• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Multiple Episodes of Mild Traumatic Brain Injury Result in Impaired Cognitive Performance in Mice

Posted on: Sunday, 15 August 2004, 06:00 CDT

Abstract

Objectives: Results from recent studies on animal models of concussion suggest that multiple, rather than single, episodes of mild traumatic brain injury result in impaired cognitive performance in mice. The objective of the present study was to administer multiple impacts to the heads of mice while directly measuring the force of the impacts to determine how these parameters are related to transient loss of consciousness, cognitive deficits, and potential neuropathologic effects. Methods: Seven-week-old male C57BL/6 mice were randomly assigned to experimental conditions involving three impacts (weight-drop method) to the head to induce mild traumatic brain injury or to sham control procedures. Some impacted (n = 10) and sham control (n = 10) mice were evaluated behaviorally and tested for spatial learning using the Morris water maze (MWM), whereas other impacted (n = 10) and sham control (n = 5) mice were used for histopathologic analysis. Results: The mean (SD) force of impact was 19 (3.5) N. Impacted mice took longer to regain consciousness compared with sham control mice (p < 0.0005). Behavioral test results showed that the groups did not differ on activity or sensorimotor tests or during cued trials in the MWM. Impacted mice exhibited impaired spatial learning performance during place trials in the MWM (p < 0.05). Silver staining revealed a contra-coup type of injury involving ventral brain structures in contact with or in close proximity to the skull. Conclusions: This multiple-impact model, delivered within a specifiable force range, results in transient, reversible loss of consciousness, a contra- coup brain injury, and cognitive impairment. Key words: traumatic brain injury; concussion; memory; cognition; mice. ACADEMIC EMERGENCY MEDICINE 2004; 11:809-819.

It is not clear whether mild traumatic brain injury (TBI) causes demonstrable long-term cognitive impairment in humans. A number of earlier epidemiologic studies suggest that this may be the case, but the use of retrospective designs, lack of appropriate control groups, and an inability to correct for several confounding factors did not allow for a straightforward interpretation of the results.1- 4 More recently, a number of studies of athletes who have experienced a sports-related concussion suggest that symptomatology and cognitive impairment can last for one week5 and, in some cases, up to three months6.

Animal models have been used to study the behavioral and neuropathologic sequelae associated with TBI.7-10 In three of these studies, in which slightly different murine models of mild TBI were used, the results suggested that mild single impacts did not produce discernible learning/memory deficits when behavioral testing was conducted more than three days after impact.8-10 In contrast, multiple mild impacts have usually been found to produce significant effects on behavior, but the nature of these changes has been inconsistent across studies. For example, DeFord et al.10 reported that multiple mild impacts produced learning/memory impairments in wild-type mice, whereas Uryu et al.9 found that wildtype mice were resistant to learning/memory deficits induced by multiple impacts, although amyloid precursor protein transgenic mice showed significant cognitive impairment during the postimpact period. Interpreting the effects of multiple impacts is further complicated by the results of a third study by Laurer et al.11 suggesting that multiple impacts induced more robust effects on motor than on cognitive functions, thus possibly compromising the measurement and interpretation of learning/memory deficits.

The widely divergent results associated with multiple mild- impact studies are in many cases most likely due to the lack of standardization of the methodology used for applying the impacts and the techniques to evaluate the behavioral and histopathologic effects. In an effort to promote the reproducibility of results from multiple mild-impact studies, we developed a murine TBI model with the following features: 1) the ability to directly measure the force of impact to the head; 2) the use of a common strain of wild-type mouse, which is the most often used background strain for transgenic/ knockout mice; 3) the ability to measure the duration of loss of consciousness immediately following the impacts; 4) the utilization of in-depth behavioral analyses to evaluate potential impairments in spatial learning, memory, and sensorimotor capabilities; and 5) an assessment of acute neuropathologic effects following the multiple impacts using a histologic technique (de Olmos cupric silver stain) that has a high signal-to-noise ratio for showing degenerating neuronal cell bodies and fibers, including terminal field degeneration.

The goal of this study was to develop a reliable and valid mouse model of mild TBI that provides information regarding the functional deficits associated with multiple episodes of mild TBI of sufficient force to induce reversible loss of consciousness and the underlying neurobiological mechanisms involved.

