Key words

1 Introduction

The central nervous system (CNS) is the most delicate and complex part of the body. The definition for the CNS is based on the glial cells that surround and support neurons. Both the brain and the spinal cord are included in the CNS, since these parts of the nervous system comprise astrocytes and oligodendrocytes. Spinal nerve roots and peripheral nerves contain Schwann cells and are therefore regarded as parts of the peripheral nervous system (PNS).

Physical trauma in the central nervous system is often the result of a number of forces in different directions and dimensions. The timeline for the interaction between the acting forces and the nervous tissue is usually very narrow and it can therefore be very difficult to assess the influence of each contributing force. In addition, human patients, which come in different ages and sizes, may have a number of additional injuries or diseases as well a genetic background that may influence the outcome of the injury.

1.1 Spinal Cord Injury (SCI)

It has been known since ancient civilizations, such as the Pharaonic Egypt, that spinal cord injury (SCI) leads to chronic loss of functions, such as paralysis. The complex cellular response and lack of successful regenerative growth in CNS was described in great detail by Ramón y Cajal [1]. Although a substantial amount of research on SCI has been conducted, there has been only limited success in the identification of new strategies to improve the regenerative response after SCI. Models for SCI usually have a more regenerative focus than the models for brain injury.

1.2 Traumatic Brain Injury (TBI)

TBI is a leading cause for death and disability in both civilian life and at the battlefield. The signature TBI has evidently changed from penetrating TBI to blast-induced TBI in recent military conflicts. All types of TBI may occur in both the civilian and military setting. However, there are some important considerations for military TBI, mostly relating to the extreme energy transfer.

  1. 1.

    TBI can be graded from mild to severe and primary or secondary. TBI can be classified as diffuse or focal, and very often in a complex mixture.

    1. (a)

      Mild TBI: A TBI is often classified as mild (concussion or commotion) if loss of consciousness or confusion is shorter than 30 min. MRI (magnetic resonance imaging) and CT (computer tomography) scans are usually normal but the patient may have headache, cognitive problems (memory problems, mood disturbance, attention deficits), and the effect on the patient can be devastating. The majority of blast-induced TBI fall into this category, although the pathophysiology is largely unknown. Cerebral concussion is often associated with other types of brain injury.

    2. (b)

      Moderate and Severe TBI: These injuries can be divided into closed head injuries and penetrating injuries. Closed head injuries may be both diffuse and focal. The penetrating TBI will always induce a focal injury and often diffuse secondary injuries.

    3. (c)

      Diffuse TBI: The most common diffuse injury is the diffuse axonal injury (DAI), which is defined as the presence of diffuse damage to axons in the cerebral subcortical parasagittal white matter , corpus callosum, brain stem, and cerebellum. This is usually the result of an acceleration/deceleration trauma. Different parts of the brain move at different speeds because of their relative density. If this is a rotational trauma, the positions of the axis of rotation will be an important factor in the injury mechanism and areas at a greater distance from this axis will sustain larger forces. This can lead to shearing injury and DAI. Beta-amyloid precursor protein (APP) has been proven to be an excellent marker for axonal injury in histology. Modern imaging techniques, such as MRI with DTI (diffusion tensor imaging), have provided improved possibilities to detect DAI.

    4. (d)

      Focal TBI: Focal brain injuries include contusions and penetrating TBI. Focal TBI is usually the result of bullet and shrapnel entrance in the skull cavity or a contact between the head and a sharp or semi-blunt object that results in skull fracture. Also acceleration/deceleration trauma can cause focal injuries to the brain.

    5. (e)

      Secondary traumatic brain damage occurs as a complication of the different types of primary TBI and includes ischemic and hypoxic damage, swelling, raised intracranial pressure, and infection. The onset of these complications varies from minutes to hours. The secondary TBI is potentially reversible with adequate treatment. Axonal damage in both DAI and focal injuries interferes with axoplasmic transport. Severe traumatic injury results in primary axonal disruption or transection termed “primary axotomy” or sets in train a series of ill-understood events culminating in secondary axonal degeneration or secondary axotomy. Thus, adequate treatment could probably limit the axonal damage.

