How and Why Study Posttraumatic Epileptogenesis in Animal Models?

Abstract

Traumatic brain injury (TBI) greatly increases the risk of medically intractable epilepsy. Several models of TBI have been developed to investigate the relationship between TBI and posttraumatic epileptogenesis. Because the incident that precipitates development of epilepsy is known, studying mechanisms of epileptogenesis, identifying biomarkers to predict PTE, and developing treatments to prevent epilepsy after TBI are attainable research goals.

Over 2 million people are treated medically each year in the United States after sustaining a traumatic brain injury (TBI) (1), and posttraumatic epilepsy (PTE), which is often intractable to medical treatment, is a common long-term consequence of TBI. Epilepsy affects 1% to 2% of the population (2), but the overall incidence of epilepsy after moderate-to-severe closed-head injury is 7% to 39% and is over 50% after penetrating injury (3–5). Approximately 20% of all symptomatic epilepsies result from TBI (6), making PTE one of the most prevalent types of epilepsy. Although a number of seizure types can develop after brain trauma, PTE manifests as temporal lobe epilepsy (TLE, either neocortical or mesial) in 35 to 62% of trauma patients (7–9). As with other forms of acquired epilepsy, spontaneous recurrent seizures associated with PTE develop with a latency ranging from weeks to many years after the initial injury. This seizure-free period after TBI is thought to represent the period of epileptogenesis, during which the brain undergoes physiological, anatomic, cellular, and molecular changes that lead to a state of chronically increased seizure susceptibility. In principle, this delay between the TBI and development of PTE also represents a window of opportunity during which strategies might be employed to inhibit the reactive plasticity in the brain that leads to PTE, but no antiepileptogenic therapies have been successfully developed to date. This review focuses on key pathophysiological changes in the brain associated with posttraumatic epileptogenesis, animal models used to study those changes, and efforts to identify predictive biomarkers of PTE and to employ candidate treatments to prevent PTE.

TBI results in acute cell death, axon injury, vascular damage, and accompanying excitotoxicity shortly after injury. Secondary damage, occurring within days of the primary injury, including or resulting from hypoxia/ischemia, delayed necrotic and apoptotic cell death, oxidative stress, gliosis, edema, blood brain barrier disruption, and inflammation may exacerbate initial damage. Putative homeostatic repair mechanisms, including angiogenesis, synaptic reorganization, and neurogenesis can engage over time and are hypothesized to compensate for functional loss resulting from the initial injury; they are also associated with PTE. By their very nature, brain injuries are highly variable across patients, and this heterogeneity is reflected in a wide variety of epileptogenic responses to injury. Injury severity and location, the presence of seizures shortly after injury, intracranial hemorrhage, cortical contusion, the level of postinjury consciousness impairment, age, and sex have all been implicated as potential factors associated with increased risk of developing PTE after brain trauma (4, 5, 10). Many of the cellular events that occur after TBI also occur after insults used to induce epilepsy in animal models, and rodent models of PTE have been developed with the goal of identifying those factors associated with the eventual development of spontaneous seizures after brain trauma. The cellular aspects of both the processes contributing specifically to epileptogenesis after brain injury and the pathologies underlying the increased propensity for seizure generation can be reasonably studied in rodents, bearing in mind that rodents model cellular aspects of human disease but probably cannot replicate all facets of PTE.

Rodent Models

Several rodent models have been employed to study cellular changes in the brain after head trauma, but development of unambiguous, spontaneous recurrent seizures after a latent period has not been documented in most animal models of TBI. That noted, increased susceptibility to evoked seizures has been identified in a few TBI models (e.g., weight drop [11]); early (<1 week) postinjury seizures have been seen in others (e.g., blast injury [12]); and epileptiform or spike-and-wave electrographic signals have been documented after still other types of brain injury (13–15). Further investigation into the relationship of these brain injuries to PTE is warranted. There are currently, however, two rodent models that recapitulate the essential progression of PTE (including development of recurrent spontaneous seizures in a significant percentage of animals after a seizure-free latent period following TBI) and that have also been replicated in at least two laboratories. Epilepsy development in the lateral fluid percussion injury (LFPI) model of closed-head TBI is well documented (16, 17). The injury is delivered through a craniotomy by a rapid fluid pulse that strikes the intact dura and moves through the epidural space, resulting in a diffuse injury, and the pressure of the fluid pulse can be increased to cause a mixed diffuse and focal injury. Severe LFPI also leads to epileptogenesis, with spontaneous seizures developing in up to 50% of rats and 3% of mice within 12 months of injury (18, 19). The seizures involve the hippocampus electrographically and are similar behaviorally to those described for rodents by Racine et al. (20). The variability of the injury (e.g., severity, structures affected) and relatively long latency period can complicate the identification of cellular mechanisms that are specific to posttraumatic epileptogenesis, but this model presents several cellular correlates of epileptogenesis that are also identified in animal models of chemically and electrically induced epilepsy as well as patient histopathology, including hippocampal neuronal death, neurogenesis, and axon sprouting.

The controlled cortical impact (CCI) model of closed-head focal TBI also results in epileptogenesis in rats (21) and mice (22, 23). In this model, an electronically controlled impactor is used to create a focal contusion injury to the brain surface through a craniotomy (24). The injury occasionally results in early seizures (<24 hours) (22, 25), and spontaneous convulsive seizures develop in 40 to 50% of mice within approximately 8 weeks after severe injury (22, 23, 26–28) and 9 to 20% of mice after more moderate injuries (18, 22, 26). Like other murine epilepsy models, mouse strain may also influence seizure phenotype and the incidence of PTE (18, 29). Seizures after CCI are similar to spontaneous behavioral and electrographic seizures that have been described in rats after LFPI (19) and in models of acquired TLE (30, 31). Seizure frequency and prevalence after CCI injury are lower than in widely utilized chemoconvulsant TLE models, but seizure onset latency is considerably shorter than after severe LFPI in rats (19). Thus, at least two rodent PTE models have been developed and replicated sufficiently to support studies aimed at identifying biomarkers of PTE and to make reasonable hypotheses regarding candidate mechanisms underlying epileptogenesis after TBI.

