-containing aCSF for 20 min. The number of NFPs was quantified. Each NFP was part of a spontaneous epileptiform event (EE) made up of one or more NFPs (see Fig. 4C). Therefore, the number of NFPs/EE was compared for control versus dysplastic cortex following the method of Roper (19). The Wilcoxon ranked-sum test (22) was used to compare the ratio, NFPs/EE, under spontaneous and stimulated conditions for control and dysplastic slices.


Figure 4. A: Electrical stimulation produces a single small negative field potential in both control and dysplastic cortex in normal artificial cerebrospinal fluid (aCSF). Twenty responses to stimulation at the layer VI–white matter junction are averaged for both a control (left) and dysplastic (right) slice. Responses were recorded from layers II/III. The recordings demonstrate a stimulation artifact followed by a small negative field potential (NFP). There were no significant differences in latency, amplitude, or duration (see text). B: In 10 μM bicuculline methiodide (BMI)-containing aCSF, electrical stimulation produced multiple NFPs from both control (left) and dysplastic (right) cortex. For example, in the upper trace from the dysplastic slice, electrical stimulation produced six NFPs within the 4-s recording period. Each 4-s episode was considered an epileptiform event (EE). As demonstrated here, NFPs/EE were greater for dysplastic slices versus control (p < 0.05; Wilcoxon ranked-sum test, values given in Table 1). C: In 10 μM BMI–containing aCSF, control (left) and dysplastic (right) cortical slices demonstrated spontaneous epileptiform events (EEs). Each EE was made up of one or more NFPs. In the data shown, the control slice had three EEs, which have one, two, and three NFPs, respectively. The dysplastic slice had seven EEs. The dysplastic group had more EEs and a greater NFPs/EE ratio for each 20-s recording period (p < 0.05; Wilcoxon ranked-sum test; values given in Table 1).

Download figure to PowerPoint

Whole-cell recording

To test the intrinsic properties of control and dysplastic cortical neurons (pyramidal cells), slices were prepared as earlier from postnatal rats ages P14–P21 (mean age for control and dysplastic, P16). Slices from this younger age were chosen because pyramidal neurons could be identified by their morphology while stable whole-cell recordings could still be obtained. At older ages, stable whole-cell recording was difficult because of the thickness of the neuropil. Coronal slices (400 μm) were prepared at the level of the somatosensory cortex in ice-cold aCSF. Slices were allowed to recover for ≥1 h in aCSF bubbled with 95% O2, 5% CO2 at RT. Slices were perfused with aCSF (∼2 ml/min) on the stage of a microscope equipped with IR-DIC optics (Olympus, Tokyo, Japan). Whole-cell recording (17) was performed with a two-stage amplifier (HEKA, Lambrecht/Pfalz, Germany) and microelectrodes (3–6 MΩ) filled with the following solution: 135 mM Kgluconate; 4 mM KCl, 2 mM NaCl, 10 mM HEPES, 1 mM EGTA, 10 mM ATP-Mg, 0.3 mM GTP-Tris, 25 mM phosphocreatine, and 0.5 mM Lucifer Yellow (pH 7.25, 295 mOsm). All chemicals were purchased from Sigma-RBI (St. Louis, MO, U.S.A.). A high concentration of ATP and phosphocreatine was used in the electrode to prevent rundown of the GABA current (23).

Each neuron was characterized by the active membrane properties of the cell (e.g., regular spiking vs. bursting) (24). Regular-spiking (RS) neurons show a rapidly adapting response to a suprathreshold step of current with spiking frequencies in the range of 25–50 Hz. Bursting neurons demonstrate clusters of spikes to a suprathreshold step of current. Fast-spiking neurons respond to suprathreshold current with a spiking rate at ∼300–400 Hz. After characterization, tetrodotoxin (0.5 μM) was added to the bath solution to inhibit spiking. Membrane resistance and capacitance were calculated by fitting a single exponential to the current response to a step of voltage (–20 mV) and deriving the access resistance Ra, the membrane resistance Rm, and the membrane capacitance Cm(25).

The current response of the cell to increasing concentrations of GABA (100, 250, 500, 1,000, 2,000 μM) also was recorded. GABA was applied by using focal pressure application (ALA Scientific, Westbury, NY, U.S.A.) at a holding potential of –20 mV. GABA currents are outward under these conditions. The reversal potential was approximately –70 mV for all cells. Pyramidal neurons in layers IV/V were identified by their characteristic morphology under IR-DIC optics. Detailed analysis of GABA responses was limited to RS pyramidal neurons. Intact neurons close to the slice surface (<20 μm) were chosen so that applied GABA would have rapid access to the cell. In addition, to be included in the study, all neurons had to have a membrane potential (Vm) less than –60 mV and a membrane resistance (Rm) >20 MΩ. The current response of these cells to GABA application was recorded both digitally (sampling rate, 50 kHz) and with a chart recorder (Gould, Essex, U.K.). In addition, the current response to glutamate application was recorded as a control at a holding potential of –60 mV. In all cells, a large inward current to glutamate application was recorded. The Hill equation was fit to the current responses after GABA application:

  • image(1)

where R is the current response of the neuron to applied GABA, A0 is the maximal current response, C is the concentration of GABA, EC50 is concentration of the GABA concentration that produces a half-maximal response, and n is the exponent that determines the steepness of the curve. A0, EC50, and n were adjusted by the fitting program until a minimum of the χ2 statistic was found.

