Full Terms & Conditions of access and use can be found athttp://www.tandfonline.com/action/journalInformation?journalCode=ibij20

Download by: [VCU Libraries Serials] Date: 12 November 2016, At: 06:24

Brain Injury

ISSN: 0269-9052 (Print) 1362-301X (Online) Journal homepage: http://www.tandfonline.com/loi/ibij20

Chronic cerebrovascular abnormalities in a mousemodel of repetitive mild traumatic brain injury

Cillian E. Lynch, Gogce Crynen, Scott Ferguson, Benoit Mouzon, Daniel Paris,Joseph Ojo, Paige Leary, Fiona Crawford & Corbin Bachmeier

To cite this article: Cillian E. Lynch, Gogce Crynen, Scott Ferguson, Benoit Mouzon,Daniel Paris, Joseph Ojo, Paige Leary, Fiona Crawford & Corbin Bachmeier (2016) Chroniccerebrovascular abnormalities in a mouse model of repetitive mild traumatic brain injury,Brain Injury, 30:12, 1414-1427, DOI: 10.1080/02699052.2016.1219060

To link to this article: http://dx.doi.org/10.1080/02699052.2016.1219060

Published online: 11 Nov 2016.

Submit your article to this journal

View related articles

View Crossmark data

Chronic cerebrovascular abnormalities in a mouse model of repetitivemild traumatic brain injury

Cillian E. Lynch 1,2,3, Gogce Crynen1,2,3, Scott Ferguson1,2,3, Benoit Mouzon1,2,3, Daniel Paris1,2,3, Joseph Ojo1,2,3,Paige Leary1, Fiona Crawford1,2,3, & Corbin Bachmeier1,2,3

1The Roskamp Institute, Sarasota, FL, USA, 2The Open University, Department of Life Sciences, Milton Keynes, UK, and 3James A. Haley Veteran’sAdministration Center, Tampa, FL, USA

Abstract

Primary objective: To investigate the status of the cerebrovasculature following repetitive mildtraumatic brain injury (r-mTBI).Research design: TBI is a risk factor for development of various neurodegenerative disorders. Acommon feature of neurodegenerative disease is cerebrovascular dysfunction which includesalterations in cerebral blood flow (CBF). TBI can result in transient reductions in CBF, withsevere injuries often accompanied by varying degrees of vascular pathology post-mortem.However, at this stage, few studies have investigated the cerebrovasculature at chronic timepoints following repetitive mild brain trauma.Methods and procedures: r-mTBI was delivered to wild-type mice (12 months old) twice perweek for 3 months and tested for spatial memory deficits (Barnes Maze task) at 1 and 6 monthspost-injury. At 7 months post-injury CBF was assessed via Laser Doppler Imaging and, followingeuthanasia, the brain was probed for markers of cerebrovascular dysfunction andinflammation.Main outcomes and results: Memory impairment was identified at 1 month post-injury and persistedas late as 6 months post-injury. Furthermore, significant immunopathological insult, reductions inglobal CBF and down-regulation of cerebrovascular-associated markers were observed.Conclusions: These results demonstrate impaired cognitive behaviour alongside chronic cere-brovascular dysfunction in a mouse model of repetitive mild brain trauma.

Keywords

Repetitive mild traumatic brain injury,cerebrovascular

History

Received 1 March 2016Revised 22 July 2016Accepted 26 July 2016Published online 1 November 2016

Introduction

The exact nature of the causal link between repetitive mild trau-matic brain injury (r-mTBI) and development of the neurodegen-erative disorder termed Chronic Traumatic Encephalopathy(CTE) is much researched yet disproportionately unexplained.CTE is characterized pathologically by diffuse deposition ofhyper-phosphorylated tau (pTau) immune-reactive neurofibrillaryand astrocytic tangles throughout the temporal and frontal corticesand propensity for accumulation of pTau species at perivascularlocations at the depths of cerebral sulci [1]. While the aforemen-tioned unique Tau pathology, observable only at post-mortem, is aprominent feature of CTE, this disease is also characterized by ahost of co-pathologies including, but not limited to, amyloidplaque formation, deposition of TAR DNA Binding Protein 43(TDP-43), sustained microglial inflammatory profile, white mat-ter axonal degeneration, and Lewy body neuritic inclusions [2].

Although originally described in a retired boxer asDementia Pugulistica [3], CTE has only recently becomefiercely researched, following many studies identifying the

disorder in individuals with a history of r-mTBI, includingreports documenting CTE in two former NFL players [4,5]and further case reports confirming CTE pathology post-mortem in athletes [6] and military personnel exposed toblast shock wave (BSW) injury [1]. Indeed, of all CTEcases diagnosed symptomatically in the past decade, 50%are seen in professional football players [7], more than 20%of which did not show CTE pathology post-mortem, despiteclinical diagnosis of disease, and 5% of which showed classicCTE hallmarks post-mortem despite being clinically asymp-tomatic for CTE [7]. Similarly, a recent report by Hazratiet al. [8] mirrors that mentioned, showing non-CTE likepathology and absence of overt Tau hyper-phosphorylationand aggregation in post-mortem brains of retired athletespreviously diagnosed with CTE and instead presenting withpathologies likening those of vascular and motor neuron dis-ease and AD [8]. Notwithstanding the fact that CTE can onlybe definitively diagnosed post-mortem [6], another confound-ing issue regarding its multi-faceted poly-pathology is the factthat many animal models of r-mTBI have failed to emulatethe signature pTau profile of the disease [9]. Owing to theaforementioned prevalence of CTE amongst professional ath-letes, there is an urgent need for a more diverse range of

Correspondence: Cillian E. Lynch, Roskamp Institute, 2040 WhitfieldAvenue, Sarasota, FL 34243, USA. E-mail: clynch@roskampinstitute.netColor versions of one or more of the figures in the article can be foundonline at www.tandfonline.com/ibij.

http://tandfonline.com/ibijISSN: 0269-9052 (print), 1362-301X (electronic)

Brain Inj, 2016; 30(00): 1414–1427© 2016 The Roskamp Institute. DOI: 10.1080/02699052.2016.1219060

diagnostic criteria in the mapping of CTE pathogenesis andalso a validated animal model of sports-related r-mTBI.

The primary effects of mTBI involve not only mechanicalneuronal insult such as diffuse axonal shearing, but alsotraumatic cerebral vascular injury (TCVI) [10]. The brain iscritically dependent on a steady cerebral blood flow [11].Both the maintenance of physiologically adequate cerebralperfusion and any changes therein due to increased CentralNervous System (CNS) metabolism or to the tight coupling ofCBF to increased neuronal activity (a process known asneurovascular coupling) are, in part, dependent on the func-tion of vascular-associated Smooth Muscle Cells (SMCs) andpericytes [11], whose dysfunction is believed to play a defin-ing role in neuropathological conditions such as AD [12].TCVI manifests as extensive pericyte pathology post-mortemin the acute and sub-acute phase following a severe TBI[13,14] and also as brain contusion and capillary compressionby astrocytic end-feet postmortem [15] and signs of Blood–Brain Barrier (BBB) penetrance [16]. The causal link of ahistory of repetitive brain trauma to disease pathogenesis isnot monogamous for CTE, as many reports have positedrepetitive injuries as a pre-disposing factor in developmentof dementia and AD [17–20]. Indeed, hypoperfusion has beennoted as a definitive risk factor for development of AD[21,22] and a rapid conversion of mild cognitive impairment(MCI) to AD with reduced CBF in human patients [23].Changes in both regional and global CBF are observed inhuman patients following a single severe TBI, with vasos-pasm accompanying reduced flow in the acute period post-injury, before normalization of CBF to basal levels [15].These perfusion-specific mTBI effects have also been recapi-tulated in animal models, with a single cortical impact mTBIsufficient to cause reduced CBF and behavioural deficit inmice [24]. All of these data indicate a distinct role for cere-brovascular dysregulation in the pathogenesis of TBI-relateddisease.

All of the data mentioned postulate cerebrovascular dys-function as a contributory factor in pathophysiology, CTE orotherwise, following r-mTBI in both humans and animalmodels. However, pre-clinical information regarding r-mTBIand the cerebrovasculature at the chronic time-point post-injury is lacking. As such, the current study demonstratesthat the repetitive injury-dependent pathology seen in agedwild type mice several months following r-mTBI is accom-panied by dysregulation of cerebral blood flow and alteredexpression of cerebrovascular markers at this chronic time-point post-injury.

