Repeated Mild Closed Head Injury Impairs Short-Term
Visuospatial Memory and Complex Learning
Michael J. Hylin, 1 Sara A. Orsi, 1 Natalia S. Rozas, 1 Julia L. Hill, 1 Jing Zhao, 1 John B. Redell, 1 Anthony N. Moore, 1 and Pramod K. Dash 1,2 Abstract Concussive force can cause neurocognitive and neurobehavioral dysfunction by inducing functional, electrophysiological, and/or ultrastructural changes within the brain. Although concussion-triggered symptoms typically subside within days to weeks in most people, in 15%20% of the cases, symptomology can continue beyond this time point. Problems with memory, attention, processing speed, and cognitive exibility (e.g., problem solving, conict resolution) are some of the prominent post-concussive cognitive symptoms. Repeated concussions (with loss or altered consciousness), which are common to many contact sports, can exacerbate these symptoms. The pathophysiology of repeated concussions is not well understood, nor is an effective treatment available. In order to facilitate drug discovery to treat post-concussive symptoms (PCSs), there is a need to determine if animal models of repeated mild closed head injury (mCHI) can mimic the neurocognitive and histopathological consequences of repeated concussions. To this end, we employed a controlled cortical impact (CCI) device to deliver a mCHI directly to the skull of mice daily for 4 days, and examined the ensuing neurological and neurocognitive functions using beam balance, foot-fault, an abbreviated Morris water maze test, context discrimination, and active place avoidance tasks. Repeated mCHI exacerbated vestibulomotor, motor, short-term memory and conict learning impairments as compared to a single mCHI. Learning and memory impairments were still observed in repeated mCHI mice when tested 3 months post-injury. Repeated mCHI also reduced cerebral perfusion, prolonged the inammatory response, and in some animals, caused hippocampal neuronal loss. Our results show that repeated mCHI can reproduce some of the decits seen after repeated concussions in humans and may be suitable for drug discovery studies and translational research. Key words: active place avoidance, axonal injury, context discrimination, neuroinammation Introduction A concussion can cause neurocognitive and neurobeha- vioral dysfunctions by inducing functional, electrophysio- logical, and/or ultrastructural changes within the brain. Although the terms concussion and mild traumatic brain injury (mTBI) are often used interchangeably, concussion is the term typically used to describe mTBI in humans. According to the CDC, an estimated 1.7 million people a year sustain a TBI in the United States, with approximately 75% of these cases considered to be mild TBI or concussions. 1 This likely underestimates the number of cases, as it is estimated that close to 40% of people who have experienced a concussion do not seek medical attention. 2 The incidence of con- cussion is high in military populations, with studies reporting that approximately 20% of Operation Enduring Freedom/Operation Iraqi Freedom (OEF/OIF) veterans have had a clinical diagnosis of concussion. 3 Prominent neurological symptoms of concussion include problems with balance, headache, vomiting, nausea, and sleep disturbances. The most frequently reported cognitive symp- toms include problems with memory, attention, processing speed, and complex learning (e.g., problem solving/conict resolution). Existing literature indicates that in the majority of patients with concussions, symptoms typically resolve within days to weeks following the injury. 4 However, in 15%20% of the cases, symp- tomology can continue beyond the acute time point. 57 While the neural mechanism(s) underlying the neurological, cognitive, and behavioral dysfunctions resulting from a single concussion are not well understood, there is less understanding of the pathobiology of repeated concussions. Repeated concussions (which includes loss or altered con- sciousness) are common to many contact sports. For example, studies examining the frequency of head injuries in American football players have reported that sustaining a concussion triples the risk of sustaining a second concussion. 8,9 Athletes who have 1 Department of Neurobiology and Anatomy, and 2 The Vivian L. Smith Department of Neurosurgery, The University of Texas Medical School at Houston, Houston, Texas. JOURNAL OF NEUROTRAUMA 30:716726 (May 1, 2013) Mary Ann Liebert, Inc. DOI: 10.1089/neu.2012.2717 716 sustained repeated concussions have been shown to have impaired visual-spatial perception and complex learning. 10 Postmortem gross anatomical analyses of the brains from athletes with a history of repeated concussions indicate enlarged lateral and third ventri- cles, atrophy of the temporal and frontal cortices, and sclerosis of the hippocampus. 11 These anatomical changes are consistent with the observed impairments of episodic memory and high level cognitive functions in persons sustaining repeated concussions. In addition, recent studies have indicated that repeated concussions or subconcussive injuries can lead to the development of chronic traumatic encephalopathy (CTE), a neurodegenerative disease characterized by tauopathy. 11,12 As repeated concussions can cause lasting and potentially dev- astating impairments, there is a need to determine if animal models of repeated mTBI can mimic the observed neurocognitive and histopathological consequences of repeated concussions in humans in order to facilitate translational research. A few experimental studies in rodents have employed closed head injury (CHI) models delivering a range of 210 impacts in order to examine the ensuing histopathological changes and learning decits. 13,14 While these studies have shown that repeated mCHI can cause spatial learning and memory decits, inammation, decreased glucose metabolism, and axonal injury, it has not been examined if repeated mCHI to rodents can cause learning impairments in cognitive tasks that re- quire greater exibility. To address this knowledge gap, we de- livered four mCHI, each 24 h apart, to mice using a controlled cortical impact (CCI) device as described previously. 1517 Con- sistent with previous reports, we observed that repeated mCHI produced no visible brain contusion, and that inammation and axonal damage were more prominent after repeated mCHI than after a single injury. Furthermore, repeated mCHI reduced cerebral perfusion, and exacerbated decits in vestibulomotor and motor function, short-term visuospatial memory, and conict learning. Methods Mild closed head injury Mild closed head injury (mCHI) was delivered to mice using a pneumatically driven CCI device, essentially as described previ- ously. 