Single exponential fits and the Hill equation fits were done by using a plotting and statistics program (Origin 6.1; Originlab, Northhamptom, MA, U.S.A.). Population statistics are given as the mean ± the standard error of the mean (SEM).


  1. Top of page
  2. Abstract

Litter size and morphologic characteristics

The litter sizes in the BCNU-exposed animals were reduced. For five consecutive control and BCNU-exposed litters, the mean litter size of the control groups was 11.8 versus 6.0 in the BCNU-exposed group (p < 0.05). BCNU-exposed pups also had a significantly lower body and brain weights than those of controls. For five randomly selected BCNU-exposed and control P0 pups, the mean body weight was 7.9 and 4.7 g, respectively (p < 0.01). Similarly, the mean wet brain weight for each group was 405 versus 187 mg, respectively (p = 0.01). By adult age (P 50), however, control and BCNU-exposed rats were found to have no significant difference in body weight. For a group of P60 randomly selected control and exposed animals used for histology (n = 3 in each group), the mean body weight was 259 g in each group. However, the difference in mean brain weight was still significant (1.9 vs. 1.25 g, respectively; p = 0.02).

Cortical histology

Examination of the brains of pups exposed to BCNU in utero demonstrated clear evidence of cortical maldevelopment at P0. The entire quadrigeminal plate is visible on the dorsal surface of the brain in these animals, indicating an overall reduction in cortical volume. These abnormalities resemble the previously reported description of MAM-exposed rats (7). On average, the brain weight of BCNU-exposed pups was reduced by ∼50%. Cresyl-violet staining reveals significant differences in cortical architecture between BCNU-exposed and control cortex (Fig. 1A). Compared with the control P0 cortex, the BCNU-exposed cortex shows a thin cortical plate and distinct clusters of neuronal elements that represent heterotopias (Fig. 1B). In the control adult rat cortex, the typical six-layered lamination pattern is apparent (Fig. 1C), whereas the BCNU-exposed cortex shows laminar disorganization and thinning (Fig. 1D). In addition, there are clusters of large neurons in both superficial and deep layers in the BCNU-exposed cortex.


Figure 1. Nissl stain of control and dysplastic cortex. A: Photomicrograph of neonatal control cortex (P0) at the approximate level of the anterior commissure showing the normal lamination and thickness of cortex. B: 1-3-Bis-chloroethyl-nitrosourea (BCNU)-exposed neonatal (P0) cortex at the same rostral–caudal level shows disruption of lamination and heterotopic areas (arrow) of neurons in cortex. C: Adult control cortex (P60) shows the normal lamination pattern and distribution of neurons. D: In the adult (P60) BCNU-exposed cortex, the cortex is thin, and there is disruption of lamination. Abnormal clusters of neurons also are apparent (arrow).

Download figure to PowerPoint


The better to characterize the effect of BCNU exposure on rat cortex, cryostat sections were stained with antibodies against molecules thought to be important in the development of cortical architecture. Both control and exposed P0 cortex were processed with the anti-nestin antibody, RT-401, which labels radial glial fibers. Prior work with irradiated and MAM-exposed cortex showed disruption of radial glial fibers (6,26). Similarly, the BCNU-exposed cortex shows patterns of disrupted radial glial fibers with RT-401 labeling (Fig. 2).


Figure 2. Immunohistochemical staining shows disruption of the normal radial glial fiber pattern and displaced CR-50+ cells in neonatal dysplastic cortex. A: RT-401 antibody staining of control neonatal (P0) cortex shows the normal pattern of radial glial fibers, which have a smooth, parallel appearance. B: In neonatal (P0) 1-3-bis-chloroethyl-nitrosourea (BCNU)-exposed cortex, RT-401 staining shows disrupted and disorganized radial glial fibers. C: In control neonatal cortex, CR-50 antibody staining shows typical Cajal–Retzius cells in layer I (arrows). D: In BCNU-exposed neonatal cortex, Cajal–Retzius cells take up their normal position in layer I, but also are found in deeper layers (arrows) among heterotopic neurons.