Materials and methods

Animals

Male and female C57BL/6 mice were housed under standardlaboratory conditions (23 ± 1°C, 50 ± 5% humidity and 12-hourlight/dark cycle) with free access to food andwater throughout thestudy. All procedures were carried out under institutional animalcare and use committee approval in accordance with the NationalInstitute of Health Guide for the Care and Use of LaboratoryAnimals.

Injury schedule

Twelve-month-old animals were randomly assigned to one oftwo groups: repetitive mild Traumatic Brain Injury (r-mTBI,delivered twice each calendar week, i.e. one hit approximatelyevery 72 hours, over a period of 3 months) or repetitive Sham(r-Sham; animals underwent the same duration and frequencyof anaesthesia as r-mTBI animals). An electromagneticimpactor (Leica Instruments; Leica, Wetzlar, Germany) wasused to generate a midline mTBI, using a 5.0 mm diameterflat face tip, 5 m s–1 strike velocity, 1.0 mm strike depth and a200 millisecond dwell time, as previously characterized [25].The mice were euthanized 7 months after the final injury/anaesthesia (22 months of age).

Assessment of cognitive function

Cognitive functionwas assessed at 1 and 6 months after the finalinjury/anaesthesia (15 months of age and 21 months of age,respectively) by use of the Barnes maze, as described previouslyby this group [26]. Researchers conducting the experimentswere blind to grouping and the Ethovision XT System(Noldus; Noldus, Wageningen, Netherlands) was used to trackand record the movement of each animal. Mice were given 90seconds to locate and enter the target box and required to remainin the target box 30 seconds prior to retrieval, regardless ofsuccess. For a series of 6 consecutive days, four trials weregiven per day, withmice starting from one of four cardinal pointson each trial. The inter-trial interval for each mouse on any givenday of acquisition was ~ 40 minutes. The maze platform andretrieval box were both cleaned thoroughly between trials so asto limit the confounding effects of scent on performance of themice during each trial. On the seventh day, a single probe triallasting 60 seconds was performed with the mouse starting fromthe centre of the maze and the target box removed. Escapelatency measured the time taken for the mouse to enter the box.

In addition to assessment of learning and spatial memoryvia use of the Barnes maze, animals were tested for anxiety-like behaviour in the elevated plus maze (EPM).

Determination of regional cerebral blood flow by laserscanner doppler imaging

For CBF measurements, 22 month-old r-Sham and r-mTBIwild type mice (at 7 months post-injury) were anaesthetizedwith a gas mixture of 3% isoflurane and 0.5 L min–1 oxygen,immobilized on a mouse stereotaxic table (Kopf Instruments;Kopf, Tujunga, CA, USA). Body temperature was monitoredvia rectal probe and maintained at 37°C using a mouse home-othermic blanket system (Harvard Apparatus; Holliston, MA,USA). An incision was made through the scalp and the skinretracted to expose the skull. The periosteal connective tissue,which adheres to the skull, was removed with a sterile cottonswab. Cortical perfusion was measured with the LaserDoppler Perfusion Imager from Moor Instruments as pre-viously described [27]. A computer-controlled optical scannerdirected a low-powered He–Ne laser beam over the exposedcortex. The scanner head was positioned parallel to the cere-bral cortex at a distance of 26 cm. The scanning procedure

DOI: 10.1080/02699052.2016.1219060 Repetitive mild traumatic brain injury 1415

took 1 minute 21 seconds for measurements of 5538 pixelscovering an area of 0.8 × 0.8 cm and six replicate images permouse were collected. At each measuring site, the beamilluminated the tissue to a depth of ~ 0.5 mm. An imagecolour-coded to denote specific relative perfusion levels wasdisplayed on a video monitor. All images were stored incomputer memory for subsequent analysis. For each animal,a square area of 0.05 cm2 (360 pixels) equally distributedbetween the right and left hemispheres was defined andapplied to each image of the series in order to measure theCBF in the entire, frontal and occipital cortex using the MoorLDI Image Processing V3.0h software. CBF was also mea-sured by manually delineating for each mouse the cortex area(0.51–0.54 cm2, corresponding to 3504–3714 pixels).Relative perfusion values for each area studied wereexpressed as arbitrary units.

Histology

Following CBF measurements at 22 months of age, theanimals were anaesthetized with isoflurane and perfusedtrans-cardially with phosphate-buffered saline (PBS), pH7.4, the brain was removed from the skull and one hemi-sphere was dissected to separate the brainstem, cerebellumand olfactory bulbs from the rest of the hemisphere forbiochemical analysis. The other hemisphere was post-fixedin 4% paraformaldehyde solution at 4°C for 48 hours andparaffin-embedded for Immunohistochemistry as previouslydescribed [26]. For each group, sagittal sections were cut at6 µm in thickness (lateral 1 in 10 series, 0.1–0.2 mm), withthe first 10 consecutive sections cut mounted one by oneonto 10 separate consecutive slides and the next 10 sectionsbelow these, resulting in 10 slides, each with two sections,per animal. Sections were deparaffinized with histoclear (15minute incubation) followed by re-hydration via decreasingethanol gradient. Sections were then incubated in 3% hydro-gen peroxide in order to quench endogenous peroxidaseactivity and then underwent citrate buffer (pH 6.0) antigenretrieval for 7 minutes. Following antigen retrieval, sectionswere blocked in 2.5% normal serum at room temperaturefor ~ 30 minutes and then incubated with primary antibodyovernight at 4°C. Sections were stained with primary anti-bodies against glial fibrillary acid protein (GFAP; DAKO,rabbit anti-GFAP polyclonal, ZO334, 1:10 000 dilution) andionized calcium binding adaptor protein 1 (Iba1; Abcam,goat anti-Iba1 polyclonal, Ab5076, 1:2500 dilution), eitheranti-rabbit or anti-goat secondary antibodies, using respec-tive Vector ABC rabbit IgG or goat IgG Vectastain reagentkits (Vector Laboratories; Bulingame, CA, USA) andrevealed using the Avidin-Biotin peroxidase DAB substratesolution (Vector DAB peroxidase substrate kit, VectorLaboratories, SK4105). Sections were counter-stained withMayer’s Hematoxylin (Sigma-Aldrich; St Louis, MO,USA). Slides were visualized with a bright field microscope(BX60; Tokyo, Japan) and digital images were visualizedand acquired using a MagnaFire SP camera (Olympus). Atleast two sections (one slide) were stained per animal foreither Iba1 or GFAP immunohistochemistry, with at leastfive non-overlapping 40× magnification images of cortexanalysed per brain section and at least two non-overlapping

60× magnification images of the Cornu Ammonis 1 (CA1)and at least six non-overlapping 60× images of the body ofthe Corpus Callosum (CC) regions of interest analysed perbrain section. The percentage area of image positive foreither GFAP or Iba1 immunoreactivity was then analysedby Zeiss Software following manual thresholding of theimage. The percentage area of individual regions of interestacross sections for each mouse were averaged and thenanalysed as biological replicates per group using unpairedStudent t-test.

Biochemistry

Western Blot analysis was carried out on brain tissue homo-genates, following homogenization of tissue by sonication inHALTTM Protease and Phosphatase Inhibitor Cocktail mPER-containing lysis buffer. Equivalent total protein amounts wereanalysed on sodium dodecyl sulphate-polyacrylamide gelelectrophoresis under denatured and reduced (and denaturedand non-reduced) conditions using 4–20% Bis-Tris pre-castGels (Biorad; Hercules, CA, USA) and electro-blotted on to apolyvinylidene difluoride (PVDF) membrane overnight at 90mA constant current. Membranes were then washed in de-ionized water, before being blocked for 1 hour at room tem-perature with 5% non-fat milk in tris buffered saline (TBS).Membranes were then incubated with primary antibodiesovernight (12–16 hours at 4°C). The following primary anti-bodies were used at the given concentrations; GFAP (DAKO,rabbit anti-GFAP polyclonal, ZO334, 1:10 000 dilution),laminin (Sigma-Aldrich, rabbit anti-mouse laminin, L9393,1:50 dilution), platelet derived growth factor receptor β(PDGFRβ; Abcam, rabbit anti-PDGFRβ monoclonal,ab32570, 1:1000 dilution), alpha smooth muscle actin(αSMA; Millipore, mouse anti-αSMA monoclonal, ASM-1,1:1000 dilution), Iba1 (Abcam, goat anti-Iba1 polyclonal,Ab5076, 1:2500 dilution) and rabbit anti-GAPDH, glyceral-dehyde 3-phosphate dehydrogenase (Sigma-Aldrich, 1:1000dilution). Membranes were washed with deionized water,incubated with their respective secondary antibody for 1hour at 4°C, washed once more and then developed usingECL chemiluminescent detection reagent (GE Life Sciences).Membranes were imaged using a Biorad ChemiDoc WesternBlot Imager and densitometry results of individual bandswere collected using ImageLab 5.2 (Biorad; Hercules, CA,USA) software. Target protein values for each lane werenormalized against densitometry values for GAPDH fortheir respective lane. Target protein values were also normal-ized against the total protein stain for their respective laneusing a commassie blue stain (data not shown) to validate theGAPDH-normalized results.