16,17 Male C57BL/6 mice (2025 g) were anesthetized ini- tially with 5% isourane in a 1:1 O 2 /air mixture, mounted on a stereotaxic frame, then maintained with a 2.5% isourane and 1:1 O 2 /air mixture via a face mask. Amidline incision was made and the soft tissue reected to expose the skull. A stereotaxic frame was used to immobilize the animal during the surgical incision and en- sure that a smooth cut was made. The mouse was then transferred to a foampad designed to make the head level with the body. At 40 sec after discontinuation of anesthesia, a single impact was applied to the skull. For this model, our objective was to deliver a mCHI to a partially awake animal as assessed by the recovery of the tail-pinch response. Because the response to a tail pinch is an abrupt move- ment, this could not be tested in the animals prior to the injury as the resultant movement would upset the positioning of the skull beneath the impactor. The 40 sec delay was based upon our prior data, which showed that uninjured mice regain their tail-pinch reex on average 40 sec after discontinuation of anesthesia. A metal impactor tip (5 mm in diameter) was driven at a velocity of 5.0m/sec to a depth of 1.0 mm. The center of the tip was located midway between lambda and bregma, over the sagittal suture. Immediately following the injury, the animals were monitored for apnea, and when normal breathing was observed, the scalp was closed using sterile surgical staples. Duration of apnea, tail pinch, and righting responses were assessed as described previously. 18,19 Animals with obvious skull fractures (approximately 25%) were excluded from this study. Groups of animals received either a single mCHI, or repeated mCHIs (once per day for 4 days). Single mCHI animals were in- jured on the day of the last injury for the repeated mCHI. In order to assess the potential consequences of repeated exposure to anes- thesia, an additional group of mice received daily anesthesia ex- posure for 3 days prior to a single mCHI on the fourth day. The rationale for using four impacts, separated by 24 h each, is based on previous studies. 13,20 In both these studies, a single mild injury was insufcient to produce reliable decits in their behavioral tasks. However, DeFord et al. observed that four injuries at 24 h intervals resulted in signicant cognitive impairments. 13 Likewise, Meehan et al. found that between 3 to 5 repeated injuries spaced every 24 h produced measurable decits in spatial memory. 20 Collectively, these studies indicated that a minimum of three to four repeated injuries spaced 24 h apart would be necessary to cause measurable cognitive decits. Sham animals went through all of the anesthesia and surgical procedures for 4 days but did not receive a mCHI. Cerebral blood perfusion imaging Cerebral perfusion was measured using a PIM3 scanning laser Doppler device (Perimed, Ardmore, PA). Mice were initially an- esthetized using 5% isourane with a 1:1 O 2 /air mixture, mounted on a stereotaxic frame, then maintained with a 2.5% isourane and 1:1 O 2 /air mixture via a nose cone. After the skull was exposed, the scan eld was set so that the center of the scan was on the midline. The scanhead was aligned via the targeting laser and a built-in CCD camera. Distance to the subject was calculated prior to every scan to ensure consistency, and respiration was monitored throughout the scanning procedure. Ten consecutive scans were performed, and the resultant perfusion units averaged to give a perfusion value for each animal. Pre-injury measurements were obtained as a baseline for each animal, and used for normalization across animals. Prior to each days injury, animals were anesthetized, and once breathing rate had stabilized, they were scanned. After completion of the scanning, the animals were removed from the stereotaxic device and injured as described above. Sham and single mCHI groups were scanned prior to injury (baseline), 2 min, 2 h, and 24 h post- injury or post-sham surgery. Repeated mCHI subjects were scan- ned prior to their rst injury (baseline) and then scanned prior to each subsequent injury. A nal scan was performed 24 h after the fourth mCHI. Vestibulomotor and motor functions Beam balance and foot-fault tasks were used to determine vesti- bulomotor and motor performances following injury, respectively. 21,22 For beam balance, animals were placed on a narrow wooden beam (0.5 cm wide) and the duration they remained on the beam was measured (for a maximum of 60 sec). Animals were given three daily trials on the beam balance task. Foot-fault was evaluated by placing the animal on a wire grid (1 1 cm) and the number of foot misplacements out of a total of 50 steps were counted. 22 A foot- fault was dened as when a front paw missed and appeared below the plane of the wire grid. The foot-fault test was repeated three times to give an average daily score. Abbreviated Morris water maze In order to assess short-term memory, we used a modied ver- sion of the Morris water maze task in which mice are trained to nd the location of a hidden platform within a single training session followed by a short-term probe trial. 23,24 Mice were given 10 consecutive training trials with an inter-trial interval (iti) of 15 min. Each trial was initiated by placing the animal into the water maze at one of four randomly chosen starting positions. The animal was allowed to search for the hidden platform for a period of 90 sec, and the time to nd the platformrecorded. If the animal failed to nd the REPEATED MILD CLOSED HEAD INJURY 717 hidden platform on any given trial, it was led there by the experi- menter. Thirty minutes following the last training trial, animals were tested in a probe trial in which the platformwas removed fromthe tank and allowed to search for a period of 60 sec. Movement within the maze was monitored using a video camera linked to tracking software (Ethovision, Noldus Information Technology, Leesbury, VA, USA). Measures of memory, including latency to rst platform crossing, number of crossings, and quadrant preference, were recorded. Context discrimination This task assesses an animals ability to distinguish between two similar contexts, and has been shown to be dependent on hippo- campal function. Contextual discrimination was assessed essen- tially as described previously. 25 Animals were pre-exposed (without shock) to two contexts sharing certain features (horizontal grid oor, background noise, animal handling to and from the room) while differing in others (differently spaced grids, scent, distal cues and oor shape). Animals were given two trials, one in each chamber, each day. Animals were placed in the shock chamber and 178 sec later, a 2 sec, 0.75 mA shock was given. In the safe chamber, animals were free to roam for 3 min and no shock was given. Animals were exposed to the shock and safe chambers once a day for 3 days. Discrimination of the two contexts was assessed by comparing the time spent freezing in each chamber during training. Active place avoidance and conict learning Different congurations of this task can be used to assess learning ranging from simple (passive avoidance) to complex (conict) learning. 26,27 In the active avoidance conguration, the animal learns to avoid a region (wedge formed form a 60 arc measured from the center) where they receive a shock. After the animal has learned to avoid this shock zone, the shock zone is moved to a new location, thereby creating a conict. An abbrevi- ated three-phase protocol, modeled after Burghardt and colleagues, was used to assess place avoidance and conict learning. 