We present here a protocol of a blast wave model for rodents to investigate neurobiological and pathophysiological effects of mild to moderate traumatic brain injury. We established a gas-driven, bench-top setup equipped with pressure sensors allowing for reliable and reproducible generation of blast-induced mild to moderate traumatic brain injury.
Traumatic brain injury (TBI) is a large-scale public health problem. Mild TBI is the most prevalent form of neurotrauma and accounts for a large number of medical visits in the United States. There are currently no FDA-approved treatments available for TBI. The increased incidence of military-related, blast-induced TBI further accentuates the urgent need for effective TBI treatments. Therefore, new preclinical TBI animal models that recapitulate aspects of human blast-related TBI will greatly advance the research efforts into the neurobiological and pathophysiological processes underlying mild to moderate TBI as well as the development of novel therapeutic strategies for TBI.
Here we present a reliable, reproducible model for the investigation of the molecular, cellular, and behavioral effects of mild to moderate blast-induced TBI. We describe a step-by-step protocol for closed-head, blast-induced mild TBI in rodents using a bench-top setup consisting of a gas-driven shock tube equipped with piezoelectric pressure sensors to ensure consistent test conditions. The benefits of the setup that we have established are its relative low-cost, ease of installation, ease of use and high-throughput capacity. Further advantages of this non-invasive TBI model include the scalability of the blast peak overpressure and the generation of controlled reproducible outcomes. The reproducibility and relevance of this TBI model has been evaluated in a number of downstream applications, including neurobiological, neuropathological, neurophysiological and behavioral analyses, supporting the use of this model for the characterization of processes underlying the etiology of mild to moderate TBI.
Traumatic brain injury (TBI) accounts for more than two million hospital visits each year in the United States alone. Mild TBI commonly resulting from car accidents, sporting events, or falls represent approximately 80% of all TBI cases1. Mild TBI is considered the ‘silent disease’ as patients often experience no overt symptoms in the days and months following the initial insult, but can develop serious TBI-related complications later in life2. Moreover, blast-induced mild TBI is prevalent among military service-members, and has been associated with chronic CNS dysfunction3,4,5,6. Due to the rising incidence of blast-related mild TBI7,8, preclinical modeling of neurobiological and pathophysiological processes associated with mild TBI has thus become a focus in the development of novel therapeutic interventions for TBI.
Historically, TBI research has primarily focused on severe forms of neurotrauma, despite the relatively lower number of severe human TBI cases. Preclinical rodent models for severe human TBI have been developed, including the controlled cortical impact (CCI)9,10 and fluid percussion injury (FPI)11 models, which are both well established to produce reliable pathophysiological effects12,13. These models have laid the groundwork for what is known today about neuroinflammation, neurodegeneration, and neuronal repair in TBI. Although considerable knowledge of the pathophysiology of TBI has been developed, there are currently no effective, FDA-approved treatments available for TBI.
More recently, the focus of TBI research has been broadened to include a wider spectrum of TBI-related pathologies with the ultimate goal of developing effective therapeutic interventions. Nevertheless, few preclinical models for mild TBI have been established that have shown measurable effects, and only a small number of studies have investigated the mild TBI spectrum2,14,15. As mild TBI accounts for the large majority of all TBI cases, reliable models of mild TBI are urgently needed to facilitate research into the etiology and neuropathophysiology of the human condition, in order to develop novel therapeutic strategies.
In conjunction with biomedical engineers and aerospace physicists, we have established a scalable, closed-head blast wave model for mild to moderate TBI. This preclinical rodent model has been specifically developed to investigate the effects of force dynamics, including blast waves and acceleration/deceleration movement, that are associated with human mild TBI obtained in military combat, sporting events, car accidents, and falls. As blast waves correlate with the force dynamics that cause mild TBI in humans, this model was designed to produce a consistent Friedlander waveform with an impulse, which is measured as pounds per square inch (psi)*millisecond (ms). The impulse level is scaled to fall below defined lung lethality curves for mice and rats in order to conduct preclinical investigations16,17,18. In addition, this model allows for investigation of coup and contrecoup injury due to rapid rotational forces of the animal’s head. This kind of injury is inherent to several types of clinical TBI presentations, including those observed in both military and civilian populations. Therefore, this versatile model fits a need that encompasses multiple clinical presentations of TBI.
