Studying Epilepsy Using The Slice Preparation

Many different types of epilepsy exist. Hence multiple models have been developed, for example, partial or generalized epilepsies and convulsive or nonconvulsive epilepsies. A similar diversity of "models" has been reported using the hippocampal slice preparation (and variations thereon), most focusing on generating activity relevant to TLE. Many different issues can be addressed in the slice preparation. A major area of investigation has been the issue of ictogene-sis, including where ictal discharges are generated, how they propagate, how they stop (or how we can stop them), and what underlying mechanisms might be involved. Another important theme involves studies of epileptogenesis, that is, what processes lead to an epileptic state following an initial insult. In pursuit of these and related questions, two general types of preparations have been used for the past three decades: (1) brain slices obtained from "normal" animals, in which in vitro manipulations have been used to generate epileptiform activities; and (2) brain slices obtained from human patients or chronically epileptic animals. Slices from human patients are difficult to obtain and study (see chapter 8), and generating epileptic animals is time consuming (as described elsewhere in this volume). Many investigators thus turned to normal brain slices to study ictogenesis. In addition to easy accessibility, the extensive (but far from complete) knowledge of the underlying network architecture of hippocampus, and of its physiological properties, has facilitated the design of such experiments and the interpretation of their results. The main drawback of these "acute" models is that little structural pathological change is apparent in these normal circuits. In contrast, most investigators believe that in temporal lobe epilepsy, the network architecture has been considerably modified (cell loss, sprouting, protein modifications etc.). The "rules" and insights derived from acute studies must therefore be reviewed with caution in trying to understand the mechanisms of chronic epilepsies.

In the following sections, I describe a number of the acute models generated in hippocampal slices and briefly comment on the use of slices to study chronic models. Before doing so, however, it is important to consider the issue of the age of the animal from which the slice is made. Rats 5 to 6 weeks old are considered adults. Before this age, the rat's neuronal networks in the hippocampus are still developing and undergo considerable modifications. During the first postnatal week in the rat, GABA acts as an excitatory neurotransmitter (Ben Ari, 2002), GABAergic interneu-rons are the source and targets of the first synapses (Gozlan and Ben Ari, 2003), and neuronal network activity results from the synergistic excitatory actions of GABA and glutamate (Ben Ari et al., 1989; Leinekugel et al., 2002). At these young ages, many connections are lacking, and neurons continue to proliferate and migrate to their proper target regions. The state of such a developing network and its functioning mode are quite different from the adult brain. This early period of rodent development corresponds roughly to the last trimester of gestation in infrahuman primates (Khazipov et al., 2001) and (perhaps) in humans (Clancy et al., 2001). Because GABA switches from a depolarizing to a hyperpo-larizing mode well before birth in infrahuman primates (Khazipov et al., 2001), studies performed in rodents during the first 2 postnatal weeks may be relevant primarily to in utero epilepsy in human. The following weeks are characterized by a refinement of connections, myelination, changes in receptor subunit composition, channel expression, etc. (Fritschy et al., 1994; Katz & Shatz, 1996; Meier et al., 2004; Ritter et al., 2002; Stellwagen and Shatz, 2002; Tansey et al., 2002). How the time points of these changes correspond to human developmental stages is unknown.

This issue of age relevance is important for many slice studies because recordings are typically carried out on slices from animals between postnatal days 14 and 21. This timing choice is dictated largely by technical considerations, an empirical consensus that it is easiest to prepare slices from these young animals and that slice viability is superior (e.g., compared with slices from more mature animals). As suggested, it is possible to argue that experiments performed on 6-week-old rodents are relevant to human (adult?) epilepsy because the hippocampal neuronal networks have been stabilized in their mature forms. However, in slice studies on rats (or mice) between the second and the sixth postnatal week, it is particularly difficult to extrapolate the data to human epileptic phenomena. Not only is "matching" ages difficult, but the situation is also complicated by the fact that maturation of different cells, different systems, and different molecular controls occur at different speeds (and span different periods) in rodents and in humans. While there is no "solution" to this problem, it is important to keep these issues in mind while designing experiments and interpreting in vitro results.

Acute Models of Epilepsy

Several in vivo models of ictal-like and inter-ictal-like activity have been developed in hippocampal slices, including slices challenged with high-frequency electrical stimulation (Somjen et al., 1985), GABAA-receptor antagonists (Schwartzkroin and Prince, 1978; Swann and Brady, 1984), kainic acid (KA) (Fisher and Alger, 1984; Westbrook and Lothman, 1983), K+ channel blockers (Galvan et al., 1982), low [Ca2+]o (Jefferys and Haas, 1982; Taylor and Dudek, 1982), no [Mg2+]o (Anderson et al., 1986) or high [K+]0 (Traynelis and Dingledine, 1988). As detailed later, some models reproduce some features of interictal activity and others produce tonic-clonic discharges. These models have several advantages. They are easy to implement, and pathological discharges emerge with high reproducibility. The slice preparation allows easy access to numerous parameters, including activity patterns of neuronal populations, propagation of signals from one subregion to the next, synaptic inputs in individually recorded cells, and pharma-cologic responsiveness of different patterns of paroxysmal discharges.