28 Each phase lasted 20 min and was separated by a 2 h period in which the animal was allowed to remain in their home cage. In the rst phase (pre-training), animals were allowed to move around the rotating platform (moving clockwise at a speed of 1 rpm) with the shock turned off. This allowed the animals to become familiar with the environment. This also served as a pre-training baseline to identify any exploratory differences between the groups. During the second phase (active avoidance), the shock zone was activated and the animal had to learn to avoid the area delineated by a 60 arc from the center of the platform, which was located in a consistent region of the rotating arena. An animals entrance into the zone resulted in a brief constant current footshock (500 ms, 60 Hz, 0.3 mA) that was scrambled across the rods lining the arena. The third phase of the task consisted of moving the shock zone 180 in order to test conict learning. The arena was located in a room that had nu- merous visual cues on the walls. Analysis of the number of times the animal entered the shock zone was computed using Track Analysis software (Tracker, Bro-Signal Group Corp., SUNY Technology Center, Brooklyn, NY). Immunohistochemistry Animals were deeply anesthetized using sodium pentobarbital. Once the animal failed to respond to foot and tail pinch, the chest cavity was opened and the animal transcardially perfused with phosphate-buffered saline (PBS) followed by 4% paraformalde- hyde in PBS. Brains were removed and post-xed overnight. Re- presentative brains were cryoprotected in a 30% sucrose solution in PBS. Coronal brain sections (40 lm) were prepared using a cryo- stat. Immunohistochemistry was performed by incubating brain sections with primary antibodies (1 lg/mL) overnight at 4C in a solution consisting of PBS with 0.25% Triton X-100 (PBST), 2.0% BSA, and 2.5% normal goat or horse serum. After extensive washing in PBST, tissue sections were incubated for 1 h in species- specic secondary antibodies conjugated to Alexa-uors. Sections were washed in PBST, mounted on microscope slides, and visu- alized using epiuorescence detection. Silver impregnation Silver staining was carried out on free-oating sections using a kit from FD Neurotechnologies (Columbia, MD), essentially as described by the vendor with one exception: the impregnation time was extended from 4 min to 6 min in order to maximize the signal- to-noise ratio. Statistical analysis Statistical comparisons were carried out using SigmaStat (Systat Software, San Jose, CA). Across group comparisons of data col- lected over time (e.g., behavioral training and blood owmeasures) were evaluated using a repeated measures two-way ANOVA, fol- lowed by post-hoc analysis. Group main, or interactions of group and time, differences were used to compare the groups. Single measure data (e.g., probe trial data) was statistically compared using an one-way ANOVA, whereas within group comparisons (e.g., quadrant preference) were tested using a repeated measures one-way ANOVA. Data were considered signicant at p < 0.05. Results The timeline of the experimental design to assess the conse- quences of repeated mCHI is shown in Figure 1A. Acute neurological changes resulting from a single and repeated mCHI Duration of apnea, and suppression of tail pinch reex and righting response were measured for sham, single, and repeated mCHI groups. A single mCHI caused apnea for 31.8 3.5sec. This duration of apnea was not signicantly altered by repeated mCHI (F (3,11) =2.277, p =0.099) (Fig. 1B). When compared to sham ani- mals, single and repeated mCHI animals demonstrated a signicant increase in the duration of suppression of the tail pinch reex (Fig. 1C) (F (2,31) =19.71, p <0.001). However, the duration of this reex did not change within the repeated mCHI group across days (F (3,11) =2.646, p =0.066). When compared to the shamanimals, the suppression of the righting responses were signicantly longer in both single and repeated mCHI animals (Fig. 1D) (F (2,31) =30.679, p <0.001), but did not differ between the two groups. Furthermore, animals that received repeated anesthesia and a single mCHI (n =9) were found to have comparable acute neurological responses to single anesthesia, single mCHI animals (n =9) with the exception of a modest decrease in suppression of the righting response in the multiple anesthesia group. Cerebral perfusion is decreased acutely after mCHI Sham and single mCHI groups (n =6/group) were scanned using a PIM3 Imager prior to injury (baseline), 2 min, 2 h, and 24 h post- injury or sham surgery. Figure 2A shows a representative picture of a mouse mounted in the stereotaxic frame with the position of the scan eld (white box) indicated. Color-coded images in Figure 2B indicate the relative cerebral perfusion of a mouse subjected to a single mCHI. Black boxes indicate the areas of the cortex used for quantication. Perfusion in the right and left hemispheres were averaged. The summary data in Figure 2C shows that cerebral 718 HYLIN ET AL. perfusion is signicantly decreased by 2min post-injury (F (3,21) =4.53, p =0.013), an effect that normalized by 24 h post-injury. Sham animals did not show any signicant change in cerebral perfusion over the monitoring period (data not shown). To determine if the resolution of cerebral perfusion is inuenced by repeated mCHI, mice (n = 12) were given four daily injuries, with cerebral perfu- sion monitored 24 h after each injury. This time point was chosen based upon the data (Fig. 2C), which indicated that cerebral per- fusion returned to baseline by 24 h after a single injury. Figure 2D shows that cerebral perfusion had normalized by 24 h after the rst, second, and third mCHI. However, after the fourth mCHI, cerebral perfusion remained signicantly decreased (F (4,44) = 3.91, p = 0.008). Respiratory rate was not signicantly different at any of the time points examined, nor was the cerebral perfusion signi- cantly altered in uninjured animals exposed to multiple anesthesia (data not shown). Repeated mCHI exacerbates vestibulomotor and motor decits Vestibulomotor response was measured daily using a beam balance task. Testing began the day after the last mCHI and con- tinued until the injured animals could perform the tasks similar to uninjured controls. Shamanimals (n =4) were able to remain on the balance beam for the entire 60 sec duration (Fig. 3A). In contrast, both single (n = 9) and repeated (n = 7) mCHI animals displayed vestibulomotor dysfunction and were incapable of balancing on the beam for the full 60 sec testing period (F (4,68) = 38.313, p < 0.001). Post-hoc analysis revealed that the repeated mCHI animals per- formed signicantly worse as compared to single mCHI animals (t (78) = 4.884, p < 0.001). The performance of both injury groups were similar those of shams by day 6. The results for the foot fault task are shown in Figure 3B (right forelimb) and Figure 3C (left forelimb). Testing was conducted on the days animals were tested