What Types Of Epileptic Activity Can Be Modelled


The organization, generation, and propagation of seizures in temporal lobe epilepsy (TLE) can be understood via the concept of the epileptic zone (Talairach et al., 1974). TLE is indeed a disease (or perhaps a set of disorders) involving neuronal networks within the temporal lobe. In TLE, seizures engage different structures, not only at their onset (ictogen-esis) but also during their course of propagation. Most past studies of TLE have concentrated on the hippocampus as the key structure in seizure generation and spread, and thus studies of hippocampal slices may provide insight into seizure activity associated with TLE. It should be noted, however, that several reports have proposed that various cortical structures participate in ictogenesis (Bartolomei et al.,

1999; Munari et al., 1994; Spencer and Spencer, 1994; Talairach et al., 1974; Wieser, 1983), with particular attention focused on the critical involvement of the entorhinal cortex (EC) in ictogenesis. The EC has extensive reciprocal connections with the hippocampus and neocortical and paralimbic regions (perirhinal cortex, temporal pole). Neuropathological (Du et al., 1995) and neuroradiologic (Bernasconi et al., 1999, 2000, 2001) data have revealed anomalies within the rhinal cortex of patients with mesial TLE, but the relationship between those observations and electrophysiologic data remains hypothetical. If extra-hippocampal structures are involved in the type of epilepsy under consideration, the isolated hippocampal slice might not be the best preparation for elucidating more complex seizure networks associated with TLE (e.g., see chapter 4 for consideration of the combined hippocampal-entorhinal slice). Nevertheless, hippocampal slices are useful for studying the propagation of seizures. With this preparation, one can examine some of the relationships associated with TLE-like discharges (e.g., the relationships between ictal and interictal activity).

Most of the extrinsic—and many of the intrinsic—connections are severed in the hippocampal slice, so it is hardly surprising that ictal-like discharges do not occur spontaneously in slices obtained from human resections or animal models. The same connectivity issues may be relevant to the paucity of spontaneous interictal-like activities in these preparations, although they can occasionally be recorded in human tissue (Cohen et al., 2002). As a consequence, ictal-or interictal-like discharges must, in most hippocampal slice experiments, be artificially induced. Several different protocols can be used, as described in the section Studying Epilepsy Using the Slice Preparation. One major drawback with artificially induced epileptiform activity in vitro is the absence of behavioral and clinical seizure correlates. To address this obvious problem, experimenters have generally tried to produce electrographic seizure equivalents in vitro. An in vitro equivalent of an in vivo TLE-like seizure should satisfy several conditions. Before its occurrence, interictal activity can (should?) be present; that is, large-amplitude field potential "spikes" (in the clinical electroencephalo-graphic [EEG] sense). Individual spikes last up to 100 ms and may occur in bursts. Just before seizure onset, these spikes may increase in amplitude, become more rhythmic (~1Hz), and display striking synchrony when recorded across different hippocampal subregions and limbic structures. There is then a flattening of the EEG (field potential) signal, and the seizure starts with a high frequency (~20Hz) discharge. The electrographic seizure then typically consists of tonic rhythmic discharge and clonic bursting phases (rhythmic spikes that increase in amplitude and then decrease in frequency). Alternatively TLE-like seizures may start independently of interictal activity; such seizures usually have the same electrographic signature with seizures lasting 30 to 60 seconds. These general descriptions, based on EEG recordings from intact animals (or humans) with TLE-associated seizures, provide the general target for experiments using slices that focus on seizure activity involving hippocampus. Can we match that description in hippocampal slices?


The method and care of slice preparation will determine the tissue's viability and in large part the validity and applicability of the experimental results obtained in the preparation. Every laboratory uses its own "magical" tricks to prepare good slices. Variations include the type of anesthetic used before decapitating the animal. Some laboratories find that slice quality is improved when animals (before decapitation) are intracardially perfused cold (~ 4° C) oxygenated aCSF in which NaCl is replaced by an equimolar concentration of sucrose or choline (Hirsch et al., 1996; Hoffman and Johnston, 1998). Other tricks include the use of gluta-matergic antagonists or low Ca2+/high Mg2+ (0.5/6.0mM) concentrations in the modified aCSF to prevent glutamate excitotoxicity and to reduce synaptic transmission. The use of prior brain perfusion becomes a critical issue when juvenile or adult brains are used (>21 days postnatal). A pump can be used for these perfusions, although a simple gravity system is sufficient. The temperature should be controlled, as a too-cold (0° C) aCSF is deleterious.

