Elsevier

Journal of Neuroscience Methods

Volume 272, 15 October 2016, Pages 50-55
Journal of Neuroscience Methods

Animal models of post-traumatic epilepsy

https://doi.org/10.1016/j.jneumeth.2016.03.023Get rights and content

Highlights

  • Post-traumatic epilepsy represents significant morbidity after brain injury.

  • Animal models include lateral fluid percussion, weight drop, & controlled cortical injury.

  • Each model has unique histopathological and behavioral characteristics.

Abstract

Post-traumatic epilepsy (PTE) is defined as the development of unprovoked seizures in a delayed fashion after traumatic brain injury (TBI). PTE lies at the intersection of two distinct fields of study, epilepsy and neurotrauma. TBI is associated with a myriad of both focal and diffuse anatomic injuries, and an ideal animal model of epilepsy after TBI must mimic the characteristics of human PTE. The three most commonly used models of TBI are lateral fluid percussion, controlled cortical injury, and weight drop. Much of what is known about PTE has resulted from use of these models. In this review, we describe the most commonly used animal models of TBI with special attention to their advantages and disadvantages with respect to their use as a model of PTE.

Introduction

Traumatic brain injury (TBI) is estimated to affect 1.7 million Americans per year and 3.2–5.3 million Americans live with long-term sequelae of TBI (Pearson et al., 2012). Post-traumatic epilepsy (PTE), defined as at least one unprovoked seizure occurring at least than one week after injury, has been shown to result in much worse functional outcome after TBI (Asikainen et al., 1999). The annual direct and indirect costs of TBI in the United States are estimated to cost roughly sixty billion dollars (Heise, 2010).

Seizures after TBI are classified according to the duration of latency before occurrence. Seizures occurring within 24 h are called immediate; seizures between 24 h and one week are early; and seizures after one week are late (Frey, 2003). The duration of latency is vitally important to prognosis. After the initial injury, also called the direct (or primary) injury, the indirect (or secondary) injury takes place as a series of biochemical and electrophysiologic process, some of which may lead to epilepsy. As a result, in contrast to most forms of epilepsy, PTE is possibly preventable during the time period of intervention if the processes that lead to epileptogenesis can be impacted. Patients with moderate to severe TBI are almost invariably seen by clinicians at the time of injury, so there is a potential opportunity to halt the secondary injury that leads to post-traumatic epilepsy.

Human PTE usually presents as spontaneous secondarily-generalized seizures in the weeks to months following TBI (Englander et al., 2003). For patients with mild to moderate TBI, the five-year incidence of PTE is roughly 1% (Annegers et al., 1998). In contrast, severe TBI is associated with an incidence of PTE 10% at five years and 16.7% at thirty years after TBI (Annegers et al., 1998). Risk factors for the development of PTE include the presence of intracranial abnormalities such as subdural hematoma or cerebral contusion (Annegers et al., 1998), particularly frontal and temporal contusions (Messori et al., 2005). Penetrating trauma is associated with a significantly increased risk of PTE (Englander et al., 2003).

There are currently no effective treatment options to prevent the development of PTE. A meta-analysis of randomized controlled trials concluded that prophylactic use of anti-epileptic drugs may be effective to prevent early seizures, but there is no evidence for prevention of late seizures and PTE (Schierhout and Roberts, 2001). However, it is likely that treatment modalities other than antiepileptic medication may reduce or eliminate the risk of PTE after severe TBI. Reliable animal models are essential to elucidate the pathophysiology of PTE and evaluate new treatment options.

Development of epilepsy after TBI in animal models reproduces the clinical features of human epilepsy. After the initial insult, there is a variable latency period after which recurrent seizures develop. It is during the latency period that epileptogenic changes are likely to occur. A variety of insults other than TBI are associated with this pattern of delayed epilepsy, including exposure of the brain to electrical stimulation or chemical toxins. In human PTE, the duration of the pre-epilepsy latency period varies widely but in most patients is less than 12 months (Englander et al., 2003).

This review will focus on the three most commonly used animal models of TBI used to evaluate PTE: lateral fluid percussion, weight drop, and controlled cortical injury. Many other models of TBI exist, but these are not used to study PTE because they do not reliably produce delayed recurrent spontaneous seizures and therefore have little pathophysiologic similarity to human PTE. Interestingly, despite the known incidence of PTE in humans following combat-related TBI, there have not yet been any reports of PTE after experimental blast injury (Kovacs et al., 2014).