in the beam balance. For 4 days fol- lowing injury, animals that received repeated mCHI performed poorly in this task, making signicantly more foot faults than sham animals (right forelimb: F (2,17) = 46.046, p <0.001 and left fore- limb: F (2,17) = 16.453, p <0.001). It is worth noting that the injury used in the current study did not result in any signicant motor asymmetry in the number of foot-faults made by the right or left forelimb (single mCHI, F (1,16) =1.208, p = 0.288; repeated mCHI, F (1,12) =3.55, p = 0.084). Further, by day 6, performance was sim- ilar among the three groups. Repeated mCHI impairs short-term visuospatial memory Because short-termmemory is often impaired in humans who have sustained a concussion, visuospatial learning and short-term memory were assessed using an abbreviated Morris water maze task on day 10, as described in the Methods section. Figure 4Aillustrates the latencies to locate the hidden platform for each group. The repeated mCHI group (n =7) performed poorly as compared to the single (n =9) or sham (n =8) groups (F (2,21) =8.770, p =0.002). When the perfor- mance of the single mCHI animals was directly compared to the sham group only (i.e., the repeated mCHI group was omitted), there was a statistically signicant interaction between the single mCHI and sham groups (F (4,60) =2.877, p =0.03). Post-hoc analysis revealed that these two groups differed on trials 36. Additional analysis of the training revealed that when compared to single mCHI and sham animals, the repeated injury group required more training before they began searching the center of the maze in order to locate the hidden platform (Figure 4B). As indicated in Figure 4C, when short-term memory was tested 30 minutes later by a probe trial, the repeated mCHI group demonstrated an impaired ability to nd the previous location of the platform. The swimming path to rst platform crossing was more circuitous in repeated mCHI animals as compared to the path taken by sham or single mCHI animals. During the probe, both the sham and single mCHI groups demonstrated a preference for the quadrant that previously contained the platform (sham: F (3,21) =11.307, p <0.001; single mCHI: F (3,24) =5.395, p =0.006), whereas the repeated mCHI group randomly searched in all four quadrants (F (3,18) =1.920, p =0.163) (Figure 4D). FIG. 1. Single and repeated mCHI worsen acute neurological functions. (A) Timeline for the experimental design. Separate groups of animals were used to assess conict learning and abbreviated Morris water maze. (B) A single mCHI (n = 10) increases the duration of apnea. For the repeated mCHI group (n =12), apnea was assessed after each injury and did not decrease signicantly across days. (C) Both the single and repeated mCHI groups demonstrated a suppression of duration of their response to a tail pinch relative to sham animals (n = 12), but did not differ from each other. (D) Both injury groups showed suppression in their righting response. REPEATED MILD CLOSED HEAD INJURY 719 FIG. 3. Single and repeated mCHI impairs vestibulomotor and motor functions. (A) Both single (n = 9) and repeated (n = 7) mCHI animals were impaired in their ability to balance on a narrow beam compared to sham animals (n = 4). Furthermore, repeated mCHI signicantly worsened performance compared to single mCHI animals. Repeated mCHI mice demonstrated signicantly more foot faults with both their (B) right forelimb and (C) left forelimb. Both motor and vestibulomotor function were normalized by Day 6 post- injury. Data is represented as mean S.E.M., *p < 0.05. FIG. 2. mCHI causes a transient decrease in cerebral perfusion that fails to recover after repeated mCHI. Cerebral perfusion was measured using a PIM3 scanning laser Doppler device. (A) Picture of a mouse head, as seen from the PIM3 imager, with the scan eld indicated by white outline. (B) Representative PIM3 scans at the indicated time points after a single mCHI (n = 8) (boxes: areas used for quantication). Quantication of cerebral perfusion (C) at acute time points following a single mCHI and (D) 24 h after each of the four repeated mCHIs (n = 12). Data is represented as mean S.E.M., *p < 0.05. Color image is available online at www.liebertpub.com/neu 720 Probe trials were further analyzed for measures of platform lo- calization. Concentric rings of increasing diameter (2X, 3X, and 4X platform diameters) centered on the platform were used to assess the latency and number of entries into the target areas. Repeated mCHI animals demonstrated higher latencies for all concentric circles (F (2,21) =4.616, p =0.022) (Fig. 5A) and made fewer circle crossings (F (2,21) = 15.852, p < 0.001) (Fig. 5B) when compared to single mCHI and shams. Analysis of the path length to the rst platform crossing revealed that repeated mCHI animals travel signicantly further prior to crossing (F (2,21) =8.312 , p =0.002) (Fig. 5C), taking a more circuitous path in nding the platform. Additionally, repeated mCHI mice spent more time away from the target as indicated by a higher mean proximity to platform (F (2,21) =9.569, p = 0.001) (Fig. 5D). However, no differences were seen in swim speed (F (2,21) = 1.554, p =0.235) or latency to nd a visible platform (F (2,21) =0.646, p =0.534). In order to examine if repeated exposure to anesthesia contributed to the observed cog- nitive decits, a group of animals were exposed daily to anesthesia for three days and received a mCHI on Day 4. A second group of animals received only a single mCHI on Day 4. No difference was observed between these two groups in either learning or short-term memory ( p > 0.05). Mild CHI impairs context discrimination The ability to discriminate between two similar contexts has been shown to require hippocampal function. As repeated mCHI mice were impaired in visuospatial learning and memory, we tested their ability to discriminate between two contexts that contained similar visual and auditory characteristics but differed in other sensory characteristics. Sham animals (n = 8) were able to distin- guish between two similar contexts after one day of training, as indicated by signicantly more freezing in the context in which the mild foot shock was delivered (F (2,14) =16.756, p = 0.001) (Fig. 6A). In contrast, both the single mCHI (n = 10) (Fig. 6B) and re- peated mCHI animals (n =7) (Fig. 6C) required an additional day of training to distinguish between the shock and safe contexts (single mCHI: F (2,18) =11.878, p < 0.001; repeated mCHI: F (2,12) = 6.343, p = 0.007). No difference was detected between performances of the single and repeated mCHI groups ( p > 0.05). However, by the FIG. 4. Repeated mCHI results in learning and short-term memory decits in a visuospatial task. Visuospatial learning and memory was assessed using an abbreviated Morris water maze task. (A) The repeated mCHI group (n =7) required signicantly longer time to nd the hidden platform across trials, indicating a learning impairment. (B) The poor acquisition of the repeated mCHI group is in part due to the animals spending more time in the periphery of the maze, dened as two body widths away from the wall (indicated by the gray annulus on inset). (C) Representative traces from an animal in each group during a probe trial given 30 min after the completion of training. (D) During the probe trial, sham (n =8) and single mCHI (n = 9) animals demonstrated a preference for the target quadrant that contained the platform (black circle in inset), whereas the repeated mCHI injury group did not have a preference for the target quadrant. Data is represented as mean S.E.M., *p <0.05. REPEATED MILD CLOSED HEAD INJURY 721 third day of training, all groups demonstrated the ability to dis- criminate between the two contexts. Repeated mCHI causes decits in active avoidance and conict learning We next examined the performance of a separate group of in- jured animals using a cognitively demanding task to assess com- plex learning. Active avoidance is a task that has been previously shown to rely upon the integrity of the hippocampus. 