METHODS

Study Design. This was a laboratory study using a model of mild TBI in mice. The Institutional Animal Care and Use Committee and Animal Studies Committee at the Washington University School of Medicine approved all experimental protocols.

Animal Subjects and Preparation. Male C57BL/6 mice, aged 7-8 weeks, were obtained from Taconic Farms (Germantown, NY). The mice were housed under standard laboratory conditions (23C 1C, 50% 5% humidity, 12-hour light/dark cycle) in groups of five with free access to food and water throughout the experiment. For the behavioral study, mice were divided into two experimental groups: impacted mice (n = 10) and sham controls (n = 10). One animal in the impacted group was killed after the third impact due to skull fracture, leaving a sample size of nine animals for analysis in the impact group. A separate group of animals was used for histopathologic analysis and randomly assigned to two groups: impacted mice (n = 10) and sham controls (n = 5).

Study Protocol.

Weight-drop (Impact) Method. The construction of the weight-drop apparatus as well as the drop masses and heights were based on previous work by Tang et al.7 and pilot studies conducted in our own laboratory. The apparatus (Figure 1) consisted of a 21-g cylindrical acrylic weight inside a plexiglass tube designed to guide the weight onto a cylindrical metallic force transducer (PCB model 209C01; PCB Piezotronics Inc., Depew, NY) that is placed centrally on the head of the mouse between and slightly caudal to the eyes. The weight was allowed to fall freely from a height of 35 cm. A counterweight was attached to the top of the acrylic cylinder to avoid rebound impacts onto the force transducer. Data from the force transducer were digitized using SigLab 20-22 (Spectral Dynamics, San Jose, CA), a dynamic signal-analyzer data acquisition hardware and software. Dental molding material was used to make a mold of the mouse body and provided firm (24.6 N/mm) and consistent support for the head and body.

Figure 1. The weight-drop apparatus used to induce mild traumatic brain injury and the measurement of force at impact. (A) A schematic diagram of the weight-drop apparatus and force measurement equipment is depicted. Force data acquisition was triggered on impact, amplified, and analyzed using SigLab 20-22, a dynamic signal analyzer running on a laptop computer. (B) A representative force pulse measured during impact of the weight drop on the mouse head.

Our protocol consisted of anesthetizing the mice by using a vaporizer that exposed the animals to 5% isoflurane (in air at 2.5 L/ min) for 60 seconds. At this pointy the mice were immediately placed in the holder and the force transducer was carefully positioned on the head. Sham controls were placed in the holder with the force transducer in position but were not subjected to any impacts. Animals in the impacted group were placed in the apparatus and subjected to a single weight drop. They were then immediately removed from the holder, placed on their backs on the laboratory bench, and timed to see how long it took to regain consciousness (see righting reflex measure as described later). The mice were exposed to three weight-drop/sham-impact sessions using a 24-hour intersession interval similar to that used in previous studies.8,9,11 Following each impact, the mice were returned to their home cages and their recovery was monitored for signs of skull fracture or severe head trauma, including seizures, paralysis, and/ or ataxia. Any mouse found to exhibit any of these symptoms was killed immediately.

Neurobehavioral Testing.

Righting Reflex Response. The latency period for each mouse to right itself from the supine position was measured from the termination of anesthesia for both sham controls and impacted mice as a means to quantify the period of unconsciousness. In a previous pilot study, some of the animals had a long lag time between arousal and attempts to right their posture. In the current study, to shorten this lag time, we provided continuous stimulation to the abdomen while the animals were in the supine position. The mice were considered conscious when they were able to orient themselves into and maintain a righted posture.

One-hour L\ocomotor Activity Testing. General activity levels were measured on day 4 (24 hours after the last weight-drop/sham- impact session) according to previously published methods.12,13 Each mouse was placed in a polystyrene cage (47.6 25.4 20.6 cm high) for one hour to assess its general activity using a computerized, high-resolution photobeam system (Hamilton Kinder LLC, Poway, CA). Movement-related variables that were used to quantify motivation and general locomotor activity included total number of ambulations, rearing, and distance traveled in the periphery of the "field" as well as entries made into, time spent in, and distance traveled in the center of the "field."