  2. 2.

    Experimental models for studies of CNS injuries

    Experimental research is a strategy to take control over confounding factors that may obscure data from real life clinical CNS injuries.

  3. 3.

    Experimental models may serve many different purposes, such as

    1. (a)

      To test new treatments for injuries.

    2. (b)

      To provide a better understanding of injury mechanisms .

    3. (c)

      To reveal thresholds for injuries.

    4. (d)

      To develop test tools and methods for clinical settings. For example to develop protocols for MRI, new biomarkers, protocols for EEG.

1.3 Model requirements

It is difficult to provide a complete list of all different models that are employed in neuroscience. We will limit the description to examples of models for traumatic injuries and therefore exclude models for stroke, neurodegenerative diseases, infections, etc. It is usually important to test if the data from the employed model can be translated and provide any conclusions for the condition that is supposed to mimic. It is also important to enable verification of any data that can be harvested from the model. Thus, a good monitoring and documentation of physical data should be provided from all new models. The model needs to be both validated and verified. Thus, to show that the model can replicate exactly what it is suggested to replicate and that the experiments can be repeated by both the scientists that have developed the model and by any other scientist that has the proper tools. This is a highly complex process, since the methods for measurement also need to be validated and verified. It can be supposed that many of the failures in translating new treatments from experimental setups to clinical practice may be due to mistakes in verification and validation of the model or the experimental setup. It has also been said that a lot of published experimental data cannot be repeated, possibly due to improper validation and description of the experimental setup.

1.4 Obstacles for Translation

  1. 1.

    There are a number of more or less obvious factors that may impede translation between an experimental model and the clinical situation—even if the model has been properly characterized:

    1. (a)

      Trauma in the clinical situation is the result of many complex forces, delivered during a short time. It is therefore extremely difficult to produce a model that can in all details simulate a clinical situation.

    2. (b)

      Severe trauma to the head or spine is usually combined with other injuries, thus a multitrauma situation that can modify the outcome of the TBI or spinal cord injury by for example a systemic inflammation, infection, and hypoxia. The importance of such contributing factors can be tested in the experimental situation, but will still provide an obstacle in translation between experiments and the clinical populations.

    3. (c)

      The time scale for biological response in experimental animals, such as rodents, is probably very different than that in humans.

    4. (d)

      The genetic variation in rodent strains is usually very limited compared to clinical populations. It is therefore possible that some results are over-amplified in the experimental situation and may not at all be relevant in a clinical situation.

2 Experimental Models of Brain Neurotrauma

Below are a set of different experimental models of brain trauma. For each model, we will try to describe the rationale (why the model was developed), the development, results obtained, and limitations with the model. We will also describe the work with validation, verification, and translation. In each section, we also give examples of alternative useful and validated models.

2.1 Spinal Injury, with Focus on Motor Axons

The pioneering studies of Ramón y Cajal demonstrated that injured neurons in the central nervous system (CNS) have a strong capacity to extend new axons into the peripheral nervous system (PNS) [1], he showed that motor axons that had been severed due to spinal cord injury had the capacity to regrow to neighboring ventral roots. The observation that neurons in the CNS could extend axons into PNS tissue was reinvestigated by Richardson and colleagues [2].

2.1.1 A Lesion in the Ventral Funiculus of the Spinal Cord

With background from the observations that motoneurons in adult cats survived after peripheral nerve injuries to a much better degree than sensory neurons, it was decided to analyze if motoneurons would survive also after very proximal injuries. The most proximal location would then be to cut the axons inside the spinal cord, but outside the motor nuclei in the grey matter. The selected site was the ventral funiculus, i.e. the white matter between the ventral horn and the transition to the spinal ventral nerve root. The experiments were performed in adult cats that were subjected to a lumbar laminectomy. The denticulte ligament at the side of the lumbosacral enlargement was divided and used to gently twist the spinal cord so that the junction between the spinal cord and the ventral root was visible. A thin fragment of a razor blade (or a specialized knife for micro-surgery) was pushed into the lateral aspect of the spinal cord segment L7 and moved longitudinally to produce a selective lesion in the ventral funiculus. The aim was to cut the motor axons at this location, without damage to the grey matter (where the cell bodies reside) or the connection between the spinal cord and the nerve root. It was observed that, although the animals were deeply anesthetized, the muscles in the hindlimb corresponding to the L7 segment (the L7 myotome) responded with contractions during this procedure. The erector spinae muscle, fascia, and skin were closed with sutures in separate layers. The animals were sacrificed after survival times from 1 to around 300 days and fixed by intravascular fixation with glutaraldehyde. Using electron microscopy, it was revealed that the lesion was located in the white matter and had spared both the grey matter and the nerve root exit zone. A large proportion of the motoneurons actually had survived the lesion (Fig. 1).

