In Vitro Models Of Corticocortical Circuitry In Epilepsy

FIGURE 4 Preparation of rat thalamocortical slices. A: The brain is placed on a ramp with the caudal end elevated above the rostral end at an angle of 10 degrees. Th, approximate location of the thalamus. B: Dorsal view showing angle of initial vertical cut made with a razor blade. C: The cut surface of the brain is then glued to the vibratome stage. Horizontal lines indicate the approximate position of collected slices. Typically two or three slices (550-650 |mm in thickness) retain full thalamocortical connectivity. D: Typical positions of recording (left, in cortical barrel field) and stimulating (right, in the external part of the ventrobasal (VB) near the border with the reticular thalamic nucleus (RTN) electrodes in a schematic thalamocortical slice. Cx, cortex; Hip, hippocampus; RTN, reticular thala-mic nucleus; VB, ventrobasal thalamus. (After Agmon and Connors, 1991; and Tancredi et al., 2000.)

FIGURE 4 Preparation of rat thalamocortical slices. A: The brain is placed on a ramp with the caudal end elevated above the rostral end at an angle of 10 degrees. Th, approximate location of the thalamus. B: Dorsal view showing angle of initial vertical cut made with a razor blade. C: The cut surface of the brain is then glued to the vibratome stage. Horizontal lines indicate the approximate position of collected slices. Typically two or three slices (550-650 |mm in thickness) retain full thalamocortical connectivity. D: Typical positions of recording (left, in cortical barrel field) and stimulating (right, in the external part of the ventrobasal (VB) near the border with the reticular thalamic nucleus (RTN) electrodes in a schematic thalamocortical slice. Cx, cortex; Hip, hippocampus; RTN, reticular thala-mic nucleus; VB, ventrobasal thalamus. (After Agmon and Connors, 1991; and Tancredi et al., 2000.)

Whereas thalamocortical involvement is implicated in most forms of typical absence seizures, in some pharmacologic feline epilepsy models, SWDs have been observed in isolated cortical circuits (Marcus and Watson, 1966; Steri-ade and Contreras, 1998), suggesting that the cortical circuitry itself is capable of generating bilaterally synchronous 3- to 5-Hz electrographic activity. As a preliminary step in testing the involvement of long-range intracortical connections in widespread neocortical synchronization, Kumar and Huguenard (2001) refined a callosal slice preparation based on earlier work of Vogt and Gorman (1982).

What Does the Corticocortical Preparation Model?

The mammalian neocortex is characterized by extensive recurrent excitatory circuitry. Roughly 85% of synapses are excitatory, and an almost equal percentage of synapses made by excitatory neurons are onto other excitatory neurons (Braitenberg and Schuz, 1991). Such neural networks with prominent positive feedback are prone to become hyperex-citable and seizure generating following disruption of inhibition. The study of epilepsy necessitates model systems in which excitatory corticocortical synapses can be studied in isolation so that response properties of individual neurons and their contribution to the overall excitability of the network can be characterized. Not only are such model systems useful in assessing pathophysiology, but they are also essential in characterizing normal physiologic properties of excitatory-to-excitatory synapses and in elucidating their receptor compositions. This background is essential for understanding the changes in synaptic connectivity during epileptogenesis, changes that are difficult to detect given the complexity of the underlying circuitry. Attempts to isolate and study synaptic excitatory responses in brain tissue have been frustrating in many models of experimental epilepsy because attempts to block inhibition invariably lead to conditions of "runaway" excitation or hyperexcitability. By contrast, studies of inhibition are more readily accomplished because excitatory receptors can be readily blocked (using glutamatergic receptor antagonists), yielding "residual" synaptic responses that are (for the most part) inhibitory responses mediated by GABA receptors. In the following sections we provide an overview of the development of a callosal slice model that lends itself quite well to the study of long-range cortical excitatory connectivity.