Quantitative assessment of S100β plasma concentration

Analysis of S-100 calcium binding protein β (S100 β) con-centration of plasma samples obtained from 22 month oldwild type r-Sham and r-mTBI mice was carried out by meansof an enzyme-linked immuno-sorbent assay (ELISA;MBS261174, MyBioSource; San Diego, CA, USA) as permanufacturer’s direction and data expressed as pico-grams(pg) per millilitre (mL).

1416 C. E. Lynch et al. Brain Inj, 2016; 30(00): 1414–1427

Statistical analysis

All data were assessed as if they were coming from a nor-mally distributed population, using skewness and kurtosis ofthe distribution. If there were statistically significant kurtosisand skewness, the data was transformed using log2 or squareroot transformation. When transformation did not yield anormally distributed data set, non-parametric Kruskal-Wallistest was used.

For Barnes Maze, normally distributed data was analysedusing parametric methods. If a given dataset was normallydistributed and variances were equal, One-way ANOVA (forprobe) or repeated measures ANOVA (for acquisition) wereused to assess significant changes due to injury. Mauchly’ssphericity test was used to evaluate sphericity in repeatedmeasures ANOVA and, if it was statistically significant,degrees of freedom were corrected with Greenhouse-Geisserestimates of sphericity. Two-way ANOVA was used to assesssignificant differences between groups on only day 6 ofacquisition trials.

For the measurement of cerebral blood flow, the data wasfirst tested for unequal variances using the O’Brien test forhomogeneity of variance. If the variances are not equal,Welch’s test was used. For the analysis of CBF measurements,

two-way ANOVA of values was run on a one-way ANOVAplatform, selecting the replicate values as block, i.e. X and Yvariables were analysed via two-way ANOVA to reveal statis-tical differences in X. Both Western Blot andImmunohistochemistry data were analysed using unpairedStudent t-test. A given effect was considered significant at p <0.05 and indicated by asterisks in the figures. Error bars repre-sent the standard error of the mean. Statistical analyses wereperformed using JMP 11.1.1 (JMP; Cary, NC, USA) and graphswere created using GraphPad Prism 5.0.

Results

Assessment of cerebral blood flow

Average global or entire cortex CBF was markedly decreasedin wild type r-mTBI mice compared to their respectiver-Sham controls (Figure 1(d), wild type r-mTBI vs r-Sham,10.66% ± 2.04%, p < 0.001). There was no significant dif-ference in average CBF between the frontal cortex regions ofr-mTBI and r-Sham mice (Figure 1(e), p > 0.05); however,there is a significant decrease in average CBF in the occipitalcortex of r-mTBI mice, compared to r-Sham controls(Figure 1(f), p < 0.001), indicating that the decrease in

Figure 1. Representative replicate images (three per mouse) of two-dimensional colour-coded microvasculature flow maps of both sham and injuredaged wild-type mice (a and b, respectively) obtained using a Moor Instruments® Laser Doppler Imager. Orientation of a representative dorsal view of amicrovasculature flow map along the posterior-anterior axis is shown in (c). Mouse skull was exposed under isoflurane anaesthetic and a bright-fieldimage taken before every repeat scan to allow for future sub-division of the entire cortex reading into frontal and occipital regions of interest relative toBregma and Lambda. Six replicate scans per mouse were analysed to give the average cerebral blood flow for the entire cortex, occipital cortex andfrontal cortex for each group (n = 6), represented here as scatter dot plot figures (d–f, respectively). The cerebral blood flow recordings were analysedusing a two-way Analysis of Variance (ANOVA) mixed model ANOVA, selecting the replicate values as block (n = 6 for each group). At 7 monthspost-injury, wild type r-mTBI mice show a statistically significant decrease in average cerebral blood flow recordings in the entire cortex of 10.66% ±2.04%, compared to age matched r-Sham controls. A statistically significant difference in cerebral blood flow is not seen between the frontal cortices ofr-Sham and r-mTBI mice; however, a statistically significant and injury-dependent decrease in occipital cortex CBF of 11.87% ± 2.33% for wild-typer-mTBI mice, compared to their r-Sham controls, is evident several months following injury (f). Wild type mice: six r-Sham, six r-mTBI. Statisticalanalysis was conducted using two-way ANOVA. Error bars represent mean ± SEM.

DOI: 10.1080/02699052.2016.1219060 Repetitive mild traumatic brain injury 1417

average global CBF readings in the wild type r-mTBI mice isbeing driven by the decrease in CBF in the occipital and notfrontal cortical regions.

Barnes maze acquisition

At both the 1 month and 6 month time-points following the finalinjury/anaesthesia, r-mTBI mice travelled a significantly greatercumulative distance than their r-Sham controls (Figures 2(a) and(b), p < 0.001). Additionally, a distinct effect of injury on cumu-lative distance travelled over time, reflected by a progressiveseparation of cumulative distance travelled by r-mTBI mice,

compared to r-sham controls, across the 6 consecutive days ofacquisition at 1 month post-injury (Figure 2(a), r-Sham vsr-mTBI; p = 0.0028) and a non-significant effect of treatmenton cumulative distance with time at 6 months (Figure 2(b),r-Sham vs r-mTBI; p = 0.07) were noted. The time-dependenteffect of injury across the 6 days of acquisitionwas not significantat 1month post-injury (Figure 2(c), p=0.08); however, this injuryby time interaction was significant at 6 months post-injury(Figure 2(d), p < 0.001), with the average velocity of r-mTBImice being statistically greater than that of r-Sham controls on day6 of acquisition at 6 months (Figure 2(d), p < 0.001) but not at 1month (Figure 2(c), p = 0.2) post-injury.

Figure 2. Evaluation of learning (acquisition) and retention of spatial memory of wild type mice using the Barnes maze at 1 and 6 months followingrepetitive mild traumatic brain injury. Mice were tested in the Barnes maze for their ability to locate a black box at the target hole. During the course ofthe 6 days of acquisition at both the 1 month sub-acute and 6 month chronic time-points post-injury, the mTBI-treated mice travelled a greater meancumulative distance to reach the target hole, compared to sham controls (a and b, p < 0.001, repeated measures ANOVA). In cumulative distance data,the injury by time interaction term was statistically significant across the 6 days of acquisition at the 1 month post-injury time-point (p < 0.01,Repeated Measures ANOVA); however, this was not seen at the 6 month time-point (p = 0.07, repeated measures ANOVA). There was also asignificant effect of injury on mean velocity between groups across all 6 days of acquisition at both the sub-acute 3 month and the chronic 6 monthtime-point (c and d; p < 0.001, repeated measures ANOVA) with a significant increase in velocity of r-mTBI mice vs r-Sham controls on day 6 ofacquisition at 6 months post-injury (d; p < 0.001, Two Way ANOVA). There was no significant effect of injury with time on mean velocity at 1 monthpost-injury (c; p = 0.08, repeated measures ANOVA) and no significant difference in velocity between groups on day 6 of acquisition at 1 month post-injury (c, p = 0.2, two-way ANOVA). Evaluation of spatial memory retention (Probe) of wild type mice using the Barnes maze, at 1 and 6 monthsfollowing r-mTBI (e and f, respectively). For the probe trial (the day immediately following the 6 consecutive days of acquisition testing), the target boxwas removed and mice were placed in the middle of the table for a single, 60-second trial. Probe test performance was significantly impaired in ther-mTBI mice at 1 month (a, p < 0.01, Welch’s Test) and 6 months (b, p < 0.01, one-way ANOVA), compared to r-Sham controls. Data are presented asMean ± Standard Error of the Mean (SEM); 19 r-Sham and 13 r-mTBI at 1 month; 18 r-Sham and six r-mTBI at 6 months. Statistical p-values onfigure represent statistical analysis only for day 6 of acquisition. The injury by time interaction values are not shown in the figure.