Sensorimotor Battery. The sensorimotor capabilities of the animals were assessed on day 8 to evaluate the direct effects of the impacts on neurologic function and to assess whether any impact- induced impairments in motivation, balance, coordination, or movement were likely to affect the spatial learning and memory performance of the mice during the Morris water maze (MWM). The battery included the following tests using previously published protocols.13,14

1. Walking initiation: each mouse was placed in the center of a square (21 21 cm) outlined on a smooth black tabletop, and the latency for initiating movement out of the square (placing all four limbs outside the square) was timed in seconds.

2. Ledge test: a mouse was timed to see how long it could maintain its balance on the edge (0.75 cm thick) of a piece of plexiglass that was oriented 90 to the floor and was 33 cm high. A score of 60 seconds was assigned if the mouse was able to traverse the entire length (51 cm) of the ledge and return to the starting place in <60 seconds.

3. Platform test: each mouse was timed (in seconds) for now long it remained on an elevated (47 cm above the floor) circular platform (1.0 cm thick; 3 cm in diameter) that had rounded edges. A maximum score of 60 seconds was assigned if the mouse remained on the platform for that time or if it climbed down the thin pole that supported the platform without falling.

4. Inclined/inverted screens: each mouse was placed on an elevated (47 cm above the floor) wire mesh grid (15 cm in width 52 cm in length; 16 squares per 10 cm) that was inclined to 60, inclined to 90, or completely inverted. For the two inclined screen tests, the mouse was placed in the middle of the screen with its head oriented downward to the floor and was timed (in seconds) for how long it took it to turn 180 and climb to the top of the apparatus. For the inverted screen test, a mouse was placed on the screen in the 60 position and then the screen was inverted and the mouse timed (in seconds) to see how long it remained on the screen.

Morris Water Maze. On day 8 of the experiment (five days after the last weight-drop/sham-impact session), the mice were evaluated on the MWM using a procedure similar to previously published methods.14 Briefly, this task involved training mice to navigate to a platform in a pool (100 cm wide 30 cm deep) of opaque water to escape out of the water. All trials were monitored by a video camera mounted above the pool, and data were recorded using a computerized tracking system (Polytrak; San Diego Instruments, San Diego, CA) that measured the escape path length (distance traveled to find the platform) and escape latency (time to find the platform). Swimming speeds were also calculated. The mice were evaluated using three different types of 60-second trials: cued (visible platform, varied location), place (submerged platform, fixed location), and probe (platform removed from fixed location). All mice were first tested (day 8) in the cued condition to evaluate for nonassociative factors (i.e., sensorimotor or visual disturbances, differences in motivation) that might affect acquisition performance during subsequent place training. During the cued trials, mice were trained to navigate to a submerged platform, the location of which was "cued" by a red tennis ball atop a rod that protruded out of the water. The mice participated in two blocks of four consecutive cued trials, with the blocks separated by two hours, during one day of testing. The location of the platform was moved to a different quadrant for each trial (i.e., each quadrant was used once for each block of trials). The number and salience of distal cues were minimized; along with the movement of the platform and the brevity of training, this reduced the possibility that associations would be made between pool locations and distal cues.

On the following day (day 9), the mice were trained in the "place" condition to evaluate spatial reference (trial-dependent) memory capabilities. This involved training the mice to learn the location of a platform submerged 1.0 cm beneath the surface of the water and located 15 cm from the wall of the pool. The location of the platform, which was no longer "cued," remained the same for all place trials. Distinct distal spatial cues were placed around the room for the place trials to facilitate association of the cues with platform locations. Acquisition training during place trials involved giving the mice two consecutive trials (60 seconds maximum) per day using pseudorandomly determined release points across the four pool quadrants. Pseudorandomization refers to the stipulation that the same release point not be used on two consecutive trials. The mice were exposed to nine days of training, yielding 18 trials. To evaluate retention capabilities, the mice were tested on a probe trial approximately one hour after completing the last place trial on the ninth training day. During the 60-second probe trial, the escape platform was removed from the pool and the mouse was released into the maze at a point diagonally opposite from the previous location of the platform. The time spent searching the target quadrant where the platform had been located and the number of crossings over the former platform location (platform crossings) were recorded.