2830 Mice (n = 6/group) were initially tested on Day 10 post-injury in an ab- breviated three-part protocol consisting of a pretraining phase, an active avoidance phase, and a conict phase. Figure 7A shows representative traces of animal movements during the pre-training, active avoidance, and conict learning phases. During the pre- training phase, all groups explored the rotating platform and fre- quently entered the inactivated shock zone. This indicated that there was no inherent bias or tendency to avoid this part of the platform prior to turning on the shock. During active avoidance training, sham and single mCHI groups demonstrated a signi- cantly lower number of entries into the shock zone. In contrast, the repeated mCHI group entered the shock zone signicantly more times than either the sham or single mCHI groups (Fig. 7B) (F (2,15) = 39.873, p < 0.001). Two hours after the completion of active avoidance training, the animals were given a conict training with the shock zone rotated 180. Repeated mCHI animals were impaired in their ability to avoid the location of the newshock zone, making more entries into it than either sham or single mCHI FIG. 5. Repeated mCHI results in short-term memory decits in a visuospatial task. The repeated mCHI group (n = 7) demonstrated an inability to learn the general location of the platform as indicated by (A) an increase in the latency to nd the platform, (B) a decrease in the number of platform crossings during a probe trial, (C) longer swimming paths for the initial platform crossing, and (D) higher mean proximity to the platform than the sham (n = 8) and single mCHI (n =9) animals. Data is represented as mean S.E.M., *p < 0.05. FIG. 6. mCHI results in an attenuated ability to discriminate between two similar contexts. Percent time spent freezing in the shock cage or safe cage after 0, 1, and 2 days of training in a context discrimination protocol. (A) Sham mice (n =8) learn to discriminate between two similar contexts by test Day 1, as indicated by enhanced fear to the shock chamber relative to the safe chamber. Both (B) single mCHI (n =10) and (C) repeated mCHI (n =7) mice require an additional day of training in order to differentiate between the two contexts. Data is represented as mean S.E.M., *p < 0.05. 722 HYLIN ET AL. animals (Fig. 7A and 7B) (F (2,15) = 3.964, p = 0.041). To examine if these decits observed in the repeated mCHI group persisted, a second group of repeated mCHI animals was generated and tested at both 10 days and 3 months post-injury. This cohort exhibited similar pretraining exploration (40.4 1.3 versus 37.0 3.2 entries, p =0.314) as well as decits in active avoidance (31.7 6.3 versus 36.3 3.6 entries, p = 0.555) and conict learning (36.6 5.2 versus 39.7 6.1, p = 0.705) at 10 days, as the repeated mCHI cohort shown in Figures 7A and 7B. At 3 months post-injury (Fig. 7C), the repeated mCHI animals were still unable to learning the shock zone as evidenced by no signicant difference in the number of shock zone entries between the testing phases (F (2,12) =1.229, p =0.327) (Fig. 7D). Figure 7E shows representative heat maps for each of the tested groups obtained during the conict training phase. The amount of time an animal spent at any particular location in the arena is indicated by color coding, with black indicating no time and red indicating a large amount of time. While both sham and single mCHI groups spent the majority of their time avoiding the shock zone learned during the active avoidance phase (indicated by red wedge), the repeated mCHI groups tended to spend a relatively larger amount of time in that location. Repeated mCHI exacerbates neuroinammation and axonal damage Following the completion of cognitive testing (on Day 14 post- mCHI), animals were euthanized, and brains were extracted to examine cortical contusion, neuronal loss, dendritic damage, axonal injury, and inammation. No visible contusion was de- tected in either the single or repeated mCHI animals. Tissue sections were prepared from representative animals in each group for immunohistochemical evaluation. No overt neuronal loss was observed in the majority of animals examined (data not shown). However, in some repeated mCHI animals (*20%), cell loss in either the CA1 or CA3 subelds was detected. Silver staining showed argyrophilic staining in the corpus callosum of mice that sustained repeated mCHI, suggestive of axonal damage (Fig. 8A). Figure 8B shows that both single and repeated mCHI increased GFAP immunoreactivity in the corpus callosum. 17 A similar in- crease in Iba1 immunoreactivity (a marker for activated micro- glia) was also seen in the corpus callosum (data not shown). Both axonal damage and neuroinammation occurred in all the mCHI animals examined. Discussion In the present study, we have examined the neurological, neu- rocognitive, and histopathological changes resulting from single and repeated mild closed head injury. Our results revealed ve key ndings: repeated mild closed head injury 1) reduces cerebral perfusion, 2) results in a lingering neuroinammation, 3) causes axonal damage, 4) exacerbates vestibulomotor and motor dys- function, and 5) impairs short-term visuospatial memory and con- ict learning. FIG. 7. Repeated mCHI impairs cognitive exibility. Active avoidance and conict learning were tested using a place avoidance task. (A) Representative path tracings for an animal in each group (n =6/group) during the pre-training, active avoidance, and conict learning phases. The shock zone is dened by the red pie wedge. Small red circles represent where the animal was shocked (or would have been shocked in the pre-training phase). (B) Summary results indicating the number of entries into the shock zone during the pretraining (shock off), active avoidance, and conict learning phases. (C) Representative path tracings for repeated mCHI animals that were tested 3 months following injury. (D) Group summary data showing the number of shock zone entries repeated mCHI animals made following injury when tested at 3 months. The number of entries did not signicantly differ from the number of entries made by repeated mCHI animals at 10 days. (E) Representative heat maps for an animal in each group during the conict training phase. The color represents the amount of time an animal spent in a particular location (blue indicates the least amount of time, red indicates the most amount of time). The red pie wedge represents the location of the shock zone that was used during active avoidance training. The sham and single mCHI groups avoided this location during the conict training, whereas the repeated mCHI group did not, indicating impaired short-term memory. Data is represented as mean S.E.M., *p <0.05. Color image is available online at www.liebertpub.com/neu REPEATED MILD CLOSED HEAD INJURY 723 Both uid percussion injury (FPI) and CCI devices have been used to cause mTBI in which the force is delivered directly to the exposed brain. 