Importance of the Corpus Callosum in a Model for the Study of Intracortical Excitability

The corpus callosum is the principal commissural pathway in the forebrain linking the two cerebral hemispheres. The cells of origin of neocortical callosal projections are almost entirely pyramidal neurons, located mainly in layers III and V; the projection terminates exclusively with excitatory asymmetric synapses on spines of pyramidal neurons in homotopic and heterotopic regions of the contralateral cortex (Akers and Killackey, 1978; Jacobson, 1965; Jacobson and Trojanowski, 1974; Seggie and Berry, 1972; Wise and Jones, 1976). Callosal projections can be reliably stimulated (Vogt and Gorman, 1982); because they are purely excitatory, activation of this pathways aids greatly in the study of excitatory-to-excitatory cortical synapses (Aram and Lodge, 1988) and intracortical excitation. Furthermore, the callosum is itself considered a primary substrate for intrahemispheric spread of discharges in generalized epileptic seizures (Gazzaniga et al., 1975; Reeves and O'Leary, 1985; Wilson et al., 1975). Clinical studies have shown that cutting of this pathway (calloso-tomy) can eliminate seizure activity or decrease its severity and frequency, suggesting that transcortical excitation can initiate or perpetuate epileptiform activity. Excitatory cor-ticocortical projections may play a crucial role in determining the strength and extent of seizure generalization, especially during critical periods in early maturation when neocortical tissue is vulnerable to epileptiform activity (Luhmann and Prince, 1990; Moshe et al., 1983; Swann et al., 1993). However, despite the long history of callosal studies in epilepsy, and a wealth of anatomic information on callosal organization, a neocortical preparation involving hemispheric connection via the corpus callosum has only recently been recognized as a viable model for studying intracortical excitation. This delay is due in part to limitations in our knowledge of the functional properties and receptor composition of the synapses made by callosal fibers and the mechanisms that control their efficacy. To overcome this deficiency, Kumar and colleagues (2001) undertook a study to assay the usefulness of callosal activation in a model system for studying intracortical excitability by (1) investigating the receptor composition of the callosal synapse; (2) characterizing the voltage dependence of the pharmacologically isolated components of excitatory post-synaptic currents (EPSCs) evoked by minimal stimulation of the callosum; (3) determining whether kinetic properties of spontaneous and evoked EPSCs recorded in these neurons can reveal differences in function of the underlying receptor populations at callosal and noncallosal synapses; and (4) identifying changes in the physiologic properties of these receptors as a function of early postnatal age (Kumar and Huguenard, 2001).

Kumar and colleagues (2001) found that midline stimulation of the corpus callosum in callosally connected acute brain sections of the rat neocortex (Figure 5) could evoke inward EPSCs in whole-cell voltage-clamped (-70mV) pyramidal neurons in layer V of the agranular frontal cortex. In the presence of the GABAa receptor antagonist picrotoxin (to isolate the EPSCs) and a low concentration of 1,2,3,4-tetrahydro-6-nitro-2,3-dioxo-benzo [f] quinoxaline-7-sulfonamid (NBQX) (0.1 ||M) (to block partially a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors and thereby prevent epileptiform activity), purely glutamatergic EPSCs were obtained. These isolated EPSCs had fixed latencies from stimulus onset, could follow stimulus trains (1-20Hz) without changes in kinetic properties, and were completely antagonized with high concentrations of NBQX (10 ||M), indicating that they were mediated by AMPA receptors. Depolarization of the recorded neurons revealed a late, slowly decaying component that reversed at ~0mV and that could be blocked by D-2 amino-5-phospho-novaeric acid (D-APV). These studies showed clearly that callosal synapses involve both AMPA and NMDA receptors.

Analysis of the callosal EPSCs revealed age-dependent changes in the composition of the underlying AMPA receptors. In particular, AMPA receptor-mediated responses in immature cortex (tissue from younger than 15 day old rats) were characterized by inward rectification—the hallmark of GluR2-deficient AMPA receptors. Direct measurement of Ca2+ permeability, another characteristic of GluR2 deficiency, confirmed this finding (Kumar et al., 2002). An interesting finding was that the functional properties of AMPA receptors were similar at callosal and noncallosal excitatory connections. By contrast, NMDA receptor-mediated