The preclinical model presented here produces reliable and reproducible pathophysiological changes associated with clinical mild TBI as demonstrated by a number of prior studies17,19,20,21,22,23. Studies with this model showed that rats subjected to a low-intensity blast wave exhibited neuroinflammation, axonal injury, microvascular damage, biochemical changes related to neuronal injury and deficits in short-term plasticity and synaptic excitability19. However, this mild TBI model did not induce any macroscopic neuropathological changes, including tissue damage, hemorrhage, hematoma and contusion19 that have been commonly observed in studies using moderate to severe invasive TBI models10,24. Previous research19,21,22,23 has shown that this preclinical model can be used to characterize neurobiological and pathophysiological processes underlying the etiology of mild and moderate TBI17,19,20,21,22,23. This model also permits for testing of new therapeutic compounds and strategies, as well as the identification of novel, suitable targets for the development of effective TBI interventions19,21,22,23.
This model was developed to investigate effects induced by blast waves as well as rapid rotational forces on molecular, cellular and behavioral outcomes in rodents. Analogous to the blast wave model presented here, a number of preclinical models has been developed that attempt to recapitulate mild to moderate TBI using gas-driven overpressure waves2,14,17,25,26,27,28. Some of the limitations of other models include: the animal is fixed to a wire-mesh gurney and the head is immobilized upon impact; the peripheral organs are exposed to the wave in addition to the brain, which creates the confounding variables of polytrauma; and the models are large and stationary, which limits changing and adapting critical parameters to better model conditions reminiscent of human TBI.
The benefits of this bench-top, gas-driven shock tube setup are its relative low-cost for acquisition and running expenses, as well as ease of installation and use. Furthermore, the setup allows for high-throughput operation and generation of controlled reproducible blast waves and in vivo outcomes in both mice and rats. In order to control for consistent test conditions (i.e., constant blast wave and overpressure) the setup is equipped with pressure sensors. The advantages of this model for TBI include scalability of the injury severity and that mild TBI is induced using a non-invasive, closed-head procedure. Peak overpressure and subsequent brain injury increase with thicker polyester membranes in a consistent scalable manner17. The ability to scale TBI severity through membrane thickness is a useful tool to determine the level, at which specific outcome measures (e.g., neuroinflammation) become evident. Providing protective shielding for the peripheral organs, also allows focused investigation into mild TBI mechanisms by avoiding or reducing confounding variables of systemic injury, such as lung- or thoracic injury. Moreover, this setup allows selecting the direction, by which the blast wave strikes/penetrates the head (i.e., head-on, side, top or underneath) and therefore different types of TBI-inducing insults can be investigated. The standard procedure to induce mild to moderate TBI described here employs side exposure to evaluate the effects of blast wave injury in combination with coup and contrecoup injury due to rapid rotational forces. Furthermore, in order to investigate exclusively blast-induced injury, top down blast wave exposure can be employed in this model.
The protocol follows the animal care guidelines of the University of Cincinnati and West Virginia University. All procedures involving animals were approved by the Institutional Animal Care and Use Committees (IACUC), and were performed according to the principles of the Guide for the Care and Use of Laboratory Animals.
1. Installation of the blast TBI setup
2. Evaluation of the setup and blast wave properties using pressure sensor recordings.
3. Preparation of experimental setup and induction of mild TBI in rodents
NOTE: Transfer rodents to holding area 30 min to 1 h prior to start of TBI experiments to acclimatize. Select holding area that is minimally affected by noise of the procedure.
4. Downstream applications for rodents exposed to blast wave/rotational forces and controls
NOTE: In previous studies, the effects of mild to moderate TBI at various time points after exposure to a blast wave and rotational forces were assessed in rodents using downstream applications, including biochemical, neuropathological, neurophysiological, and behavioral analyses19.
