INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Disorders of cortical development are a major cause of epilepsy in adults and children, comprising 40% of cases of medically refractory epilepsy in children who require surgical treatment (Farrell et al. 1992). Despite this fact, relatively little is known about how such disorders occur and the mechanistic relationship between these structural abnormalities and epilepsy. Neuronal heterotopia is a condition where neuronal cell bodies reside in an abnormal location within the brain. In the cerebral hemisphere, neuronal heterotopia may be subcortical, residing in the cerebral white matter, or periventricular, residing adjacent to the lateral ventricle. Both types have a strong association with epilepsy (Barkovitch and Kjos 1992; Dubeau et al. 1995). Two human syndromes of diffuse heterotopia, subcortical band heterotopia and X-linked periventricular heterotopia, have been linked to two different genes that reside on the X chromosome; DCX and FLN-1, respectively (des Portes et al. 1998; Eksioglu et al. 1996; Fox et al. 1998). However, the etiology of most sporadic and focal cases of subcortical and periventricular heterotopia is still unknown.

The function and physiological properties of heterotopic neurons are not well understood. Metabolic studies and electrical recordings from human heterotopic gray matter show that the neurons are active and can generate electrical seizure activity; but their local and long-range connections are difficult to study in humans. Several animal models have shown that heterotopic neurons can form long-range connections with ipsilateral and contralateral neocortex, thalamus, spinal cord and other brain structures (Colacitti et al. 1998; Jensen and Killackey 1984; Schottler et al. 1998; Yurkewicz et al. 1984). This raises the possibility that seizure activity generated in heterotopic gray matter could be transmitted to the rest of the brain and produce clinical seizures.

Exposure of fetal rats to external irradiation results in diffuse cortical dysplasia, subcortical and periventricular heterotopia, hippocampal heterotopia, and agenesis or hypoplasia of the corpus callosum (Cowan and Geller 1960; Riggs et al. 1956). They also demonstrate spontaneous electrographic seizures in vivo (Kondo et al. 2001). In this paper we use the term "heterotopic cortex" to describe masses of gray matter that reside deep to the normal location of the neocortex and that are clearly separated from the normotopic cortex by a layer of white matter. We use the term "normotopic, dysplastic cortex" to refer to the disorganized gray matter that resides just below the pia in the normal location of the neocortex. We have used the irradiated rat to study properties of excitatory and inhibitory synapses in subcortical heterotopic gray matter to better understand how these areas may be involved in the generation of seizures. Similar to a previous study in normotopic dysplastic cortex in irradiated rat (Zhu and Roper 2000), we report a selective impairment of inhibition in heterotopic gray matter.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals and irradiation

Pregnant rats with known insemination times were obtained from Harlan Sprague-Dawley (Indianapolis, IN). The day of insemination was designated embryonic day 0 (E0). Irradiation was performed on E17. Pregnant rats were placed in a well-ventilated Plexiglass box and exposed to 225 cGy of external -irradiation from a linear accelerator source. Control litters were obtained and housed in an identical fashion but not exposed to radiation. Offspring from 11 irradiated and 9 control mothers were used for experiments. All animals were maintained on 12-h light/dark cycles and were provided food and water ad libitum. All procedures used in the study adhered to guidelines approved by the Institutional Animal Care and Use Committee at the University of Florida.

Brain slice preparation

Coronal brain slices were obtained from 20- to 28-day-old rats using procedures described previously (Roper et al. 1997b). The rat was anesthetized by the inhalation of isoflurane and decapitated by a guillotine, and the brain was rapidly removed. Four hundred-micrometer-thick coronal brain slices were cut at the rostrocaudal level of the anterior commissure using a Vibratome (Technical Products International, St. Louis, MO). Slices were incubated on cell culture inserts (8 µm pore diam, Becton Dickinson, Franklin Lakes, NJ), covered by a thin layer of artificial cerebrospinal fluid (ACSF) containing (in mM) 124 NaCl, 26 NaHCO₂, 1.25 NaH₂PO₄, 2.5 KCl, 1 CaCl₂, 6 MgCl₂, and 10 D-glucose, and surrounded by humidified 95% O₂-5% CO₂ atmosphere at room temperature (22°C). After at least 1 h of incubation, the slice was transferred to a submerged recording chamber with continuous flow (2 ml/min) of ACSF as described above except for 2 mM CaCl₂ and 1 mM MgCl₂ and gassed with 95% O₂-5% CO₂ giving pH 7.4. All experiments were carried out at room temperature (22°C).