FIGURE 5 The callosal slice preparation and schematic. A1-A5: low-powered images of a coronal cortical section, obtained with a 10x objective, depicting placement of a bipolar stimulating electrode (S in A1) at different locations (•) along the dorsal—ventral extent of the callosum. The corresponding whole-cell responses of a layer V pyramidal neuron (R in A1), evoked by stimulation of the callosum at these locations, are shown as insets. Traces are averaged responses evoked at a holding potential of -70 mV; stimulus intensity was constant throughout. Patch pipette in A1 is visible during whole-cell recording from pyramidal neuron on the right (R, scale bar is 40 ||m). Note that responses were especially robust for one near-midline stimulus location (* in A2). Postsynaptic currents (PSCs) with longer latency and similar kinetic properties could also be evoked from the contralateral hemisphere (A5). Typically the recording region was closer to midline, in the frontal-cortex areas 1 and 2 (Fr1 and Fr2, A1). B: schematic of the slice preparation, showing typical placement of the stimulating (S) and recording (R) electrodes (S1 and S2 for callosal and local intracortical stimulation, respectively) on the slice. Anatomic structures that can be easily recognized are labeled: CC, corpus callosum; CA1 and CA3 regions of the hippocampus; neocortical laminae indicated by Roman numerals. (Reprinted with permission from the American Physiological Society, 2001.)

response properties were pathway-specific, suggesting differences in subunit composition of the underlying receptors (Kumar and Huguenard, 2003).

Corticocortical Slice Preparation General Considerations

The primary consideration in making acute callosal slices for studying corticocorticol synapses is to retain and optimize intrahemispheric connectivity. All experiments were carried out in coronal sections of rat (Sprague-Dawley) brains, retaining both hemispheres and the callosal tract. Preserving callosal connectivity in brain slices is complicated owing to the curvature of the corpus callosum. We recommend using straightforward coronal sections (300350 |mm thick) obtained on a vibratome. We have found that during slicing the presentation angle of the sectioning blade is a critical factor for optimal slice health and connectivity; an angle of 18 degrees relative to the plane of the section is optimal. Using coronal slices prepared in this manner, we have been able to evoke responses in neurons located as lateral as the somatosensory cortex via midline stimulation of the callosum. In general, however, we prefer recording closer to the midline—from the agranular frontal cortex— to maximize callosal connectivity (Figure 6). The thickness of the slice is also a critical factor in determining connectivity; the range of slice thickness we recommend is chosen as a compromise so that we can maintain connectivity but still record from visually identified neurons deep in the tissue. Another important parameter that affects connectivity is the age of the experimental animals. We have used animals in three different developmental age groups for studying the ontogeny of callosally evoked responses: neonates (P6-7), in which callosal fibers just approach their final cortical target lamina; juveniles (P12-16), in which neocortical synaptogenesis approaches adult levels of maturity; and young adults (P20-28), in which synaptogenesis is almost complete and myelination has begun to develop. It is worth noting that these demarcations of age groups overlap a period of early postnatal maturation between P11-20, during which there is transient but well-defined temporal manifestation of strong excitatory polysynaptic activity (NMDA receptor-mediated) leading to an enhanced sensitivity for epileptogenesis.

Although the callosal model represents long-range intra-hemispheric connections, layer V pyramidal neurons also receive a diverse set of short-range intracortical excitatory afferents that arise locally from neurons in close vicinity to the cells of interest (Burkhalter and Charles, 1990). These latter fibers can also be activated concomitantly with the cal-losal afferents by using stimulating electrodes placed intra-cortically in close proximity to the recorded neurons (either off- or on-column), to study any pathway-specific differences

FIGURE 6 Orthograde transport of biocytin reveals callosal projection in coronal slices. A: A biocytin crystal (~0.5 mm diameter) was placed in layer V of the contralateral cingulate cortex (right side, crystal placement not shown in this figure). The slice was incubated for 3 hours in an interface chamber and then processed to obtain a final horseradish peroxidase reaction product. The arrows in A indicate labeling of the callosal fibers. The square area marked in A is expanded in B. B: A dark-field image showing the callosal fibers coursing upward through the ipsilateral layer V. (See color insert.)