14,31 While mild FPI and CCI have been shown not to cause overt brain damage, these injuries do result in impaired plasticity, axonal damage, neuroinammation, and spatial learning and memory decits. 3234 For example, Spain et al. have observed that mild FPI in mice causes evolving axonal damage in addition to a spatial learning impairment. 35 Recently, it has been shown that mild FPI in rats alters hippocampal place cell ring; this may be an underlying mechanism for the spatial learning impairment. 34 Mild CCI injury in rats also triggers an inammatory response and ax- onal damage. 32,36 Although these models have yielded interesting results on the pathophysiology of mild injury, they are not typically used to model repeated mTBI. In order to cause closed head injury in rats and mice, investi- gators have employed the weight-drop method, CCI device, or projectiles, and have examined the ensuing pathophysiology. 13,37,39 For example, Creed and associates employed the CCI device to cause closed head injury by delivering the impact directly to the skull. 38 This injury increased cerebral edema and neurodegenera- tion in the cortex and dentate gyrus, and caused axonal damage in the corpus callosum. Daily testing of animals in the Morris water maze showed impaired learning on Days 13 post-injury but not when tested on Days 46 post-injury. Chen and colleagues em- ployed a projectile-based approach in which a sealed microfuge tube containing dry ice was used to propel the cap into a rats head. 39 The use of a helmet and adjustment of the distance between the sealed tube and rat head resulted in a mild TBI with no gross brain pathology or hemorrhage. In these animals, sensory motor abnormalities were detected 14 hours post-injury using a Catwalk test, but no cognitive impairments were observed using the standard Morris water maze (MWM). We employed an abbreviated MWM task to examine short-term visuospatial learning and memory. Using this task, we found that a single mCHI impairs spatial learning as compared to sham animals when tested on Day 10 post- injury. Furthermore, when short-term memory for platform local- ization was assessed using concentric circle analysis (Fig. 5B), single mCHI animals showed poor recall. These results suggest that the abbreviated Morris water maze task may be a more sensitive measure for learning and short-term memory than the standard spaced version of this task. It remains unresolved if there is an upper limit (or threshold) for the number of concussions a person can sustain and still make a full recovery. A limited number of clinical and experimental studies have attempted to examine this issue. 15,20,40 It has been reported that neurocognitive and neurobehavioral decits do not differ among athletes who sustained 1 or 2 concussions. 41 However, it appears that more than three mild head injuries may cause irre- versible decits. 42 Based on these and other ndings, the American Academy of Neurology recommends termination of the season for an athlete after the third concussion within that same season. Using the weight drop model, DeFord and associates examined cognitive impairments resulting from four concussive injuries to mice de- livered 24 hours apart. 13 Four repeated weight drops of 100 g or 150 g masses from a height of 40 cm did not cause overt brain damage, but mice had impaired acquisition in the MWM task. Whalen and colleagues, using a weight drop model, reported that ve closed head injuries, each separated by a day, impaired learning and memory. However, a single closed head injury did not cause any signicant cognitive decit. 20 Recently, Crawford and col- leagues used the CCI device to deliver ve closed head injuries to mice and reported motor decits in the rotarod task and cognitive decits in the Barnes maze that occurred in the absence of overt hippocampal cell loss. 17 Consistent with these studies, we observed that four repeated mCHIs exacerbated performance in the beam balance and foot-fault tasks. Furthermore, short-term learning and memory were also impaired in mice sustaining four repeated mCHIs. At the beginning of MWM training, all animals tend to swim adjacent to the perimeter of the tank looking for an escape route. Sham and single mCHI animals quickly learned to abandon this strategy in favor of an open water search strategy in order to nd the hidden platform (Fig. 4B). In contrast, the repeated mCHI mice required more training trials before they switched to this open water search strategy, though once they had switched to this strategy they used it for the remainder of the training trials. While we cannot rule out that enhanced thigmotaxic-like behaviors FIG. 8. mCHI results in axonal disruption and inammation. Representative photomicrographs of (A) silver staining revealed an increase in silver impregnated bers in the corpus callosum of the single mCHI group and to a greater degree in the repeated mCHI group (n = 6/group). (B) Images of the corpus callosum demonstrating that repeated mCHI increases inammation as indicated by an increase in GFAP (astrocytes) immunoreactivity. Color image is available online at www.liebertpub.com/neu 724 HYLIN ET AL. contribute to the decits we observed, it appears that the decit is more related to both an inability to switch the search strategy and to localize the position of the hidden platform. These decits were detected in the absence of demonstrable hippocampal pathologies in the majority of animals, with only 20% of repeated injury mice having overt hippocampal neuron loss at the magnitude of injury employed in this study. Although we cannot exclude the possibility that some animals had cell loss that was more distributed in nature and not visibly obvious, it appears that the percentage of animals with cell loss is likely to depend on the magnitude of the impact force delivered to the skull. Consistent with this, we have observed that impact forces that cause skull fractures reproducibly cause visible cortical and hippocampal cell loss (data not shown). Pattern separation is a neural process by which similar but not identical experiences, events, or spatial maps are transformed into discrete non-overlapping representations, allowing for cognitive exibility. 