Electrophysiological recording

At the time of recording, areas of subcortical heterotopic gray matter were identified using a 2.5× objective and light microscopy (Fig. 1A). Whole cell recordings were made from pyramidal neurons in heterotopic cortex of irradiated rats and layer II/III of the neocortex in control rats under visual control using infrared differential interference contrast (IR-DIC) videomicroscopy with a fixed-stage microscope equipped with a 40×, 0.8 W water-immersion lens (Zeiss, Oberkochen, Germany). Pyramidal neurons were identified by their triangular somata (Fig. 1A), a single apical dendrite, and regular spiking pattern in response to a depolarizing current pulse. Patch electrodes had a resistance of 3-5 M when filled with intracellular solution. Electrodes were filled with (in mM) 120 K-gluconate, 8 NaCl, 10 HEPES, 4 MgATP, 0.3 Na₃GTP, and 0.2 EGTA (pH 7.3 with KOH, osmolarity 290-300 mOsm) for recording excitatory postsynaptic currents (EPSCs) or (in mM) 125 CsCl, 10 HEPES, 4 EGTA, 4 MgCl₂, 4 MgATP, and 0.3 Na₃GTP (pH 7.3 with CsOH, osmolarity 290-300 mOsm) for recording inhibitory postsynaptic currents (IPSCs). Neurons were voltage-clamped at 68 mV (for EPSCs) or 60 mV (for IPSCs) using an Axopatch 1D amplifier (Axon Instruments, Foster City, CA). Series resistance was 12-23 M, and cells were rejected if they changed >10% throughout the recording session. All EPSCs (spontaneous, miniature, and evoked) were recorded with the GABA_A receptor antagonist, picrotoxin (50 µM, Sigma) in the bath solution and all IPSCs (spontaneous, miniature, and evoked) were recorded in the presence of the AMPA and N-methyl-D-aspartate (NMDA) receptor antagonists, 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, 10 µM; Tocris Cookson, Ballwin, MO) and D-2-amino-5-phosphonopentanoic acid (AP5, 50µM; Tocris Cookson). TTX (1 µM; Sigma) was added to the bath solution for recording miniature EPSCs and IPSCs.

View larger version (28K): [in this window] [in a new window]   Fig. 1. Whole cell recordings from a pyramidal neuron in heterotopic cortex. A: in the left panel, heterotopic gray matter (arrow) is seen in a cortical slice from an irradiated rat using a 2.5× objective. In the right panel, a pyramidal neuron (arrow) within the heterotopic cortex is seen using infrared differential interference contrast (IR-DIC) microscopy and a 40× objective. B: responses from a typical pyramidal neuron in heterotopic cortex (left) and from a normal layer II/III pyramidal neuron (right) to intracellular current pulses. Both neurons show a regular spiking pattern in response to a suprathreshold current pulse. C: representative recording of spontaneous inhibitory postsynaptic currents (sIPSCs) in a heterotopic neuron (top) that were completely blocked by picrotoxin (PTX, 50 µM), an antagonist of the GABAA receptor (bottom). D: representative recording of spontaneous excitatory postsynaptic currents (sEPSCs) in a heterotopic pyramidal neuron (top) that were blocked by CNQX (10 µM), an antagonist of the AMPA receptor (bottom).

To evoke monosynaptic EPSCs or IPSCs, a glass electrode (3-5 M) filled with ACSF was placed inside of the heterotopia or in layer II/III of control neocortex 50-100 µm away from the cell bodies (in the direction of the apical dendrite, this would be toward the pial surface for control neocortical pyramidal cells) that were being recorded. Five-pulse trains at 10 and 20 Hz were used to elicit IPSCs and EPSCs, respectively. We chose 20-Hz stimulation for EPSCs because Varela et al. (1999) showed that this frequency produces robust short-term plasticity (STP) in cortical neurons. We used 10-Hz stimulation for IPSCs because STP effects were still robust at this frequency, but the longer duration of evoked IPSCs led to subsequent stimuli occurring on the tail of the preceding IPSC at higher frequencies. We chose five pulses based on the findings of Varela et al. (1999) that multiple stimuli (5-10) may give a better indication of steady-state performance than paired-pulse stimuli. The interval between trains was 10 s. Thirty to 40 trains at each frequency were given to each cell and the responses were averaged. Monosynaptic currents were identified by their constant latency and a single peak for all five responses. Stimulation strength was adjusted to induce the first response with the amplitude between 50 and 100 pA for EPSCs and 100 and 200 pA for IPSCs.

