- Top of page
- Material and Methods
- Supporting Information
Purpose: Periventricular nodular heterotopia (PNH) is, in humans, often associated with difficult-to-control epilepsy. However, there is considerable controversy about the role of the PNH in seizure generation and spread. To study this issue, we have used a rat model in which injection of methylazoxymethanol (MAM) into pregnant rat dams produces offspring with nodular heterotopia-like brain abnormalities.
Methods: Electrophysiologic methods were used to examine the activity of the MAM-induced PNH relative to activity in the neighboring hippocampus and overlying neocortex. Recordings were obtained simultaneously from these three structures in slice preparations from MAM-exposed rats and in intact animals. Bath application or systemic injection of bicuculline was used to induce epileptiform activity.
Key Findings: In the in vitro slice, epileptiform discharge was generally initiated in hippocampus. In some cases, independent PNH discharge occurred, but the PNH never “led” discharges in hippocampus or neocortex. Intracellular recordings from PNH neurons confirmed that these cells received synaptic drive from both hippocampus and neocortex, and sent axonal projections to these structures—consistent with anatomic observations of biocytin-injected PNH cells. In intact animal preparations, bicuculline injection resulted in epileptiform discharge in all experiments, with a period of ictal-like electrographic activity typically initiated within 2–3 min after drug injection. In almost all animals, the onset of ictus was seen synchronously across PNH, hippocampal, and neocortical electrodes; in a few cases, the PNH electrode (histologically confirmed) did not participate, but in no case was activity initiated in the PNH electrode. Interictal discharge was also synchronized across all three electrodes; again, the PNH never “led” the other two electrodes, and typically followed (onset several milliseconds after hippocampal/neocortical discharge onset).
Significance: These results do not support the hypothesis that the PNH lesion is the primary epileptogenic site, since it does not initiate or lead epileptiform activity that subsequently propagates to other brain regions.
Medically intractable forms of epilepsy are often associated with abnormal brain development (Schwartzkroin & Walsh, 2000; Porter et al., 2002; Schwartzkroin et al., 2004). Abnormalities in neuronal migration, where newly developed neurons fail to reach their proper destinations, result in aberrant brain structure, often termed cortical dysplasia (CD). Dysplastic cell clusters and/or surrounding structures (e.g., in neocortex or hippocampus) are often identified as the sites in which epileptic activities are generated, and invasive surgical resection of these “lesions” has been used to provide seizure control (Hirabayashi et al., 1993; Francione et al., 2003; Russo et al., 2003; Aghakhani et al., 2005; Stefan et al., 2007). The condition of periventricular nodular heterotopia (PNH), in which a collection of neurons protrude into the ventricular space, has been described as a genetic disorder associated with intractable seizure activity (Guerrini, 2005; Sheen et al., 2005; Cardoso et al., 2009). Although the PNH lesion is relatively rare, it serves as a “model” of other epilepsy-associated migration disorders, and provides a useful context in which to explore the relative epileptogenicity of the aberrant structure (the nodule) versus the relatively “normal” surrounding tissue (Battaglia et al., 2006).
Direct recording from human PNH is invasive, and electrographic data from human patients provide only indirect evidence of the involvement of these aberrant structures in the generation of seizure activity (Dubeau & Tyvaert, 2010). Functional imaging approaches have been useful for gaining additional insight into the relationship between the nodular lesions and surrounding tissue (Archer et al., 2010). But given the difficulty of drawing “cause–effect” conclusions from these studies, animal models of PNH may provide valuable insights into the role of the nodular lesion in seizure generation.
Several animal models have been developed to study neuronal migration disorders. Chemical treatments [e.g., using the DNA-alkylating agent, methylazoxymethanol acetate (MAM)] (e.g., Baraban & Schwartzkroin, 1995, 1996; Germano & Sperber, 1997; Chevassus-au-Louis et al., 1999a,b; Castro et al., 2001), in utero irradiation (Roper et al., 1997), genetic manipulations of important cell development genes (Lee et al., 1997), and cortical freeze-lesions (Jacobs et al., 1996, 1999; Scantlebury et al., 2004; Takase et al., 2009) have provided valuable experimental tools to study different aspects of the basis of epileptic activities in neuronal migration disorders.