The scalability of the blast wave setup was tested using three different membrane thicknesses, 25.4, 50.8 and 76.2 μm. Peak pressure levels were assessed at the head placement area and the exit of the shock tube apparatus using piezoelectric pressure sensors (see Figure 1 & Figure 2). Peak pressures increase in concordance with membrane thickness at both sensor locations (Figure 3A,B), demonstrating that the peak pressure is scalable in nature. This property of the setup can be exploited to calibrate the system and assess its scalability as described in step 2.3.
In order to evaluate effects of blast-induced TBI in vivo, adult, 3-month-old, male, wild-type C57Bl/6J mice were exposed to blast waves produced by this setup (Figure 1 & Figure 2) using the protocol described here. First, the effects of blast waves produced with two different membrane thicknesses (50.8 and 76.2 μm) or sham treatment on righting reflex time (RRT) were assessed (Figure 4A). The latency of the mice to fully right themselves (4 paws on the ground) after anesthesia is determined here as RRT. The mice were anesthetized using isoflurane (consistent, short and mild anesthesia) and then underwent TBI induction or sham treatment. Immediately following injury, mice were allowed to recover and time to regain righting reflex was recorded. Mice that were exposed to a blast wave produced with the 76.2 μm membrane exhibited a significant increase in RRT as compared to sham controls that underwent the same anesthesia procedure (Figure 4A), suggesting that this blast wave induces loss of consciousness. In contrast, mice exposed to a blast wave from the 50.8 μm membrane exhibit no significant increases in RRT (Figure 4A), indicative of mild form of TBI. Rupture of a standard 76.2 μm polyester membrane results in the rapid generation of a short duration blast wave of approximately 160 psi of overpressure (Figure 3C), which the left side of the subject’s cranium is exposed to during the experimental procedure.
The short-term physiological effects occurring after the exposure to blast wave and rotational forces in rodents are currently not well characterized. To delineate the acute effects of blast wave exposure and rotational forces from this model, we assessed core body temperature regulation and body weight. The temperature and body weight of adult, 3-month-old, male wild-type C57Bl/6J mice were recorded following TBI induction. Baseline core body temperature and body weight were recorded in the mice prior to TBI procedure or sham treatment. Exposure to a blast wave produced with the 76.2 μm membrane significantly decreased the body temperature during the first hour in TBI-induced mice as compared to their sham controls (Figure 4B), indicative of a significant physiologic effect produced by TBI induction. Consistently, mice subjected to TBI using 76.2 µm membranes exhibited an acute, time dependent yet significant reduction in total body weight one-day post-TBI compared to sham (Figure 4C).
In order to examine the impact of TBI on behavioral outcomes, the effect of blast-induced TBI on acute locomotor activity was analyzed (Figure 4D). Adult, 3-month-old, male C57Bl/6J mice underwent TBI induction using 76.2 µm membrane or sham treatment and locomotor activity was monitored for 30 minutes three hours post-TBI. Exposure to a blast wave produced with the 76.2 μm membrane resulted in an acute, significant decrease in locomotor activity (Figure 4D).
Figure 1: Setup of murine blast wave model. (A–C) Representative images of the setup of the blast wave model for mice. Side view of the setup (A). Top view of the setup (B). 1, gas cylinder with a high flow gas regulator; 2, 9.53 mm high pressure hydraulic line and quick connect male and female attachments; 3, driver section of the shock tube; 4, driven section of the shock tube; 5, PVC pipe shield; 6, head placement area; 7, polyester membrane. The individual parts of the setup are installed on machine slide tables allowing for precise positioning of driver (3) and driven sections (4) in relation to subject undergoing injury induction. (C) Top view of setup with pressure sensor placements. Three sensors are located in one plane at the exit of the shock tube, 120 degrees apart (S1 – S3), to monitor the blast wave properties during the TBI induction. One sensor is installed at the head placement area (S4). Please click here to view a larger version of this figure.