Data acquisition and analysis

Data were acquired using pClamp 8 software. The recordings were started 10 min after accessing the cell to allow for stabilization of spontaneous synaptic activity. The recordings were analyzed only when there was no significant change in the frequency or the amplitude of spontaneous responses or in the series resistance (change <10%) during the 5-min recording. Analysis of spontaneous and miniature currents (sEPSC, sIPSC, mEPSC, and mIPSC) were based on 5-min continuous recordings from each cell. Events were detected using the Mini Analysis Program (Synaptosoft, Leonia, NJ) with parameters optimized for each cell and then visually confirmed prior to analysis. The peak amplitude was measured from each event and then averaged: 10-90% rise time, 90-37% decay time, and width at half-maximal amplitude (half-width) were measured based on the average of all events aligned by rise phase. Overlapping events were excluded from analysis. Action potential (AP) duration in the voltage trace was measured from its starting point to where it repolarized to baseline (Xiang et al. 1998). All results are reported as mean ± SE. The unpaired Student's two-tailed t-test was used to compare group results unless otherwise indicated. Statistical significance was defined as P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subcortical, heterotopic gray matter in slices from irradiated rats was first identified with a 2.5× objective (Fig. 1A). Heterotopic cortex was defined as a mass of gray matter that resided deep within the subcortical white matter. It was clearly separated from the overlying normotopic, dysplastic cortex by a layer of white matter. It routinely occurred deep to the dorsomedial frontoparietal cortex. Individual pyramidal neurons within the heterotopic gray matter were visualized using a 40× objective (Fig. 1A). Pyramidal neurons in heterotopic cortex were initially selected by their triangular cell body and single apical dendrite and then further identified by their regular spiking pattern in response to a depolarizing current pulse (Fig. 1B). Initial characterization of the cells was carried in current-clamp mode and a series of current pulses were applied (from 5 to 15 pA, with increments of 5 pA). The average AP duration in heterotopic pyramidal neurons was 6.2 ± 0.3 ms (n = 20), and the average amplitude of spikes was 87.5 ± 5.1 mV (n = 30, Fig. 1B). These properties were not significantly different from layer II/III pyramidal cells of control neocortex (AP duration = 6.0 ± 0.2 ms, n = 20; AP amplitude = 88.5 ± 4.6 mV, n = 30, P > 0.05). The duration of AP is consistent with that of pyramidal cells reported in rat visual cortex (Xiang et al. 1998). The average resting membrane potential of heterotopic pyramidal neurons was 68 ± 2 mV (n = 30), and this was not significantly different from layer II/III pyramidal neurons in control animals (70 ± 3 mV, n = 30, P > 0.05). Average input resistance of heterotopic pyramidal neurons (194 ± 23 M, n = 30) was not different from controls (190 ± 18 M, n = 30, P > 0.05).

Frequencies of sIPSCs and mIPSCs are decreased in heterotopia

Whole cell voltage-clamp recordings were made from pyramidal neurons in heterotopic gray matter and layer II/III pyramidal cells of control neocortex. sIPSCs were recorded at a holding potential of 60mV in the presence of CNQX (10 µM) and AP-5 (50 µM) using a CsCl-based internal solution. Under these conditions, sIPSCs were recorded as inward currents and were blocked completely by picrotoxin (50 µM, Fig. 1C). The presence of sIPSCs in heterotopic pyramidal cells indicated that the GABAergic network still exists and is functional in heterotopic gray matter. Figure 2A shows sIPSCs from individual heterotopic and control pyramidal neurons. The frequency of sIPSCs in heterotopic pyramidal cells was reduced significantly compared with control pyramidal cells. As shown in Fig. 3A, the average frequency of sIPSCs from all heterotopic pyramidal neurons was 4.1 ± 0.4 Hz (n = 18) compared with control pyramidal neurons (7.4 ± 0.6 Hz, n = 18, P < 0.05). This is similar to the reduction in the frequency of sIPSCs that was reported in pyramidal neurons from normotopic, dysplastic cortex of irradiated rats (Zhu and Roper 2000).