Fluid percussion injury is by far the best-described and most widely studied animal model of PTE (Pitkanen et al., 2009). Variations of this technique have been described by multiple authors, but the most common modern use is based off the work by Dixon et al. (1987) and McIntosh et al. (1989) in the late 1980s. The technique involves two steps: (1) craniotomy for exposure of the dura and implantation of an adapter to transduce the energy, and (2) induction of injury using an energy wave propagated along a fluid column and focused on the exposed dura. For the first step, general anesthesia is induced and a small craniotomy is created in the parietal bone. The defect can be either midline or lateral to midline, and each produces a distinct pattern of deficits. The skull defect is fashioned to the same diameter of the tip of a syringe adapter, often a Luer-lock device. This adapter is then secured using methyl-methacrylate or a similar cement compound. For the second step, a specially-designed fluid percussion apparatus is used that consists of a pendulum that strikes a pressure transducer that consists of a container of sterile water (Fig. 1). The pressure wave is transmitted through this medium where it connects to the previously implanted adapter and creates a pressure wave on the surface of the dura. The severity of injury can be adjusted by increasing the pendulum force and titrated to either a specific pressure (often between 1–2 atmospheres) or titrated to produce a specific “righting reflex” time for recovery of consciousness after injury (often between 2–10 min for mild to moderate injuries). The injury is induced as they begin to awaken but before they are fully conscious. Control subjects usually undergo implantation but not injury.

One variation on this technique is the rapid lateral fluid percussion injury. This was first used by Goodrich et al. in 2013 and was later described in more detail (Hameed et al., 2014). For this variation, the Luer lock is tightly approximated to the dural surface and the animals are injured immediately without permanent implantation of the Luer lock. This reduces the total operative time, decreases exposure to anesthesia, and avoids the recovery period required after implantation in the traditional lateral fluid percussion technique. Although studies using this technique are limited, it appears to produce similar behavioral and histopathological findings as seen in the two stage lateral fluid percussion injury.

The lateral fluid percussion model has a number of strengths. The ability to precisely set the severity of injury affords repeatability across experiments and across different laboratories. The craniotomy is not technically difficult and, after some practice, can be performed in a rapid fashion. The pulse duration, magnitude of force, and relative mortality are similar to those of human TBI (Lindgren and Rinder, 1966). Fabrication of the fluid percussion apparatus requires attention to detail; however it uses inexpensive materials and off-the-shelf electronics. It shares histopathological similarities with human TBI, with evidence of local tissue damage with a surrounding area of inflammatory changes that may relate to the development of PTE. Behavioral and cognitive deficits appear to what is seen in human TBI, with difficulties with balance (Dixon et al., 1987) and multiple tests of cognition (Pierce et al., 1998). When severe injury is induced, cerebral contusions are formed which are known to be a risk factor for development of PTE. Also, the presence of pre-injury cortical dysplasia in rat model substantially increases the risk of developing PTE after fluid percussion injury, which implies that TBI may increase the risk of developing epilepsy in susceptible individuals (Nemes et al., 2016).

There are some limitations of the lateral fluid percussion model for the study of PTE. For unclear reasons, the severity of TBI required to produce PTE in rats is substantially higher than in humans: the level of injury required to produce PTE also leads to a mortality of roughly 30% (Kharatishvili et al., 2006). Also, even at relatively severe injury intensity, only a proportion of animals develop PTE after fluid percussion injury, so a large number of subjects is often required to produce enough epileptic animals to study. Another significant issue is the relative location of the hippocampus in rodents. In its original description (Dixon et al., 1987), the craniotomy and injury were located midline between lambda and bregma. The lateral fluid percussion (McIntosh et al., 1989) displaces the craniotomy laterally to the parietal bone to better mimic the rotational force vector that is likely pathophysiologically important in human TBI. However, as opposed to its relatively deep position in humans, the hippocampus in rodents is fairly superficial, which often leads to disproportionate structural injury to the hippocampal formation following TBI. Finally, the mechanism of injury may not be an accurate replication of human TBI since a craniotomy is performed for induction of injury directly to dura.