Figure 2: Schematic of murine overpressure shock tube. Precision-machined shock tube is made from high-tensile steel. Internal space of the driver section is angled at 6 degrees. Internal diameter of driver and driven section is 37 mm. Mating surfaces of driver-driven sections are precision-machined to ensure complete seal. The entire shock tube is industrially clamped to a machine slide table to ensure solid mounting and consistency of blast wave generation. At the exit of the driven section holes are drilled (in one plane, 120° apart) to install the three pressure sensors (indicated by *). Please click here to view a larger version of this figure.
Figure 3: Pressure recordings from murine blast wave setup. (A,B) Peak pressure is scalable and dependent upon polyester membrane thickness. Pressure sensors were used to record peak pressures produced by the shock tube with helium gas and polyester membranes of 25.4, 50.8 or 76.2 μm thickness. (A) At the head placement area, the mean peak pressure produced with 25.4 µm membranes was 428 ± 15.9 kPa, with 50.8 µm membranes 637 ± 21.4 kPa and with 76.2 µm membranes 1257 ± 40.7 kPa (SEM, n = 7-12, one-way ANOVA followed by post-hoc Dunnett’s comparison test, *** P ≤ 0.001). (B) At the exit of the shock tube, the mean peak pressure recorded with 25.4 µm membranes was 164 ± 11.7 kPa, with 50.8 µm membranes 232 ± 11.7 kPa and with 76.2 µm membranes 412 ± 11.0 kPa (SEM, n = 7-12, one-way ANOVA followed by post-hoc Dunnett’s comparison test, ** P ≤ 0.01, *** P ≤ 0.001). (C) Representative graph of the pressure recording from the sensor at the head placement area (incident sensor) using a 76.2 µm membrane. The waveform is similar to that of a Friedlander wave, scaled in time/duration for murine subjects. (D) Representative graph of the pressure recording from 3 distinct sensors located at end of the driven section to determine the linearity/phase of waveform within the driven section. All three sensors (located 120 degrees apart) show a similar rise/fall duration indicating that the waveform leaving the driven section is similar in cross-section within the driven section. The blast wave was generated using a 76.2 µm membrane. Please click here to view a larger version of this figure.
Figure 4: Acute in vivo effects of blast-induced TBI. (A) Moderate TBI, but not mild TBI increases righting reflex time (RRT). Adult, 3-month-old, male, wild-type C57Bl/6J mice were subjected to TBI procedures using the shock tube with helium gas and polyester membranes of 50.8 or 76.2 μm thickness or sham treatment. Immediately following injury or sham treatment, mice were allowed to recover and RRT was recorded. TBI induction with 50.8 μm membrane or sham treatment exhibited comparable levels of RRT. In contrast, TBI induction using a 76.2 μm membrane increases RRT, indicative of a loss of consciousness induced by the blast wave with the 76.2 μm membrane (SEM, n = 4-10, Sham RRT = 35.6 ± 2.0 s, 50.8 µm membrane RRT = 43.0 ± 4.3 s and 76.2 µm membrane RRT = 254.0 ± 40.2 s, one-way ANOVA followed by post-hoc Dunnett’s comparison test, *** P ≤ 0.001). (B) Moderate TBI significantly and transiently reduces core body temperature. Adult, 3-month-old, male, wild-type C57Bl/6J mice were subjected to TBI induction with 76.2 µm membranes or sham treatment. Their core body temperature was recorded for two hours. Baseline core body temperature was recorded prior to TBI induction. Blast-induced TBI with 76.2 µm membranes is associated with a significant drop in core body temperature within the first hour post-TBI. (SEM, n = 10, two-way repeated measures ANOVA, followed by post-hoc Bonferroni’s multiple comparison tests, ** P ≤ 0.01, *** P ≤ 0.001). (C) Moderate TBI results in a transient reduction in body weight. Adult, 3-month-old, male C57Bl/6J mice were subjected to TBI procedures using 76.2 µm membranes or sham treatment. Subsequently, body weights were recorded for 5 days. Total body weight was significantly reduced one-day post-TBI (SEM, n = 7, two-way repeated measures ANOVA followed by post-hoc Bonferroni’s multiple comparison tests, * P ≤ 0.05). (D) Moderate TBI results in acute reduction in locomotor activity. Adult, 3-month-old, male C57Bl/6J mice were subjected to TBI procedures using 76.2 µm membranes or sham treatment. Three hours post-TBI locomotor activity was tracked for 30 minutes and quantified using video tracking software (SEM, n = 9-11, unpaired two-tailed t-test, ** P = 0.01). Please click here to view a larger version of this figure.
