A Time and Place for Everything: Early-Life Seizures at Different Developmental Epochs Have Distinct Effects on Adult Hippocampal Structure and Function

Commentary

The developing brain is a unique substrate for seizure activity. Multiple lines of evidence, including epidemiologic studies (1), as well as in vivo and in vitro laboratory work (2), have demonstrated an increased propensity for seizures in early life. The extent of the long-term impact of these early-life seizures on brain development remains a topic of debate. Multiple animal studies have demonstrated long-lasting functional and anatomic changes as a consequence of early-life seizures, while other studies have demonstrated resistance to neuronal damage and cell loss in younger animals (3, 4). The extent to which these cellular and molecular alterations relate to seizure-induced changes in human behavior and cognition is unclear.

Under normal conditions, activity-dependent plasticity of the developing nervous system allows the refinement of immature connections into adult networks (5). Seizure-induced excessive, synchronous network activity may influence these processes to pathologically remodel developing networks. A better understanding of the long-term effects of the complex interplay between seizures and the inherent plasticity of the developing nervous system may help inform clinical decision making and anticipatory guidance for childhood seizures. Sayin and colleagues have undertaken their study in an effort to methodically investigate the long-term structural and physiologic effects of seizures during development. Seizures were induced in rats at specific developmental stages, from the immediate postnatal period through young adulthood. Three different methods of seizure induction were used: systemic injection of kainic acid (KA, a glutamatergic agonist), hyperthermic seizures, and injection of pentylenetetrazole (PTZ, a proconvulsant CNS stimulant). The use of multiple seizure models, encompassing both status epilepticus and repeated brief seizures, may help improve the translatability of these results, particularly since the significant findings of this study were consistent across seizure models. After seizure induction, experimental outcomes were measured in animals once they reached adulthood. Paired-pulse inhibition (PPI), a nonspecific measure of network functional properties that may correlate with GABA_A receptor-dependent inhibition, was measured in hippocampal slices. Ex vivo MRI and diffusion tensor imaging (DTI) were obtained to assess gross and microstructural parameters, and the response of adult animals to kindling or further seizure induction was also assessed.

Of interest, seizure-induced changes in adult functional and structural parameters could be separated into three cohorts based on the age at which the seizures were induced, regardless of induction method, seizure duration, or severity. Animals that experienced seizures from postnatal day 1 through 20, corresponding to a period from birth to early childhood in humans, demonstrated long-lasting increased PPI in the dentate gyrus and subtle microstructural differences in the corpus callosum and dentate gyrus without gross structural abnormalities. Similarly, animals that experienced seizures after postnatal day 30, a juvenile to adult phase of development, also demonstrated increased PPI in the dentate gyrus, and showed brain structural changes including atrophy on MRI and an alteration in the primary orientation of the dentate hilus on DTI. In contrast to these two groups, animals that experienced seizures between postnatal day 20 and 30, corresponding to childhood to adolescence in humans, demonstrated a reduction or complete loss of PPI in the dentate gyrus, without significant imaging abnormalities. The authors surmise that this may suggest a long-lasting decrease in GABAergic inhibition, although with the caveat that data from these PPI experiments do not provide definitive information about the underlying mechanism. Further seizures induced in adulthood in this cohort did not further change PPI as it did in control animals, suggesting a loss of the capacity for further seizure-induced plasticity in adulthood after seizures between postnatal day 20 and 30.

Synaptic plasticity in the CNS is not limited to the developing brain; the ability of experience to modify the strength of particular connections throughout an organism's life-span underlies the ability to learn and adapt. However, unchecked plasticity is undesirable from the standpoint of functional stability. Without compensatory mechanisms to limit excessive shaping and rewiring of circuits, the positive feedback mechanisms of long-term potentiation and long-term depression could lead to excessive hyper- or hypoexcitability over time. Homeostatic mechanisms operate to constrain plasticity and preserve stable networks (6). During development, the level of homeostatic control varies to allow periods of increased activity-dependent structural and functional plasticity – termed critical periods (7) – allowing appropriate shaping of developing networks by experience and the environment. This phenomenon has been best characterized in sensory systems; for example, the change in responsiveness of the primary visual cortex that leads to amblyopia occurs when visual input is disrupted during a critical period for that system. It is interesting to consider seizures during development in such a context; excessive, synchronous network activity occurring during a window of increased plasticity could have profound developmental consequences. Sayin and colleagues propose that the different structural and functional responses to seizures induced between postnatal day 20 and 30 reflects such a critical period for hippocampal and limbic networks. The author's interpretation of the data fits within the conceptual framework of critical periods, wherein the brain is particularly vulnerable to the effect of seizures during a developmental epoch with sharply delimited temporal borders. However, further behavioral and physiologic evidence of lasting seizure-dependent changes in hippocampal function is needed to establish the 20 to 30 day window as a critical period. It is also important to know whether ongoing seizures differ in the distinct cohorts and whether these influence the outcomes following status epilepticus evoked at different time points.

While correlation between rodent postnatal development and that of humans is inexact at best, the identification of a similar critical period in human development would have multiple implications for the clinical care of patients with seizures. Further research into how seizures interact with developmental critical periods could inform open questions about seizures in early life, such as which children are at risk for mesial temporal sclerosis after febrile status epilepticus (8), or how aggressively neonatal seizures should be treated (9). From a basic science perspective, understanding the excitatory and inhibitory factors gating this critical period could shed light on mechanisms underlying epileptogenesis. While such causative mechanisms were beyond the scope of this study, the description of a possible hippocampal critical period and the effect of seizures during this period on adult hippocampal circuit properties and plasticity is an important contribution to the literature and raises questions that merit further exploration.