1418 C. E. Lynch et al. Brain Inj, 2016; 30(00): 1414–1427

Barnes maze probe

Probe test performance was profoundly impaired in r-mTBImice, compared to sham controls, at both 1 month (Figure 2(e), p < 0.01) and 6 months (Figure 2(f), p < 0.01) post-injury. At the 1 month time point, the r-Sham mice displayeda statistically significant increase in frequency of pokes intothe target hole during the 60 second probe trial, compared tor-mTBI mice, with r-Sham mice having a 98.8% ± 22.9%increased frequency in nose-poke in target hole, compared tor-mTBI mice at 1 month post-injury (p = 0.0119, One-WayANOVA, data not shown). There was no statistically signifi-cant difference in mean velocity, distance travelled or totalerrors made between groups during the probe trial at either 1month or 6 months post-injury (p < 0.05, data not shown).

Elevated plus maze

Following probe testing at 1 month post-injury, r-mTBI andr-Sham mice were assessed for anxiety-like behaviour via theElevated Plus Maze (EPM, Figure 3). There was no signifi-cant difference in percentage time spent in the open arms byr-mTBI mice, compared to r-Sham controls (Figure 3(a), p >0.05), indicating a lack of disinhibition at 1-month post-injuryin r-mTBI mice and, so, similar levels of anxiety betweenboth r-Sham and r-mTBI animals at this time-point.Additionally, there was no significant difference in the num-ber of entries of r-mTBI mice in to the open arms (Figure 3(b), p > 0.05), compared to r-Sham controls.

GFAP immunostaining and biochemistry

Immunohistochemical analysis of glial fibrillary acidic pro-tein (GFAP) immunostaining was carried out on the brainregions underlying the impact site, specifically the somato-sensory and primary motor cortices, Corpus Callosum (CC)and Cornu Ammonis 1 (CA1) of the Hippocampus, of ther-Sham and r-mTBI mice from which Cerebral Blood Flowrecordings were obtained. There was no statistically signifi-cant difference in the percentage area of GFAP-positive label-ling in images taken of the cortices of r-Sham (0.1523% ±0.05%) vs r-mTBI (0.1234% ± 0.03%) animals (Figure 4(a), p= 0.89). The lack of injury-effect on GFAP labelling in the

cortex was not accompanied by the presence of hypertrophicGFAP-positive cells, but instead by a morphologically quies-cent GFAP-positive cell phenotype in both r-Sham andr-mTBI mice (Figure 4(a), boxed inset of cell in magnifiedareas indicated by arrows). Pursuant to the immunopatholo-gical GFAP labelling in the cortex, western analysis of thecontralateral brain homogenate did not demonstrate a statis-tically significant difference in GFAP expression betweenr-Sham and r-mTBI mice at 7 months post-injury (Figure 4(b), P=0.57). Immunohistochemical analysis of the CA1revealed no significant difference in GFAP immunoreactivitybetween r-Sham and r-mTBI animals (Figures 5(a, b and i); p= 0.41); however, there was a noticeable increase in astro-gliosis in the CC of r-mTBI mice (Figure 5(d), 9.05% ±0.68% of area) compared to that of r-Sham controls(Figure 5(c), 5.69% ± 1.14% of area); however, this was notsignificant (Figure 5(j), p = 0.0519).

Iba1 immunohistochemistry and biochemistry

The degree of glial reactivity throughout the superficial layersof the cortices, the CA1 and the CC of all wild type micefrom which CBF recordings were taken, was assessed viaimmunostaining of brain sections with a macrophage/micro-glial marker antibody specific to ionized calcium bindingadaptor molecule 1 (Iba1). Immunohistochemical analysis ofthe Cortex (Figure 6(a)) revealed no significant difference inIba1 staining between the cortices of r-Sham (0.48% ±0.04%) and r-mTBI (0.67% ± 0.09%) mice (Figure 6(a), p= 0.065). Furthermore, it appeared that cells positive for Iba1in the superficial layer of cortex of either r-Sham or r-mTBIanimals were rounded, but not amoeboid in shape, displayedlong thin processes and a spherical shaped cell soma(Figure 6(a), boxed inset of cell in magnified areas indicatedby arrows), indicating an unreactive phenotype.Immunohistochemical evaluation of Iba1 expression in thecortex of these animals was confirmed by Western BlotAnalysis (Figure 6(b)), which showed no significant differ-ence in brain homogenate Iba1 levels between r-Sham andr-mTBI mice (Figure 6(b), p = 0.67) at 7 months post-injury.Immunohistochemical analysis of the CA1 revealed no sig-nificant difference in Iba1 immunoreactivity between r-Sham(Figure 5(e)) and r-mTBI (Figure 5(f)) animals (Figure 5(k), p

Figure 3. Evaluation of anxiety using the elevated plus maze (EPM) at 1 month post-injury. (a) Injured animals spent on average 79% more time in theopen arms of the EPM compared to r-Sham controls; however, this trend in behaviour was not statistically significant (p > 0.05, unpaired Student t-test). Additionally, r-mTBI animals exhibited ~ 26% greater frequency of entries into the open arms of the maze, compared to r-Sham controls;however, this result was not significant (b, p > 0.05, unpaired Student t-test). Data are presented as mean ± standard error of the mean (SEM); 19 r-Sham and 13 r-mTBI at 1 month.

DOI: 10.1080/02699052.2016.1219060 Repetitive mild traumatic brain injury 1419

= 0.89); however, there was a significant increase in Iba1levels in the CC of r-mTBI mice (Figure 5(h), 1.25% ± 0.01%of area) compared to that of r-Sham (Figure 5(g), 0.88% ±0.07% of area) control mice (Figure 5(l), p < 0.01).

Biochemical analysis of vascular markers

Immunoblot analysis of the blood vessel markers laminin,PDGFRβ and αSMA was carried out in wild type r-Shamand r-mTBI mice. The antibody chosen for the experimentsherein (Rabbit Anti-laminin, Sigma-Aldrich, L9393) wasraised against the α (440 kDa) and β/γ (220 kDa) chains ofpurified mouse laminin; however, this antibody does notdetect the α1 chain on western blots consistent with previousreports [28] and so immunoblot analysis detected a 200–220kDa protein band appropriate to be the size of β/γ chains.

Western analysis of Wild Type brain homogenate revealed astatistically significant increase in levels of Laminin, PDGFRβ,but not αSMA, in r-mTBI mice, compared to r-Sham controls(data not shown, p < 0.01; PDGFRβ, p < 0.01; αSMA, p =0.3). However, as mural cell expression is dependent on vesseldensity, PDGFRβ and αSMA expression was normalized to thelaminin/GAPDH ratio for each lane in the gel, which is por-trayed in the graphs seen in Figures 7(b) and (c)). The values for

PDGFRβ and αSMA, when normalized to their respectiveLaminin proportions, showed a statistically significant decreasein PDGFRβ and αSMA r-mTBI cortex of 30% and 33%,respectively, compared to r-Sham controls (PDGFR, Figure 7(b), p < 0.01; αSMA, Figure 7(b), p < 0.05).

BBB penetrance: ELISA measurement of S100β andbiochemical analysis of occludin

Quantitative analysis of plasma at 7 months post-injury (22months of age) via ELISA revealed no significant differencein peripheral plasma concentration of the S100βmarker of BBBrupture/leakage between r-Sham and r-mTBI mice at 7 months(22 months of age) post-injury (Figure 8(a), p > 0.05).Expression of the tight junction protein occludin was not sig-nificantly decreased in r-mTBI mice, compared to r-Sham con-trols, at 7 months post-injury (Figure 8(b), p > 0.05), indicatingno disruption of BBB penetrance at this time-point post-injury.