Models of post-traumatic epilepsy are generally mutually exclusive in experimental design. However, this should not be generalized to the assumption that traumatic and non-traumatic models of epilepsy are mutually exclusive in experimental design. The significant breadth and depth of research in non-traumatic animal models of epilepsy provide researchers with techniques to further probe animal models of post-traumatic epilepsy. An excellent example of this was work by Zanier et al. (2003). Given the well-established presence of significant glutamate currents following TBI, a mild TBI was induced with lateral fluid percussion and then kainic acid was administered an hour after injury. When compared to controls, this combination confirmed that even mild TBI creates significant vulnerability to a secondary injury, even if it is not a second traumatic injury. Injured animals exposed to kainic acid had decreased seizure threshold, greater cell death in CA3 (the most susceptible subfield to glutamate excitotoxicity), and worsened post-traumatic epilepsy. Given the well-established data on lateral fluid percussion, it provides an excellent model for combinations other epilepsy models as well as other models of secondary injury.

The impact-acceleration model of diffuse traumatic brain injury, also known as the weight drop model, was originally described by Marmarou in 1991. Following the induction of general anesthesia, the animal is positioned on a foam block with a cutout shaped to the animal’s head (Fig. 2). This setup is designed to provide a consistent position of the animal’s body and head relative to the weight drop apparatus. Depending on the desired severity of injury, a pre-selected weight is dropped from a pre-selected height using a release device. The weight strikes the animal’s head and not only causes a focal injury, but causes an acceleration-deceleration of the head into the foam apparatus.

The clear advantage of the weight drop model is its simplicity. The apparatus itself is easily fabricated using inexpensive supplies. Some researchers surgically expose the surface of the skull and, using cement, secure a metallic impact surface to control the exact location that force is applied. However, in the absence of this choice, an animal has to be anesthetized, carefully positioned in the apparatus, injured, and then recovered in a matter of minutes. Without the need for a craniotomy in a stereotactic frame, the throughput of this model is vastly superior. It also minimizes animal exposure to anesthetic agent.

While the weight drop model leads to many of the known sequelae of human TBI, it produces PTE only at very high intensities, at which a majority of animals do not survive. This is likely explained by the fact that the weight drop model produces a much more widespread and diffuse injury pattern with less focal cortical damage. Direct comparison with the lateral fluid percussion model shows significantly less focal injury patterns (Hallam et al., 2004), and this correlates with milder cognitive and behavioral deficits and absence of PTE.

Controlled cortical impact was first described by Lighthall in 1988. Similar to fluid percussion injury, this technique originated as a model of TBI that has been adapted for studying PTE. This technique consists of inducing general anesthesia and creating a small craniotomy in a location similar to that of the lateral fluid percussion injury model. A computer-controlled pneumatic impactor is then used to strike the dura directly (Fig. 3). Injury severity is worsened by increasing the depth that the impactor displaces cortex. The precise administration of force directly to the cortex mimics focal injury and removes a potential source of error with regard to the integrity and relative position of the adapter-dural interface in fluid percussion. This technique requires anesthetic to only be administered once, eliminating the potential neuroprotective influence of multiple inhaled anesthetics (Kitano et al., 2007). Given the relative simplicity of this method, it is easier to use this technique with different animal species and it has been performed in sheep (Lewis et al., 1996) and mice (Smith et al., 1995). The latter is clearly beneficial due to the ability to obtain knock-out mice to test hypotheses.

Controlled cortical impact closely resembles penetrating trauma and is much more likely to produce PTE than the other two techniques. Depending on the severity of injury, controlled cortical impact produces PTE in 5–40% of experimental subjects (Bolkvadze and Pitkanen, 2012, Hunt et al., 2009). Controlled cortical impact is also unique in that it frequently causes seizures in the so-called “early” window (1–7 days after TBI), (Bolkvadze and Pitkanen, 2012, Hunt et al., 2009, Nilsson et al., 1994) which is a common feature of human TBI but uncommonly seen in other animal models. The injury pattern in cortical impact is almost completely focal as opposed to mixed focal and diffuse injury pattern in fluid percussion and the mostly diffuse injury pattern in weight drop, so less severe overall injury is necessary to produce cortical disruption necessary for epileptogenesis. This difference likely highlights a key difference between the PTE and the other sequela of traumatic brain injury; focal cortical injury is much more likely to lead to development of epileptogenic foci whereas a more diffuse injury process likely underlies the non-epileptic changes after TBI.