Download figure to PowerPoint

Antibody against the reeler gene–related antigen, CR-50, labels Cajal–Retzius cells in layer I of neonatal cortex. The reeler gene product, reelin, labeled by the CR-50 antibody, is thought to be important in establishing cortical lamination (27). We used the CR-50 antibody to characterize the location of CR-50+ cells in neonatal control and BCNU-exposed cortex. Control neonatal cortex shows CR-50+ cells located in layer I with the typical “torpedo” morphology of Cajal–Retzius cells (28,29). In BCNU-exposed cortex, CR-50 antibody labels this layer I population of cells, as well as a population of cells with similar morphology in deeper layers. These cells may represent inappropriately situated Cajal–Retzius cells (Fig. 2).

In several models of dysplastic cortex, decreased numbers of inhibitory interneurons have been noted compared with those in controls. This observation was made in both freeze-lesioned (3) and γ-irradiated models (12). Antibodies against the calcium-binding proteins, calbindin, calretinin, and parvalbumin, identify populations of inhibitory interneurons (30–32). We used antibodies against these calcium-binding proteins on control and BCNU-exposed cortex to investigate whether there was a gross reduction in the inhibitory interneuron population in BCNU-exposed cortex. Figure 3 shows representative sections from adult control and BCNU-exposed cortex. There was no clear difference in the density of calbindin-, parvalbumin-, or calretinin-positive cells in control or BCNU-exposed cortex.


Figure 3. Immunohistochemical staining with anti-calbindin (CB) and anti-parvalbumin (PV) demonstrates little difference between the adult control and dysplastic cortex. A: Anti-calbindin antibody staining of adult (P60) control cortex shows interneurons in the more superficial cortical layers. B: 1-3-Bis-chloroethyl-nitrosourea (BCNU)-exposed adult cortex (P60) shows a pattern of anti-calbindin antibody staining similar to control. C: Anti-parvalbumin antibody staining of control adult cortex shows a diffuse pattern of interneuron staining throughout the cortex. D: The pattern of anti-parvalbumin antibody staining seen in dysplastic cortex is similar to that in control.

Download figure to PowerPoint

Field recording

Before analyzing the properties of individual neurons in dysplastic (BCNU-exposed) cortex, we sought to characterize further the physiologic behavior of the neuronal population. It has been shown that dysplastic cortex produced by in utero treatment with γ-irradiation is hyperexcitable under conditions of partial GABAA-receptor blockade (19). We sought to perform a similar analysis on adult cortex exposed in utero to BCNU. Adult (>P50) cortical slices were prepared from control and dysplastic brains at the level of the anterior commissure. A stimulating electrode was used to deliver a suprathreshold stimulus to the WM, and the population response was recorded in the superficial cortical layers (layers II/III). There were no consistent differences in the responses of control and dysplastic cortex under these conditions. The responses typically consisted of a single NFP. The responses to 20 episodes of stimulation were averaged, and the mean latency, amplitude, and duration of the NFP were calculated for each slice (Fig. 4A). There was no statistical difference in these parameters between the responses of the control (mean, 16.6 ms; –0.305 mV, 182 ms) and dysplastic slices (mean, 19.3 ms; –0.241 mV, 125 ms).

Next slices were perfused with 10 μM BMI-containing aCSF for 15 min. After this exposure, responses to the same stimulus were recorded. Under conditions of GABAA-receptor blockade, a prolonged NFP, part of an epileptiform event (EE), was elicited from cortical slices. Sometimes, this EE was made up of more than one NFP (Fig. 4B). Ten control and 10 dysplastic cortical slices from seven control rats (five male and two female) and six BCNU-exposed rats (four male and two female) were tested. For each slice, the number of NFPs per EE (NFPs/EE) was calculated and averaged across 20 episodes. Between episodes, the slice was allowed to recover for 30 s. The data are summarized in Table 2. The mean number of NFPs/EE was 1.33 for control slices and 1.75 for dysplastic cortex. Because there could be no less than one NFP per EE and the data were not Gaussian distributed, the Wilcoxon ranked-sum test was used to test the statistical significance of the data. With this test, the difference in the NFPs/EE between control and dysplastic cortex was significant (p = 0.032) (Table 1).

Table 2.  Membrane properties of control versus dysplastic cortical neurons (pyramidal type)
  1. Membrane properties do not differ between pyramidal neurons from control and dysplastic tissue. Whole-cell recordings were performed on 10 pyramidal neurons from control tissue and 11 pyramidal neurons from dysplastic tissue. Ra, access resistance; Rm, membrane resistance; Cm, membrane capacitance as described in Methods. The mean, median, SEM (standard error of the mean), maximum, and minimum are given for each parameter. The significance of a two-sided t test comparing the mean of each parameter for control versus dysplastic neurons is given by the p value. A p value <0.05 indicates statistical significance.