Discussion

Understanding the initial primary injuries and evolving second-ary injury cascades behind repetitive head trauma is paramount

Figure 4. Astrogliosis in the cortex of animals 7 months following r-mTBI. (a) There was no difference in expression of GFAP within the cortex of ther-mTBI group when compared to their sham counterpart (p = 0.89, unpaired Student t-test). (b) Western Blot analysis of brain homogenates showed nosignificant difference in GFAP expression between groups at 7 months post-injury (p = 0.57, unpaired Student t-test). Boxed inset in (a) shows 60×magnification of area indicated by the arrow. Astrocytes in the superficial layer of cortex of both r-sham and r-mTBI animals appeared to bemorphologically quiescent, and opposed to hypertrophic, in shape. Immunohistochemistry: six r-Sham, six female; and six r-mTBI, three male, threefemale. Values represent percentage area of image analysed. Error bars represent ± SEM, scale bars in images represent 20 μm in length.

1420 C. E. Lynch et al. Brain Inj, 2016; 30(00): 1414–1427

to the development of new therapies to combat disorders likeCTE in the human population. Using a mouse model of repeti-tive injury, this study demonstrates that repetitive injury, given ata frequency similar to that endured by professional athletes,results in profound impairment of global cerebral blood flow.This impairment in CBF was associated with persistent cogni-tive deficits evident at 1 month and 6 months post-injury, asdemonstrated by impaired acquisition and consolidation of spa-tial memory in the Barnes Maze. The diminished CBF record-ings and behavioural effect identified here following r-mTBI arein accordancewith CBF impairment observed following a singlemTBI [24] in rodent models of TBI and prior reports from thelab describing neurobehavioural deficits at chronic time-pointsfollowing repetitive injury [26].

Laser Doppler Flowmetry (LDF) is a non-invasive method toestimate the relative blood perfusion in the microcirculation.The LDF technique used here is not only useful in that is areal-time measurement of blood flow in these mice, but alsopertinent due to the recent use of a similar technique (transcra-nial Doppler ultrasonography) investigating sports-related inju-ries in humans, a TBI demographic which this animal model

emulates in terms of the frequency of hits sustained over aprolonged period of time. Len et al. [29] recently demonstratedan impaired Cerebrovascular Reactivity (CVR) response in ath-letes followingmTBI for days following the injury. CVR reflectsthe capacity of blood vessels to dilate and is an important markerof brain vascular reserve in humans [30]. Recently, recordings ofregional and global CBF from retired military personnel showsveterans having suffered a TBI to exhibit a significantly lowerCBF recording than non-TBI controls [31]. Owing to the failureof many animal models of repetitive injury to recapitulate thekind of extensive p-tau immunoreactivity, neurofibrillary tan-gles (NFTs) and amyloid burden seen in the human populationand indeed the fact that, as mentioned, tau/amyloid pathologypost-mortem may not be the best indication of development ofCTE, the recapitulation of CBF impairment in both athletes andretired military personnel by this r-mTBI mouse model may beof great translational significance. Adding further credence tothe diagnostic and pre-clinical relevance of CBF impairment inrepetitive injury-induced pathogenesis, a recent report by Amenet al. [32] demonstrated global and regional CBF impairments inNational Football League (NFL) players, as measured by Single

Figure 5. Assessment of glial cell reactivity in the cortex of animals 7 months following injury. (a) Immunohistochemistry of brain sections did notreveal a significant difference in glial cell count, as assessed by Iba1 immunostaining, between the cortex of injured animals and sham controls post-injury (p = 0.065, unpaired Student t-test). (b) Western Blot analysis of brain homogenates showed no significant difference in Iba1 expressionbetween groups at 7 months post-injury (p = 0.67, unpaired Student t-test). Boxed inset in (a) shows 60× magnification of area indicated by the arrow.Glia in the superficial layer of cortex of both r-sham and r-mTBI animals appeared to be small and rounded, but not amoeboid in shape, with long, thinprocesses, indicative of a non-reactive microglial state. Immunohistochemistry: six r-Sham, six female; six r-mTBI, three male and three female.Biochemical analysis: six r-Sham, six female; six r-mTBI, three male and three female. Values expressed as percentage of area of image analysed. Errorbars represent ± SEM, Scale bars in images represent 20 μm in length.

DOI: 10.1080/02699052.2016.1219060 Repetitive mild traumatic brain injury 1421

Photon Emission Computed Tomography (SPECT) imaging,compared to healthy control individuals. The perfusion studiescoincidewith these findings as both a global reduction of CBF inr-mTBI mice, compared to control animals, and also brainregion-specific alterations in the aftermath of brain traumawere observed.

In order to ascertain whether the observed changes in CBFin the mouse model of r-mTBI were due to alterations invessel status, biochemical analysis of the protein laminin, aknown component of the vascular basement membrane, wasconducted [33]. Significant increases were observed inLaminin expression at 7 months following injury in r-mTBImice. Recent studies describe angiogenesis and vasculogen-esis in both human patients and animal models following TBIand stroke [34]. Angiogenesis involves the proliferation ofendothelial cells and sprouting of micro-vessels, both culmi-nating in a gross increase in intra-cerebral vascular density[35]. Indeed, neovascularization at a chronic time-point post-injury has been seen at as late a time-point as 9 months post-injury in a Lateral Fluid Percussion injury mTBI model in the

rat [36]. This evidence of chronic increases in vascularizationof upwards of 28% (as detected using arterial spin labellingMRI) are not unlike observations made here of increasedexpression of laminin in r-mTBI mice.

Pursuant to determining the culpability of changes invessel density as a cause of reduced CBF, this study alsoinvestigated the possibility of vascular-related mural cellinvolvement in this phenomenon. Leptomeningeal arteriesand pre-capillary arterioles are covered by SMCs, whichexpress the contractile component αSMA, wrap around thevessel with a band-like circumferential morphology and arethought to be principally responsible for vasomotion andregional CBF flux [11]. Capillaries are bereft of SMCs andare instead covered by pericytes, expressing high constitu-tive levels of PDGFRβ [11]. The involvement of SMCs andpericytes in CBF regulation [11,37–45] has been well char-acterized and, as such, these markers have been chosen forinterrogation of mural cell involvement in r-mTBI inducedCBF dysregulation. In order to represent the observedmural cell marker changes as a function of vessel density,

Figure 6. Evaluation of astrogliosis and glial cell reactivity by expression of glial fibrillary acidic protein (GFAP) and ionized calcium binding adaptorprotein 1 (Iba1), respectively, in the CA1 and corpus callosum (CC) of wild type r-Sham and r-mTBI mice at 7 months post-injury (GFAP (a–d); Iba1(e–h)). There was no significant difference in expression of GFAP-positive cells between wild type r-Sham (a) and r-mTBI (b) mice in the CA1 of thehippocampus at 7 months post-injury (i, p = 0.78, Unpaired Student t-test); however, analysis of GFAP immunoreactivity in the CC revealed amoderate increase in astrogliosis in the CC of r-mTBI mice (d, 9.05% ± 0.68% of area) compared to that of r-Sham controls (c, 5.69% ± 1.14% ofarea), although this was not significant (j, p = 0.0519, unpaired Student t-test). Analysis of Iba1 expression in the CA1 (e and f) and CC (g and h) ofwild type mice 7 months post-injury revealed no significant difference in the degree of gliosis in the CA1 (k, p > 0.05, Unpaired Student t-test);however, there was a significant increase in the degree of Iba1 immuno-positive staining in the CC of r-mTBI mice (1.25% ± 0.01% of area) comparedto r-Sham controls (0.88% ± 0.07% of area) at 7 months post-injury (l, p < 0.05, unpaired Student t-test). GFAP Immunohistochemistry: six r-Shamand six r-mTBI. Iba1 Immunohistochemistry: four r-Sham, four Female; six r-mTBI, three male and three female. Values expressed as a percentage ofarea of image analysed. Error bars represent ± SEM, scale bars in images represent 60 μm in length for images taken at 40× magnification and 100 μmin length for images taken at 20× magnification.