Data Analysis. The data were analyzed by analysis of variance (ANOVA) models, which often contained one between-subjects variable group (impacted vs. sham) and one within-subject variable such as blocks of trials or impact/sham sessions. The Huynh-Feldt adjustment of [alpha] levels was used for all within-subject effects containing more than two levels to protect against violations of the sphericity/ compound symmetry assumptions underlying repeated-measures ANOVA models. Contrasts involving pairwise comparisons were conducted following a significant main effect of group or a significant interaction involving the group variable. Bonferroni correction was used to maintain prescribed [alpha] levels at 0.05 as a function of the number of comparisons performed, and p-values for a given comparison had to be less than these corrected levels to be considered significant. An ANOVA containing one between-subjects variable (group) was used to analyze the data from each of the tests within the sensorimotor battery as well as the locomotor activity measures.

Histology. A separate set of animals (N = 15) that was divided into impacted (n = 10) and sham (n = 5) groups was subjected to a weight-drop procedure identical to that used with the mice that served as subjects in the behavioral experiments. Twenty-four hours after the third impact was delivered, the mice were subjected to histopathologic procedures, which are part of the de Olmos cupric silver-staining technique, to evaluate for any brain damage that may have resulted from the impacts using our previously published methods in mice.13 We have shown in drug neurotoxicity studies in mice that the de Olmos technique reliably stains populations of degenerating neurons that are identical to those stained by hematoxylin and eosin (H&E) but provides a higher signal-to-noise ratio than H&E, which makes detection of degenerating neurons easier.14 The de Olmos technique also stains fibers as well as terminal fields, which represents a decided advantage over H&E, and has been used successfully to characterize damage accruing from head trauma in infant rats.15 The procedure involves deeply anesthetizing the mice with chloral hydrate and killing them by transcardial perfusion using a 1% solution of paraformaldehyde in tris- (hydroxymethyl)-aminomethane (TRIS) buffer for 8 minutes. Brains were removed and postfixed in the perfusate for 24 hours and then sectioned (50 m) on a microslicer throughout the anterior-posterior extent of the brain. Sections were processed for cupric silver staining (de Olmos technique) according to previously published methods in mice.14 Eight coronal sections were selected from each mouse at the following levels: 2.0, 1.2, 0.8, -0.8, -2.0, -2.8, - 3.3, and -4.2 mm relative to bregma.

RESULTS

Force Data. The results of a one-way repeated-measures ANOVA showed that, in general, the mean forces of impact differed significantly across the three impacts: F (2,18) = 5.597, p = 0.013 (see Table 1). Subsequent pairwise comparisons showed that, although the measured force of the first impact was substantially greater than that measured during the second impact, the resulting p-value (F [1,9] = 6.27, p = 0.034) was not less than the Bonferroni- corrected level (p = 0.025) and thus did not quite reach statistical significance using this conservative approach. The force of the third impact was not significantly different from that of the second impact.

Righting Reflex Recovery. Results of the ANOVA conducted on righting times (Figure 2) showed that there was a significant main effect of group (F [1,18] = 82.08, p < 0.0005), indicating that, in general, the impacted mice took significantly longer to right themselves compared with the sham controls. Subsequent pairwise comparisons showed that the impacted mice had significantly longer righting times than the sham controls during each of the three weight-drop/sham\-impact sessions (p < 0.0005; Bonferroni corrected a = 0.017). The results also showed a significant within-subject effect of repeated (weight-drop/sham-impact) sessions (F [2,36] = 7.22, p = 0.002) and a significant treatment by sessions interaction (F [2,36] = 4.671, p = 0.016). Additional comparisons showed that righting reflex latencies were significantly increased in both groups from the first to the second session (p < 0.015; Bonferroni- corrected [alpha] = 0.025) but not from the second to the third session, and the significant interaction indicates that these changes in latencies were different in the two groups across sessions.