We present here a preclinical mild TBI model that is cost-effective, easy to set up and execute, and allows for high-throughput, reliable, and reproducible experimental outcomes. This model provides protective shielding to peripheral organs to allow for focused investigation into mild TBI mechanisms while limiting the confounding variables of systemic injury. In contrast, other blast models are known to inflict damage to peripheral organs2,39,40. Another advantage of this model is its capability to deliver the blast wave from any desired angle compared to the fixed position in other blast models40. This allows for focused anatomical studies to better understand brain vulnerability.
In order to study human blast-related TBI, a relevant model for TBI should produce biomechanical forces comparable to those experienced by subjects during TBI induction. A clinically relevant model should also induce neurobiological, pathophysiological and behavioral outcomes observed in subjects suffering from mild TBI. In previous studies, the blast wave model presented here has been thoroughly examined17,19,21, and numerous biophysical and neurobiological aspects reminiscent of human TBI, including blast wave dynamics and forces, neuroinflammation, axonal injury and microvascular damage have been evaluated. These studies have provided evidence that this preclinical blast wave model for TBI produces reliable and reproducible neurobiological and pathophysiological changes associated with clinical TBI.
Furthermore, with the increased incidence of mild blast TBI within the military population7,8, this versatile rodent model for mild human TBI provides researchers with a valuable tool to investigate processes underlying blast-related TBI and explore novel therapeutic strategies. For example, our model demonstrates neurovascular complications, and highlights the importance for vascular intervention as a promising therapeutic approach22,23,35. Consistently, other preclinical models of blast TBI have also produced neurovascular effects associated with neurodegeneration and behavioral deficits2,25,40,41,42,43.
Based on previous research19,21,22,23, we have established that the blast wave model presented here may be well suited for the investigation into the pathophysiology and etiology of human concussion. Most preclinical TBI models do not permit head movement44 even though the biomechanical properties associated with rapid head acceleration/deceleration are a predictive factor for the development of a concussion in humans45,46. Consistent with the model described herein, Goldstein and colleagues14 showed that rapid head movement induced by blast forces are a prerequisite for the induction of behavioral deficits, possibly due to rotational forces and shearing. A better understanding of the pathophysiological changes that occur in mild TBI and in response to concussion would also help to determine clinical biomarkers and identify novel targets for the development of treatments for TBI.
Little is known about the pathophysiological changes and the disease progression following repetitive mild TBI (e.g., repetitive concussion experienced in sports). This preclinical model permits the study of repetitive mild TBI with little to no mortality. In contrast, some TBI models inflict severe injuries, and therefore it is often difficult, or inhumane, to induce further injury. In addition, severe injuries are often irreparable and the detection of subtle physiological changes may be precluded. This model also allows for the scalable investigation of various inter-injury intervals; a critical parameter for repetitive mild TBI that requires further characterization. After TBI, a CNS injury response is triggered that helps to protect brain integrity and prevent widespread neuronal cell death. The injury response may be, indeed, significantly impacted by the induction of another injury within a short time point after the initial injury. This model permits the investigation of the inter-injury interval, which is an important aspect of clinical trial design for repetitive mild TBI. Moreover, this scalable model allows for a rapid high-throughput workflow, which facilitates investigation of multiple parameters simultaneously, as well as the evaluation of therapeutic activity of novel interventions.