FIGURE 6 Orthograde transport of biocytin reveals callosal projection in coronal slices. A: A biocytin crystal (~0.5 mm diameter) was placed in layer V of the contralateral cingulate cortex (right side, crystal placement not shown in this figure). The slice was incubated for 3 hours in an interface chamber and then processed to obtain a final horseradish peroxidase reaction product. The arrows in A indicate labeling of the callosal fibers. The square area marked in A is expanded in B. B: A dark-field image showing the callosal fibers coursing upward through the ipsilateral layer V. (See color insert.)

in synaptic properties. Previous work suggests that although functional attributes of AMPA receptors are uniform among synaptic connections onto layer V pyramidal neurons (i.e., callosal and noncallosal excitatory inputs onto these cells are indistinguishable), properties of NMDA receptors at these inputs can be functionally distinct (Kumar and Huguenard, 2003).

How Sections Are Prepared

Standard techniques for preparing and maintaining neo-cortical slices are employed (reviewed in Kumar and Huguenard, 2001). Briefly, Sprague-Dawley rats of the desired age are anesthetized (50mg/kg pentobarbital), decapitated, and the brain rapidly removed and transferred to a chilled (4° C) low Ca2+, low Na+ slicing solution equilibrated with a 95%: 5% mixture of O2 and CO2. The brain is subsequently blocked and coronal slices, 350mm thick, are prepared on a vibratome with the blade set at the appropriate cutting angle. All slices are incubated at 32° C in oxygenated artificial cerebrospinal fluid (ACSF) for 1 hour before being transferred to the recording chamber. Generally, four to six slices are obtained per animal; slices remain viable for up to 6 hours.

Electrophysiologic Techniques, Data Collection, and Analysis

Recordings are obtained at 32 ± 1°C from layer V pyramidal neurons in the agranular frontal cortex (Paxinos and

Watson, 1997). Patch recording electrodes (1.2-2 |mm tip diameters; 3-6MW) typically contained the following electrolyte (in mM): for voltage-clamp experiments, 120 cesium gluconate, 1 MgCl2, 1 CaCl2, 11 KCl, 10 HEPES, 2 NaATP, 0.3 NaGTP, 1 QX-314, and 11 EGTA (pH 7.3 was corrected with Cs-OH, 290 mOsm); for current-clamp experiments, 105 potassium gluconate, 30 KCl, 10 HEPES, 10 phospho-creatine, 4 MgATP, and 0.3GTP (adjusted to pH 7.3 with KOH). Slices are maintained in oxygenated (95% O2-5% CO2) ACSF, and drugs and chemicals are applied via the perfusate (2ml/min) or through a local perfusion system that allows fast exchange of media at the level of the synapse (Kumar et al., 2002). Concentric bipolar electrodes (e.g., CB-XRC75 from Frederick Haer Company, Bowdoinham, ME), with 75-| m-tip diameters, are positioned on the callosal tract or intracortically in close proximity to the recorded neuron; constant current pulses (50-300 |ms in duration and 100-500 ||A in amplitude) are applied at low frequencies (0.1-0.3 Hz). Callosal stimulation activates fibers potentially in both orthodromic and antidromic directions, each of which in turn activates monosynaptic excitatory connections onto the recorded pyramidal neuron (Kumar and Huguenard, 2001). Thus this model consists of activating a well-defined, relatively homogeneous population of intracortical excitatory connections. Stimulation parameters are determined by increasing current strength until post-synaptic responses can be evoked (minimal stimulation); stimulus intensity is held constant at about 1.2 times the threshold for obtaining a detectable response throughout the duration of the experiment (thresholds are characterized by a large proportion of failures). Characterization of synaptic currents mediated by AMPA and NMDA receptors requires the isolation of purely excitatory monosynaptic responses with minimal contribution from polysynaptic events, such as those arising from feed-forward excitation. Toward this end, the minimal stimulation paradigm outlined previously can be used in conjunction with a pharmacologically controlled blockade of inhibition. Callosally evoked responses can be compared with EPSCs evoked during stimulation of local excitatory circuitry in the vicinity of the recorded neuron to determine differences in the biophysical properties of the underlying receptor subtypes and the input specificity of their responses. Evoked responses are recorded with a patch amplifier filtered at 2 to 5 kHz and digitized at 10kHz.