1422 C. E. Lynch et al. Brain Inj, 2016; 30(00): 1414–1427

the values of PDGFRβ and αSMA were normalized to therelative amount of laminin in each sample and a significantdecrease in levels of both PDGFRβ and αSMA wasobserved in r-mTBI mice, compared to r-Sham animals. Itmay be worth noting that this study examined whole-brainhomogenate and, thus, it is impossible to surmise whetheror not the reductions in mural cell markers seen herein areconfined to either large leptomeningeal and intraparenchy-mal arteries, smaller arterioles or both. With the exceptionof aforementioned work by Dore-Duffy et al. [38], theexpression of αSMA has not been extensively researchedwith respect to mTBI. However, it has been demonstratedthat intracerebral and leptomeningeal SMCs taken from ADpatient brains express significantly greater levels of αSMA,compared with age-matched control subject samples, when

normalized to β-tubulin [46], a housekeeping protein whoseexpression was found not to be altered by the AD pheno-type [37,46]. Similarly, Chow et al. [46] reported anincrease in αSMA, as opposed to the decrease cataloguedby this group; however, their study [46] did not normalizeαSMA expression to the associated vasculature, as has beendone in the current report, but instead to the individualvesicular SMC β-tubulin levels. Furthermore, recent studiesby Merlini et al. [47] demonstrated an early-onset BraakTau-dependent, amyloid pathology-independent decrease inlarge leptomeningeal artery αSMA expression in the brainsof human AD patients, with smaller arterioles exhibiting asignificant decrease in αSMA at later disease progression.Akin to this approach, the relative αSMA expression wasobtained by normalization to vessel wall levels [47]. These

Figure 7. Western immunoblot analysis of brain homogenates of wild type r-Sham and r-mTBI mice. (a) Expression of the neurovascular-associatedbase-membrane protein laminin was significantly increased in the brains of r-mTBI mice (0.18 ± 0.02, Arbitrary units (AU)) compared to r-Shamcontrols (0.26 ± 0.03, AU, p < 0.01, unpaired Student t-test). (b) Levels of platelet derived growth factor receptor β (PDGFRβ) were significantlydecreased in the cortex of r-mTBI mice compared to that of r-Sham controls (r-mTBI; 11.52 ± 0.61 AU vs r-Sham; 16.63 ± 0.07, p < 0.01, unpairedStudent t-test) when normalized to laminin expression throughout the brain. (c) Expression of the contractile-associated protein α-smooth muscle actin(αSMA) was significantly lower in r-mTBI cortex compared to r-Sham control cortex (5.67 ± 0.49 AU for r-Sham, vs 3.85 ± 0.32 for r-mTBI, p <0.05, unpaired Student t-test), when normalized to laminin expression throughout the brain. Six r-Sham and six r-mTBI. Values expressed as AU. Errorbars represent ± SEM. All densitometry values for laminin for individual bands were normalized to the glyceraldehyde 3-phosphate dehydrogenase(GAPDH) value for their respective lane and this ratio used for statistical analysis. Values for individual lanes for PDGFRβ and αSMA were bothnormalized to the laminin/GAPDH ratio of the same lane for statistical analysis and generation of graphs (b) and (c).

DOI: 10.1080/02699052.2016.1219060 Repetitive mild traumatic brain injury 1423

results may in part explain the hypoperfusion deficits seenwith early AD [21–23] and provide rationale for how thereduced αSMA expression observed in this report may becausative of the reduced CBF noted in r-mTBI mice.

Although some studies have designated capillaries and, bydefault, pericytes as the principal contributors to regionalCBF both in normal physiology [40,43,48] and followingcerebral ischaemia [40,42,44], others [11,39,41] have catalo-gued pericytes as unimportant in these regards. However, asmentioned earlier, pericytes have been implicated in TBI-induced hypoperfusion in both human patients [15] and alsoin animal model studies, one such being a recent report byDore-Duffy et al. [38], showing the release of endothelin-1(ET1), a potent vasoconstrictor, from damaged pericytes 4hours following a closed head injury. Inhibition of ET1 in-vivo has also been shown to both alleviate impairment inautoregulation in the piglet cerebrovasculature [49] and torestore CBF levels in the rat [50] following mTBI and eleva-tion of ET1 levels in CSF is shown to be correlated withunfavourable outcomes in children following severe TBI [51].DeGracia et al. [52] have shown a robust stress response,identified by Heat Shock protein 70 (HSP-70) immunofluor-escence, confined to the capillary bed of the mouse micro-vasculature and coinciding with endothelial and pericyte celldeath, following a single mTBI. Using quantitative autoradio-graphy, Bell et al. [37] have demonstrated that knockdown inPDGFRβ signalling leads to pericyte deficiency-inducedimpairments in CBF and BBB integrity in adult mice, inaddition to age-dependent neuronal degeneration and novelobject recognition memory deficits. As opposed to the 10%reduction in global CBF recorded in r-mTBI mice, compared

to r-Sham controls, in this study, Bell et al. [37] observedCBF reductions of up to 50% in the cortex and hippocampusof 14–16 month old mutant mice. However, the decrease inPDGFRβ expression and any consequent effects on CBF seenin this study are not directly relatable to those seen by Bellet al. [37]. First, the increases in laminin expression in ther-mTBI mice in this study may reflect injury-induced angio-genesis and/or thickening of laminae, whereas Bell et al.documented a significant decrease in microvessel length anddecreased laminin expression, as a result of PDGFRβ defi-ciency. Second, the mice used in the aforementioned studyexhibited impaired PDGFRβ signalling from birth, asopposed to an injury-induced depletion later in life as withthese animals. However, in spite of these caveats, it is clearthat disruption of pericytes, be it through direct ablation ofthe cells or encroached PDGFRβ activity, can dramaticallyimpair CBF in mouse models. Further research is merited todiscover whether the decrease in PDGFRβ following r-mTBIreflects reduced pericyte density or mere alterations in theexpression of this receptor.

These findings, alongside observations of reduced mural cellmarker expression following r-mTBI, indicate a role for muralcells in the pathophysiological profile following brain trauma.Select representative images from the superficial layers of cortexof both wild type r-Sham and r-mTBI mice show non-hyper-trophic astroglia and microglia with small and round cell bodiesand many dendritic processes, indicative of a resting or unreac-tive state of both astrocytes and microglia [53]. There is also anincrease in GFAP and Iba1 immunoreactivity in the CC, data inagreement with previous studies from this group investigatingastrocytic and glial immunoreactivity at this time-point post

Figure 8. Evaluation of BBB leakage of r-Sham and r-mTBI mice at 7 months post-injury (22 months of age). (a) Assessment of peripheral plasmaconcentrations of the glial-associated Ca2+ binding protein S100β at 7 months post-injury in wild type r-Sham vs r-mTBI mice. There was nostatistically significant effect of repetitive injury on peripheral plasma concentration of S100β at the chronic 7-month post-injury time-point (p = 0.77,Unpaired Student t-test), as assessed by ELISA. Six r-Sham and six r-mTBI. (b) Biochemical analysis of brain homogenates revealed no significantdifference in expression of the tight junction protein occludin between r-Sham and r-mTBI mice at 7 months post-injury (p > 0.05, unpaired Studentt-test). Values expressed as pico-gram per millilitre of plasma sample. Error bars represent ± SEM. ELISA; six r-Sham and five r-mTBI. Western BlotAnalysis; five r-Sham and five r-mTBI, values expressed as AU. Error bars represent ± SEM. All densitometry values for individual bands werenormalized to the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) value for their respective lane and this ratio used for statistical analysis.

1424 C. E. Lynch et al. Brain Inj, 2016; 30(00): 1414–1427

injury [26]. The fact that there is neither a discernable increase inGFAP/Iba1 immunoreactivity in the cortex, nor a reactive phe-notype of these cells would suggest that glial-associated neuro-inflammation may not be a factor with regards to the injury-induced hypoperfusion seen at 7 months post-injury in thesestudies. Instead, the reduced CBF observed following r-mTBI ismore likely due to changes in the expression profile of thecerebrovascular mural cell markers.