TABLE 1. Mean Force Measurements for Impacted Mice

Figure 2. Impacted mice took significantly longer to regain consciousness following the weight drop compared with sham controls. A significant main effect of group (F (1,18) = 82.08, *p < 0.0005) and subsequent significant pairwise comparisons showed that the righting reflex recovery times were significantly greater for the impacted mice compared with the. sham controls for each weight-drop/ sham-impact session (*p < 0.0005) and that the righting reflex recovery times increased from the first session to the second (^sup [dagger]^p < 0.015) Put not thereafter. All p-values for the comparisons between groups as well as those from the within-group comparisons for the impacted mice across sessions exceeded Bonferroni-corrected levels (i.e., p < 0.017 and p < 0.025, respectively).

One-hour Locomotor Activity Testing and Sensorimotor Battery. Although there was a tendency for the impacted mice to be less active on a number of the activity variables (Table 2), the results of ANOVAs conducted on the one-hour activity data showed that there were no significant differences between groups on any of the activity measures. Similarly, group performances were not different on the walking initiation, ledge, platform, or inclined/inverted screen tests within the sensorimotor battery (Table 3). In summary, these results suggest that multiple impacts did not produce significant effects on movement-related variables or sensorimotor capabilities.

TABLE 2. Activity and Locomotor Results According to Group

Morris Water Maze. The path length data from the cued trials (Figure 3A) showed that the impacted and sham control groups performed similarly, with the possible exception of the third block of trials, and that both groups showed substantial improvement over time.

The cued escape latency data (data not shown) were virtually identical to the path length results. The nonsignificant effects of group and the group by blocks of trials interactions formally document the lack of differences between groups in terms of both path length and latency. An ANOVA of the swimming speed data (not shown) indicated that the two groups also did not differ on this variable. These results are consistent with those from the one-hour activity test sensorimotor battery, suggesting that multiple impacts did not produce nonassociative effects (e.g., sensorimotor or visual disturbances) that might affect subsequent performances on the place (spatial learning) trials. In contrast to the cued trials data, the sham controls seemed to perform better than the impacted mice during the place condition in terms of path length (Figure 3B) and latency (not shown), and this appeared to be the case beginning with the initial phase of training (block 1).

After the second block of trials, performance levels in both groups began to gradually improve over time, although the sham controls generally performed at higher levels until the last block of trials. Significant main effects of group resulting from the ANOVAs conducted on the place trials data confirmed that the impacted mice had significantly longer mean escape path lengths (F [1,17] = 5.10, p = 0.036) and mean latencies (F [1,17] = 5.18, p = 0.036) compared with sham controls. In terms of swimming speeds exhibited during the place trials (Figure 3C), the sham controls tended to swim faster than the impacted mice, although an ANOVA on these data showed a marginally nonsignificant effect of group (p = 0.074) pertaining to this variable.

TABLE 3. Sensorimotor Results According to Group

Analyses of the probe trial data (Figure 3D and E) showed that the two groups did not differ in terms of their retention performance with regard to time spent in the target quadrant where the platform had been located or in terms of the times they swam over the former platform location (platform crossings). In addition, Figure 3D shows that both groups exhibited a spatial bias toward spending more time searching the target quadrant versus the other pool quadrants. Although the groups did not differ in retention performance, Figure 3F shows that the impacted mice swam a significantly shorter total distance during the probe trial compared with the sham controls (F [1,17] = 6.43, p = 0.021). Post hoc, we evaluated possible differences in swimming speeds during the last block of trials of the place condition that preceded the probe trial. The impacted mice were found to swim at significantly slower speeds than the sham controls during the last block of trials. These slower swimming speeds of the impacted mice did not prevent them from performing as well as the sham mice on the retention measures tested during the probe trial or on the last block of place trials.