Identifying Biomarkers of PTE

In most cases, TBI patients—at least those suffering mild or moderate injuries—do not develop seizures later in life, implying that there are qualitative or quantitative differences between brain injuries that lead to PTE and those that do not, or that some patients are predisposed (or resistant) to PTE development (e.g., genetics, prior injury). Treatments to prevent seizures in PTE patients are often unsuccessful, but because the initial event that precipitates epileptogenesis is identifiable and there is a delay from insult to first seizure of weeks to years, opportunities may exist to prevent postinjury epileptogenesis. There are no current treatments to prevent epileptogenesis after brain injury, and this is due, in part, to an incomplete understanding of the process by which TBI leads to PTE in some cases but not in others. Biomarkers specific for and predictive of PTE would be useful in identifying patients for which treatments designed to prevent epilepsy after TBI could be applied. A major challenge to this strategy lies in identifying biomarkers for PTE development that are unique from those that occur ubiquitously after TBI. Potential biomarkers hypothesized to predict which injuries will result in epileptogenesis include, for example, electrographic expression of high-frequency oscillations (HFOs) (32) or imaging of glucose metabolism or hippocampal surface shape (17). The predictive value of other changes that occur early after TBI and may be related to eventual epilepsy development (including measures of axon damage, blood or CSF markers of inflammation, or cognitive behavior) is difficult to assess as TBI often results in their expression independent of later epilepsy development. Retrospective analyses of potential biomarkers specific to PTE in patients is complicated by the fact that many outcomes associated with TBI are also associated with seizures, so the seizures themselves could induce the biological phenotype that is hypothesized to predict epileptogenesis. Animal models of PTE that can be tested at various times after injury—but before the first seizure—might be more effective in identifying candidate biomarkers of posttraumatic epileptogenesis. Studies focused on predictive biomarkers of PTE (versus TBI and seizure mechanisms) have been limited, but this is a promising area of future research.

Convergence of Cellular Changes in PTE and TLE Models

Convergent evidence from rodent models suggests several cellular changes that occur in both PTE and other acquired epilepsies, including status epilepticus–induced TLE. Understanding the mechanisms underlying these changes could, conceivably, prompt a mechanistic approach to identifying predictive biomarkers of PTE as well as the causative events that trigger epileptogenesis. Most PTE studies have focused on the hippocampal formation in TBI models because this structure is affected in TLE and often in PTE. Selective cell loss in the hilus, including a large percentage of somatostatinergic hilar inhibitory interneurons, is associated with development of both TBI- and status epilepticus–induced epilepsy in rodent models (28, 33, 34). A variable degree of mossy fiber sprouting and synaptic reorganization in the dentate gyrus, as well as axon sprouting onto hilar interneurons has also been observed in concert with development of PTE and other forms of acquired epilepsy (19, 22, 26, 29, 35, 36). Mossy fiber sprouting is localized to areas near the injury epicenter after focal injury (29) and is less severe and sparser after TBI than, for example, in chemoconvulsant models of TLE, but its presence, however limited, is a consistent feature of both models. Other cellular outcomes identified in both TLE and PTE models include an increase in postinjury granule cell progenitor proliferation and a reorganization of GABA_A receptors in granule cells (33, 37–40). Increased adult neurogenesis has been hypothesized to contribute to aberrant mossy fiber connectivity and an eventual increase in seizure susceptibility but may also contribute to cognitive recovery shortly after TBI (40, 41). Reorganization of GABA_A receptors might also serve a compensatory function, particularly since model- and location-dependent increases or decreases in receptor function have been observed (37–39). As for most other outcomes, cellular changes are correlated with epileptogenesis, but definitive studies to determine whether there is a causal role for these events in PTE development are lacking.

Even in the absence of identified PTE-specific biomarkers, recent studies have focused on preventing epileptogenesis or modifying disease progression after TBI, using cellular correlates of epilepsy or seizures as major outcome measures and comparing results to those obtained in other models of symptomatic epilepsy. Inhibition of the mechanistic (mammalian) target of rapamycin (mTOR) pathway has provided positive results in rodent models of both PTE and TLE (23, 28, 33, 42), although a more recent study found that mTOR inhibition suppressed some cellular correlates of TLE development (i.e., mossy fiber sprouting) but did not prevent seizures (43). In another example, JAK/STAT inhibition suppressed epileptogenesis in a TLE model (44), but not in a model of PTE (27). These and similar studies highlight both the promise for developing therapies to prevent posttraumatic epileptogenesis and also the difficulties in developing such treatments.

Conclusions

Epileptogenesis is highly amenable to study after TBI because the incident precipitating the development of epilepsy is defined, and the latency period between the injury and first seizure allows a potential window of opportunity to interfere with the processes that lead to increased seizure susceptibility. Current rodent models imperfectly replicate human PTE, but they recapitulate several cellular aspects of epilepsy and may be useful in identifying cellular mechanisms of posttraumatic epileptogenesis. Refinement and standardization of the models to make them more consistent should provide insight into the key mechanisms linking brain injury to epilepsy development. Given the variability of injuries after TBI and the relative nescience regarding the basic mechanisms underlying posttraumatic epileptogenesis, it seems unlikely that a single “magic bullet” preventative measure—or even identification of a single predictive biomarker—will be found soon. Developing treatments to prevent epileptogenesis after TBI, however, appears feasible and will likely be vitally informed by animal models of PTE.