Experimental Neurology 248 (2013) 520–529

Contents lists available at ScienceDirect

Experimental Neurology

j ourna l homepage: www.e lsev ie r .com/ locate /yexnr

Modeling clinically relevant blast parameters based on scaling principlesproduces functional & histological deficits in rats

Ryan C. Turner a,b, Zachary J. Naser a,b, Aric F. Logsdon b,c, Kenneth H. DiPasquale b,c, Garrett J. Jackson a,b,Matthew J. Robson b,c, Robert T.T. Gettens d, Rae R. Matsumoto b,c, Jason D. Huber b,c, Charles L. Rosen a,b,⁎a Department of Neurosurgery, West Virginia University, School of Medicine, Morgantown, WV, USAb Center for Neuroscience, West Virginia University, School of Medicine, Morgantown, WV, USAc Department of Basic Pharmaceutical Sciences, West Virginia University, School of Pharmacy, Morgantown, WV, USAd Department of Biomedical Engineering, College of Engineering, Western New England University, Springfield, MA, USA

⁎ Corresponding author at: West Virginia University SchNeurosurgery, OneMedical Center Drive, Suite 4300, HealtMorgantown, WV 26506-9183, USA. Fax: +1 304 293 48

E-mail address: crosen@hsc.wvu.edu (C.L. Rosen).

0014-4886/$ – see front matter © 2013 Elsevier Inc. All rihttp://dx.doi.org/10.1016/j.expneurol.2013.07.008

a b s t r a c t

a r t i c l e i n f o

Article history:Received 25 March 2013Revised 28 June 2013Accepted 12 July 2013Available online 20 July 2013

Keywords:Blast injuryShock tubeBlast durationTraumatic brain injuryNeural injury

Blast-induced traumatic brain injury represents a leading cause of injury in modern warfare with injury patho-genesis poorly understood. Preclinical models of blast injury remain poorly standardized across laboratoriesand the clinical relevance unclear based upon pulmonary injury scaling laws. Models capable of high peak over-pressures and of short duration may better replicate clinical exposure when scaling principles are considered. Inthisworkwe demonstrate a tabletop shock tubemodel capable of high peak overpressures and of short duration.By varying the thickness of the polyester membrane, peak overpressure can be controlled. We used membraneswith a thickness of 0.003, 0.005, 0.007, and 0.010 in to generate peak reflected overpressures of 31.47, 50.72,72.05, and 90.10 PSI, respectively. Blast exposurewas shown to decrease total activity and produce neural degen-eration as indicated by fluoro-jade B staining. Similarly, blast exposure resulted in increased glial activation asindicated by an increase in the number of glial fibrillary acidic protein expressing astrocytes compared to controlwithin the corpus callosum, the region of greatest apparent injury following blast exposure. Similarfindingswereobserved with regard to activated microglia, some of which displayed phagocytic-like morphology within thecorpus callosum following blast exposure, particularly with higher peak overpressures. Furthermore, hematoxy-lin and eosin staining showed the presence of red blood cells within the parenchyma and red, swollen neuronsfollowing blast injury. Exposure to blast with 90.10 PSI peak reflected overpressure resulted in immediate mor-tality associated with extensive intracranial bleeding. This work demonstrates one of the first examples of blast-induced brain injury in the rodent when exposed to a blast wave scaled from human exposure based on scalingprinciples derived from pulmonary injury lethality curves.

© 2013 Elsevier Inc. All rights reserved.

Introduction

Blast-induced traumatic brain injury has been described as the hall-mark injury of modern warfare (Elder and Cristian, 2009; Nakagawaet al., 2011) with ~25,000 members of the armed services (U.S. andCoalition forces) injured or killed by explosive devices in Iraq andAfghanistan conflicts in the past decade (Champion et al., 2009). Theprevalence of blast exposure is associated with an immense financialand societal burdenwith estimates of 13–22% of combat veterans havingsustained a traumatic brain injury (TBI) during time in service (Panzeret al., 2012; Schneiderman et al., 2008; Terrio et al., 2009). Consequently,there is a clear clinical need for increased understanding of blast injurypathogenesis and the development of improved therapeutics for the

ool of Medicine, Department ofh Sciences Center, PO Box 9183,19.

ghts reserved.

treatment, or prevention, of blast-induced brain injury. Increasing effortsto study blast injury has resulted in the propagation of numerous pre-clinical models of blast injury with the shock tube representing themost widely used model type. Despite the extensive use of shock tubemodels, model parameters such as size of the driver and driven sections,membranematerial and thickness, and gas used seem to vary across lab-oratories. As such, the shock waves produced vary significantly in peakoverpressure and duration.

Recent work by Panzer and colleagues has identified potential dis-crepancies between many preclinical models and real-world recordedblast parameters with models often exposing small rodents to longduration blasts (Panzer et al., 2012). Assuming the same scaling princi-ples apply as those discovered with regard to pulmonary blast injury,many current rodentmodels, inwhich overpressure exposure durationsfrequently exceed 4 ms and in some cases as much as 10 ms or more,may be subjecting animals to the equivalent of a human exposed to det-onation of 27,000-kg of TNT from over 70 m away (Panzer et al., 2012).In other words, many preclinical models may be exposing animals to

521R.C. Turner et al. / Experimental Neurology 248 (2013) 520–529

the equivalent of long-duration nuclear blasts rather than the morecommonhigh explosive blasts (Bass et al., 2012). Therefore, a preclinicalblast model capable of delivering high peak overpressures over a shortduration may be of value, particularly should similar scaling laws existin the brain, as is the case in pulmonary-based work.

In this work we demonstrate the ability to model blast injury withclinically relevant peak overpressure and duration based on scalinglaws from humans to rodents with a novel, tabletop shock tube model.Furthermore, we show the ability to model blast injury of varying sever-ity with different polyester membranes that result in reflected peakoverpressures ranging from approximately 30–90 PSI while consistentlybeing of short duration (b3 ms). Animals exposed to blast-induced trau-matic brain injury across this range of pressures exhibit alterations intotal activity as well as clear histological changes as determined utilizingcommon markers of neural injury.