Fig. 1
figure 1

A schematic representation of the ventral funiculus lesion model. The lesion area in the ventral funiculus of the spinal cord is penetrated by regrowing motor axons, surrounded by a thin sheet or glial tissue. The extracellular room is expanded and contains large amounts of matrix molecules such as laminin and collagen, the blood–brain barrier has a more or less permanent defect

Full size image

It was shown with electron microscopy and intracellular labeling with horseradish peroxidase that axons from such surviving motoneurons penetrated the scar tissue in the ventral funiculus of the spinal cord and entered the ventral roots [3–5]. The regrowing axons penetrate through a highly unusual CNS environment which lacks a blood–brain barrier (BBB) function [6] but possesses a high content of cells bearing neurotrophin receptors [7–9] and matrix molecules [10, 11].

2.1.2 Avulsion Injuries: The Clinical Use of the Observed Regenerative Response

A clinical counterpart to this injury is a ventral root avulsion at the border between the CNS and PNS, typically caused by a high-energy trauma such as a motorbike accident causing excessive trauma to the shoulder and head resulting in stretching and rupturing of ventral roots. Ventral root avulsion is not followed by spontaneous regrowth, since the avulsed roots are widely separated from the spinal cord inside the subarachnoid space or even pulled to a position outside the vertebral channel. Replantation of avulsed spinal ventral roots into the spinal cord (Fig. 2) has been shown to enable significant and useful regrowth of motor axons in both experimental animals and human clinical cases [12–15]. Gene expression array studies show that replantation of the ventral root can initiate a rapid upregulation in the expression of genes associated with neurite growth, in contrast to neurons subjected to avulsion only [16]. Thus, the replantation can influence gene programs in the spinal cord.

Fig. 2
figure 2

A schematic representation of replantation of avulsed ventral roots. The ventral roots are avulsed at the border between CNS and PNS (spinal cord surface). The rootlets are then replanted into the ventro-lateral surface of the spinal cord. Regrowth into the ventral root can occur both directly within the spinal cord and by growth in the pia mater. After reaching the ventral root, motor axons can continue and reinnervate denervated muscle

Full size image

The results from such treatment in humans are less successful in older patients and good reinnervation in distal muscles like in the hand is seldom possible. In a recent case report, Carlstedt et al. [14] described a preadolescent boy with complete brachial plexus avulsion injury that was treated by replantation of five ventral roots. Shoulder muscle recovery started 8–10 months after the spinal cord operation. At 12–15 months, elbow function began to recover, followed 2 years postoperatively by forearm, wrist, and intrinsic hand muscle activity. This led to recovery of hand motor function without restoration of sensation. Bilateral motor cortex activity and activation of the sensory cortex on use of the affected hand was demonstrated by fMRI . The findings of that case study suggested that the restored hand function might rely on cortical sensory programs established before the injury. Although the time needed for recovery seems very long, it cannot be excluded that the final result is dependent on several biological programs that are elicited in the early acute stage.

The method for producing nerve root injuries has a significant role for the outcome of the injury. Rhizotomy (division of the root) leads to a milder representation of the spinal cord trauma that occurs after “true” avulsion injury [17]. Shortland and colleagues employ a model for extra-vertebral avulsion (without need for laminectomy) [18]. However, for replantation surgery laminectomy is necessary. One limitation with most animal studies for replantation is that the injury and the replantation surgery are performed at the same time, whereas delayed surgery is the typical clinical situation.