Electrical Stimulation-Induced Afterdischarge

Ictal-like afterdischarges displaying tonic- and clonic-like phases can be evoked following tetanic stimulation (40-50 pulses at 100 Hz) applied in CA1 stratum radiatum (Figure 2A). The tonic-clonic phase appears after several stimulations, suggesting a long-term potentiation-like effect (Rafiq et al., 1993). The afterdischarge is composed of several epochs, with a sequence of slow and fast oscillations

FIGURE 2 A: Progression of afterdischarge waveform and duration, recorded in the CA1 region in response to stimulation of the Schaffer collaterals. (Adapted from Rafiq et al., 1993.) The graph at the right shows the gradual increase in afterdischarge duration with repeated stimuli. B: Interictal-like and ictal-like activities recorded in the disinhibited CA1 minislice. Left panel: Simultaneous intracellular recordings of a pyramidal cell soma (top) and dendrite (bottom), showing an interictal discharge (shown at a faster time scale below) followed by an ictal-like discharge (duration, 7 seconds). Upper right panel: Experimental design showing the cuts performed to isolate the CA1 region and the positions of the recording electrodes. Lower right panel: Bar graph showing the duration of ictal-like activity as a function of the g-aminobutyric acid (GABA) antagonist used. Bic, bicuculline; PTX, picrotoxin; GBZ, gabazine 100 mm. (Adapted from Karnup and Stelzer, 2001.) C: Simultaneous field and intracellular recordings in the CA1 region showing interictal-like activity induced by 1 mm kainic acid. (Adapted from Fisher and Alger, 1984.) D: Simultaneous field and whole-cell patch-clamp recordings in the CA3 region, showing ictal-like activity induced by 100 mm 4-AP in toto. (Adapted from Luhmann et al., 2000.)

FIGURE 2 A: Progression of afterdischarge waveform and duration, recorded in the CA1 region in response to stimulation of the Schaffer collaterals. (Adapted from Rafiq et al., 1993.) The graph at the right shows the gradual increase in afterdischarge duration with repeated stimuli. B: Interictal-like and ictal-like activities recorded in the disinhibited CA1 minislice. Left panel: Simultaneous intracellular recordings of a pyramidal cell soma (top) and dendrite (bottom), showing an interictal discharge (shown at a faster time scale below) followed by an ictal-like discharge (duration, 7 seconds). Upper right panel: Experimental design showing the cuts performed to isolate the CA1 region and the positions of the recording electrodes. Lower right panel: Bar graph showing the duration of ictal-like activity as a function of the g-aminobutyric acid (GABA) antagonist used. Bic, bicuculline; PTX, picrotoxin; GBZ, gabazine 100 mm. (Adapted from Karnup and Stelzer, 2001.) C: Simultaneous field and intracellular recordings in the CA1 region showing interictal-like activity induced by 1 mm kainic acid. (Adapted from Fisher and Alger, 1984.) D: Simultaneous field and whole-cell patch-clamp recordings in the CA3 region, showing ictal-like activity induced by 100 mm 4-AP in toto. (Adapted from Luhmann et al., 2000.)

followed by a silent postictal depression and a secondary discharge. This pattern is similar to that found in afterdis-charges evoked in vivo (Bragin et al., 1997). The mechanisms underlying physiologic oscillations often involve interneurons (Freund and Buzsaki, 1996), and so after-discharges constitute a good model for studying synchronization via local circuits (including the role of N-methyl-D-aspartate [NMDA] receptors) (Stasheff et al., 1985, 1993), interneurons and depolarizing GABA (Bracci et al., 1999; Fujiwara-Tsukamoto et al., 2003, 2004; Kaila et al., 1997; Staley et al., 1995; Velazquez and Carlen, 1999; Whittington et al., 1997) as well as ephaptic interactions (Bracci et al., 1999). Combining electrophysiology (intra-cellular recording) and morphology (intracellular labeling) allows the investigator to dissect out key features of specific hippocampal sub-networks (Fujiwara-Tsukamoto et al., 2004). This model of electrically induced afterdischarge has been discussed as potentially relevant to ictal discharges in human temporal lobe epilepsy (Stasheff et al., 1985). As mentioned previously, these experiments are performed on hippocampal slices from normal animals, and translating the results to chronic epilepsy is somewhat dangerous. For example, CA1 stratum oriens interneurons are directly involved in the synchronization process during the afterdischarge (Fujiwara-Tsukamoto et al., 2004); yet a large number of these interneurons are lost in chronic epilepsy (Cossart et al., 2001; Dinocourt et al., 2003). Nevertheless, it is intriguing to note that NMDA receptors and depolarizing GABA play an important role in evoked afterdischarges (Staley et al., 1995; Stasheff et al., 1985). In chronic epilepsy, the NMDA receptor-dependent component of synaptic transmission is considerably increased (Turner and Wheal, 1991) and fast GABAergic neurotransmission becomes, in part, excitatory (Cohen et al., 2002).