In utero exposure of rats to MAM results in reproducible brain abnormalities, including microcephaly, heterotopic cell clusters in hippocampus and cortex, and abnormally located cell clusters in the periventricular region (Singh, 1977; Baraban & Schwartzkroin, 1996; Chevassus-au-Louis et al., 1998; Sancini et al., 1998; Tschuluun et al., 2005). Previous MAM studies have focused primarily on heterotopia within the hippocampus and cortex, using coronal rat slices. Tract-tracing methods were used to demonstrate abnormal reciprocal connectivity between heterotopia and ipsilateral and contralateral cortical and hippocampal regions (Colacitti et al., 1998), and abnormal connectivity has been described in the malformed tissue (e.g., tangential cortico-cortical axon bundles, partial deafferentation of neocortical heterotopia, and abnormal ectopic mossy fiber projections within the hippocampus) (Chevassus-au-Louis et al., 1999a).
The periventricular localization of dysplastic cell clusters resembles PNH as seen in patients with intractable epilepsy (Aghakhani et al., 2005; Stefan et al., 2007; Giannopoulos et al., 2008). Clinical studies suggest that the PNH may serve as the initiator of epileptic activities in these patients—or at least contribute to a more widespread “epileptic” network (Aghakhani et al., 2005; Tassi et al., 2005; Stefan et al., 2007). Data from clinical imaging studies (Aghakhani et al., 2005; Tassi et al., 2005; Tyvaert et al., 2008; Archer et al., 2010) as well as results from experimental animal studies (Chevassus-au-Louis et al., 1998, 1999b; Colacitti et al., 1998) suggest that PNH-like lesions are functionally integrated into the surrounding cortical and hippocampal tissue. Based on our previous studies (Tschuluun et al., 2005), we have suggested that the PNH structure provides aberrant connectivity between normally unconnected brain regions, and thus creates a novel pathway by which epileptiform activity can more easily spread/generalize to relatively normal brain regions.
In the current experiments, we have used the MAM rat model of PNH to address the following questions: (1) Does the PNH initiate epileptiform discharges in the brain of MAM rats? (2) Does the PNH integrate structurally into neighboring tissue so as to facilitate the propagation of the epileptiform discharges? We used both in vitro (slice) and in vivo (intact rat) electrophysiologic recording approaches to examine these issues. In acute brain slice preparations, the PNH did not initiate (i.e., lead) epileptiform discharges (either ictal or interictal electrographic activities), although the nodule appeared capable of generating independent electrographic “bursts.” We also found that PNH neurons are reciprocally integrated into the neighboring cortical and hippocampal networks, and thus can serve as functional “shortcuts” for epileptiform discharges initiated in the hippocampus to propagate into the cortex. Recordings from the intact rat brain containing PNH-like lesions confirmed our general results from slice experiments: The PNH never initiated epileptiform discharges when the rat was injected with bicuculline, but usually participated (i.e., generated synchronized discharge) in events initiated in ipsilateral hippocampus or neocortex.
- Top of page
- Material and Methods
- Supporting Information
In these studies, we have attempted to address the question of if/how activity in a discrete dysplastic lesion affects epileptogenicity. Using a rat model of PNH, we have found that the aberrant structure does not appear more excitable than the surrounding tissue (hippocampus and neocortex), and does not appear to initiate or lead epileptiform activities. In both in vitro slice and in vivo intact rat preparations, epileptiform activity in the nodule was generally synchronized with discharge in surrounding tissue, but did not serve as a “driver” for either ictal or interictal events. These electrophysiologic results appear within a context of obvious structural connectivity between the nodule and surrounding brain tissue. The slice data further suggest that the aberrant nodule provides a bridge that facilitates communication between brain regions that in normal brain do not communicate directly. It should be emphasized, however, that these data do not address the questions of whether (1) the presence of the nodule affects general brain epileptogenicity, or (2) to what extent the surrounding tissue has become structurally (and functionally) abnormal as a result of the developmental process that also gives rise to the nodule. These latter questions require additional experimental investigation.
PNH constitutes a subclass of cortical dysplasia with well-defined pathologic features (Aghakhani et al., 2005; Tassi et al., 2005). Although seizures are highly correlated with this usually genetically determined brain disorder, clinical studies have not provided clear results regarding the role of the PNH in the initiation and spread of the electric activity during seizures (Tassi et al., 2002, 2005; Aghakhani et al., 2005; Stefan et al., 2007). Common findings from clinical studies are generally consistent with our experimental results; that is: (1) The PNH generates electrical ictal activity closely synchronized with cortical and hippocampal discharges; (2) the PNH is sometimes capable of generating independent interictal-like discharges; and (3) it is anatomically (and functionally) connected to neighboring cortical structures (Aghakhani et al., 2005; Stefan et al., 2007).