View larger version (24K): [in this window] [in a new window]   Fig. 2. Characterization of spontaneous IPSCs recorded over a 5-min period from representative pyramidal neurons from control and heterotopic cortex showing decreased frequency of sIPSCs in the heterotopic neuron. A: representative recordings of sIPSCs from pyramidal neurons in control (left) and heterotopic cortex (right). B: waveforms of sIPSCs from control (left) and heterotopic (right) cortex averaged from all sIPSCs recorded in each cell over a 5-min period. C: histogram showing number of sIPSCs over time during a 5-min recording from control (left) and heterotopic (right) neurons showing a consistently decreased frequency of sIPSCs over time in the heterotopic pyramidal neuron (bin size = 10 s). D: cumulative probability curve showing an increased inter-event interval (decreased frequency) in sIPSCs from the representative heterotopic neuron compared with the representative control pyramidal neuron (Kolmogorov-Smirnov test, P < 0.05). E: cumulative probability curve showing no difference in amplitude of sIPSCs between the heterotopic and control neurons (Kolmogorov-Smirnov test, P > 0.05).

View larger version (30K): [in this window] [in a new window]   Fig. 3. Averaged data of sIPSCs from all pyramidal neurons in control and heterotopic cortex showing a decreased frequency of sIPSCs in heterotopic cortex. A: average frequency of sIPSCs (right) was reduced in heterotopic cortex. B: average amplitude of sIPSCs (right) was not different between the 2 groups. In A and B, mean values from individual neurons are shown in left panels to show distribution of values across cells. C: comparison of average 10-90% rise time, 90-37% decay time, and half-width shows no difference in kinetics of sIPSCs between the 2 groups. *P < 0.05.

Although the frequency of sIPSCs in heterotopic neurons was reduced, sIPSC amplitude was not significantly different from that in control. As shown in Fig. 3B, the average amplitude of sIPSCs from heterotopic pyramidal neurons was 32.9 ± 2.7 pA (n = 18), and in control pyramidal neurons it was 34.9 ± 2.2 pA (n = 18, P > 0.05). We also compared the kinetic properties of sIPSCs including 10-90% rise time, 90-37% decay time, and half-width between heterotopic and control neurons. The average rise time, decay time, and half-width of sIPSCs from heterotopic neurons were 2.4 ± 0.2, 12.2 ± 0.6, and 13.5 ± 0.8 ms (n = 18), respectively. They were not significantly different from control values (2.5 ± 0.2, 12.5 ± 0.4, and 14.0 ± 0.5 ms, respectively, n = 18, Fig. 3C, P > 0.05).

We recorded mIPSCs by adding TTX (1µM) to the bath solution (Fig. 4). Similar to sIPSCs, the frequency of mIPSCs was significantly reduced in heterotopic pyramidal neurons. As shown in Fig. 5A, the average frequency of mIPSCs in cells from heterotopic gray matter was 2.2 ± 0.4 Hz (n = 15), significantly smaller than control values (3.3 ± 0.3 Hz, n = 16, P < 0.05). As for sIPSCs, the average amplitude of mIPSCs showed no difference between heterotopic and control neurons. It was 31.3 ± 1.9 pA in heterotopic cells (n = 15) and 33.2 ± 1.5 pA in control neurons (n = 16, Fig. 5B). The three parameters of kinetic properties also showed no difference between heterotopic and control cortex. The rise time, decay time, and half-width were 2.2 ± 0.2, 9.8 ± 0.5, and 11.4 ± 0.4 ms, respectively, in heterotopic neurons (n = 15) and 2.4 ± 0.2, 10.2 ± 0.6, and 12.1 ± 0.5 ms, respectively, in control neurons (n = 16, Fig. 5C).

View larger version (26K): [in this window] [in a new window]   Fig. 4. Characterization of miniature IPSCs (mIPSCs) recorded over a 5-min period from representative pyramidal neurons from control and heterotopic cortex showing decreased frequency of mIPSCs in the heterotopic neuron. A: representative recordings of mIPSCs from pyramidal neurons in control (left) and heterotopic cortex (right). B: waveforms of mIPSCs from control (left) and heterotopic (right) cortex averaged from all sIPSCs recorded in each cell over a 5-min period. C: histogram showing number of mIPSCs over time during a 5-min recording from control (left) and heterotopic (right) neurons showing a consistently decreased frequency of mIPSCs in the heterotopic pyramidal neuron (bin size = 10 s). D: cumulative probability curves showing an increased inter-event interval (decreased frequency) in mIPSCs from the representative heterotopic pyramidal neurons compared with the representative control neuron (Kolmogorov-Smirnov test, P < 0.05). E: cumulative probability curves show no difference in amplitude of mIPSCs between representative heterotopic and control neurons (Kolmogorov-Smirnov test, P > 0.05).