This is to our knowledge, the only example of an experimental model that has been translated into practical clinical use for treatment of a traumatic injury that involves axonal growth in the spinal cord. Many models for SCI seem to simplifications of the usual clinical type of SCI. Maybe it would be of benefit to develop models for spinal cord contusion that could represent a better representation of actual clinical situations, with high-energy transfer, rapid deformation of vertebra and incomplete injury to the spinal cord.

2.2 TBI in the Military Setting: Modeling Primary Blast TBI

The use of Improvised Explosive Devices (IED) in contemporary warfare has changed the scene and spectrum of TBI at the battlefield. At the same time, new equipment for body protection has increased the survival rate after TBI at the battlefield. It may be assumed that several mechanisms contribute to the injury. TBI has been identified as major health problem in military personnel returning from service. The injuries range from severe multitrauma to a number of mild TBI that still has to be settled. Propagation of blast waves is very complex. It could involve both direct propagation through the skull and indirect propagation via blood vessels and it is obvious that a systemic response comes along with the blast TBI. It is difficult to identify a reliable borderline between mild blast TBI and posttraumatic stress syndrome (PTSD). Many of the symptoms are similar and many patients might suffer from both TBI and PTSD. The energy transfer during an explosion is extremely complex and it is difficult to predict how energy is propagated into the body and absorbed. Exposure data from clinical situations at the battlefield are usually lacking. One way to generate a better understanding of the mechanisms of TBI after a blast exposure is to perform controlled experiments in animals. Many clinical TBI cases have a complex mixture of diffuse and focal injuries, which are complicated by secondary injury events. Individuals exposed to a blast often suffer from multiple injuries, i.e., pulmonary lesions or amputations. Severe blast-related TBI with brain edema and vascular spasm [19] can be assumed to be the result of a combination of more than one injury mechanism (see Fig. 3).

Fig. 3
figure 3

A schematic representation of different mechanisms for blast-induced neurotrauma. Previously published in [94]

Full size image
  1. 1.

    One way to understand the effects of a blast wave is to divide the mechanism into

    1. (a)

      Primary blast: Effects of the primary blast wave, thus the propagation of a supersonic pressure transient with short duration. The threshold for injuries is determined by factors such as peak pressure, duration, and shape of the wave.

    2. (b)

      Secondary effects of blast, i.e., due to the impact of flying objects, such as shrapnel fragments, which can generate penetrating injuries. The proportion of such injuries was larger in previous conflicts and seems to have been reduced by improvements in helmet construction.

    3. (c)

      Tertiary effects of blast, i.e., the result of acceleration movements, which may result in tissue shearing and diffuse injuries, such as diffuse axonal injuries (DAI).

    4. (d)

      Quaternary effects of blast, the result of heat, smoke, or emission of electromagnetic pulses (EMP).

  2. 2.

    Open field exposure

    Large-scale experiments during the 1950s in desert areas and ponds generated fundamental data for effects of blast with simple wave forms. Large numbers of animals of different species were subjected to open field exposure to blast. These experiments determined thresholds for mortality and injuries such as bleeding in air-filled organs such as the lungs and intestines. The potential effects on the central nervous system were generally not assessed. These experiments provided the Friedländer type of wave and dose–response curves (the Bowen curves) were determined [20–22]. Large amounts of explosives are needed and dosimetry can be difficult. Good control of the physiology of the experimental animals and proper tissue collection is usually not possible in outdoor conditions in combination with a large number of animals in a single experiment. However, open field experiments may allow for more realistic experiments with large animals that are more similar in size to humans. It also makes it possible to use waveforms relevant for simulation of IED, for example reflection from the ground or vehicles. Examples of new models with open field exposures aimed to produce mild TBI include the Combat Zone-like blast scenery for mice [23] and a primate model [24].

  3. 3.