28,43,44 A conict-based place avoidance task can assess cognitive exibility. Bergold and colleagues were the rst inves- tigators to use the active place avoidance task to showthat mild CCI injury in rats caused decits in the ability to perform this task. 26 Similar to that seen previously, we observed that repeated mCHI mice performed poorly in the active avoidance version of the task. Likewise, when the location of the shock zone was moved in order to increase the need for cognitive exibility, the performance of repeated mCHI animals was impaired as compared to sham ani- mals. These decits appear to persist as we have seen decits at least 3 months after repeated mCHI. Histopathological assessment showed no visible brain contu- sion, consistent with a classication of mild injury. However, recent imaging studies have begun to indicate that even mild brain injuries can give rise to axonal damage that can be detected using diffusion tensor imaging (DTI). For example, decreased frac- tional anisotropy (FA) is often observed in the genu of corpus callosum and in the internal and external capsules. 45 While de- creased FA is commonly seen in the chronic stages of mTBI, increased FA (possibly resulting from axonal edema) has been observed in the fornix 16 days after injury that correlated with decreased ANAM score. 46 To assess axonal injury in single and repeated mCHI animals, we performed silver staining, which revealed degenerating processes in the corpus callosum that ap- peared to worsen in animals subjected to repeated mCHI. Inter- estingly, the immunoreactivities of GFAP (and Iba-1, not shown) were also found to be increased in the corpus callosum, suggesting that ongoing neuroinammation may be associated with the ob- served axonal damage. These changes are similar to those re- ported by Mouzon et al., who examined the histopathological changes resulting from ve repeated mCHI, each separated by 2 days. 17 While these studies highlight some of the pathological changes associated with repeated mCHI, a spectrum of patho- logical changes ranging from tissue loss to subtle neuronal dys- function are likely to be observed, depending on injury severity, injury location, age and gender. Future studies will be required to examine the inuence of these variables. In addition to the abovementioned variables, injury interval is thought to be a dening factor in the degree and types of decits observed following repeated mild injuries. In a recent study, Meehan et al. used the weight drop model to examine the effect of time interval between repeated mCHI in mice on neurocognitive function. 20 Mice that received ve mCHI injures delivered at one day or one week intervals had impaired spatial memory. In contrast, when the interval was increased to one month, no difference was detected between injured and sham animals, suggesting a window of vulnerability of at least one week. Using a similar weight drop model, Vagnozzi et al. examined the window of metabolic vul- nerability to a second mTBI in rats. 40 This study showed that maximum reduction of brain levels of ATP, NAA, NAAG, NAD, and acetyl CoAresulted when the interval for the second impact was separated by 3 days as compared to 1, 2, 4, or 5 days. 40,47 In addition, a secondary injury occurring shortly after the rst mild injury has been shown to prolong this period of metabolic suppression, and correlates with a signicant increase in hippocampal cell death. 48 A recent study by Prins et al. shows that when a second mCHI is delivered 24 hours, but not 3 days, after the rst injury, a larger and more persistent suppression of cerebral glucose metabolism is ob- served. 14 Clinical studies have employed neuropsychological test- ing to assess the duration of vulnerability. A few clinical studies have reported that, if a concussion is sustained prior to resolution of the symptoms of the rst concussion have resolved, this can prolong brain metabolic changes and cause fatality. 49 This condition is re- ferred to as second impact syndrome (SIS) and has only been re- ported in young adults. 50 While the occurrence and the mechanism underlying SIS are controversial, it has been hypothesized that impaired cerebral autoregulation and brain edema may be under- lying mechanisms. Our cerebral perfusion results showed that a single mCHI transiently decreased cerebral blood ow which re- covered by 24 h post-injury. However, following the fourth closed head injury, cerebral perfusion remained signicantly suppressed 24 h later. At this time, we are uncertain if this reduction of cerebral perfusion may be due to impaired autoregulation. Future studies will address the impact of repeated mCHI on vascular autoregulation, and on metabolic disturbances in order to help dene the period of vulnerability. Acknowledgment This work was supported by The Institute for Rehabilitation and Research/Mission Connect; The Vivian L. Smith Foundation and the Gilson-Longenbaugh Foundation. Author Disclosure Statement No competing nancial interests exist for any author. References 1. Faul M, Xu L, and Wald MM. (2010). Traumatic brain injury in the United States: Emergency department visits, hospitalizations, and deaths 2002-2006. Atlanta (GA): Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. 2. Sosin DM, Sniezek JE, and Thurman DJ. (1996). Incidence of mild and moderate brain injury in the United States, 1991. Brain Inj 10: 4754. 3. Warden D. (2006). Military TBI during the Iraq and Afghanistan wars. J Head Trauma Rehabil 21:398402. 4. dHemecourt P. (2011). Subacute symptoms of sports-related con- cussion: Outpatient management and return to play. Clin Sports Med 30:6372, viii. 5. Bigler ED. (2008). Neuropsychology and clinical neuroscience of persistent post-concussive syndrome. J Int Neuropsychol Soc 14:122. 6. Dikmen S, Machamer J, Fann JR, and Temkin NR. (2010). Rates of symptom reporting following traumatic brain injury. J Int Neu- ropsychol Soc 16:401411. 7. Middleboe T, Andersen HS, Birket-Smith M, and Friis ML. (1992). Minor head injury: Impact on general health after 1 year. A pro- spective follow-up study. Acta Neurol Scand 85:59. 8. Guskiewicz KM, Weaver NL, Padua DA, and Garrett WE, Jr. (2000). Epidemiology of concussion in collegiate and high school football players. Am J Sports Med 28:643650. 9. Guskiewicz KM. (2003). Assessment of postural stability following sport-related concussion. Curr Sports Med Rep 2:2430. REPEATED MILD CLOSED HEAD INJURY 725 10. Matser JT, Kessels AG, Jordan BD, Lezak MD, and Troost J. (1998). Chronic traumatic brain injury in professional soccer players. Neu- rology 51:791796. 11. McKee AC, Cantu RC, Nowinski CJ, et al. (2009). Chronic traumatic encephalopathy in athletes: Progressive tauopathy after repetitive head injury. J Neuropathol Exp Neurol 68:709735. 12. Stern RA, Riley DO, Daneshvar DH, Nowinski CJ, Cantu RC, and McKee AC. (2011). Long-term consequences of repetitive brain trauma: Chronic traumatic encephalopathy. PMR 3:S460S467. 13. DeFord SM, Wilson MS, Rice AC, et al. (2002). Repeated mild brain injuries result in cognitive impairment in B6C3F1 mice. J Neuro- trauma 19:427438. 