View larger version (26K): [in this window] [in a new window]   Fig. 5. Averaged data of mIPSCs from all pyramidal neurons in control and heterotopic cortex showing a decreased frequency of mIPSCs in heterotopic cortex. A: average frequency of mIPSCs (right) was reduced in heterotopic cortex. B: average amplitude of mIPSCs (right) was not different between the 2 groups. In A and B, mean values from individual neurons are shown in left panels to show distribution of values across cells. C: comparison of averaged 10-90% rise time, 90-37% decay time, and half-width shows no difference in kinetics of mIPSCs between the 2 groups. *P < 0.05.

Short-term plasticity of evoked IPSCs in heterotopia

We have shown in previous experiments that the GABAergic circuitry was still functional in heterotopic cortex, but the frequencies of sIPSCs and mIPSCs were significantly reduced. One important aspect of neuronal function is synaptic plasticity, the change of synaptic strength to in response to repetitive activation. STP of inhibitory postsynaptic potentials (IPSPs) and IPSCs have been reported in both neocortex and hippocampus. The most common property of STP in IPSCs is depression, a decrease in amplitude with subsequent stimuli (Hefft et al. 2001, Jensen et al. 1999, Thomson et al. 1996, Varela et al. 1999, Xiang et al. 2002). To explore STP of evoked IPSCs, we recorded evoked IPSCs by placing a stimulating electrode close to the recorded cell in control and heterotopic cortex and delivered trains of five pulses at 10 Hz. STP was evaluated by normalizing the amplitude of all subsequent responses to the first response. As shown in Fig. 6, the evoked IPSCs that followed the first one were all depressed in both control and heterotopic cortex. In control neocortex (n = 7), the amplitudes of the second to the fifth stimulus (relative to the first response) were 73 ± 7%, 61 ± 6%, 54 ± 5%, and 52 ± 4%, respectively. Short-term depression was not different in heterotopic cortex where the percentage values for the second to the fifth pulse were 70 ± 4%, 66 ± 5%, 60 ± 7%, and 51 ± 6%, respectively (n = 15, P > 0.05). We were not able to analyze STP in IPSCs from some control neurons because high-frequency sIPSCs contaminated evoked IPSCs, making it impossible separate the two currents. Therefore STP was analyzed in neurons from control neocortex that had sIPSC frequency <10 Hz.

View larger version (15K): [in this window] [in a new window]   Fig. 6. Characterization of short-term plasticity of evoked IPSCs in pyramidal neurons from control and heterotopic cortex. A: representative traces (an average of 10 successive responses at 0.1 Hz) from a representative control (top) and heterotopic (bottom) pyramidal neuron. Each train contains 5 pulses at 10 Hz. B: averaged data from all recorded cells shows short-term depression of evoked IPSCs in both control () and heterotopic () pyramidal neurons. The amplitudes of the 2nd to 5th responses were normalized to that of the 1st response.

No change in spontaneous and miniature EPSCs in heterotopia

K-gluconate-based internal solution was used to record sEPSCs in the presence of picrotoxin (50 µM), an antagonist of the GABA_A receptor. As shown in Fig. 1D, sEPSCs were abolished with the addition of the AMPA receptor antagonist, CNQX (10 µM). TTX (1 µM) was added to the bath solution for recording mEPSCs. The average frequency and amplitude of sEPSCs were not different between heterotopic neurons, 2.1 ± 0.2 Hz and 12.9 ± 0.3 pA (n = 19), and control neurons, 1.9 ± 0.2 Hz and 13.6 ± 1.3 pA (n = 18, P > 0.05). The frequency and the amplitude of mEPSCs were 1.9 ± 0.4 Hz and 11.9 ± 0.2 pA in heterotopic neurons (n = 15) and 1.4 ± 0.2 Hz and 12.4 ± 0.5 pA in control neurons (n = 15). Again, there was no significant difference in these values between heterotopic and control neurons (P > 0.05). There was no difference in sEPSC kinetics (rise time, decay time, and half-width) between heterotopic (2.9 ± 0.1, 7.1 ± 0.5, and 9.1 ± 0.4 ms, respectively; n = 19) and control (2.7 ± 0.1, 6.1 ± 0.3, and 8.5 ± 0.2 ms, respectively; n = 18, P > 0.05) pyramidal neurons. There was no difference in mEPSC kinetics (rise time, decay time, and half-width) between heterotopic (2.6 ± 0.2, 6.1 ± 0.3, and 9.0 ± 0.3 ms, respectively; n = 15) and control (2.3 ± 0.1, 5.6 ± 0.2, and 8.3 ± 0.3 ms, respectively; n = 15, P > 0.05) pyramidal neurons.