    Shock tubes with compressed air or gas

    Shock tubes employ compressed air or gas, rather than explosives. Systems with compressed air were used already in the 1950s [25]. Most systems comprise two chambers, separated by a membrane (Fig. 4). Compressed gas is loaded into one of these chambers (the overpressure chamber or the driver section), which is separated from the other chamber, referred to as the main section or the driven section, by a diaphragm. The main section is usually several meters long. The object, i.e., the experimental animal, is positioned somewhere in the main section and the operator can rupture the diaphragm. The compressed gas enters the main section and simulates a propagating blast wave. One advantage associated with this type of shock tube is the absence of quaternary blast effects as well as other disadvantages of explosives. There are a number of modifications of the shock tube design and there seems to be a need to calibrate the different systems. Well-documented modern shock tubes can for instance be found at the Walter Reed institute [26], Wayne State University [27], and the US Naval Medical Research Center (Maryland, Silver Spring, USA) [28, 29]. One sophisticated shock tube system has been installed at the Applied Physics Laboratory at Johns Hopkins University [30]. This is a modular, multi-chamber shock tube capable of reproducing complex shock wave signatures. The instrumentation should allow direct measurement and calculation of the various shock loading characteristics, including static pressure , total pressure, and overpressure impulse.

    Fig. 4
    figure 4

    A schematic representation of a typical over pressure shock tube composed of two chambers separated by a membrane. Compressed air/gas in the driver section is released as a pressure wave to the other (driven) compartment by the controlled puncture of the membrane. Previously published in [94]

    Full size image
  4. 4.

    A Swedish Blast tube for studies on primary blast

    In 1949, Carl-Johan Clemedson published a thesis titled “An experimental study on air blast injuries” in 1949 at the university in Uppsala [31]. He continued his research at FOA (Swedish Defence Research Establishment) using compressed air [25] and finally a newly constructed blast tube [32] in which a charge of plastic explosive (pentaerythritol tetranitrate PETN) was used. The system was composed of a cylindrical 400 mm wide and 565 mm long cast iron tube, with a cone shaped tip where the charge is placed. The wall thickness at the tip is 100 mm. At the other end, the iron tube is elongated to a total length by an extension tube of steel (wall thickness 10 mm). The open end of the extension tube was by a steel disc connected to a pendulum in Clemedson’s initial experiments. At the conical end of the tube, there is a threaded screw plug for the insertion of the charge that is mounted on the detonator. The charge is fixed at a distance of 100 mm from the inner tip. Clemedson and his coworkers published a number of studies on muscle tissue [33] as well as vascular and respiratory effects of blast in the rabbit [34]. After some time, this work was extended to include the central nervous system [35] and the cerebral vasculature [36]. The blast tube (without the pendulum door—which results in a more complex waveform) was modified for work with rodents by Anders Suneson. Annette Säljö used this system in her thesis work [37–39]. The ignition system was later changed to a non-electric ignition (NONEL from Nobel). The anesthetized rat is mounted in the blast tube at a distance of 1 m from the charge. The PETN generates a pulse with rapid raise time and very short duration (<0.3 ms). An increase in the charge has a direct and proportional effect on the peak pressure, but only a small effect on the duration. Secondary reflections are limited. By moving the object to the orifice of the tube (1.5 m from the charge) the peak pressure drops to about half, whereas the duration is increased somewhat. All rat experiments have been performed at the 1 m distance from the charge (Fig. 5), which has been varied from 0.5 to 5 g PETN (Spherical charges of Swedish army plastic explosive m/46, containing 86 % pentaerythritol tetranitrate PETN and mineral oil, where used. In the following text, the weights of the plastic explosive charges are given in gram). We have previously reported that the peak pressure during detonation of 2.5 g PETN would be 260 kPa and nearly 600 kPa with a 5 g charge. To get representative recordings of the peak pressure during a detonation of explosives is connected to substantial methodological difficulties due to the extreme requirement for good dynamics. Recent recordings with a set of open silicon piezoresistive sub-miniature pressure transducers (Entran Sensors & Electronics) indicate a peak pressure exceeding 10 bar during detonation of 5 g PETN, as recorded in front of a rat dummy at the 1 m position in the tube. A charge of 2.5 g in an unprotected animal (side-on at 1 m from the charge) results in about 50 % lethality and at 3 g the majority of the exposed animals die from pulmonary bleedings. On the other hand, the lethality is less than 10 % at 2 g PETN. In recent experiments we have used a full protection (rigid steel tube) of the body, except for the head and had no lethality with 5 g charges.