One of the main limitations of animal models of PTE is that only a relatively small proportion of animals develop PTE. Non-traumatic animal models of epilepsy have a relatively high incidence of post-intervention epilepsy with reliable induction of multiple seizures per day. By contrast, experimental PTE is associated with a much lower seizure frequency and duration. In one study of PTE using continuous intracranial monitoring for 11 months after TBI, the average rate of seizures was one every 1–2 weeks (Kharatishvili et al., 2006). In another study of PTE using the lateral fluid percussion injury model, 94% of experimental animals showed electroencephalogram (EEG) evidence of non-convulsive seizures, but these generally lasted less than ten seconds (Campbell et al., 2014). Unlike many other experimental models of epilepsy, experimental PTE often leads to generalized epilepsy that is relatively easy to evaluate using video monitoring to establish frequency and severity of seizure activity (Racine, 1972).

Another potential limitation of animal models is the location of the temporal lobe in rodents. The relatively mesial location of the hippocampus in humans means that it is potentially protected in comparison to animal models, where it is significantly more superficial. Although the significance of this anatomical difference is difficult to accurately define, it certainly raises a concern when applying a traumatic force. While less frequently subjected to direct trauma in humans, the impact of the temporal tip onto the sphenoid wing frequently causes contusions of the anterior temporal tip. While certainly different from a pathophysiologic perspective when compared to the focal injuries seen in animal models, it should be noted that a similar pathology is often seen in severe human TBI.

Section snippets

Conclusions

The three most commonly used animal models of TBI provide researchers with experimental models that replicate different features of human PTE. The best studied is the lateral fluid percussion injury model, although significant severity is required and only a minority of animals develop PTE, so a large number of animals are generally required. The weight drop model produces an injury even more diffuse than fluid percussion and PTE is therefore rarely induced. The controlled cortical impact model

References (30)

  • J.N. Campbell et al.

    Traumatic brain injury causes a tacrolimus-sensitive increase in non-convulsive seizures in a rat model of post-traumatic epilepsy

    Int. J. Neurol. Brain Disord.

    (2014)
  • C.E. Dixon et al.

    A fluid percussion model of experimental brain injury in the rat

    J. Neurosurg.

    (1987)
  • L.C. Frey

    Epidemiology of posttraumatic epilepsy: a critical review

    Epilepsia

    (2003)
  • G.S. Goodrich et al.

    Ceftriaxone treatment after traumatic brain injury restores expression of the glutamate transporter, GLT-1, reduces regional gliosis, and reduces post-traumatic seizures in the rat

    J. Neurotrauma

    (2013)
  • T.M. Hallam et al.

    Comparison of behavioral deficits and acute neuronal degeneration in rat lateral fluid percussion and weight-drop brain injury models

    J. Neurotrauma

    (2004)
  • Cited by (27)

    • Contusion brain damage in mice for modelling of post-traumatic epilepsy with contralateral hippocampus sclerosis: Comprehensive and longitudinal characterization of spontaneous seizures, neuropathology, and neuropsychiatric comorbidities

      2022, Experimental Neurology
      Citation Excerpt :

      Animal models of TBI have been used for decades to better understand how the underlying mechanisms of brain injury can trigger PTE. The controlled cortical injury (CCI), fluid percussion injury (FPI), and weight-drop models are the most used rodent models of TBI (Ostergard et al., 2016). There are advantages to each of these models.

    • Knowledge atlas of post-traumatic epilepsy research: Based on citespace visualization analysis

      2021, Epilepsy Research
      Citation Excerpt :

      Moreover, as PTE research models are diverse, the mechanisms are also varied, which further alters their focuses. Thus, examining whether the evidence from different animal models could be applied to different common causes of epilepsy after acute injury in humans and the translational research of epileptic prevention is imperative (Klein et al., 2018; Ostergard et al., 2016). Interestingly, Rodgers et al. detected the same spike-wave discharge (SWD) in FPI epilepsy model and normal rats, which strengthens the connection between acquired epilepsy and primary hereditary epilepsy and validates the FPI model of acquired epilepsy (Rodgers et al., 2015).

    • Animal models of post-traumatic epilepsy and their neurobehavioral comorbidities

      2021, Seizure
      Citation Excerpt :

      Even though this model leads to many of the known features of human TBI, it only produces PTE at very high intensities and survivability of the experimental animal is low. This is probably due to the diffuse injury produced by the weight-drop model instead of a focal cortical damage [113]. Rodents are deeply anesthetized, placed into a stereotaxic frame, and the skull is exposed with or without a craniectomy (closed-head injury).

    • Complications of traumatic brain injury

      2020, Brain Injury Medicine: Board Review
    View all citing articles on Scopus
    View full text