Ra control (MΩ)
Ra dysplastic (MΩ)22.52.836.612.50.133
Rm control (MΩ)100.020.1252.013.2
Rm dysplastic (MΩ)79.614.7168.09.00.416
Cm control (pF)216.141.0551.057.0
Cm dysplastic (pF)265.633.6516.0129.00.358
Vm control (mV)−67.82.1−60−82
Vm dysplastic (mV)−68.71.1−63−740.704
Table 1.  Number of negative field potentials per epileptiform event in bicuculline (10 μM) for control versus dysplastic cortex
  1. Dysplastic tissue demonstrates more negative field potentials per epileptiform event than does control. Data shown are for 10 control and 10 dysplastic cortical slices. Max, the maximal number of negative field potentials per epileptiform event observed across these populations; min, the minimum observed; p, p value of the Wilcoxon ranked-sum test used to assess whether the responses of the control and dysplastic populations were different; p < 0.05, statistical significance; SEM, standard error of the mean.

Control (stimulated)1.330.1561
Dysplastic (stimulated)1.750.22710.032
Control (spontaneous)1.310.1561
Dysplastic (spontaneous)2.230.551210.028

After 20 episodes of stimulation, the control and dysplastic cortical slices were perfused with 10 μM BMI-containing aCSF for 20 min, and spontaneous activity was recorded. Under conditions of partial GABA blockade, cortical slices will periodically exhibit a spontaneous EE. Spontaneous events were recorded, and the mean number of NFPs per EE was calculated. There were more spontaneous events in the dysplastic slices than the control slices, and this was statistically significant (mean control, 13.5, vs. mean dysplastic, 29.1; p = 0.014, t test). Furthermore, the number of NFPs per EE was greater in the dysplastic group, 2.26, versus the control group, 1.31, and this difference also was significant with the Wilcoxon ranked-sum test (p = 0.028; Table 1). Recordings of spontaneous events from control and dysplastic slices are shown in Fig. 4C.

Whole-cell recording

Having noted hyperexcitablity of the overall neuronal population in dysplastic cortex, we wished to examine the passive membrane properties and the physiology of GABA inhibition in individual neurons from dysplastic cortex and control tissue. To obtain stable whole-cell recordings, it was necessary to use younger animals (mean age, P16) than for the field-recording experiments. However, we were still able to identify easily the deeper cortical layers and to differentiate pyramidal cells at these ages (Fig. 5A). Gigaseal patches were obtained from 10 pyramidal neurons (five animals) in the deep layers (IV/V) of control cortex and 11 pyramidal cells (six animals) from the deep layers of dysplastic cortex. Under voltage-clamp conditions, the cells were first classified as neurons because of the presence of action-potential currents, and under current clamp, current steps of 250-ms duration were delivered to elicit repetitive action-potential firing (Fig. 5B). The analysis of control and dysplastic neurons was limited to RS pyramidal cells (24). A single exponential was fit to the response to a step of voltage (–20 mV), and the membrane parameters calculated: access resistance (Ra), membrane resistance (Rm), and membrane capacitance (Cm). The passive membrane properties for both populations are summarized in Table 2. There was no statistical difference in any of these parameters between the control and dysplastic populations. In addition, there was no significant difference in the mean membrane potential (control, –67.8 mV, vs. dysplastic, –68.7 mV).


Figure 5. Whole-cell recording in control and dysplastic cortical slices. A: Layer IV/V pyramidal cells were identified in both control and dysplastic cortex by their appearance under IR-DIC optics and with epifluorescence after filling with Lucifer yellow (LY). B: Voltage recordings in current-clamp mode were used to identify neurons (see text). Only regularly spiking (RS) pyramidal cells were examined. This example shows the appearance and active membrane properties of an RS pyramidal cell from dysplastic cortex (P16). A suprathreshold step of current lasting 250 ms evoked a train of action potentials at a rate of ∼25 Hz. The spiking rate rapidly adapts, as signified by the shorter initial interspike interval.

Download figure to PowerPoint

GABA and glutamate responses

Both glutamate and GABA were used to characterize responses to neurotransmitters. At a holding potential of –60mV, application of glutamate (600 μM) gave a large inward current. There was no statistical difference between the responses of pyramidal neurons from control cortex and dysplastic cortex to the application of glutamate: mean response control, 1,254 ± 228 pA; mean response dysplastic, 911 ± 120 pA; p = 0.2.

Control and dysplastic neuronal responses also were recorded after application of increasing concentrations of GABA (100–2,000 μM). At a holding potential of –20 mV, GABA was applied for 5 s and then washed for 30 s before the application of the next sample. Responses from control and dysplastic neurons to each concentration of GABA were charted (Fig. 6A).


Figure 6. Outward currents measured from pyramidal cortical neurons from control and dysplastic slices in response to increasing concentrations of γ-aminobutyric acid (GABA; 100–2,000 μM). The holding potential was –20 mV to generate robust outward currents under these conditions. Responses are shown to increasing concentrations of GABA (see Methods). Note that the control neuron plateaus to a near-maximal response at a lower concentration of GABA than does the neuron from dysplastic cortex. B: The Hill equation was fit to the GABA responses of layer IV/V pyramidal neurons from control and dysplastic cortex. The fit of the Hill equation is shown to pyramidal neuron responses in control (solid circles) and dysplastic cortex (open circles) Each graph consists of the normalized, averaged responses of 10 neurons (error bars, one standard error of the mean). Solid line, fit of the Hill equation (Eq. 1) to control pyramidal neuron responses (EC50, 252 μM; n, 1.05); dashed line, fit of the Hill equation to pyramidal neuron responses from dysplastic cortex (EC50, 387 μM; n, 1.24). Note that the EC50 is higher in the pyramidal neurons from dysplastic cortex. The Hill parameters averaged for a population of 10 control neuron responses and 11 dysplastic neurons are shown in Table 1 (see text).