Short-term plasticity of evoked EPSCs is altered in heterotopic cortex

We explored STP of evoked EPSCs in heterotopic and control pyramidal neurons by using a protocol similar to the one described for evoked IPSCs, except that the frequency of pulses in each train was 20 Hz. As shown in Fig. 7, evoked EPSCs of control pyramidal neurons showed facilitation at 20 Hz. Facilitation occurred at all pulses in the train (i.e., the 2nd to the 5th pulse), but the second pulse showed the most pronounced increase. The percentage increase of the amplitudes for the second pulse to fifth pulse (compared with the 1st response) was 135 ± 11%, 128 ± 9%, 125 ± 5%, and 119 ± 6%, respectively (n = 17). However, in heterotopic neurons, the average second response was not significantly different from the first, and the third through fifth responses showed significant depression. The percentage of the amplitudes for the second pulse to fifth pulse to that of the first response was 104 ± 5%, 95 ± 8%, 78 ± 7%, and 72 ± 9% (n = 17). These results indicate that STP of evoked EPSCs is altered in pyramidal neurons from heterotopic gray matter. This is similar to findings in STP of evoked EPSCs in normotopic, dysplastic cortex from irradiated rats (Chen and Roper 2001).

View larger version (16K): [in this window] [in a new window]   Fig. 7. Short-term plasticity of evoked EPSCs is altered in heterotopic cortex. A: representative traces of evoked EPSCs of pyramidal neurons from control (top) and heterotopic (bottom) cortex. B: averaged data from all recorded cells show that control pyramidal neurons demonstrate facilitation (). However, heterotopic neurons () show depression in response to the 3rd, 4th, and 5th stimuli of the train.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This is the first report of reduced inhibition in heterotopic gray matter in the in utero irradiation model of cortical dysgenesis. These changes are similar to those that were previously reported in normotopic dysplastic cortex in this model (Zhu and Roper 2000). The specific reductions in sIPSC and mIPSC frequency without changes in amplitude or current kinetics suggest a presynaptic locus for this impairment. Previous studies in this model have shown that heterotopic gray matter is composed of pyramidal and nonpyramidal neurons that lack any laminar organization or consistent spatial orientation (Smith et al. 1999). They also showed that inhibition was present but made no quantitative comparisons of measures of inhibition between heterotopic and control cortex.

The current findings are similar to previous findings in normotopic dysplastic cortex from this model (Zhu and Roper 2000). In that study, mIPSC frequency but not amplitude was reduced in pyramidal neurons from dysplastic cortex when compared with pyramidal neurons from control neocortex. This indicated a presynaptic locus for the impairment in inhibition and was supported by an earlier immunohistochemical study that showed a selective reduction in the density of parvalbumin- and calbindin-immunoreactive (putative inhibitory) neurons in normotopic dysplastic cortex from irradiated rats (Roper et al. 1999). Although parvalbumin- and calbindin-immunoreactive neurons are present in heterotopic gray matter in this model (Roper et al. 1999), their numbers have not been quantified. The current study found a lower frequency for sIPSCs and mIPSCs in control pyramidal neurons compared with the previous study by Zhu and Roper (2000). This may be an age-related effect. The current experiments were carried out on 20- to 28-day-old rats and the previous study used 28- to 35-day-old rats. The frequency of sIPSPs and mIPSCs in cortical pyramidal neurons increases throughout postnatal development (Dunning et al. 1999; Luhmann and Prince 1991). Therefore the earlier age of the rats in this study may explain the reduced frequency of sIPSCs and mIPSCs compared with our previous study.