    Fig. 5
    figure 5

    Pictures showing the blast tube that was constructed by the Swedish scientist Clemedson in the 1950s. This system may be one of the oldest systems that still are in use. Previously published in [94]

    Full size image

    In spite of the substantial lethality with this system (in unprotected animals), structural changes in the brain seem very limited or absent. The cell death in the brain that was reported by Säljö and coworkers [39], employing a similar protocol, has not been possible to verify. Instead, there is a significant cell death and gene expression changes [40, 41] in the inner ear. The hippocampus of the brain of rats exposed to a 2 g charge has been subjected to examination of the gene expression by use of Affymetrix Rat Gene Arrays. In such experiments, the expression of all known genes can be evaluated at the same time. It was found that about 100 genes had a significant expression change 24 h survival after the blast. The majority (76 genes) had a decrease in their expression whereas an increased expression was found in 39 genes [42]. Affected systems include synaptic transmission and neurogenesis. Both unprotected rats exposed to detonation of 2 g PETN and rats with the torso protected by a rigid steel tube during exposure to a 5 g blast have been examined. Taken together, the lack of structural damage combined with some functional changes indicates that the primary blast in this model generates a mild TBI. Future work with the blast tube will include repeated exposures and variations in body position. The blast tube in the current setting creates a very simple and short pulse that may imitate the situation at short distance with an open field charge. This pulse form is probably not at all similar to the situation in a protected vehicle that is hit by a road bomb. It should therefore be of interest to modify the length of the blast tube and introduce reflecting objects that could create a more complex pulse form. One aspect with this type of blast tube is the heat and gas emission that is a result of the use of explosives instead of compressed air. Thus, quaternary blast is added to the primary blast. Acceleration movements are however limited, by the montage of the animal, and impacts of flying objects do not occur. Secondary and tertiary blast is instead imitated and better controlled in other models.

  5. 5.

    Breacher studies

    “Breachers” are military or law enforcement personnel that are routinely exposed to low-level blast during training. Such repeated exposure has been associated with symptoms similar to that of sports concussion. The physical parameters of the blast exposure can be monitored during training in a much better way than in battle-field situations. This creates a situation that is similar to an experimental model and collection of samples, such as biomarkers [43], can therefore be compared to animal experiments for primary blast.

3 Models for Acceleration TBI

Although TBI can be associated with skull fractures, it commonly occurs without fractures [44]. About 40 % of all TBI patients admitted to hospitals are non-focal injuries [45] and are usually referred to as distributed brain injuries (DBI).

At least four categories of DBI can be identified: diffuse axonal injury (DAI) ; diffuse hypoxic, anoxic, or ischemic injury; diffuse swelling; and diffuse vascular injury. DAI is the most common type of DBI and commonly results in unconsciousness or death [44, 46]. The DAI pathology, which is characterized by perturbations to the axoplasmic transport along the length of axons [47], is likely to cause axonal swelling or degeneration which can reduce the functionality or disconnect the axons from their existing networks [48]. It has been reported that DAI commonly are localized in the subcortical white matter, grey–white matter interface and corpus callosum [44, 49], and at points of attachment, such as cranial nerves [50].

DBI is commonly a result of inertial induced loads; intracranial motions arise when the skull is accelerated and the brain mass, due to its inertia, lags behind or continues its motion relative to the skull. These inertia-induced loads are most common in rapid head rotations [51] which often occur in fall accidents, traffic accidents, and military assaults. Mechanical and mathematical models have been used to show that these inertia-induced loads produce strains in the brain tissue and that these strains cause the following neurological deficiencies [48, 52–58].

An attractive approach to study DBI pathology and its associated injury mechanism and threshold would be to reconstruct well-documented accident cases in which the patient is slightly injured. The real life accidents are however commonly rather complex, the injuries are regularly sever, and the patient suffers from a multitude of injuries. Therefore, anesthetized animals have been used in the past, and are in use, to study DBI and DAI .