Download figure to PowerPoint

The maximal current at each concentration from multiple applications (typically three) was averaged, and the Hill equation (Eq. 1) was fit to the data (Fig. 6B). There was no statistical difference in the maximal current, A0, or the exponent, n, for the neurons from control and dysplastic cortex (Table 3). However, EC50 was significantly greater in the dysplastic (400 μM) than in the control (261 μM) populations (p = 0.026). This result indicates a decreased sensitivity to GABA inhibition in pyramidal neurons in the dysplastic cortex.

Table 3.  Hill equation parameters fit to the responses of cortical neurons to GABA (control vs. dysplastic)
  1. Dysplastic tissue exhibits an upward shift in EC50. Whole-cell recording was performed on 10 pyramidal neurons from control and 11 pyramidal neurons from dysplastic cortex. The Hill equation was fit to the responses of these neurons to increasing concentration of γ-aminobutyric acid (GABA). The parameters are A0, EC50, n are given in the Hill equation (Eq. 1). The mean, SEM, maximum, and minimum are given for each parameter in both the control and dysplastic groups. P indicates the significance level of a two-sided t test comparing the parameters for both the control and dysplastic groups. Note that EC50,dysplastic and EC50,control are significantly different (p < 0.05).

A0,control (pA)2,524.7233.13,840.11,226.6
A0,dysplastic (pA)2,806.8263.63,848.81,451.50.437


  1. Top of page
  2. Abstract

Cortical dysplasia

CD, defined as an area of disrupted cortical architecture, can result from abnormal cortical development or postnatal injury. From clinical specimens, it is clear that CD is associated with epilepsy (33,34). Heterotopic clusters of neurons, dyslamination, and abnormal cells have all been found in surgical specimens obtained after resection of epileptogenic cortex (1). The exact mechanisms that underlie the epileptogenicity of the cortex in CD are unclear. However, CD itself is not a single disorder but can be due to the phenotypic expression of multiple derangements of cortical development. MAM exposure and γ-irradiation both mimic CD by affecting the proliferating pool of neuronal and glial precursors when introduced in utero on E15–E17 in the rat (4,6,7). We demonstrate here that the DNA-alkylating agent, BCNU, also affects these progenitor cells when given in utero. Our histologic results show that this insult leads to characteristics of CD: disruption of cortical lamination, heterotopias, and abnormal neurons.

Immunohistochemistry has shown that radial glial fibers and Cajal–Retzius cells are disrupted in irradiation and MAM models, respectively (6). Similarly, we have found that BCNU exposure leads to abnormal positioning of Cajal–Retzius cells and disruption of radial glial fibers. These features may underlie or contribute to the observed cortical maldevelopment. The characteristic columnar and laminar organization of neocortex is thought to be established largely by the action of these two cell types, both of which are present transiently during early cortical development (35–38). Radial glial fibers form a scaffold for migrating neurons and recently have been shown also to serve as neuronal precursors in the developing cortex (39–41). Disruption of radial glial function could therefore produce abnormalities of neuronal production as well as abnormalities of neuronal migration.

The radial organization of the cortex is critically dependent on Cajal–Retzius cell function, as demonstrated by the migration defects seen in the mutant mouse, reeler, that result from absence of the reelin protein secreted by Cajal–Retzius cells (42). However, radial glial cells also are significantly altered in reeler cortex, being poorly differentiated and nonradially oriented (43,44), suggesting that there may be a relation between Cajal–Retzius cell function and radial glia. The possibility of a functional relation between these two cell types is supported by several additional observations including experiments showing that ectopic Cajal–Retzius cells transplanted into adult cerebellum induce host Bergmann glia to assume a more primitive radial glial phenotype (45). Furthermore, ablating Cajal–Retzius cells in neonatal cortex disrupts the density and organization of radial glial processes (46), and a variety of antimitotic treatments applied to early developing cortex alter both the location of Cajal–Retzius cells and the morphology of radial glia (26,47). Thus the cortical migration defects observed here may derive from BCNU-induced injury either to radial glia or Cajal–Retzius cells or to the coordinated function of both.

Epileptogenic cortex

Several studies have attempted to analyze neuronal properties in regions of dysplastic cortex resected from patients with epilepsy (48–50). These studies encountered special problems with the preparation and maintenance of mature human cortex after surgical resection. Experimental models of CD can help formulate hypotheses that may be testable with limited surgical material. Here we have described a novel model in which CD is obtained.