The majority of cortical interneurons are generated in the ganglionic eminence and have a long, tangential migratory pathway to arrive in the cortical plate (Anderson et al. 1997; Tamamiki et al. 1997). Some interneurons migrate down into the ventricular zone before migrating into the cortical plate during normal cortical development (Nadarajah et al. 2002). Therefore it is not surprising that some inhibitory interneurons would make their way into periventricular heterotopic gray matter. However, it was unknown how interneurons would be affected in the heterotopic gray matter of irradiated rats. Previous studies showed the presence of interneurons in heterotopic gray matter of irradiated rats (Roper et al. 1999) and spontaneous IPSCs in heterotopic pyramidal neurons (Smith et al. 1999). The current results confirm that interneurons are able to form functional connections with the pyramidal neurons in heterotopic cortex; however, there is a relative impairment in their function. The current study cannot tell if this is due to a loss of interneurons and their presynaptic terminals, as appears to be the case for the normotopic dysplastic cortex, or if there is some other alteration of presynaptic release mechanisms that might result in a decreased mIPSC frequency. We found short-term depression of evoked IPSCs in both heterotopic and control pyramidal neurons. This is similar to previous reports of in control neocortex (Thomson et al. 1996) and suggests a high initial release probability in the GABAergic presynaptic terminals in both heterotopic and control neocortex. Therefore at the current level of analysis, changes in initial release probability would not be a good explanation for the reduction in spontaneous and miniature IPSCs that we have found in heterotopic neurons. However, additional studies are needed to address this question.

This study found no differences in frequency, amplitude, or kinetics of sEPSCs or mEPSCs in heterotopic and control pyramidal neurons. This is similar to previous findings in normotopic dysplastic cortex in this model with the exception that amplitude and frequency of sEPSCs (but not mEPSCs) were increased in dysplastic pyramidal cells in that study (Zhu and Roper 2000). These findings suggest that the most pronounced alterations in cortical circuitry reside within the GABAergic system. This does not rule out the possibility of significant changes in excitatory neurotransmission that might not be detectable with the methods used in this study. For instance, if a subset of excitatory synapses were altered, a modest but important effect might be missed when analyzing all spontaneous excitatory currents as a group.

Short-term synaptic plasticity of EPSCs was altered in heterotopic neurons. Excitatory synapses in control pyramidal neurons showed facilitation while those in heterotopic neurons showed depression. Our finding of facilitation in control, layer II/III pyramidal neurons from 21- to 28-day-old rats supports studies that showed a switch from depression to facilitation of excitatory postsynaptic potentials (EPSPs) between 2 and 4 wk of age (Reyes and Sackmann 1999). This suggests that excitatory synapses in heterotopic neurons may preserve a more immature form of STP. Whether this represents a delay or a permanent alteration will require additional testing at older ages.

There are several factors that affect STP. The initial release probability at the presynaptic terminal is an important determinant of STP. In the hippocampus, excitatory synapses with a high initial release probability show depression while those with a lower initial release probability show facilitation and this correlates with the size of the readily releasable pool of synaptic vesicles in the presynaptic terminal (Dobrunz and Stevens 1997). This would suggest that excitatory synapses in heterotopic cortex have a higher initial release probability than those in control neocortex. A wide variety of neurotransmitters and neuromodulators acting through presynaptic receptors can also influence STP in neocortical neurons (Wu and Saggau 1997) and alterations in these systems could be contributing to our findings. It is also important to note that STP in neocortical neurons can vary depending on the source of the presynaptic input (Gil et al. 1997). Since specific presynaptic sources were not identified with our stimulation protocols, our results may represent an average of these responses. Postsynaptic factors may also affect STP. In neocortical multipolar cells, use-dependent relief of calcium-permeable AMPA receptor blockade by polyamines enhances facilitation of EPSPs and counteracts presynaptic factors that promote depression (Rozov and Bernashev 1999). However, this effect was not seen in neocortical pyramidal neurons since they lack the calcium-permeable AMPA receptor. The current results show a clear abnormality in STP of excitatory synapses in heterotopic cortex but the underlying mechanisms and implications of these changes for cortical function require further study.