Methods

Animals

All procedures involving live animals were approved by the Institu-tional Animal Care and Use Committee of West Virginia University andwere performed according to the principles of the Guide for the Care andUse of Laboratory Animals. This worked used thirty-eight 350 g maleanimals acquired from Hilltop Lab Animals (Hilltop Lab Animals, Inc.).Animals were acclimated for 1 week prior to experimental use andwere housed under 12-hour light/12-hour dark conditions with foodand water available ad libitum.

Design of blast model

A 4-piece, machined aluminum shock tube apparatus was con-structed anddrivenusing compressednitrogengas. Thedriver anddrivensections were separated by clear polyester membranes (Ridout Plastics

Fig. 1. Experimental setup utilized in this work demonstrating the novel shock tube apparatusutilizes a tapered driver section to reducewave reflection and alterations of the shockwave by thproducemembrane rupture and therefore, reduced overpressure duration and impulse. The abdthe animal has been inverted for illustration purposes but the blast was delivered to the right s

Co.) of varying thickness (0.003″–0.010″) to achieve a range of peakoverpressure exposures. A tapered design of the driver section (Fig. 1)was included in an effort to minimize wave reflection and alterations ofthe blast wave by the expansion wave. This feature is similar in natureto blast tubes utilizing explosives in which the explosive is often det-onated in a conical or parabolic-shaped driver section. The driver sectionwas kept short in length (overall length = 6.7″; taper length = 4.63″;diameter = 2.8″) in an effort to reduce the amount of gas required formembrane rupture and subsequent blast duration, regardless of mem-brane thickness. The overpressure duration and impulse are reduced asdescribed previously due to the shortened driver section (Bass et al.,2012). For additional details concerning design of the blast model, pleasesee Supplemental information.

Blast waveform acquisition & analysis

Shock wave pressures were detected using piezoelectric sensors(PCB Piezotronics) that were placed in both reflected and incident posi-tions at the exit of the shock tube (Fig. 1). Data was acquired andprocessed using the National Instruments input module (NI 9223)connected to the National Instruments receiver USB chassis (CDAQ-9171). Signals were recorded using a custom-designed Labview 12.0program and the NI data acquisition system (National Instrument)with a sampling rate of 500 kHz as described previously (Ahlers et al.,2012; Chavko et al., 2011).

Blast exposure

Prior to blast exposure, animals were anesthetized by intraperitone-al injection of ketamine (90 mg/kg; Webster Veterinary) and xylazine(5 mg/kg; Webster Veterinary). Animals were oriented with the longaxis of the animal perpendicular to the blast front. In other words, theblast was delivered side-on to the head only, with the wave first en-countering the right side of the head (direct) prior to passing through

for simulating blast-induced brain injury and data acquisition procedures. The shock tubeeexpansionwave. The driver sectionwas kept short to reduce thevolumeof gas required toomen and thorax of animalswere protectedusing a rigid apparatus. Note: the orientation ofide of the head initially rather than the left as pictured.

522 R.C. Turner et al. / Experimental Neurology 248 (2013) 520–529

the skull and exiting the left side (indirect), as the thorax and abdomenwere protected using rigid shielding, creating an experimental setup(Fig. 1) similar to that used by Shridharani et al. (2012). This setupallows the head to accelerate and decelerate freely, consistent withcommon battlefield exposures (Goldstein et al., 2012). With the animalplaced directly at the exit of the tube, the effect of expansion waves onthe planar conditions of the blast front is minimized as described byBass et al. (2012). Animals (n = 38) were divided into the followinggroups: control (n = 10), 0.003 membrane blast exposure (n = 6),0.005membrane blast exposure (n = 10), 0.007membrane blast expo-sure (n = 9), and 0.010 membrane blast exposure (n = 3). Immedi-ately following blast exposure, animals were returned to a holdingcage equipped with a homeothermic heating blanket and rectal ther-mometer to maintain body temperature at 37 °C. Once basic reflexeswere restored, animals were returned to the home cage.

Total activity measurements

Animal activity was measured using an automated activity monitor-ing system (San Diego Instruments). Rats were given 1 h to acclimateto the testing room prior to initiation of testing. Each testing chamberconsisted of a 16 × 16 photobeam array surrounding a Plexiglass cham-ber to record lateralmovements. A separate 8 photobeam array to detectrearing activity was located above the 16 × 16 array. Activity was quan-tified over a 30 minute period 24 h post-blast exposure with total activ-ity, ambulatory, andfine activity reported based ondefault settings of thedata acquisition software (PAS, San Diego Instruments).

Histological preparations

At 72 hpost-blast exposure, a total of 24 animals (6 control and6 fromeach blast exposure group) were anesthetized as described above andperfused transcardially with cold 0.9% saline followed by 10% formalinfor a total of 10 min. The brain was then extracted and placed into fresh10% formalin for a minimum of 24 h. After 24 h of fixation, the brainwas blocked into sections and paraffin embedded as previously described(Turner et al., 2012). Briefly, tissues were processed using the Tissue-TekVIP 5 automatic tissue processor (Sakura Finetek) and embedded in par-affin using the Tissue-Tek TEC 5 embedding system (Sakura Finetek). Tis-sues were sliced using a Leica RM2235 microtome (Leica Microsystems)and slices mounted on slides for staining.

Standard protocols were utilized for staining with hematoxylin andeosin (H&E), fluorojade-B (Millipore), which was used as a marker ofneural degeneration, glial fibrillary acidic protein (Dako), which wasused as a marker of reactive astrocytes, and ionized calcium bindingadapter molecule (Wako), which was used as a marker of activatedmicroglia.

Images were acquired from the S1BF region of the cortex, corpuscallosum, striatum, hippocampus, and cerebellum. Imaging was per-formed using a Zeiss Axio Imager 2 for all brightfield and fluorescentimages shown.

Histological quantification

Stereology and the optical fractionator technique were used toquantify histological results as previously described (Schmitz and Hof,2005; Smith et al., 2012; Turner et al., 2012). Briefly, a region of interestencompassing the corpus callosum was drawn at low power using anOlympus AX70 microscope and StereoInvestigator software. The soft-ware then selected random 75 μm counting frames with a depth of6 μm, and the object of interestmarked by an observed blinded to treat-ment. The volume of the region of interest previously identified is thendetermined by the software and the number of cells marked by theobserver returned.