Neither control nor dysplastic cortex generated spontaneous epileptiform events without the presence of BMI. However, under conditions of partial GABAA-receptor blockade (10 μM BMI) (51,52), the population of neurons in this dysplastic cortex was found to express hyperexcitable characteristics similar to those found in irradiated tissue. Although the irradiated cortex studied by Roper et al. (19) expressed more robust hyperexcitability than that found in BCNU-exposed cortex, we found a statistically significant increase in number and duration of epileptiform events compared with those in control. This difference could be due to either abnormal intracortical connections or alterations in the properties of single neurons. Chevassus-au-Louis et al. (53,54) showed that abnormal intracortical connections in the MAM model may support hyperexcitability. As a starting point, we wished to examine the properties of single neurons in the BCNU-exposed dysplastic cortex.

GABA alterations in CD

With whole-cell recording techniques, we found no significant differences between the passive membrane properties of control and dysplastic neurons. This is similar to results from human dysplastic cortex that showed no differences in these properties (49). Similarly, we found no difference in the responses to applied glutamate, the most abundant excitatory neurotransmitter in the mammalian cortex. However, the response to the inhibitory neurotransmitter, GABA, was found to be altered in neurons from dysplastic cortex. The half-maximal concentration, EC50, of GABA was increased in these neurons, indicating decreased sensitivity to GABA. Similarly, Gibbs et al. (23,55,56) found altered GABA responses in neurons isolated from epileptic tissue. These changes may contribute to the hyperexcitability observed in dysplastic tissue.


  1. Top of page
  2. Abstract

We have introduced a rat model of CD that reproduces the properties of disrupted cortical architecture and abnormal neurons in cortical tissue. We also found that the dysplastic tissue exhibits both hyperexcitability and decreased sensitivity to inhibition. Both of these alterations shift the normal balance of excitation and inhibition. The normal mammalian cortex is maintained in a state that is only weakly susceptible to epileptic events. However, in CD, these alterations may predispose the tissue to hypersynchronous firing and clinical seizures. Further examination of this model based on the disruption of neuronal and glial precursors by the alkylating agent, BCNU, may help identify the molecular mechanisms responsible for these alterations in excitation and inhibition.

Acknowledgment: We thank Stephen Noctor, Tamily Weissman, and Brian Clinton for many helpful suggestions and discussions. This work was supported by grants from the NIH (NS 35710) and the March of Dimes Birth Defects Foundation.