The next step is to remove the brain from the skull as quickly as possible. The detail of this manipulation depends somewhat on the animal species. In rats the difficulty of this process increases with age because the bone becomes thicker and harder. Once the bone covering the cerebellum is removed, it is possible to make a cut along the midline with sharp scissors. Rongeurs or pliers can be used to remove the top part of the skull in one rotating movement. The dura should then be carefully cut so that the brain can be safely extracted using a scoop. The intact brain is then placed into a beaker of ice-cold oxygenated, modified aCSF. This procedure should not take more than 15 to 20 seconds (from time of decapitation). Although it will be more cumbersome, the whole extraction procedure can be performed with the animal's head submerged in ice-cold oxygenated, modified aCSF.

The extracted brain is then prepared for the slicing procedure. Standard procedures for cutting hippocampal slices in rats include removing the cerebellum and the most frontal regions of cortex. The hemispheres are then typically separated. If a "chopper" is used, the hippocampus is removed from the brain and placed on a cutting stage. Caution must be used to limit the mechanical stresses and stretches. Most hippocampal slice studies use transverse sections, that is, slices cut perpendicular to the longitudinal axis of the hippocampus. Because of the curvature of the hippocampus (Figure 1A), it is not possible to obtain transverse slices from the whole hippocampus without artificially stretching the structure. When a vibroslicer is used to cut slices (Figure 1B), the investigator may leave the hippocampus embedded within the whole brain hemisphere and glue the hemisphere onto a support block. Although leaving the hippocampus protected inside the hemisphere reduces the number of slices that can be made in the "transverse" orientation, this procedure offers the advantage of lesser manipulation of the hippocampal structure before slicing. Further, by cutting the brain surface (to be glued to the support block) at an appropriate angle, the experimenter can manipulate the orientation of the slice with respect to the longitudinal axis of the hippocampus and thus "expose" hippocampal substructures of particular interest (e.g., for recording from pyramidal cell dendrites) (Hoffman et al., 1997).

The slicing procedure itself is a critical step. A vibroslicer with minimal Z-deflection and a glass or sapphire blade should be used for best results. Optimal slicing parameters (e.g., cut speed, amplitude, frequency of oscillation of the blade) depend on the preparation. Tissue from adult animals cut differently from tissue from immature animals because of the presence of choroid tissue, myelin, etc. Tissue from rats and mice cut somewhat differently from, for example, primate hippocampus, not least of all because of the different size of the structure in these different species. The issue of cutting parameters becomes particularly important when preparing slices from the human brain (following extraction of an epileptic hippocampus during surgery for medically intractable TLE), as the density and rigidity of the tissue may be very different according to the underlying reorganizations (sclerosis, gliosis, etc.). Cutting slices from immature brain presents the investigator with other types of challenges. Because the immature rat hippocampus easily "falls apart" when sliced, it is preferable to keep it embedded within the rest of the hemisphere during cutting, and to

Transverse Hippocampal

22.02 Copyi^ht John Wiley and Sons, 1990

FIGURE 1 A: Schematic showing the three-dimensional organization of the hippocampus within a rodent brain. B: Schematic showing key features of the slice-cutting procedure. The cortex (not shown) containing the hippocampus is glued on an agar block. Transverse slices are limited to the ventral part of the hippocampus. C: Typical transverse hippocampal slice showing the major cell regions within a slice. The hippocampal slices typically consist of the dentate gyrus (with granule cells), the CA3 and CA1 pyramidal cell regions, and the subiculum. Thick lines represent the distribution of the somata of the principal neurons (granule cells in the dentate gyrus and pyramidal cells in the CA3 and CA1 regions).

22.02 Copyi^ht John Wiley and Sons, 1990

FIGURE 1 A: Schematic showing the three-dimensional organization of the hippocampus within a rodent brain. B: Schematic showing key features of the slice-cutting procedure. The cortex (not shown) containing the hippocampus is glued on an agar block. Transverse slices are limited to the ventral part of the hippocampus. C: Typical transverse hippocampal slice showing the major cell regions within a slice. The hippocampal slices typically consist of the dentate gyrus (with granule cells), the CA3 and CA1 pyramidal cell regions, and the subiculum. Thick lines represent the distribution of the somata of the principal neurons (granule cells in the dentate gyrus and pyramidal cells in the CA3 and CA1 regions).

use a very slow forward speed and a high-frequency oscillation. Whatever the type of tissue, a general "rule of thumb" is to use a slow forward speed and large lateral amplitude and high frequency oscillation. However, it is important to invest the necessary time in establishing the appropriate set of parameters for a given preparation (and slicing equipment). As stressed already, all the experimental procedures and data depend on the quality of the slices.