Quantitative real-time PCR

A total of 8 animals were used for quantitative real-time PCR and in-cluded 4 control animals and 4 animals exposed to blast using the 0.005membrane. Total RNA was extracted from the brain tissue (cortex, stri-atum, hippocampus, and cerebellum) of saline perfused animals usingTrizol reagent (Invitrogen) according to the manufacturer's instruc-tions. Sample concentration was determined by spectral absorptionand the purity confirmed using a 260/280 ratio of 1.8–2.0. Samples ofcDNAwere prepared via reverse transcription using commercially avail-able kits (Applied Biosystems). Sample reactions included MultiScribeTMReverse Transcriptase and random primers, with thermal cycle con-ditions set as follows: step 1 at 25 °C for 10 min, step 2 at 37 °C for120 min, step 3 at 85 °C for 5 s, and step 4 at 4 °C for 10 min. For PCRamplification, TaqMan® Universal PCR Master Mix and the followingprobes were obtained from Applied Biosystems: 18 s (Hs99999901_s1)for use as an endogenous control gene, GFAP (Rn00566603_m1), andIba-1 (Rn00574125_g1). The reaction mixture was prepared based onthe manufacturer's instructions and the following thermal cycling con-ditions used: initial holding at 50 °C for 2 min to allow for optimalAmpErase® UNG activity, followed by a first denaturing step at 95 °C for10 min, then 45 cycles at 95 °C for 15 s, and at 60 °C for 1 min. Changesin gene expressionwere determined using theΔΔCTmethod and a thresh-old value of 0.2. Threshold cycle values of each gene were normalized to18 s rRNA.

Data analysis

Data were analyzed using GraphPad Prism 5.0 (GraphPad Software,Inc.). A one-way ANOVA with posthoc Tukey was used to compareacross control and various blast-exposed groups. For gene expressionwork a student's t-test was utilized for comparing control and theblast-exposure group (0.005 membrane). A P b 0.05 was consideredstatistically significant for all data analyzed.

Results

The model is capable of producing a characteristic Friedlander-type blastacross a range of peak overpressures with short duration of exposure

The ideal blast wave in an open area, also known as the Friedlandercurve, consists of a near-instantaneous pressure rise followed by an ex-ponential decay of the overpressure (Bass et al., 2012). With all mem-branes tested utilizing the shock tube apparatus (Fig. 1), ideal blastwaves were obtained with a representative curve shown (Fig. 2A).Peak overpressure increased in a linear fashion with membrane thick-ness with both reflected (r2 = 0.92) and incident (r2 = 0.93) sensororientations allowing for the simulation of multiple peak overpressuresin the experiment (Figs. 2B, C). Notably, the duration of overpressurewas approximately 2 ms (Figs. 1 and 2A) regardless ofmembrane thick-ness, consistentwith the notion that driver section length is the primarydeterminant of duration and impulse.

High peak overpressure blast, even of short duration, is associated withhigh mortality and extensive intracranial bleeds. While reflected pres-sures of 31.47 ± 3.049 and 50.72 ± 3.368 PSI failed to produce anymortality, exposure to 72.05 ± 1.212 PSI blast was associated withmortality in 3 of 9 experimental animals (33%) while 90.10 ± 1.030PSI blast exposures produced mortality in all 3 animals tested (100%).Surviving animals failed to exhibit large hematomas at 72 h post-blastbut evidence of microvascular dysfunction and the potential for intra-parenchymal bleeds was present. Control animals and those exposedto 31.47 ± 3.049 PSI blast failed to display the gross appearance of anyparenchymal blood following perfusion (Figs. 3A and B, respectively).Animals exposed to survivable moderate peak overpressures in thiswork did display small hematomas immediately following blast expo-sure as evident in the non-saline perfused brain (Fig. 3C), consistent

Fig. 2.A.) Sample shockwave recording (reflected sensor) representing a classical Friedlander wave of short duration. B.) Peak reflected overpressure recorded as a function ofmembranethickness. C.) Peak incident overpressure recorded as a function of membrane thickness.

Fig. 3. Gross examination of single blast-exposed brains. Blast was administered to the right side of the head (direct) but produced more severe injury in the left hemisphere (indirect).A.) Brain of a perfused control animal, B.) 31.47 ± 3.049 PSI (reflected) blast-exposed brain at 72 h post-blast following perfusion demonstrating no apparent hematomas or gross changes,C.) presence of vascular disruption on the indirect side (solid arrows), visible in a non-perfused animal, immediately following exposure to 50.72 ± 3.368 PSI (reflected) blast, D.) at 72 hpost-blast, 72.05 ± 1.212 PSI (reflected) blast exposure results in the appearance of parenchymal blood and an overall husky appearance on the indirect side following perfusion (arrowswith dashed line), E.) immediately following 90.10 ± 1.030 PSI (reflected) blast, clear evidence of bleeds on the indirect-exposed cortex (arrows with dashed circles), and F.) immediatelyfollowing 90.10 ± 1.030 (reflected) PSI blast, brainstem bleeds were apparent on both direct and indirect sides and associated with mortality (arrows with dashed circles).

523R.C. Turner et al. / Experimental Neurology 248 (2013) 520–529

Fig. 4. Blast exposure, particularly at high peak overpressures, was associated with microvascular damage and neural injury on H&E staining. A.) Presence of red blood cells within thecortical parenchyma at 72 h following 72.05 ± 1.212 PSI (reflected) blast on the indirect side. B.) Swollen, red neuron at 72 h following 72.05 ± 1.212 (reflected) PSI blast. Red neuronstypically are associated with anoxic/ischemic injury indicating the presence of secondary damage, likely related to microvascular dysfunction in the cortex on the indirect side, followingblast injury.