GABAergic Disinhibition

Pharmacologic blockade of GABAa receptors has been used extensively to study epileptiform activity in vitro. In the absence of fast GABAergic neurotransmission, synchronized inter-ictal-like bursts can occur (Miles and Wong, 1986, 1987; Schwartzkroin and Prince, 1978; Wong et al., 1986). Not surprisingly, these bursts depend on fast gluta-matergic neurotransmission to activate (alpha) amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/KA and NMDAreceptors (Simpson et al., 1991; Williamson and Wheal, 1992). In general, the generation of ictal-like events in vitro requires additional pharmacologic manipulations (at least in mature tissue), such as elevating [K+]o (Traub et al., 1996). However, ictal-like events can be recorded in slices from the immature brain (Khalilov et al., 1997; Swann and Brady, 1984). Further, ictal-like activity is present in the dis-inhibited CA1 minislice of adult guinea pigs (Figure 2B) (Karnup and Stelzer, 2001). The usefulness of the disinhib-ited hippocampal slice model, at least for studying ictogen-esis, is questionable because GABAergic neurotransmission appears to be quite robust (even increased) in chronic temporal lobe epilepsy (Bernard et al., 2000; Cossart et al., 2001; Esclapez et al., 1997). However, blocking fast GABAergic neurotransmission may be a useful approach to unravel changes in excitatory circuits in chronic epilepsy (Esclapez et al., 1999; Meier and Dudek, 1996; Patrylo and Dudek, 1998).

Kainic Acid-Induced Epileptiform Activity

Bath application of kainic acid KA induces spontaneous interictal-like activity in hippocampal slices (Figure 2C). KA has multiple presynaptic and postsynaptic effects, and so the underlying bases for this KA effect remains unclear (Ben Ari and Cossart, 2000; Huettner, 2003). KA can directly depolarize neurons that express KA receptors, such as CA3 pyramidal cells (Westbrook and Lothman, 1983) and interneurons (Cossart et al., 1998; Frerking et al., 1998), leading them to generate action potentials. KA also has a multiplicity of presynaptic effects on glutamatergic and GABAergic terminals (Huettner, 2003) as well as on ionic channels (Melyan et al., 2002). Despite its epileptogenicity in vivo (Ben Ari and Cossart, 2000) and its ability to generate g-oscillations (40 Hz) at nanomolar concentration in vitro (Buhl et al., 1998; Fisahn et al., 1998), KA has not been extensively used to study the initiation and propagation of interictal-like discharges in the hippocampal slice preparation. A study performed in the intact (in toto) hippocampus (in vitro) described the development of a "mirror focus" in contralateral (naive) hippocampus when the ipsilateral hippocampus was challenged with KA (Khalilov et al., 2003).

Blocking K+ Channels

Blocking K+ channels with 10 to 30 mm 4-aminopyridine (4-AP) produces ictal-like discharges in the olfactory cortex (Galvan et al., 1982). In the hippocampus, 50mm 4-AP depolarizes neurons (Perreault and Avoli, 1989) and induces interictal-like (Voskuyl and Albus, 1985) or ictal-like (Chesnut and Swann, 1988) discharges (Figure 2D). This model can be used to study the propagation of paroxysmal discharges between different regions (Luhmann et al., 2000), the role of GABAergic neurotransmission in such discharges (Perreault and Avoli, 1992), the transition between interic-tal- and ictal-like discharges (Dzhala and Staley, 2003b), and synchronization mechanisms (Netoff and Schiff, 2002). However, it appears particularly useful in combined hippocampus—EC slices (see chapter 4).

Low Extracellular Ca2+

Lowering [Ca2+]o (nominally to 0mM) abolishes neurotransmission and results in spontaneous paroxysmal discharges (Figure 3A) (Jefferys and Haas, 1982; Taylor and Dudek, 1982). These events arise focally and spread through the CA1 region (Konnerth et al., 1984). This model may be particularly valuable for the study of synchronization mechanisms (Bikson et al., 2003) and for investigating interventions to abort/stop these discharges (Ghai et al., 2000).

Epileptiform Discharges in the Absence of [Mg2+]o

Removing Mg2+ from aCSF results in the appearance of spontaneous interictal-like discharges in slices from juvenile/adult animals (Anderson et al., 1986). This model could be clinically relevant because low levels of Mg2+ have been associated with human epilepsy (Durlach, 1967). Removing [Mg2+]o allows NMDA receptors to respond directly and robustly to glutamate excitation (Figure 3B) (Mody et al., 1987). This model allows for the study of propagation of

Was this article helpful?

0 0

Post a comment