The rotational weight drop model that [59] was developed by Marmarou and coworkers [60, 61] has generated very important data on development of diffuse brain injuries, including an improved understanding of diffuse axonal injury [47]. One of the advantages/disadvantages with this model is that it combines DAI with a contusion injury. Andersen [62] developed a model in which a stunner was used to accelerate the head. Both sever local brain damage and axonal injury was produced by this model. Both the forth-mentioned models have been found to be less useful for threshold studies on DAI and when DAI is to be studied separately.

Several models were developed in the past to study DAI using primates [44, 63], miniature swine [59], and rabbits [64, 65]. Few of these are in use today due to ethical considerations, or a lack of appropriate methods to assess the effects of the trauma, or due to excessive meningeal bleedings.

Models developed for rats provide several advantages over models using large animals. Ellingson et al. [66] and Fijalkowski et al. [67] exposed rats at low to medium severity rotational accelerations in the coronal plane. Despite the higher accelerations, the rats suffered from classical concussion injuries with minimal histological abnormalities.

3.1 A Rotational Acceleration Model

A rotational acceleration model aimed for studies into diffuse brain injuries and mTBI is described in detail by Davidsson [68]. It is designed to produce brain injuries by sagittal plane rearward rotational acceleration. This model allows for the production of graded injury and studies into injury thresholds.

In brief, the skull of an anesthetized adult Sprague–Dawley rat, weighing from 280 to 450 g, is tightly secured to a rotating bar. The bar is impacted by a striker that causes the bar and the animal head to rotate rearward; the acceleration phase lasts 0.4 ms and is followed by a rotation at constant speed and a gentle deceleration when the bar makes contact with a padded stop (Fig. 6). By adjusting the air pressure in the rifle used to accelerate the striker, rotational acceleration between 0.3 and 2.1 Mrad/s2 can be produced.

Fig. 6
figure 6

Diagrams. Top row: skull cap and attachments plate and oblique view of test device (crossbeam removed for visibility). Bottom row: oblique view of test device and side view of the head with central nervous system schematically depicted (screws and accelerometer not depicted) (Previously published in [94])

Full size image

Numerous combinations of trauma levels, posttrauma survival times, brain and serum retrieval, and tissue preparation techniques were adopted to characterize this model. The trauma cause subdural bleedings in animals exposed to severe trauma although such hematoma can occasionally be found at rather low acceleration levels [69]. Staining brain tissue with amyloid precursor protein (APP) antibodies, which expose axons with reduced axioplasmic transport, revealed widespread axonal injuries (AI) in the frontal region of the corpus callosum, the upper and lower borders of the corpus callosum, and sometimes stretched into the caudate putamen. In the mid brain region APP-stained axons were found in the same regions as well as in the thalamus and hippocampus regions. Further, APP-stained axons were found in some of the tracts in the brain stem. Staining brain tissue with FD Neurosilver that detects degenerating axons confirmed the injury pattern [69]. For older subjects, those weighing on average 700 g, fewer positive axons appeared in the corpus callosum in animals while injured axons were found in clusters elsewhere: in the internal capsules, inside and in the vicinity of the anterior commissure, and in the structures between the lateral ventricles [70]. In addition to these areas, positive axons were spread out and frequent in structures close to the centerlines and the skull bases. The observed AI was apparent only when the rotational acceleration level was moderate (1.1 Mrad/s2) and above [71]. Older animals required higher accelerations to exhibit similar injury levels as younger animals [70]. Only limited signs of contusion injury were observed following trauma. Macrophage invasions, glial fibrillary acidic protein redistribution or hypertrophy, and blood–brain barrier (BBB) changes were unusual. S100B serum analyses indicate that blood vessel, axonal, and glial cell injuries occur following moderate level of trauma despite absence of obvious BBB injuries [72]. Affymetrix gene arrays showed changes in the expression in a large number of gene families including cell death, inflammation, and neurotransmitters in the hippocampus 24 h after moderate to severe trauma [42].

In conclusion, the signature injury with this rotational model is diffuse axonal injuries in the corpus callosum, subcortical white matter , and the brain stem. The absence of cell death and excessive bleedings indicate that this is a mild TBI and effects on behavior are indeed limited [72]. Thus, this model can add knowledge about mechanisms and thresholds for acceleration-induced mild TBI and such data can be relevant for the understanding of consequences of tertiary blast.