One limitation of this model is the inability to control the properties of the blast wave between the tube exit and the animal’s head. Although the blast wave is turbulent upon exit from the shock tube, the outcome measures are still reliable and reproducible with a consistent positioning of the rodent’s head18. Therefore, it is important to keep the experimental settings (i.e., head position and distance from shock tube exit) constant between all studies. In order to optimize model design and protocol, waveform dynamics between the tube exit and the head placement area have been measured (Figure 3) and modeled using numerical simulations18. Future projects will integrate finite element modeling to determine how force dynamics transfer from the skull to meninges, to cerebrospinal fluid, and finally into the brain tissue. The complex interplay of force dynamics and biophysics and resulting physiological responses are important areas in TBI research that have been so far underexplored.
In summary, we present here a protocol and visualized experiment of a blast wave injury model that has been developed to investigate the effects of mild TBI. The collective experience of engineers, physicians, and biomedical scientists contributed to the optimization of its biophysical/physiological validity and neurobiological relevance. This model has been thoroughly validated and has already produced meaningful results, especially in understanding early dynamics of mild TBI17,19,20,21,22,23. Exploiting this preclinical model to further study mild TBI will significantly advance our understanding of the pathophysiology and etiology of TBI and contribute to the development of novel interventions for the benefit of patients suffering from TBI.
The authors have nothing to disclose.
We thank R. Gettens, N. St. Johns, P. Bennet and J. Robson for their contributions to the development of the TBI model. NARSAD Young Investigator Grants from the Brain & Behavior Research Foundation (F.P. and M.J.R.), a Research Grant from the Darrell K. Royal Research Fund for Alzheimer’s Disease (F.P.) and a PhRMA Foundation Award (M.J.R.) supported this research. This work was supported through pre-doctoral fellowships from the American Foundation for Pharmaceutical Education (A.F.L and B.P.L.).
3/8 SAE High Pressure Hydraulic Hose | Eaton Aeroquip | R2-6-6-36M | Available from Grainger |
3/8'' Quick Connect Female Plugs | Karcher | KAR 86410440 | |
3/8'' Quick Connect Male Plugs | Karcher | KAR 86410440 | |
ANY-maze video tracking software | Stoelting Co. | ANY-maze software | |
Clear Mylar membrane | ePlastics.com | POLYCLR0.003 | http://www.eplastics.com/Plastic/Clear_Polyester_Film/POLYCLR0-003; Clear Mylar membrane is sold in various thicknesses. All are sold by vendor listed above. |
Compound Slide Table (X2) | Grizzly Industrial | G5757 | |
Deadman Gas Control Ball Valve | Coneraco Inc. | 71-502-01 | "Apollo", Available from Grainger |
Driver and driven section (murine) | own design/production | n/a | For further information please contact the authors |
Driver and driven section (rat) | own design/production | n/a | For further information please contact the authors |
Ear Muffs | 3M | 37274 | Available from Grainger |
Gas Regulator – Hi Flow 3500-600-580 | Harris | 3003539 | |
Helium Gas | AirGas | HE 300 | Tanks are available in various sizes |
Inhalation Anesthesia System | VetEquip | 901806 | |
Input Module | National Instruments | NI 9223 | |
Isoflurane | Baxter | NDC 10019-360-40 | Ordered by veterinarian |
Laboratory Timer/Stopwatch | Fisher Scientific | 50-550-352 | |
Labview version 12.0 | National Instruments | Data Acquistion Software | |
Magnetic Dial Indicator/Micrometer | Grizzly Industrial | G9849 | |
MATLAB | MathWorks | Software for pressure recording analysis | |
Oxygen Regulator | Medline | HCS8725M | |
PC for Data Processing | Dell | ||
Polyvinylchloride Tubing – 25.4 mm | FORMUFIT | P001FGP-WH-40×3 | |
Pressure sensors | PCB Piezotronics | 102A05 | |
Receiver USB Chassis | National Instruments | DAQ-9171 | |
Sensor Signal Conditioner | PCB Piezotronics | 482C series | |
Stainless NSF-Rated Mounting Table | Gridmann | GR06-WT2448 | |
T Handle Allen Wrench – 3/16'' | S&K | 73310 |