Solutions

The low-Ca2+, low Na+ slicing solution consists of (in mM): 234 sucrose, 11 glucose, 24 NaHCO3, 2.5 KCl, 1.25 NaHaPO4, 10 MgSO4, and 0.5 CaCla. The composition of the ACSF used in the incubation and perfusion (bathing medium) of the slices during recording is as follows: 126 NaCl, 26 NaHCO3, 2.5 KCl, 1.25 Na^PO4, 2 MgCl2, 2 CaCl2, and 10 glucose, pH 7.4. The following are bath-applied as required for specific protocols: APV, 40 ||M; NBQX, 0.05-0.1 ||M; picrotoxin (PTX), 50 |M. Biocytin (0.1%) can be included in the patch solution in experiments in which histologic processing is required.

Using the Corticocortical Model: An Example

Intrinsic excitatory connectivity is the primary mediator for the generation and spread of epileptiform activity within the neocortex. During maturation, there is a critical period during which normal cortical tissue appears to be most epileptogenic (Luhmann and Prince, 1990; Moshe et al., 1983; Swann et al., 1993). This maturational dynamic presents an opportunity to study changes in the functional properties of intrinsic excitatory connections and the cellular mechanisms that lead to epileptiform activity. To this end, we have used the callosal model to determine how the ontogeny of excitatory connections compares with the development of intrinsic neuronal excitability and epilepto-genicity within the frontal cortex. Given that the incidence of seizure generalization is critically dependent on the maturation of corticocortical connections, we used animals from developmentally distinct age groups to assay the ontogeny of synaptic responses in visually identified layer II/III and V pyramidal neurons. We then determined whether changes in the postsynaptic properties of these excitatory connec tions lead to alterations in synaptic efficacy, thereby making the tissue more likely to engage in epileptiform activity. The goals of these studies included the hope that our findings would aid in the identification of candidate cellular mechanisms that produce changes in synaptic efficacy, thereby providing useful information about the role of excitation in the developmental regulation of epileptiform events. We also anticipated that these studies would help address two critical basic neurobiological questions: (1) Do maturational changes in corticocortical synaptic efficacy parallel the transient manifestation of the period of hyperexcitability in the neocortex? (2) Do the functional properties and relative contribution of the different excitatory receptors (AMPA vs. NMDA) change during this critical period?

Using the callosal-connected cortical model, we discovered that synaptic AMPA receptors of excitatory layer 5 pyramidal neurons in the rat neocortex were deficient in GluR2 in early development (before P16), as evidenced by their inwardly rectifying current-voltage relationship, blockade of AMPA receptor-mediated EPSCs by external and internal polyamines, permeability to Ca2+, and GluR2 immunoreactivity. Our results indicated that neocortical pyramidal neurons underwent a developmental switch in the Ca2+ permeability of their AMPA receptors through an alteration of the AMPA subunit composition (Kumar et al., 2002). This finding led us to explore the general question of whether all cortical glutamatergic receptors change their subunit composition during maturation? If that were so, we could begin to understand how these alterations influence the development of intrinsic neuronal excitability and epileptogenicity. Early low expression of GluR2 subunits or the failure to incorporate the GluR2 subunit into AMPA receptors during the developmental critical period might underlie increased seizure susceptibility of the immature brain (Moshe et al., 1983; Schwartzkroin and Prince, 1980; Sensi et al., 1999). Indeed, one study (Sanchez et al., 2001) demonstrated that hypoxia-induced seizures in neonatal rats (P10-12) are linked with maturational and seizure-induced changes in AMPA receptor composition and function, particularly involving the GluR2 subunit (see Chapter 25). These results indicated that seizures induce an increased expression of Ca2+-permeable AMPA receptors and an increased capacity for AMPA receptor-mediated epilepto-genesis. The AMPA receptor switch described in our study occurs approximately midway through the period of maximum synaptogenesis in rats (P11-20) (Sutor and Luhmann, 1995) and results in changes in the functional properties of AMPA receptors that depend on the presence or absence of GluR2. Failure to switch the subunit composition of these receptors to incorporate GluR2 at this juncture might indeed have deleterious consequences for neocortical excitability (Feldmeyer et al., 1999; Pellegrini-Giampietro et al., 1997).

Was this article helpful?

0 0

Post a comment