4 Models for Penetrating Ballistic TBI

A penetrating ballistic TBI occurs when objects impact the head and penetrate skin, skull and meninges and cause injury directly to the brain tissue . In contrast to closed head injury, penetrating ballistic TBI involves direct laceration of brain tissue, often complicated by secondary effects such as hemorrhage, edema, inflammation, and higher risk of coagulopathy [73]. The Vietnam head injury study has generated a lot of extremely valuable data on the long-term effects of penetrating TBI [74–76]. The Vietnam head injury study included baseline (pre-injury) data and a fairly homogenous group of patients (in terms of age and type of injury). Thus, interesting possibilities for translation emerge if this type of injury can be properly represented in experimental models.

A number of models have been developed to study penetrating ballistic TBI using cat [77], dog [78], monkey [79], and sheep [80]. None of these models are currently in routine use. To date, there are only two models in use and these both use rodents. In one of these a designed probe is inserted in the brain at the desired location and rapid inflation of an attached balloon is used to mimic the temporary cavity caused by energy dissipation from the bullet [81]. In the other model, a probe penetrates the brain at high velocity [82]. When shooting this probe into gel samples large temporary cavities are formed that are correlated to both speed of penetration and the shape of the penetrating object. The latter model will be presented in greater detail below.

4.1 Penetration TBI at High Velocity

A midline incision was made through the skin and periosteum of the anesthetized rat head, and a burr hole of 2.75 mm in diameter was drilled with its center slightly lateral and posterior to bregma. The rat was thereafter placed in a stereotactic frame and positioned so that a probe, commonly 2 mm in diameter with a spherical tip, was positioned directly above the dura exposed by the burr hole. This probe was fitted into a holder and guided by a narrow tube (Fig. 7). A lead pellet was accelerated by air pressure from a specially designed air rifle and impacted the probe. The probe penetrates 5–6 mm into the brain at high speed; 90 m/s at the time it starts to enter the brain.

Fig. 7
figure 7

The penetration rig (Previously published in [94])

Full size image

The injury caused severe damage to the lateral and medial parietal cortecies, corpus callosum, hippocampus, and several parts of the posterior thalamus [82]. After a few days after trauma, a large cavity was formed in the brain. The penetration trauma also causes hemorrhages, blood–brain barrier breakdown, neurodegenerations , gliosis, transient changes in several behavior tests, and persistent deficiency of reference memory. There are significant changes in gene expression, both in the cerebral cortex that surrounds the penetration and in the hippocampus [42] and leakage of S-100B into serum [83]. The inflammatory response in the cortex includes an activation of the terminal pathway of the complement system [84]. This model has also been used to characterize changes in BDNF and neurotrophin receptors after penetrating injury [85] in order to provide a correlate to data on the importance of BDNF gene polymorphism for the outcome of penetrating TBI in the Vietnam Head Injury study [86].

5 Concluding Remarks

Far to many experimental studies end up with data that cannot be translated into clinical work. Translation can fail if the models are not properly validated and provide a good representation of the clinical injury they are supposed to model. Translation can also be impeded by large differences in the methods for outcome measure [87]. In order to facilitate the translation of experimental data, it may be important to include outcome recordings that could be used also in clinical work, such as serological biomarkers and MRI [88]. One additional strategy is to develop and use Finite Element Models of the animal and the model that is used to load the animal during the experiment. One model for the rat head has recently been employed to model rotational acceleration injury [89]. Such work can enable scaling and comparison with human clinical cases of TBI [56, 90, 91]. The choice of the animal species or strain can obviously have a significant impact on the outcome of the injury. Differences in body size and the geometry of the skull can be assumed to represent critical factors in experimental design. For example, experiments with rotational acceleration are very dependent on the distance to the axis of rotation, thus a larger brain may be far less resistant to rotational injury. Different rat strains may exhibit different inflammatory responses and reactions to TBI.

Validated and verified experimental models are necessary for successful identification of new drug candidates for treatment of TBI or SCI [92]. Differences in time scales between experimental animals, such as rodents, and humans are difficult to represent and compensate for in experimental work [93].