One of the best-studied animal models of PNH is the MAM-treated rat (Baraban & Schwartzkroin, 1995, 1996; Germano & Sperber, 1997; Germano et al., 1998; Tschuluun et al., 2005; Marchi et al., 2006; Harrington et al., 2007). MAM rats rarely exhibit spontaneous seizures (but see Harrington et al., 2007), but they show lower threshold for epileptogenic insults (Baraban & Schwartzkroin, 1996; Germano & Sperber, 1997; Germano et al., 1998). Our previous in vitro study on tissue from MAM rats (Tschuluun et al., 2005) suggested that dysplastic neurons in hippocampal heterotopia contribute to the spread of the epileptic discharges in brain slices. Based on these data, we postulated that although the dysplastic lesion may not itself be hyperexcitable, it provides a “low resistance bridge” that allows a more efficient, faster spread of epileptiform discharge, and, therefore, facilitates synchronization of aberrant discharge patterns. The initiator of that abnormal activity, however, remains to be determined.
The aberrant developmental program in MAM-exposed rats (and in human with genetic mutations that give rise to dysplastic structures) not only generates a “lesion” with obvious anatomic abnormality, but also gives rise to surrounding tissue with more subtle abnormalities (Castro et al., 2001; Garbelli et al., 2009). In the current study, we found the nodule was capable of independent discharge but did not appear to drive epileptiform discharge in surrounding tissue, consistent with previous work from our own lab and from others (Aghakhani et al., 2005; Tschuluun et al., 2005; Tyvaert et al., 2008). Further, inhibition is clearly relatively “intact” in the nodule; histology/immunocytochemistry reveals γ-aminobutyric acid (GABA)ergic inhibitory interneurons in the PNH, and stimulation of cortex and hippocampus elicits not only excitatory but also inhibitory synaptic events in PNH neurons. Therefore, it may be that surrounding tissue constitutes the initiator/driver, as has been proposed for the freeze-lesion model of epileptogenicity associated with abnormal structure (Jacobs et al., 1996, 1999; Scantlebury et al., 2004).
There is, of course, the question of if/how well this animal model (or any animal model) mimics the pathophysiology of human PNH. Most etiologic studies of epilepsies associated with PNH abnormalities have focused on genetic causes (e.g., Guerrini, 2005; Parrini et al., 2006; Ferland et al., 2009; Sprour et al., 2011). The MAM model represents, however, an extrinsic insult—perhaps more akin to developmental insults such as in utero stroke. Montenegro et al. (2002) suggested that prenatal injury may play a significant role in the genesis of development abnormalities; in almost half of their study group, prenatal events were identified that might have contributed (along with genetics) to the pathogenesis of cortical malformations. Another issue for interpreting the MAM rat model data arises from the simple geometrical differences between rat and human brain. In the rat, the PNH is physically close to both cortex and hippocampus, and thus connectivity can be “easily” established. In contrast, in the human brain, the PNH may be distant, especially from the hippocampus—and the hippocampus may, therefore, not be so involved in PNH-associated seizure generation (Li et al., 1997). The clinical literature suggests that not only may PNH pathology be heterogeneous (Sprour et al., 2011), but so also is seizure onset; therefore, in some patients, seizures may in fact originate directly from the nodules (Kothare et al., 1998; Scherer et al., 2005). Indeed, in a small number of patients with PNH, selective thermocoagulation of nodules has resulted in reduced epileptogenicity (Catenoix et al., 2008)—certainly suggestive of a causal role for PNH in seizure generation.
Epilepsy associated with periventricular nodular heterotopia often presents as medically intractable (Li et al., 1997; Guerrini, 2005; Poduri et al., 2005; Guerrini & Marini, 2006; Parrini et al., 2006). As has been shown for tubers in tuberous sclerosis complex, surgical removal of the aberrant structure may serve to control seizure activity. However, this empirical result does not “prove” that the lesion is epileptogenic, since such resection not only removes the purported “epileptic focus” but also disrupts a potential “epileptic network” (as is also the case for temporal lobe resection). Furthermore, such clinical interventions invariably also remove surrounding tissue, which may play a major role in epileptogenicity. The question of whether the dysplastic lesion is somehow responsible for brain epileptogenicity, therefore, remains an unresolved. Surgical interventions that specifically target the nodule may provide empirical evidence that guides therapeutic intervention (e.g., see Catenoix et al., 2008). If, however, it is not the features of the nodule per se that underlie epileptic function, but rather the associated abnormalities in surrounding tissue (e.g., aberrant connectivity, altered inhibitory function, etc.—see Jacobs et al., 1999; Schwarz et al., 2000), then such surgical intervention may be fruitless. It is, therefore, important for future research to provide more revealing insight into the relationship between abnormal structure and abnormal function.