524 R.C. Turner et al. / Experimental Neurology 248 (2013) 520–529

with the work of Reneer et al. (2011). Similarly, blast exposure resultedin the presence of intraparenchymal blood, even upon perfusion(Fig. 3D). Mortality was associated with the production of intracra-nial bleeds in the cortex, particularly on the indirect side as well as inthe brainstem (Figs. 3E, F). Similarfindings have been observed utilizing

Fig. 5. Blast exposure produced neural degeneration within the S1BF region of the cortex at 72blast on the direct side, B.) 50.72 ± 3.368 PSI blast on the direct side, C.) 31.47 ± 3.049 PSI blasFluoro-Jade B (FJB) positive neurons using stereology demonstrated a graded response based oP b 0.0001) and indirect (F(3,23) = 110.70, P b 0.0001) sides (E and F, respectively). Post hoccontrol with 31.47 ± 3.049 PSI on the indirect side (q = 4.51, P b 0.05), 50.72 ± 3.368 PSI72.05 ± 1.212 PSI on both direct (q = 15.13, P b 0.001) and indirect (q = 23.76, P b 0.0001)

explosives such as TNT in rat models in which animals placed in closerange exhibit increased mortality rates associated with brain contu-sions, lacerations, and hematomas (Cheng et al., 2010).

Blast-induced brain injury results in the presence of intraparenchymalred blood cells, swollen neurons, and activated microglia. To further

h post-exposure as indicated with fluoro-jade B staining. A.) 31.47 ± 3.049 PSI (reflected)t on the indirect side, and D.) 72.05 ± 1.212 PSI blast on the indirect side. Quantification ofn peak overpressure exposure in the S1BF region of the cortex on direct (F(3,23) = 47.03,analysis showed that blast exposure resulted in increased neural degeneration relative toon both direct (q = 10.29, P b 0.0001) and indirect (q = 13.69, P b 0.0001) sides, andsides.

525R.C. Turner et al. / Experimental Neurology 248 (2013) 520–529

examine the effect of blast exposure on neural injury andmicrovasculardysfunction, particularly considering the presence of hematomas andparenchymal blood on gross examination, brain sections underwent he-matoxylin and eosin (H&E) staining. The intraparenchymal blood wasconfirmed with H&E staining and red blood cells have clearly passedinto the parenchyma and are no longer confined to the vasculaturewithin the cortex (Fig. 4A). Evidence of neural injury and swollen, redaxons was apparent in some animals as well within the cortex (Fig. 4B).The presence of red neurons indicates a potential anoxic/ischemic insult(de Lanerolle et al., 2011), consistent with the presence of microvasculardysfunction observed following blast exposure (Svetlov et al., 2012). Thepresence of activated microglia, capable of releasing an array of inflam-matory mediators, was also seen with blast exposure and prompted fur-ther investigation of astro- and microgliosis.

Neural degeneration occurs in a dose-response manner consistent withblast severity. Blast-induced brain injury resulted in increased neuraldegeneration measured using Fluoro-Jade B staining in the S1BF regionof the cortex on both the direct (F(3,23) = 47.03, P b 0.0001) andindirect (F(3,23) = 110.7, P b 0.0001) sides when analyzed by one-way ANOVA (Figs. 5A–D). Post hoc tests revealed that blast-induced

Fig. 6. Blast exposure produced extensive astrocyte reactivity within the corpus callosumA.) Control (no blast exposure) on the indirect side, B.) 31.47 ± 3.049 PSI blast on the directhe indirect side. Quantification of glial fibrillary acidic protein (GFAP) positive astrocytessure in the corpus callosum on the direct (F(3,23) = 26.08, P b 0.0001) and indirect (F(3,2exposure resulted in increased astrocyte reactivity relative to control with 50.72 ± 3.368 P(q = 11.28, P b 0.0001) and indirect (q = 6.03, P b 0.01) sides.

brain injury produced increases in neural degeneration compared tocontrol at 31.47 ± 3.049 PSI on the indirect side (q = 4.51, P b 0.05),50.72 ± 3.368 PSI on both direct (q = 10.29, P b 0.0001) and indirect(q = 13.69, P b 0.0001) sides, and 72.05 ± 1.212 PSI on both direct(q = 15.13, P b 0.001) and indirect (q = 23.76, P b 0.0001) sides(Figs. 5E–F). Neural degeneration following blast exposure wasalso observed within other brain regions including the striatum, hip-pocampus, cerebellum and regions surrounding the corpus callosum(Supplemental information).

Blast injured brains demonstrate increased glial activation with increas-ing peak over-pressure. Reactive astrocytes, identified by GFAP staining,were elevated significantly within the corpus callosum on the direct(F(3,23) = 26.08, P b 0.0001) and indirect (F(3,23) = 8.03, P b 0.001)sides of the brain following blast-induced neurotrauma, consistent withfindings from other blast injury models (Svetlov et al., 2012), when ana-lyzed using one-way ANOVA (Figs. 6A–D). Post hoc tests showed thatblast injury resulted in elevated astrocyte reactivity at 50.72 ± 3.368PSI on the indirect side (q = 5.13, P b 0.01) and at 72.05 ± 1.212PSI on both the direct (q = 11.28, P b 0.0001) and indirect (q = 6.03,P b 0.01) sides (Figs. 6E–F). Similar findings were observed concerning

at 72 h post-exposure as indicated with glial fibrillary acidic protein (GFAP) staining.t side, C.) 72.05 ± 1.212 PSI blast on the direct side, and D.) 72.05 ± 1.212 PSI blast onusing stereology demonstrated a graded response based on peak overpressure expo-3) = 8.03, P b 0.001) sides (E and F, respectively). Post hoc analysis showed that blastSI on the indirect side (q = 5.13, P b 0.01) and at 72.05 ± 1.212 PSI on both the direct

526 R.C. Turner et al. / Experimental Neurology 248 (2013) 520–529

microglial activation, identified with Iba-1 staining, with increased levelsof microglial activation relative to control on both direct (F(3,23) =83.12, P b 0.0001) and indirect (F(3,23) = 25.37, P b 0.0001) sides(Figs. 7A–D). Again, althoughmicroglial presence was detected through-out the brain and appeared increased throughout, microglial activationwas most prominent throughout the corpus callosum and internal cap-sule, consistent with other studies (de Lanerolle et al., 2011). Notably,microglia appeared more rounded or amoeboid, typical of phagocyticmicroglia, within these white matter tracts consistent with white matterdamage observed in human subjects following trauma (Johnson et al.,2013). Previous reports in swine exposed to blast and assessed primarilyat 2 weeks failed to identify extensive microglial activation, a potentialreason for the discrepancy between studies (de Lanerolle et al., 2011).Post hoc tests revealed increased microglial activation following31.47 ± 3.049 PSI blast on the indirect side (q = 4.71, P b 0.05),50.72 ± 3.368 PSI blast on both direct (q = 8.37, P b 0.0001) and indi-rect (q = 7.85, P b 0.0001) sides, and 72.05 ± 1.212 PSI blast onboth direct (q = 20.49, P b 0.0001) and indirect (q = 11.92,P b 0.0001) sides as well (Figs. 7E–F). The clear neuroinflammatoryresponse, characterized by a graded glial response based on peak over-pressure exposure, observed in this work is consistentwith the findings