14. Prins ML, Alexander D, Giza CC, and Hovda DA. (2013). Repeated mild traumatic brain injury: Mechanisms of cerebral vulnerability. J Neurotrauma 30:3038. 15. Prins ML, Hales A, Reger M, Giza CC, and Hovda DA. (2010). Re- peat traumatic brain injury in the juvenile rat is associated with in- creased axonal injury and cognitive impairments. Dev Neurosci 32:510518. 16. Uryu K, Laurer H, McIntosh T, et al. (2002). Repetitive mild brain trauma accelerates Abeta deposition, lipid peroxidation, and cognitive impairment in a transgenic mouse model of Alzheimer amyloidosis. J Neurosci 22:446454. 17. Mouzon BC, Chaytow H, Crynen G, et al. (2012). Repetitive mild traumatic brain injury in a mouse model produces learning and memory decits accompanied by histological changes. J Neurotrauma 29:27612773. 18. Dixon CE, Clifton GL, Lighthall JW, Yaghmai AA, and Hayes RL. (1991). A controlled cortical impact model of traumatic brain injury in the rat. J Neurosci Methods 39:253262. 19. Hovda DA, Yoshino A, Kawamata T, Katayama Y, and Becker DP. (1991). Diffuse prolonged depression of cerebral oxidative metabo- lism following concussive brain injury in the rat: A cytochrome oxi- dase histochemistry study. Brain Res 567:110. 20. Meehan WP, III, Zhang J, Mannix R, and Whalen MJ. (2012). In- creasing recovery time between injuries improves cognitive outcome after repetitive mild concussive brain injuries in mice. Neurosurgery 71: 885892. 21. Zhao J, Pati S, Redell JB, Zhang M, Moore AN, and Dash PK. (2012). Caffeic acid phenethyl ester protects blood-brain barrier integrity and reduces contusion volume in rodent models of traumatic brain injury. J Neurotrauma 29:12091218. 22. Dash PK, Orsi SA, Zhang M, et al. (2010). Valproate administered after traumatic brain injury provides neuroprotection and improves cognitive function in rats. PLoS One 5:e11383. 23. Blum S, Moore AN, Adams F, and Dash PK. (1999). A mitogen- activated protein kinase cascade in the CA1/CA2 subeld of the dorsal hippocampus is essential for long-term spatial memory. J Neurosci 19:35353544. 24. Guzowski JF, and McGaugh JL. (1997). Antisense oligodeox- ynucleotide-mediated disruption of hippocampal cAMP response el- ement binding protein levels impairs consolidation of memory for water maze training. Proc Natl Acad Sci USA 94:26932698. 25. Frankland PW, Cestari V, Filipkowski RK, McDonald RJ, and Silva AJ. (1998). The dorsal hippocampus is essential for context discrim- ination but not for contextual conditioning. Behav Neurosci 112: 863874. 26. Abdel Baki SG, Kao HY, Kelemen E, Fenton AA, and Bergold PJ. (2009). A hierarchy of neurobehavioral tasks discriminates between mild and moderate brain injury in rats. Brain Res 1280:98106. 27. Cimadevilla JM, Fenton AA, and Bures J. (2001). New spatial cog- nition tests for mice: passive place avoidance on stable and active place avoidance on rotating arenas. Brain Res Bull 54:559563. 28. Burghardt NS, Park EH, Hen R, and Fenton AA. (2012). Adult-born hippocampal neurons promote cognitive exibility in mice. Hippo- campus 22:17951808. 29. Cimadevilla JM, Fenton AA, and Bures J. (2000). Functional inacti- vation of dorsal hippocampus impairs active place avoidance in rats. Neurosci Lett 285:5356. 30. Cimadevilla JM, Wesierska M, Fenton AA, and Bures J. (2001). In- activating one hippocampus impairs avoidance of a stable room- dened place during dissociation of arena cues from room cues by rotation of the arena. Proc Natl Acad Sci USA 98:35313536. 31. Aiguo W, Zhe Y, and Gomez-Pinilla F. (2010). Vitamin E protects against oxidative damage and learning disability after mild traumatic brain injury in rats. Neurorehabil Neural Repair 24:290298. 32. Redell JB, Moore AN, Grill RJ, Jr, et al. (2012). Analysis of functional pathways altered following mild traumatic brain injury. J. Neuro- trauma 30:752764. 33. Zhao Z, Loane DJ, Murray MG, Stoica BA, and Faden AI. (2012). Comparing the predictive value of multiple cognitive, affective, and motor tasks after rodent traumatic brain injury. J Neurotrauma 29: 24752489. 34. Eakin K, and Miller JP. (2012). Mild traumatic brain injury is asso- ciated with impaired hippocampal spatiotemporal representation in the absence of histological changes. J Neurotrauma 29:11801187. 35. Spain A, Daumas S, Lifshitz J, et al. (2010). Mild uid percussion injury in mice produces evolving selective axonal pathology and cognitive decits relevant to human brain injury. J Neurotrauma 27:14291438. 36. Levin H, and Robertson CS. (2013). Mild TBI in translation. J Neu- rotrauma 30:610617. 37. Yuen TJ, Browne KD, Iwata A, and Smith DH. (2009). Sodium channelopathy induced by mild axonal trauma worsens outcome after a repeat injury. J Neurosci Res 87:36203625. 38. Creed JA, DiLeonardi AM, Fox DP, Tessler AR, and Raghupathi R. (2011). Concussive brain trauma in the mouse results in acute cog- nitive decits and sustained impairment of axonal function. J Neuro- trauma 28:547563. 39. Chen Z, Leung LY, Mountney A, et al. (2012). A novel animal model of closed-head concussive-induced mild traumatic brain injury: De- velopment, implementation, and characterization. J Neurotrauma 29:268280. 40. Vagnozzi R, Tavazzi B, Signoretti S, et al. (2007). Temporal window of metabolic brain vulnerability to concussions: Mitochondrial-related impairmentPart I. Neurosurgery 61:379388. 41. Macciocchi SN, Barth JT, Littleeld L, and Cantu RC. (2001). Mul- tiple concussions and neuropsychological functioning in collegiate football players. J Athl Train 36:303306. 42. Macciocchi SN, Barth JT, and Littleeld LM. (1998). Outcome after mild head injury. Clin Sports Med 17:2736. 43. Sahay A, Scobie KN, Hill AS, et al. (2011). Increasing adult hippo- campal neurogenesis is sufcient to improve pattern separation. Nat- ure 472:466470. 44. Leutgeb JK, Leutgeb S, Moser MB, and Moser EI. (2007). Pattern separation in the dentate gyrus and CA3 of the hippocampus. Science 315:961966. 45. Singh M, Jeong J, Hwang D, Sungkarat W, and Gruen P. (2010). Novel diffusion tensor imaging methodology to detect and quantify injured regions and affected brain pathways in traumatic brain injury. Magn Reson Imag 28:2240. 46. Yallampalli R, Wilde EA, Bigler ED., et al. (2013). Acute white matter differences in the fornix following mild traumatic brain injury using diffusion tensor imaging. J Neuroimag 23:224227. 47. Giza CC, and Hovda DA. (2001). The neurometabolic cascade of concussion. J Athl Train 36:228235. 48. Aoyama N, Lee SM, Moro N, Hovda DA, and Sutton RL. (2008). Duration of ATP reduction affects extent of CA1 cell death in rat models of uid percussion injury combined with secondary ischemia. Brain Res 1230:310319. 49. Barkhoudarian G, Hovda DA, and Giza CC. (2011). The molecular pathophysiology of concussive brain injury. Clin Sports Med 30:33. 50. Doolan AW, Day DD, Maerlender AC, Goforth M, and Gunnar BP. (2012). A review of return to play issues and sports-related concus- sion. Ann Biomed Eng 40:106113. Address correspondence to: Pramod K. Dash, Ph.D. Department of Neurobiology and Anatomy The University of Texas Medical School at Houston PO Box 20708 Houston, TX 77225 E-mail: p.dash@uth.tmc.edu 726 HYLIN ET AL.