Another factor which may contribute to alterations in CBF isloss of blood–brain barrier (BBB) integrity. BBB disruptions caninclude mild and transient openings or loss of tight junctionsaltogether, pronounced brain pathology and persistence of neuro-logical deficit, via increased extravasation of peripherally-derivedimmune cells into the intraparenchymal space and impaired reg-ulation of transendothelium and paracellular transport of ions andmolecules [54,55]. Increased plasma concentrations of moleculessuch as the glial-derived calcium-binding protein S100B, whichare typically not found in the periphery on account of theirresidency in the CNS [16], are indicators of BBB penetrance inboth human patients and animal models of mTBI [16,56].Evidence for the immediate injurious effects of mTBI on theBBB include a 2013 study by Abdul-Muneer et al. [56], showingthat youngmale Sprague-Dawley rats exposed to a single 123KPaintensity mild Blast Shock Wave (BSW) exposure injury causedupregulation of markers of oxidative and nitrosative stress in braincapillaries, which progressed into BBB disruption, as reflected bydown-regulation of the tight junction protein occludin andreduced histological expression of PDGFRβ, all at 1–24 hourspost-injury [56]. The acute effects of this r-mTBI on BBB perme-ability were also accompanied by an elevation of plasma concen-trations S100β and neuron-specific protein enolase for severalhours following injury [56]. In order to assess the integrity ofthe BBB at the chronic 7 month time-point post-injury in thesemice, plasma concentration of S100β were measured via ELISA.There was no significant increase in peripheral levels of S100β inr-mTBImice, as compared to r-Sham controls, inferring that BBBpenetrance and/or vascular rupture at this time-point was unaf-fected by repetitive injury. In addition, no significant changeswere observed in occludin, indicating that endothelial cell tightjunction integrity is also unscathed at this time-point. As appreci-able changes were not observed in either S100β or occludin, it isunlikely the chronic impairment of CBF seen in this animalmodelis due to changes in BBB integrity.

Conclusion

A previous report from this laboratory demonstrated r-mTBIpathology in conjunction with marked and persistent beha-vioural deficits at chronic time-points following five repeti-tive injuries [26]. This study describes a separate injuryparadigm of greater frequency, occurring over a longer periodof time. This r-mTBI model more closely represents theoccurrence of injuries endured by athletes or military person-nel over the course of an entire career. In the r-mTBI mousemodel, an injury-dependent effect was observed of r-mTBI onglobal CBF readings and a down-regulation in mural cellmarkers several months after the final injury. Moreover, thisstudy demonstrates marked behavioural deficits and neuroin-flammatory hallmarks consistent with the pathology observedin other mouse TBI models and human TBI cases. Future

work will explore these concepts further in order to charac-terize the nature of the cerebrovasculature following braintrauma and elucidate the association between vascular dys-function and brain pathology after repetitive injuries to thebrain.

ORCID

Cillian E. Lynch http://orcid.org/0000-0002-4980-0354

Declaration of interest

The authors report no conflicts of interest. This research wassupported by the Roskamp Foundation and by the ChronicEffects of Neurotrauma Consortium (CENC - VA I01RX001774 and DoD W81XWH-13-2-0095). Dr. Crawford is aVA Research Career Scientist.

References1. Goldstein LE, Fisher AM, Tagge CA, Zhang XL, Velisek L, Sullivan

JA, Upreti C, Kracht JM, EricssonM,WojnarowiczMW, et al. Chronictraumatic encephalopathy in blast-exposed military veterans and a blastneurotrauma mouse model. Science Translational Medicine2012;4:134ra60.

2. McKee AC, Daneshvar DH, Alvarez VE, Stein TD. The neuro-pathology of sport. Acta Neuropathologica 2014;127:29–51.

3. Millspaugh JA. Dementia pugilistica. US Naval Medicine Bulletin1937;35:261–297.

4. Omalu BI, DeKosky ST, Hamilton RL, Minster RL, Kamboh MI,Shakir AM,Wecht CH. Chronic traumatic encephalopathy in a nationalfootball league player: part II. Neurosurgery 2006;59:1086–1092; dis-cussion 1092–1093.

5. Omalu BI, DeKosky ST, Minster RL, Kamboh MI, Hamilton RL,Wecht CH. Chronic traumatic encephalopathy in a National FootballLeague player. Neurosurgery 2005;57:128–134; discussion 128–134.

6. McKee AC, Stern RA, Nowinski CJ, Stein TD, Alvarez VE,Daneshvar DH, Lee HS, Wojtowicz SM, Hall G, Baugh CM,et al. The spectrum of disease in chronic traumatic encephalopathy.Brain 2013;136:43–64.

7. Gardner A, Iverson GL, McCrory P. Chronic traumatic encephalo-pathy in sport: a systematic review. British Journal of SportsMedicine 2014;48:84–90.

8. Hazrati LN, Tartaglia MC, Diamandis P, Davis KD, Green RE,Wennberg R, Wong JC, Ezerins L, Tator CH. Absence of chronictraumatic encephalopathy in retired football players with multipleconcussions and neurological symptomatology. Frontiers in HumanNeuroscience 2013;7:222.

9. Ojo JO, Mouzon BC, Crawford F. Repetitive head trauma, chronictraumatic encephalopathy and tau: Challenges in translating frommice to men. Experimental Neurology 2016;275:389–404.

10. Kenney K, Amyot F, Haber M, Pronger A, Bogoslovsky T, MooreC, Diaz-Arrastia R. Cerebral vascular injury in traumatic braininjury. Experimental Neurology 2015;87:95–110.

11. Hill RA, Tong L, Yuan P, Murikinati S, Gupta S, Grutzendler J.Regional blood flow in the normal and ischemic brain is controlledby arteriolar smooth muscle cell contractility and not by capillarypericytes. Neuron 2015;87:95–110.

12. Sagare AP, Bell RD, Zhao Z, Ma Q, Winkler EA, Ramanathan A,Zlokovic BV. Pericyte loss influences Alzheimer-like neurodegen-eration in mice. Nature Communication 2013;4:2932.

13. Castejon OJ. Ultrastructural pathology of cortical capillary peri-cytes in human traumatic brain oedema. Folia Neuropathologica2011;49:162–173.

14. Castejon OJ. Ultrastructural alterations of human cortical capillarybasement membrane in human brain oedema. FoliaNeuropathologica 2014;52:10–21.

15. Ostergaard L, Engedal TS, Aamand R, Mikkelsen R, Iversen NK,Anzabi M, Naess-Schmidt ET, Drasbek KR, Bay V, Blicher JU,et al.: All authors needed Capillary transit time heterogeneity and

DOI: 10.1080/02699052.2016.1219060 Repetitive mild traumatic brain injury 1425

flow-metabolism coupling after traumatic brain injury. Journal ofCerebral Blood Flow Metabolism 2014;34:1585–1598.

16. Berger RP, Adelson PD, Pierce MC, Dulani T, Cassidy LD,Kochanek PM. Serum neuron-specific enolase, S100B, and myelinbasic protein concentrations after inflicted and noninflicted trau-matic brain injury in children. Journal of Neurosurgery 2005;10312Suppl):61–68.

17. Gardner RC, Burke JF, Nettiksimmons J, Goldman S, Tanner CM,Yaffe K. Traumatic brain injury in later life increases risk forParkinson disease. Annals of Neurology 2015;77:987–995.

18. Gardner RC, Yaffe K. Epidemiology of mild traumatic brain injury andneurodegenerative disease. Molecular and Cellular Neuroscience2015;66B:75–80.

19. Plassman BL, Havlik RJ, Steffens DC, Helms MJ, Newman TN,Drosdick D, Phillips C, Gau BA, Welsh-Bohmer KA, Burke JR, et al.Documented head injury in early adulthood and risk of Alzheimer’sdisease and other dementias. Neurology 2000;55:1158–1166.

20. Schofield PW, Logroscino G, Andrews HF, Albert S, Stern Y. Anassociation between head circumference and Alzheimer’s disease ina population-based study of aging and dementia. Neurology1997;49:30–37.

21. Ruitenberg A, den Heijer T, Bakker SL, van Swieten JC, KoudstaalPJ, Hofman A, Breteler MM. Cerebral hypoperfusion and clinicalonset of dementia: the Rotterdam Study. Annals of Neurology2005;57:789–794.

22. Snowdon DA, Greiner LH, Mortimer JA, Riley KP, Greiner PA,Markesbery WR. Brain infarction and the clinical expression ofAlzheimer disease. The Nun Study. JAMA 1997;277:813–817.

23. Hirao K, Ohnishi T, Hirata Y, Yamashita F, Mori T, Moriguchi Y,Matsuda H, Nemoto K, Imabayashi E, Yamada M, et al. Theprediction of rapid conversion to Alzheimer’s disease in mildcognitive impairment using regional cerebral blood flow SPECT.Neuroimage 2005;28:1014–1021.

24. Ahn MJ, Sherwood ER, Prough DS, Lin CY, DeWitt DS. Theeffects of traumatic brain injury on cerebral blood flow and braintissue nitric oxide levels and cytokine expression. Journal ofNeurotrauma 2004;21:1431–1442.