Fig. 7. Blast exposure produced microglial changes within the corpus callosum at 72 h post-eA.) Control (no blast exposure) on the direct side, B.) 50.72 ± 3.368 PSI blast on the direct sidindirect side. Quantification of ionized calcium-binding adapter molecule 1 (Iba-1) positive mexposure in the corpus callosum on the direct (F(3,23) = 83.12, P b 0.0001) and indirect (Fblast exposure resulted in increased microglial activation relative to control with 31.47 ± 3.04P b 0.0001) and indirect (q = 7.85, P b 0.0001) sides, and 72.05 ± 1.212 PSI blast on both dire

by Cernak and colleagues in which mild and moderate blast exposureproduced increases in GFAP, aswell as ED1, amarker ofmicroglia, at nu-merous timepoints (Cernak et al., 2011). Importantly, activatedmicrog-lia in other blast exposure studies have been associatedwith changes inneuron appearance such as darkened dendrites (Kaur et al., 1995) andin vitrowork has demonstrated a mild activation of microglia followingoverpressure exposure with pulse duration altering microglial geneexpression (Kane et al., 2012).

Gene expression of GFAP (Fig. 8A) was increased at 24 h followingblast exposure in the cortex (t = 2.412, P b 0.05) but not in the stria-tum (t = 0.022, P N 0.05), hippocampus (t = 1.084, P N 0.05), or cere-bellum (t = 0.634, P N 0.05). The lack of difference in GFAP geneexpression in regions outside of the cortex was in contrast to the visibledifferences seen in these regions at 72 h based on immunohistochemis-try (Supplemental information). The reason for this discrepancy is notentirely clear but may be due to consideration of the entirety of brainsections for gene expression studies when immunohistochemistry re-veals findings largely located at the interface of regions and the gray-white matter junction. Another possibility explaining the discrepancybetween immunohistochemical and gene expression findings is thetime of analysis. Only one time point for each assay was included and

xposure as indicated with ionized calcium-binding adapter molecule 1 (Iba-1) staining.e, C.) 31.47 ± 3.049 PSI blast on the indirect side, and D.) 72.05 ± 1.212 PSI blast on theicroglia using stereology demonstrated a graded response based on peak overpressure

(3,23) = 25.37, P b 0.0001) sides (E and F, respectively). Post hoc analysis showed that9 PSI blast on the indirect side (q = 4.71, P b 0.05), 49 PSI blast on both direct (q = 8.37,ct (q = 20.49, P b 0.0001) and indirect (q = 11.92, P b 0.0001) sides as well.

Fig. 8. Effect of 50.72 PSI blast exposure on gene expression at 24 h post-injury. A.) Gene expression of glial fibrillary acidic protein (GFAP) was increased within the cortex (t = 2.412,P b 0.05) but not in the striatum, hippocampus, or cerebellum. B.) Gene expression of ionized calcium binding adapter molecule (Iba-1) was not changed within any of the brain regionsstudied at 24 h post-blast.

527R.C. Turner et al. / Experimental Neurology 248 (2013) 520–529

therefore, changes in gene expression could have occurred rapidly fol-lowing blast and already declined to baseline by 24 h. No difference inIba-1 gene expression (Fig. 8B) was detected in the cortex (t = 0427,P N 0.05), striatum (t = 0.900, P N 0.05), hippocampus (t = 0.380,P N 0.05), or cerebellum (t = 0.448, P N 0.05), consistent with immu-nohistochemical findings (Supplemental information).

Blast-induced traumatic brain injury produces deficits in total activityregardless of severity. Total activity of the experimental subjects was de-creased at 24 h post-blast in all injured groups. The total activity levelpost-injury did not vary based on injury severity with all injury groupsdemonstrating similar impairments (Fig. 9A). The deficits in total activ-ity were due to a decrease in ambulatory activity associated with blastexposure (Fig. 9B). Blast exposure did not alter the measures of fineactivity (Fig. 9C).

Fig. 9. Effect of blast exposure on total, ambulatory, and fine activity levels at 24 h post-injury.with post hoc analysis identifying a reduction relative to control with 31.47 ± 3.049 PSI blast (blast (q = 5.40, P b 0.05). B.) Ambulatory activity was reduced with blast exposure (F(3,23)31.47 ± 3.049 PSI blast (q = 4.59, P b 0.05), 50.72 ± 3.368 PSI blast (q = 5.34, P b 0.05), and 7differences across any group comparisons (F(3,23) = 0.17, P N 0.05).

Discussion

Efforts to model blast injury have been plagued by the lack of stan-dardization across laboratories and the appropriate scaling of blastparameters to the animal species used based on data obtained fromhuman studies. Preclinical studies are varied with regard to animalplacement within or outside of the shock tube, orientation to the blastfront, protection of the thorax, and blast parameters, namely peak over-pressure and duration of exposure. Studies in soldiers revealed commonIED exposures producing blast injury result from detonation of 105-mmand 155-mm artillery rounds at a distance of 5–10 m. The use of theConventional Weapons Effects Program (CONWEP) (Hyde, 2004) al-lows for interpretation of these exposures in relation to blast wave pa-rameters with a peak overpressure of 50–1000 kPA (7.25–145.03 PSI)

A.) Total activity was significantly altered with blast exposure (F(3,23) = 6.55, P b 0.05)q = 4.50, P b 0.05), 50.72 ± 3.368 PSI blast (q = 5.28, P b 0.05), and 72.05 ± 1.212 PSI= 6.74, P b 0.05) with post hoc analysis identifying a reduction relative to control with2.05 ± 1.212 PSI blast (q = 5.47, P b 0.05) exposures. C.) Fine activity failed to demonstrate

528 R.C. Turner et al. / Experimental Neurology 248 (2013) 520–529

and an exposure duration of 2–10 ms (Panzer et al., 2012). As such, uti-lizing these parameters and scaling appropriately to preclinical animalspecies used may represent a more clinically relevant blast exposureand allow for improved elucidation of blast-induced injury pathophysi-ology and evaluation of proposed therapeutics.