  1. Top of page
  2. Abstract
  • 1
    Mischel PS, Vinters HV. Neuropathology of developmental disorders associated with epilepsy. In: J Engel Jr., TA Pedley, Editors. Epilepsy: Lippincott Williams & Wilkins, 1999.
  • 2
    Goldring S, Gregorie EM. Surgical management of epilepsy using epidural recordings to localize the seizure focus. Review of 100 cases. J Neurosurg 1984;60(3):45766.
  • 3
    Jacobs KM, Kharazia VN, Prince DA. Mechanisms underlying epileptogenesis in cortical malformations. Epilepsy Res 1999;36(2–3):16588.
  • 4
    Chevassus-au-Louis N, Baraban SC, Gaiarsa JL, Ben-Ari Y. Cortical malformations and epilepsy: new insights from animal models. Epilepsia 1999;40(7):81121.
  • 5
    Jacobs KM, Gutnick MJ, Prince DA. Hyperexcitability in a model of cortical maldevelopment. Cereb Cortex 1996;6(3):51423.
  • 6
    Roper SN. In utero irradiation of rats as a model of human cerebrocortical dysgenesis: a review. Epilepsy Res 1998;32(1–2):6374.
  • 7
    Germano IM, Sperber EF. Transplacentally induced neuronal migration disorders: an animal model for the study of the epilepsies. J Neurosci Res 1998;51(4):47388.
  • 8
    Sancini G, Franceschetti S, Battaglia G, Colacitti C, Di Luca M, Spreafico R, Avanzini G. Dysplastic neocortex and subcortical heterotopias in methylazoxymethanol-treated rats: an intracellular study of identified pyramidal neurones. Neurosci Lett 1998;246(3):1815.
  • 9
    Smith BN, Dudek FE, Roper SN. Synaptic responses of neurons in heterotopic gray matter in an animal model of cortical dysgenesis. Dev Neurosci 1999;21(3–5):36573.
  • 10
    Baraban SC, Wenzel HJ, Hochman DW, Schwartzkroin PA. Characterization of heterotopic cell clusters in the hippocampus of rats exposed to methylazoxymethanol in utero. Epilepsy Res 2000;39(2):87102.
  • 11
    Rosen GD, Jacobs KM, Prince DA. Effects of neonatal freeze lesions on expression of parvalbumin in rat neocortex. Cereb Cortex 1998;8(8):75361.
  • 12
    Roper SN, Eisenschenk S, King MA. Reduced density of parvalbumin- and calbindin D28-immunoreactive neurons in experimental cortical dysplasia. Epilepsy Res 1999;37(1):6371.
  • 13
    Calabresi P, Chabner BA. Chemotherapy of neoplastic diseases. Principles of Pharmacology, 1999.
  • 14
    Colacitti C, Sancini G, Franceschetti S, Cattabeni F, Spreafico R, Di Luca M, Battaglia G. The methylazoxymethanol (MAM) treated rat as an animal model for human development brain dysgeneses: morphological features. In: RSpreafico, G.Avanzini, FAndermann, Editors. Abnormal Cortical Development and Epilepsy, London: John Libbey & Co., 1999: 109120.
  • 15
    Layton PB, Greenberg HS, Stetson PL, Ensminger WD, Gyves JW. BCNU solubility and toxicity in the treatment of malignant astrocytomas. J Neurosurg 1984;60(6):11347.
  • 16
    Bolam JP. Preparation of central nervous system tissue for light and electron microscopy. In: JPBolam, Editor. Experimental Neuroanatomy: A Practical Approach, New York: Oxford University Press, 1992: 129.
  • 17
    Blanton MG, Lo Turco JJ, Kriegstein AR. Whole cell recording from neurons in slices of reptilian and mammalian cerebral cortex. J Neurosci Methods 1989;30(30):20310.
  • 18
    Baraban SC, Schwartzkroin PA. Electrophysiology of CA1 pyramidal neurons in an animal model of neuronal migration disorders: prenatal methylazoxymethanol treatment. Epilepsy Res 1995;22(2):14556.
  • 19
    Roper SN, King MA, Abraham LA, Boillot MA. Disinhibited in vitro neocortical slices containing experimentally induced cortical dysplasia demonstrate hyperexcitability. Epilepsy Res 1997;26(3):4439.
  • 20
    Chagnac-Amitai Y, Connors BW. Synchronized excitation and inhibition driven by intrinsically bursting neurons in neocortex. J Neurophysiol 1989;62(5):114962.
  • 21
    Chagnac-Amitai Y, Connors BW. Horizontal spread of synchronized activity in neocortex and its control by GABA-mediated inhibition. J Neurophysiol 1989;61(4):74758.
  • 22
    Rosner B. Fundamentals of Biostatistics. 1995, Belmont: Duxbury Press.
  • 23
    Gibbs JW, 3rd, Sombati S, DeLorenzo RJ, Coulter DA. Physiological and pharmacological alterations in postsynaptic GABA(A) receptor function in a hippocampal culture model of chronic spontaneous seizures. J Neurophysiol 1997;77(4):213952.
  • 24
    Connors BW, Gutnick MJ. Intrinsic firing patterns of diverse neocortical neurons. Trends Neurosci 1990;13(3):99104.
  • 25
    Sakmann B. Capacitance Measurements. In: BSakmann, Editor. Single Channel Recording, New York: Plenum Press, 1995.
  • 26
    Noctor SC, Palmer SL, Hasling T, Juliano SL. Interference with the development of early generated neocortex results in disruption of radial glia and abnormal formation of neocortical layers. Cereb Cortex 1999;9(2):12136.
  • 27
    Ogawa M, Miyata T, Nakajima K, Yagyu K, Seike M, Ikenaka K, Yamamoto H, Mikoshiba K. The reeler gene-associated antigen on Cajal-Retzius neurons is a crucial molecule for laminar organization or cortical neurons. Neuron 1995;14: 899912.
  • 28
    Konig N, Hornung JP, Van der Loos H. Identification of Cajal-Retzius cells in immature rodent cerebral cortex: a combined Golgi–EM study. Neurosci Lett 1981;27(3):2259.