Slice thickness is another critical parameter, and it usually represents a compromise between the desired amount of intrinsic connectivity and the experimenter's ability to provide adequate oxygenation to the core of the slice. Slices from adult tissue are rarely thicker than 400 mm. In the immature brain, this limit is less critical; the whole hippocampus can survive for 48 hours in vitro (Khalilov et al., 1997).

Not all the slices made from a given hippocampus are useful for a given set of experiments. Which slices are chosen depends on the goals of the study. For example, the connectivity patterns in the ventral, dorsal, and midhip-pocampus are very different (Moser and Moser, 1998; Witter and Groenewegen, 1984). Physiologic features may also vary according to the slice location (along the longitudinal axis) and orientation (Ferbinteanu and McDonald, 2001).

Slices are normally cut into cold aCSF and then transferred (for "storage") into a holding chamber at room temperature. Slices needed for study are transferred to the recording chamber. The choice of interface versus submerged chamber system is another critical issue, and each offers its own advantages. For example, extracellular space (ES) is reduced when an interface chamber is used, thus increasing the ephaptic interactions, a potentially important issue in epilepsy studies (Schuchmann et al., 2002). Ictal-like discharges are more easily obtained in interface versus submerged chambers. All perfusion systems should allow for a continuous flow of aCSF into the recording chamber. The perfusion speed appears to be a critical parameter because biologically relevant hippocampal rhythms can be recorded only when high flow rates (i.e., >4ml/minvs. the usual 1.5ml/min) are used (Hajos et al., 2004; Wu et al., 2002). Differences in such apparently subtle parameters may explain interlaboratory discrepancies, for example, when using interface versus submerged chambers (Bracci et al., 1999). A higher flow rate (with O2-saturated aCSF) is usually needed for submerged chambers in which all oxygenation is derived from the medium. In contrast, much of the oxygen supply to slices in interface chambers is derived from the O2-saturated gas environment above the tissue.

A typical aCSF formulation consists of (in millimolar): NaCl 125, KCl 2.5, CaCl2 2, MgCl2 2, Na^PO4 1.25, NaHCO3 25, and dextrose 10. The aCSF is equilibrated with 5% CO2/95% O2 gas which is bubbled through the solution. The concentrations of K+, Ca2+, and Mg2+ are particularly variable (within 1 mM or so) from laboratory to laboratory.

These concentrations are important because even minor changes in these ionic concentrations will alter cell and network excitability, synaptic transmission, and intracellular processes. As pointed out earlier, typical slice aCSF does not contain many metabolites that are usually contained in the CSF of intact animals.

To be "more relevant" to physiologic conditions, slice experiments should be performed close to physiologic temperature (~35° C). Below 30° C, transporters, ionic channels, etc., have different biophysical properties. Working close to physiologic temperature, however, raises a number of challenges. First, there is often a "dead space" between the system that warms aCSF (e.g., a temperature-regulated water bath) and the recording chamber. Therefore, aCSF will have to be warmed to a higher temperature than the target, raising the risk of precipitation (e.g., Ca2+) and driving out needed oxygen. One can reduce the dead space by using a system close to the recording chamber. However, such a system is normally of smaller capacity and must also be set to a high level because the time that the aCSF spends in the warming apparatus is so short. The best technical solution is to use two warming systems: a temperature-regulated bath to prewarm aCSF and a small system close to the chamber for final adjustment. These temperature control issues are particularly difficult when high flow rates are used.

There are two good tests to check the condition of a slice. The first one involves a visual inspection with infrared microscopy. The surface of the slice should be planar, and the neurons at the surface should be healthy, that is, having no "balloon-like" somata (with a clearly visible nucleus) and no dying neurons (with a strong black-and-white contrast soma contour). The second test involves determining whether oscillations can be induced (Hajos et al., 2004). This latter assessment can be performed even when "blind" cellular or field recording techniques are used.

Was this article helpful?

0 0

Post a comment