The concept of scaling blast injury to animal models is based largelyupon previous works assessing pulmonary injury and primary blastsurvival across animal species. Richmond and colleagues identified, invarious species, the effect of blast duration on lethality with longerduration blasts requiring a lower peak overpressure to producemortal-ity when compared to short duration blasts that required a significantlyhigher peak overpressure (Richmond et al., 1968). A clear relationshipbetween mass and blast parameters associated with lethality exists,particularly with regard to the more clinically relevant short durationblast (Richmond et al., 1968). In fact, classical work by Bowen and col-leagues investigated pulmonary injury in relation to blast exposureand created a model system in which scaling is based upon the heightof burst divided by the cubed root of the mass (Bowen et al., 1968).This has been simplified to an equation utilizing the cubed root of thereference mass (human) divided by the mass of the experimental sub-ject (Bass et al., 2012; Panzer et al., 2012; Shridharani et al., 2012).The scaling equations developed have displayed significant utility inthe development of injury risk functions based upon the high level ofagreement with pre-existing experimental data concerning blast sur-vival and injury risk. As such, while scaling of blast injury and associatedblast parameters based on subject weight has not been related to neuralinjury (Park et al., 2011), scaling represents a logical step forward inmoreclosely replicating clinically relevant exposures in laboratory models(Panzer et al., 2012; Rafaels et al., 2012; Risling and Davidsson, 2012).

To address the question of how scaling affects blast-induced brain in-jury a shock tube capable of producing a range of peak overpressureswithshort durationswas designed and utilizedwithin this work. Furthermore,the model was designed to simulate open-field blast with the rat placedoutside of the shock tube in order to increase clinical relevance. Notably,the model produced clear neuropathological changes indicative of neuraldegeneration and glial reactivity. These findings were consistent with thework by Goldstein and colleagues in which single-blast exposure pro-duced neurodegenerative changes consistent with chronic traumatic en-cephalopathy (CTE) (Goldstein et al., 2012) as well as other workutilizing scaling principles (Shridharani et al., 2012). Similarly, neuropath-ological changes were observed throughout the cortex as previously de-scribed (Goldstein et al., 2012) but were most prominent with thecorpus callosum and internal capsule, consistent with previous findingsin other blast injury models (Park et al., 2011). It remains unclear whyspecific brain regions may be more susceptible to blast-induced brain in-jury but one logical explanation is that of spallation. Spallation refers tothe interface-based disruption that occurs between tissues of differentdensities upon a compression wave in the denser medium reflecting atthe interface resulting in displacement and fragmentation of the densermedium into the less dense medium (Cernak and Noble-Haeusslein,2010; Covey and Born, 2010; Yeh and Schecter, 2012). Spalling has beenidentified as a leading cause of endothelial injury and activation of mi-croglia (Duckworth et al., 2012), consistent with our findings. Similarly,endothelial damage results in a loss of tight junction integrity and conse-quently the potential for increased blood–brain barrier permeability andmicrovascular injury (Duckworth et al., 2012). Notably, increased BBBpermeability has been found as late as one month following blast-inducedbrain injury in a preclinical rodentmodel, indicating the presenceof lasting microvascular dysfunction (Rubovitch et al., 2011). Further-more, blast-induced BBB disruption has been observed to occur moreextensively on the contralateral side, as well as at fluid-tissue interfacessuch as that surrounding the lateral ventricles (Readnower et al., 2010),consistent with the findings reported within this study in which the con-tralateral side was more severely injured in comparison to the ipsilateralside based on standardmarkers of neural degeneration, astrocyte reactiv-ity, and microglial activation (Garman et al., 2011).

In addition to the neuropathological findings, the model utilized inthiswork produced deficits in total activity at 24 h post-blast, regardlessof experimental group. How blast-injury models affect behavioral func-tions in rodents is poorly understood but of clear interest based on clin-ical findings associated with blast that include retrograde amnesia,confusion, loss of consciousness, depression, post-traumatic stress dis-order (PTSD), vertigo, balance disorders, anxiety, apathy, and difficultyconcentrating and performing executive functions (Magnuson et al.,2012; Risling and Davidsson, 2012; Rubovitch et al., 2011; Scheibelet al., 2012; Taber et al., 2006). Previous reports have demonstratedimpairment on the accelerating rotarod following direct cranial blast in-jury (Kuehn et al., 2011) and on the balance beam followingwhole bodyblast exposure (Ahlers et al., 2012) while others have shown thatshielding of the torso prevents neural injury and associated behavioraldeficits following blast (Koliatsos et al., 2011). While the significanceof these findings is unclear and how these findings translate from onemodel to another, let alone to the clinical population, there is a clearneed for further studies in standardized, clinically-relevant models.

It is clear that future studies, in addition to understanding the effectof scaled blast exposure on functional ability and behavior-relatedchanges, need to better characterize the inflammatory processes associ-ated with blast exposure and how these relate to neural injury andneurodegeneration.While it is unlikely blast exposure can be complete-ly prevented, particularly in civilians exposed to terrorist-type attacks, itmay be possible to improve therapeutic options. The need for improvedtherapeutics is particularly clear in the military in which evacuation ofblast-exposed troops is often delayed by hours to several days due tothe nature of conflict. As such, secondary responses within the brainoften occur andmay represent a potential target as described previously(Dalle Lucca et al., 2012).

Acknowledgments

The authors would like to thank Nicholas St. John for contributing tothe design of the blast model, Brandon P. Lucke-Wold for his assistancewith preliminary experiments, andMs. Diana Richardson of the NationalInstitute of Occupational Safety and Health (NIOSH) for her assistancein tissue preparation. The authors would also like to thank Mr. JamesEdward Robson and Mr. Peter Bennett for their assistance in modelconstruction. The authors are also grateful to Dr. Kymberly Gyure forher assistance with neuropathological examination. This work was sup-ported by grant funding provided to Ryan C. Turner (5T32GM08174),Jason D. Huber (5RO1-NS061954), and Rae R. Matsumoto (DA013978)by theNational Institutes of Health. TheWest Virginia University Depart-ment of Neurosurgery also funded this work. Funding was also providedto Dr. Robert T.T. Gettens by the American Society for Quality, BiomedicalDivision in the form of the Dr. Richard J. Schlesinger Grant for construc-tion of the blast model.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.expneurol.2013.07.008.