  • 29
    Konig N, Schachner M. Neuronal and glial cells in the superficial layers of early postnatal mouse neocortex: immunofluorescence observations. Neurosci Lett 1981;26(3):22731.
  • 30
    Prince DA, Jacobs K. Inhibitory function in two models of chronic epileptogenesis. Epilepsy Res 1998;32(1–2):8392.
  • 31
    Schwarz P, Stichel CC, Luhmann HJ. Characterization of neuronal migration disorders in neocortical structures: loss or preservation of inhibitory interneurons? Epilepsia 2000;41(7):7817.
  • 32
    Hannan AJ, Servotte S, Katsnelson A, Sisodiya S, Blakemore C, Squier M, Molnar Z. Characterization of nodular neuronal heterotopia in children. Brain 1999;122(Pt 2):21938.
  • 33
    Barth PG. Disorders of neuronal migration. Can J Neurol Sci 1987;14(1):116.
  • 34
    Barkovich AJ, Gressens P, Evrard P. Formation, maturation, and disorders of brain neocortex. AJNR Am J Neuroradiol 1992;13(2):42346.
  • 35
    Lambert de Rouvroit C, Goffinet AM. A new view of early cortical development. Biochem Pharmacol 1998;56(11):14039.
  • 36
    Super H, Soriano E, Uylings HB. The functions of the preplate in development and evolution of the neocortex and hippocampus. Brain Res Brain Res Rev 1998;27(1):4064.
  • 37
    Marin-Padilla M. Cajal-Retzius cells and the development of the neocortex. Trends Neurosci 1998;21(2):6471.
  • 38
    Frotscher M. Dual role of Cajal-Retzius cells and reelin in cortical development. Cell Tissue Res 1997;290(2):31522.
  • 39
    Noctor SC, Flint AC, Weissman TA, Dammerman RS, Kriegstein AR. Neurons derived from radial glial cells establish radial units in neocortex. Nature 2001;409(6821):71420.
  • 40
    Malatesta P, Sgado P, Caneparo L, Barsacchi G, Cremisi F. In vivo PC3 overexpression by retroviral vector affects cell differentiation of rat cortical precursors. Brain Res Dev Brain Res 2001;128(2):1815.
  • 41
    Malatesta P, Gotz M, Barsacchi G, Price J, Zoncu R, Cremisi F. PC3 overexpression affects the pattern of cell division of rat cortical precursors. Mech Dev 2000;90(1):1728.
  • 42
    D'Arcangelo G, Miao GG, Chen SC, Soares HD, Morgan JI, Curran T. A protein related to extracellular matrix proteins deleted in the mouse mutant reeler. Nature 1995;374(6524):71923.
  • 43
    Hunter-Schaedle KE. Radial glial cell development and transformation are disturbed in reeler forebrain. J Neurobiol 1997;33(4):45972.
  • 44
    O'Brien TF, Steindler DA, Cooper NG. Abnormal glial and glycoconjugate dispositions in the somatosensory cortical barrel field of the early postnatal reeler mutant mouse. Brain Res 1987;429(2):30917.
  • 45
    Soriano E, Alvarado-Mallart RM, Dumesnil N, Del Rio JA, Sotelo C. Cajal-Retzius cells regulate the radial glia phenotype in the adult and developing cerebellum and alter granule cell migration. Neuron 1997;18(4):56377.
  • 46
    Super H, Del Rio JA, Martinez A, Perez-Sust P, Soriano E. Disruption of neuronal migration and radial glia in the developing cerebral cortex following ablation of Cajal-Retzius cells. Cereb Cortex 2000;10(6):60213.
  • 47
    Roper SN, Abraham LA, Streit WJ. Exposure to in utero irradiation produces disruption of radial glia in rats. Dev Neurosci 1997;19(6):5218.
  • 48
    Avoli M, Bernasconi A, Mattia D, Olivier A, Hwa GG. Epileptiform discharges in the human dysplastic neocortex: in vitro physiology and pharmacology. Ann Neurol 1999;46(6):81626.
  • 49
    Dudek FE, Wuarin JP, Tasker JG, Kim YI, Peacock WJ. Neurophysiology of neocortical slices resected from children undergoing surgical treatment for epilepsy. J Neurosci Methods 1995;59(1):4958.
  • 50
    Aitken PG, Breese GR, Dudek FF, Edwards F, Espanol MT, Larkman PM, Lipton P, Newman GC, Nowak TS, Jr., Panizzon KL, et al. Preparative methods for brain slices: a discussion. J Neurosci Methods 1995;59(1):13949.
  • 51
    Connors BW, Malenka RC, Silva LR. Two inhibitory postsynaptic potentials, and GABAA and GABAB receptor-mediated responses in neocortex of rat and cat. J Physiol 1988;406: 44368.
  • 52
    Gutnick MJ, Connors BW, Prince DA. Mechanisms of neocortical epileptogenesis in vitro. J Neurophysiol 1982;48(6):132135.
  • 53
    Chevassus-Au-Louis N, Jorquera I, Ben-Ari Y, Represa A. Abnormal connections in the malformed cortex of rats with prenatal treatment with methylazoxymethanol may support hyperexcitability. Dev Neurosci 1999;21(3–5):38592.
  • 54
    Chevassus-au-Louis N, Represa A. The right neuron at the wrong place: biology of heterotopic neurons in cortical neuronal migration disorders, with special reference to associated pathologies. Cell Mol Life Sci 1999;55(10):120615.
  • 55
    Gibbs JW, 3rd, Morton LD, Amaker B, Ward JD, Holloway KL, Coulter DA. Physiological analysis of Rasmussen's encephalitis: patch clamp recordings of altered inhibitory neurotransmitter function in resected frontal cortical tissue. Epilepsy Res 1998;31(1):1327.
  • 56
    Gibbs JW, 3rd, Shumate MD, Coulter DA. Differential epilepsy-associated alterations in postsynaptic GABA(A) receptor function in dentate granule and CA1 neurons. J Neurophysiol 1997;77(4):192438.