In this study, pyramidal cells in heterotopic gray matter were compared with layer II/III pyramidal neurons in control neocortex. The rationale for this comparison lies in the timing of the radiation-induced injury. Most layer V pyramidal neurons are generated on E16 and E17 and have arrived in the cortical plate by the time of the injury. Layer II/III neurons are generated on E18 and E19, after the radiation injury (Ferrer et al. 1993). Since radiation damages radial glia fibers in this model (Roper et al. 1997a), the pyramidal neurons that are "trapped" in their periventicular location and go to form the heterotopic gray matter are most likely those that were destined to become supragranular neurons. Results from Jensen and Killackey (1984) also support this concept. When rats are irradiated on E15 or E16 (using the nomenclature adopted in the current study), neurons with corticospinal projections (characteristic of layer V pyramidal cells) are found in heterotopic gray matter. But when irradiated on E17 or E18, the corticospinal neurons are found only in the normotopic cortex, not in heterotopic cortex. This suggests that the deeper layered neurons arrive in the cortical plate before the injury on E17 and that the remaining pyramidal cells in the heterotopia were those destined to be layer II/III neurons.

Although there is a strong clinical association between neuronal heterotopia and epilepsy, the exact role of heterotopic cortex in seizure generation is still unknown. Irradiated rats demonstrate spontaneous electrographic seizures in vivo (Kondo et al. 2001), but the site of seizure onset is unknown in these animals. Possibilities include normotopic dysplastic cortex, heterotopic cortex, and the hippocampus. Several studies in human heterotopic cortex suggest that the misplaced neurons may have the capacity to generate seizures. Epileptiform activity has been recorded from heterotopic cortex in patients with surgically implanted depth electrodes (Dubeau et al. 1995). Positron emission tomography and single photon emission computed tomography studies have also shown heterotopic cortex to be metabolically active (De Volder et al. 1994; Matsuda et al. 1995). However, one study in human heterotopia suggested a limited number of nerve fibers projecting outside of the heterotopic gray matter (Hannan et al. 1999). Interestingly, this study also reported that calretinin-immunoreactive interneurons were morphologically less complex with less extensive processes in heterotopic cortex compared with those in normotopic cortex from the same specimen. If these alterations were associated with a reduction in presynaptic terminals arising from these interneurons, it would produce physiological results similar to those reported in the current study.

Although this study is the first to demonstrate reduced frequency of IPSCs in heterotopic cortex, a number of animal models have shown interesting abnormalities in displaced neurons. Exposure of fetal rats to the cytotoxic agent, methylazoxymethanol (MAM), produces cortical dysplasia and heterotopic neurons in the hippocampus and adjacent to the lateral ventricles (Colacitti et al. 1998). Heterotopic hippocampal neurons have properties similar to layer II/III cortical neurons (Chevassus-Au-Louis et al. 1998b), and these cells form an abnormal functional bridge between the adjacent hippocampal and neocortical circuits (Chevassus-Au-Louis et al. 1998a). Neurons from areas of periventricular and hippocampal heterotopia also show an abnormal propensity for intrinsic bursting (Baraban and Schwartzkroin 1995; Sancini et al. 1998). Heterotopic pyramidal neurons in the hippocampus of MAM-treated rats lack functional A-type Kv4.2 potassium channels and this may promote hyperexcitability in these neurons (Castro et al. 2001). As in irradiated rats, studies in MAM-treated rats have documented long-range connections between heterotopic gray matter and normal cortical and subcortical targets (Colacitti et al. 1998; Yurkewicz et al. 1984). Tish rat is a spontaneous mutant that has unprovoked epileptic seizures and a large, tubular mass of heterotopic cortex that lies in the white matter beneath the frontoparietal cortex of each cerebral hemisphere (Lee et al. 1997). Heterotopic neurons in this model also show reciprocal connections with cortical and subcortical brain structures (Schottler et al. 1998). Metabolic studies show that the heterotopic cortex is active during in vivo seizures but in vitro slice studies suggest that epileptiform activity may actually originate in the overlying, normotopic cortex and spread to the heterotopic cortex through synaptic transmission (Chen et al. 2000).

This study has extended our findings of impaired inhibition in animal model dysplastic and heterotopic cortex. Coupled with our previous findings from normotopic dysplastic cortex (Zhu and Roper 2000), it demonstrates that a single in utero injury can have profound and long-lasting effects on cortical function. It also confirms that the GABAergic system appears to have a particular susceptibility to this deleterious effect. We still do not know if this is a direct effect from the original, radiation-induced injury or if it is a downstream effect on subsequent cortical development; but the disruption of inhibitory function appears to be pervasive in both normotopic and heterotopic cortex in this model. This may have important implications for a number of abnormalities of brain function, including epilepsy, that are seen in humans with disorders of cortical development.