References

Ahlers, S.T., Vasserman-Stokes, E., Shaughness, M.C., Hall, A.A., Shear, D.A., Chavko, M.,McCarron, R.M., Stone, J.R., 2012. Assessment of the effects of acute and repeated expo-sure to blast overpressure in rodents: toward a greater understanding of blast and thepotential ramifications for injury in humans exposed to blast. Front. Neurol. 3, 32.

Bass, C.R., Panzer, M.B., Rafaels, K.A., Wood, G., Shridharani, J., Capehart, B., 2012. Braininjuries from blast. Ann. Biomed. Eng. 40, 185–202.

Bowen, I.G., Fletcher, E.R., Richmond, D.R., Hirsch, F.G., White, C.S., 1968. Biophysicalmechanisms and scaling procedures applicable in assessing responses of the thoraxenergized by air-blast overpressures or by nonpenetrating missiles. Ann. N. Y. Acad.Sci. 152, 122–146.

Cernak, I., Noble-Haeusslein, L.J., 2010. Traumatic brain injury: an overview of pathobiologywith emphasis on military populations. J. Cereb. Blood Flow Metab. 30, 255–266.

529R.C. Turner et al. / Experimental Neurology 248 (2013) 520–529

Cernak, I., Merkle, A.C., Koliatsos, V.E., Bilik, J.M., Luong, Q.T., Mahota, T.M., Xu, L., Slack, N.,Windle, D., Ahmed, F.A., 2011. The pathobiology of blast injuries and blast-inducedneurotrauma as identified using a new experimentalmodel of injury inmice. Neurobiol.Dis. 41, 538–551.

Champion,H.R., Holcomb, J.B., Young, L.A., 2009. Injuries fromexplosions: physics, biophysics,pathology, and required research focus. J. Trauma 66, 1468–1477 (discussion 1477).

Chavko, M., Watanabe, T., Adeeb, S., Lankasky, J., Ahlers, S.T., McCarron, R.M., 2011. Rela-tionship between orientation to a blast and pressure wave propagation inside the ratbrain. J. Neurosci. Methods 195, 61–66.

Cheng, J., Gu, J., Ma, Y., Yang, T., Kuang, Y., Li, B., Kang, J., 2010. Development of a rat modelfor studying blast-induced traumatic brain injury. J. Neurol. Sci. 294, 23–28.

Covey, D.C., Born, C.T., 2010. Blast injuries:mechanics andwounding patterns. J. Surg. Orthop.Adv. 19, 8–12.

Dalle Lucca, J.J., Chavko, M., Dubick, M.A., Adeeb, S., Falabella, M.J., Slack, J.L., McCarron, R.,Li, Y., 2012. Blast-inducedmoderate neurotrauma (BINT) elicits early complement ac-tivation and tumor necrosis factor alpha (TNFalpha) release in a rat brain. J. Neurol.Sci. 318, 146–154.

de Lanerolle, N.C., Bandak, F., Kang, D., Li, A.Y., Du, F., Swauger, P., Parks, S., Ling, G., Kim,J.H., 2011. Characteristics of an explosive blast-induced brain injury in an experimentalmodel. J. Neuropathol. Exp. Neurol. 70, 1046–1057.

Duckworth, J.L., Grimes, J., Ling, G.S., 2012. Pathophysiology of battlefield associated trau-matic brain injury. Pathophysiology 20, 23–30.

Elder, G.A., Cristian, A., 2009. Blast-related mild traumatic brain injury: mechanisms ofinjury and impact on clinical care. Mt. Sinai J. Med. 76, 111–118.

Garman, R.H., Jenkins, L.W., Switzer III, R.C., Bauman, R.A., Tong, L.C., Swauger, P.V., Parks,S.A., Ritzel, D.V., Dixon, C.E., Clark, R.S., Bayir, H., Kagan, V., Jackson, E.K., Kochanek,P.M., 2011. Blast exposure in rats with body shielding is characterized primarily bydiffuse axonal injury. J. Neurotrauma 28, 947–959.

Goldstein, L.E., Fisher, A.M., Tagge, C.A., Zhang, X.L., Velisek, L., Sullivan, J.A., Upreti, C., Kracht,J.M., Ericsson, M., Wojnarowicz, M.W., Goletiani, C.J., Maglakelidze, G.M., Casey, N.,Moncaster, J.A., Minaeva, O., Moir, R.D., Nowinski, C.J., Stern, R.A., Cantu, R.C., Geiling, J.,Blusztajn, J.K., Wolozin, B.L., Ikezu, T., Stein, T.D., Budson, A.E., Kowall, N.W., Chargin,D., Sharon, A., Saman, S., Hall, G.F., Moss, W.C., Cleveland, R.O., Tanzi, R.E., Stanton, P.K.,McKee, A.C., 2012. Chronic traumatic encephalopathy in blast-exposedmilitary veteransand a blast neurotrauma mouse model. Sci. Transl. Med. 4, 134ra60.

Hyde, D.W., 2004. Conventional Weapons Effects Program. United States Army Corps ofEngineers, Vicksburg, MS.

Johnson, V.E., Stewart, J.E., Begbie, F.D., Trojanowski, J.Q., Smith, D.H., Stewart, W., 2013.Inflammation andwhitematter degeneration persist for years after a single traumaticbrain injury. Brain 136, 28–42.

Kane,M.J., Angoa-Perez,M., Francescutti, D.M., Sykes, C.E., Briggs, D.I., Leung, L.Y., VandeVord,P.J., Kuhn, D.M., 2012. Altered gene expression in culturedmicroglia in response to sim-ulated blast overpressure: possible role of pulse duration. Neurosci. Lett. 522, 47–51.

Kaur, C., Singh, J., Lim,M.K., Ng, B.L., Yap, E.P., Ling, E.A., 1995. The response of neurons andmicroglia to blast injury in the rat brain. Neuropathol. Appl. Neurobiol. 21, 369–377.