25. Mouzon B, Chaytow H, Crynen G, Bachmeier C, Stewart J, MullanM, Stewart W, Crawford F. Repetitive mild traumatic brain injuryin a mouse model produces learning and memory deficits accom-panied by histological changes. Journal of Neurotrauma2012;29:2761–2773.

26. Mouzon BC, Bachmeier C, Ferro A, Ojo JO, Crynen G, Acker CM,Davies P, Mullan M, Stewart W, Crawford F. Chronic neuropatho-logical and neurobehavioral changes in a repetitive mild traumaticbrain injury model. Annals of Neurology 2014;75:241–254.

27. Paris D, Quadros A, Humphrey J, Patel N, Crescentini R, CrawfordF, Mullan M. Nilvadipine antagonizes both Abeta vasoactivity inisolated arteries, and the reduced cerebral blood flow in APPswtransgenic mice. Brain Research 2004;999:53–61.

28. Halfter W, Dong S, Schurer B, Osanger A, Schneider W, Ruegg M,Cole GJ. Composition, synthesis, and assembly of the embryonicchick retinal basal lamina. Developmental Biology 2000;220:111–128.

29. Len TK, Neary JP, Asmundson GJ, Goodman DG, Bjornson B,Bhambhani YN. Cerebrovascular reactivity impairment after sport-induced concussion. Medical Science and Sports Exercise2011;43:2241–2248.

30. Yezhuvath US, Lewis-Amezcua K, Varghese R, Xiao G, Lu H. Onthe assessment of cerebrovascular reactivity using hypercapniaBOLD MRI. NMR Biomedicine 2009;22:779–786.

31. Ponto LL, Brashers-Krug TM, Pierson RK, Menda Y, Acion L,Watkins GL, Sunderland JJ, Koeppel JA, Jorge RE. Preliminaryinvestigation of cerebral blood flow and amyloid burden inVeterans with and without combat-related traumatic brain injury.Journal of Neuropsychiatry and Clinical Neurosciences2016;28:89–96.

32. Amen DG, Willeumier K, Omalu B, Newberg A, Raghavendra C,Raji CA. Perfusion neuroimaging abnormalities alone distinguishnational football league players from a healthy population. Journalof Alzheimers Disease 2016;53:237–241.

33. Searcy JL, Le Bihan T, Salvadores N, McCulloch J, Horsburgh K.Impact of age on the cerebrovascular proteomes of wild-type andTg-SwDI mice. PLoS One 2014;9:e89970.

34. Xiong Y, Mahmood A, Chopp M. Angiogenesis, neurogenesis andbrain recovery of function following injury. Current Opinion inInvestigative Drugs 2010;11:298–308.

35. Prakash R, Carmichael ST. Blood-brain barrier breakdown andneovascularization processes after stroke and traumatic braininjury. Current Opinions in Neurology 2015;28:556–564.

36. Hayward NM, Tuunanen PI, Immonen R, Ndode-Ekane XE,Pitkanen A, Grohn O. Magnetic resonance imaging of regionalhemodynamic and cerebrovascular recovery after lateral fluid-per-cussion brain injury in rats. Journal of Cerebral Blood FlowMetabolism 2011;31:166–177.

37. Bell RD, Winkler EA, Sagare AP, Singh I, LaRue B, Deane R,Zlokovic BV. Pericytes control key neurovascular functions andneuronal phenotype in the adult brain and during brain aging.Neuron 2010;68:409–427.

38. Dore-Duffy P, Wang S, Mehedi A, Katyshev V, Cleary K, TapperA, Reynolds C, Ding Y, Zhan P, Rafols J, et al. Pericyte-mediatedvasoconstriction underlies TBI-induced hypoperfusion.Neurological Research 2011;33:176–186.

39. Fernandez-Klett F, Offenhauser N, Dirnagl U, Priller J, LindauerU. Pericytes in capillaries are contractile in vivo, but arteriolesmediate functional hyperemia in the mouse brain. Proc Natl AcadSci U S A 2010;107:22290–22295.

40. Hall CN, Reynell C, Gesslein B, Hamilton NB, Mishra A,Sutherland BA, O’Farrell FM, Buchan AM, Lauritzen M, AttwellD. Capillary pericytes regulate cerebral blood flow in health anddisease. Nature 2014;508:55–60.

41. Kornfield TE, Newman EA. Regulation of blood flow in the retinaltrilaminar vascular network. Journal of Neuroscience 2014;34:11504–11513.

42. O’Farrell FM, Attwell D. A role for pericytes in coronary no-reflow.Nature Reviews: Cardiology 2014;11:427–432.

43. Peppiatt CM, Howarth C, Mobbs P, Attwell D. Bidirectional con-trol of CNS capillary diameter by pericytes. Nature 2006;443:700–704.

44. Yemisci M, Gursoy-Ozdemir Y, Vural A, Can A, Topalkara K,Dalkara T. Pericyte contraction induced by oxidative-nitrative stressimpairs capillary reflow despite successful opening of an occludedcerebral artery. Nature Medicine 2009;15:1031–1037.

45. Yamin R, Morgan KG. Deciphering actin cytoskeletal function inthe contractile vascular smooth muscle cell. Journa of Physiology2012;590:4145–4154.

46. Chow N, Bell RD, Deane R, Streb JW, Chen J, Brooks A, VanNostrand W, Miano JM, Zlokovic BV. Serum response factor andmyocardin mediate arterial hypercontractility and cerebral bloodflow dysregulation in Alzheimer’s phenotype. Proceedings of theNational Academy of Sciences (USA) 2007;104:823–828.

47. Merlini M, Wanner D, Nitsch RM. Tau pathology-dependent remodel-ling of cerebral arteries precedes Alzheimer’s disease-related microvas-cular cerebral amyloid angiopathy. Acta Neuropathologica2016;131:737–752.

48. Chaigneau E, OheimM,Audinat E, Charpak S. Two-photon imaging ofcapillary blood flow in olfactory bulb glomeruli. Proceedings of theNational Academy of Sciences (USA) 2003;100:13081–13086.

49. Armstead WM, Bohman LE, Riley J, Yarovoi S, Higazi AA, CinesDB. tPA-S(481)A prevents impairment of cerebrovascular autore-gulation by endogenous tPA after traumatic brain injury by upre-gulating p38 MAPK and inhibiting ET-1. Journal of Neurotrauma2013;30:1898–1907.

50. Petrov T. Amelioration of hypoperfusion after traumatic braininjury by in vivo endothelin-1 knockout. Canadian Journal ofPhysiology and Pharmacology 2009;87:379–386.

51. Salonia R, Empey PE, Poloyac SM, Wisniewski SR, Klamerus M,Ozawa H, Wagner AK, Ruppel R, Bell MJ, Feldman K, et al..Endothelin-1 is increased in cerebrospinal fluid and associated withunfavorable outcomes in children after severe traumatic braininjury. Journal of Neurotrauma 2010;27:1819–1825.

52. DeGracia DJ, Kreipke CW, Kayali FM, Rafols JA. Brain endothelialHSP-70 stress response coincides with endothelial and pericyte deathafter brain trauma. Neurological Research 2007;29:356–361.

53. Torres-Platas SG, Comeau S, Rachalski A, Bo GD, Cruceanu C,Turecki G, Giros B, Mechawar N. Morphometric characterization ofmicroglial phenotypes in human cerebral cortex. Journal ofNeuroinflammation 2014;11:12.

1426 C. E. Lynch et al. Brain Inj, 2016; 30(00): 1414–1427

54. Abbott NJ, Patabendige AA, Dolman DE, Yusof SR, Begley DJ.Structure and function of the blood-brain barrier. Neurobiology ofDisease 2010;37:13–25.

55. Obermeier B, Daneman R, Ransohoff RM. Development, mainte-nance and disruption of the blood–brain barrier. Nature Medicine2013;19:1584–1596.

56. Abdul-Muneer PM, Schuetz H, Wang F, Skotak M, Jones J,Gorantla S, Zimmerman MC, Chandra N, Haorah J. Induction ofoxidative and nitrosative damage leads to cerebrovascular inflam-mation in an animal model of mild traumatic brain injury inducedby primary blast. Free Radical Biology and Medicine2013;60:282–291.

DOI: 10.1080/02699052.2016.1219060 Repetitive mild traumatic brain injury 1427