Koliatsos, V.E., Cernak, I., Xu, L., Song, Y., Savonenko, A., Crain, B.J., Eberhart, C.G.,Frangakis, C.E., Melnikova, T., Kim, H., Lee, D., 2011. A mouse model of blast injuryto brain: initial pathological, neuropathological, and behavioral characterization.J. Neuropathol. Exp. Neurol. 70, 399–416.

Kuehn, R., Simard, P.F., Driscoll, I., Keledjian, K., Ivanova, S., Tosun, C., Williams, A.,Bochicchio, G., Gerzanich, V., Simard, J.M., 2011. Rodent model of direct cranial blastinjury. J. Neurotrauma 28, 2155–2169.

Magnuson, J., Leonessa, F., Ling, G.S., 2012. Neuropathology of explosive blast traumaticbrain injury. Curr. Neurol. Neurosci. Rep. 12, 570–579.

Nakagawa, A., Manley, G.T., Gean, A.D., Ohtani, K., Armonda, R., Tsukamoto, A., Yamamoto,H., Takayama, K., Tominaga, T., 2011. Mechanisms of primary blast-induced traumaticbrain injury: insights from shock-wave research. J. Neurotrauma 28, 1101–1119.

Panzer, M.B., Matthews, K.A., Yu, A.W., Morrison III, B., Meaney, D.F., Bass, C.R., 2012. Amultiscale approach to blast neurotrauma modeling: part I — development of noveltest devices for in vivo and in vitro blast injury models. Front. Neurol. 3, 46.

Park, E., Gottlieb, J.J., Cheung, B., Shek, P.N., Baker, A.J., 2011. A model of low-level primaryblast brain trauma results in cytoskeletal proteolysis and chronic functional impair-ment in the absence of lung barotrauma. J. Neurotrauma 28, 343–357.

Rafaels, K.A., dale' Bass, C.R., Panzer, M.B., Salzar, R.S., Woods, W.A., Feldman, S.H.,Walilko,T., Kent, R.W., Capehart, B.P., Foster, J.B., Derkunt, B., Toman, A., 2012. Brain injury riskfrom primary blast. J. Trauma Acute Care Surg. 73, 895–901.

Readnower, R.D., Chavko, M., Adeeb, S., Conroy, M.D., Pauly, J.R., McCarron, R.M., Sullivan, P.G.,2010. Increase in blood–brain barrier permeability, oxidative stress, and activatedmicrog-lia in a rat model of blast-induced traumatic brain injury. J. Neurosci. Res. 88, 3530–3539.

Reneer, D.V., Hisel, R.D., Hoffman, J.M., Kryscio, R.J., Lusk, B.T., Geddes, J.W., 2011. Amulti-mode shock tube for investigation of blast-induced traumatic brain injury.J. Neurotrauma 28, 95–104.

Richmond, D.R., Damon, E.G., Fletcher, E.R., Bowen, I.G., White, C.S., 1968. The relationshipbetween selected blast-wave parameters and the response of mammals exposed toair blast. Ann. N. Y. Acad. Sci. 152, 103–121.

Risling, M., Davidsson, J., 2012. Experimental animal models for studies on the mecha-nisms of blast-induced neurotrauma. Front. Neurol. 3, 30.

Rubovitch, V., Ten-Bosch, M., Zohar, O., Harrison, C.R., Tempel-Brami, C., Stein, E., Hoffer,B.J., Balaban, C.D., Schreiber, S., Chiu, W.T., Pick, C.G., 2011. A mouse model of blast-induced mild traumatic brain injury. Exp. Neurol. 232, 280–289.

Scheibel, R.S., Newsome, M.R., Troyanskaya, M., Lin, X., Steinberg, J.L., Radaideh, M., Levin,H.S., 2012. Altered brain activation in military personnel with one or more traumaticbrain injuries following blast. J. Int. Neuropsychol. Soc. 18, 89–100.

Schmitz, C., Hof, P.R., 2005. Design-based stereology in neuroscience. Neuroscience 130,813–831.

Schneiderman, A.I., Braver, E.R., Kang, H.K., 2008. Understanding sequelae of injurymech-anisms and mild traumatic brain injury incurred during the conflicts in Iraq andAfghanistan: persistent postconcussive symptoms and posttraumatic stress disorder.Am. J. Epidemiol. 167, 1446–1452.

Shridharani, J.K., Wood, G.W., Panzer, M.B., Capehart, B.P., Nyein, M.K., Radovitzky, R.A.,Bass, C.R., 2012. Porcine head response to blast. Front. Neurol. 3, 70.

Smith, D.W., Bailes, J.E., Fisher, J.A., Robles, J., Turner, R.C., Mills, J.D., 2012. Internal jugularvein compression mitigates traumatic axonal injury in a rat model by reducing theintracranial slosh effect. Neurosurgery 70, 740–746.

Svetlov, S.I., Prima, V., Glushakova, O., Svetlov, A., Kirk, D.R., Gutierrez, H., Serebruany, V.L.,Curley, K.C., Wang, K.K., Hayes, R.L., 2012. Neuro-glial and systemic mechanismsof pathological responses in rat models of primary blast overpressure compared to“composite” blast. Front. Neurol. 3, 15.

Taber, K.H., Warden, D.L., Hurley, R.A., 2006. Blast-related traumatic brain injury: what isknown? J. Neuropsychiatry Clin. Neurosci. 18, 141–145.

Terrio, H., Brenner, L.A., Ivins, B.J., Cho, J.M., Helmick, K., Schwab, K., Scally, K., Bretthauer,R., Warden, D., 2009. Traumatic brain injury screening: preliminary findings in a USArmy Brigade Combat Team. J. Head Trauma Rehabil. 24, 14–23.

Turner, R.C., Naser, Z.J., Bailes, J.E., Smith, D.W., Fisher, J.A., Rosen, C.L., 2012. Effect of sloshmitigation on histologic markers of traumatic brain injury: laboratory investigation.J. Neurosurg. 117, 1110–1118.

Yeh, D.D., Schecter, W.P., 2012. Primary blast injuries—an updated concise review. WorldJ. Surg. 36, 966–972.