Figure 3. Spatial learning performance of the impacted mice was impaired during place trials in the Morris water maze. (A) An analysis of variance (ANOVA) on the path length data from the cued (visible platform) trials showed that the performances of the impacted and sham control groups were not significantly different and that both groups showed substantial improvement over time. Each block of trials contains data from two trials. These data suggest that nonassociative influences (e.g., visual or sensorimotor disturbances) did not likely have an effect on the performance of the impacted mice on subsequent place trials, (B) A significant main effect of group resulting from an ANOVA conducted on the place trials data (three trials per block) showed that, in general, the impacted mice required significantly longer path lengths (F (1,17) = 5.51, p = 0.031) to find and escape onto the submerged platform. Note, however, that the impacted mice performed at control-like levels by the last block of trials. The results of ANOVA on swimming speeds during place trials (C) and on probe trial measures (time in target quadrant (D) and platform crossings (E)) suggested that, in general, the two groups did not differ significantly on these variables. Note that both groups showed a "spatial bias" for searching the target quadrant (D), suggesting that both groups exhibited retention for the platform location. The dotted line represents the amount of time expected based on chance alone. (F) This graph shows that the impacted mice swam a significantly shorter distance (*p = 0.021) than the sham controls, indicating that the impacted mice swam at slower speeds than the sham controls when they were required to swim for the entire 60-second trial. This did not prevent the impacted mice from performing as well as the sham controls in terms of time spent in the target quadrant or platform crossings, as previously described.

Histopathology. Light-microscopic evaluation of the silver- stained material showed damage primarily in the rostral half of the brain along the ventral surface, including structures that were in contact with or in close proximity to the skull, suggesting a contra- coup type of insult resulting from compression and movement of the ventral brain against the rough contours of the basilar skull (Figure 4). The optic tract (Figure 5), anterior olfactory nuclei, and lateral olfactory tract (Figure 4) were the foci of the impact- induced neuro-degeneration in this region. Damage was also observed along the innervation pathways emanating from these two sources. For example, in the visual pathway, terminal field argyrophilia was observed in the lateral geniculate and superior colliculi in most of the impacted mice, with variable damage noted on different sides of the brain (Figure 5C and E).

Figure 4. Multiple impacts produced damage in the most dorsal and ventral parts of the rostral half of the mouse brain. (A) A relatively small amount of argyrophilic soma and fibers was observed in the retrosplenial cortex (underlying the point of impact) 24 hours after the third impact. This silver staining is likely due to direct compression of the skull overlying this area. The impacts also resulted in a contra-coup type of brain damage (B-D) where structures that were in contact with or close proximity to the base of the skull were affected. (B) For example, robust silver staining is seen in extended portions of the lateral olfactory tract in this image. (D) This panel shows a higher magnification of the boxed area in B. Besides the deep staining of the lateral olfactory tract, argyrophilic profiles including soma (arrow) and puncta are also present in the neuropil of deeper structures such as the amygdala. (C) Damage was also observed in more anterior olfactory regions, including the anterior olfactory nuclei and tenia tecta, where the lesion had the appearance of a large infarct. Neurodegenerative changes were also found in layer 1 of the piriform cortex, olfactory tubercle, and lateral olfactory tract at this level. E and F show no evidence of neurodegeneration in level-matched sections in the sham controls. AONm = anterior olfactory nucleus, medial part; AONv = anterior olfactory nucleus, ventral part; COAa = cortical nucleus of the amygdala; lot = lateral olfactory tract; oc = optic chiasm; PIR = piriform cortex; RSPd = retrosplenial cortex, dorsal part; TTv = tenia tecta, ventral part. Scale bars: A = 100 m B = 1 mm; C and E = 150 m D and F = 100 m.

Damage to the anterior olfactor\y nuclei often had the appearance of a large infarct (Figure 4C), and corresponding somal and fiber staining was also observed in portions of the lateral olfactory tract (Figure 4B and D). The neocortex generally exhibited a lack of argyrophilia, with the exception of some occasional somal and fiber staining in the cingulate/ retrosplenial cortex (Figure 4A), an area that was directly under the point of impact to the skull. This damage was fairly restricted and not consistently observed in the brain sections sampled from each animal. In general, much of the brain seemed relatively unaffected by the impacts. There was no evidence of neurodegeneration on silver staining in level-matched sections from the sham controls (Figure 4E and F and Figure 5B, D, and F).

Figure 5. Multiple impacts produced damage to the visual system. (A) Contra-coup damage in the optic tract is shown in this micrograph. The damage in the optic tract was typically mild to moderate, and several different argyrophilic profiles could be observed such as degenerating fibers (arrowheads) as well as ovoid and other irregularly shaped bodies, It was not clear whether the latter two profiles were degenerating glial elements and/or myelin- related debris. Although damage to the optic tract tended to be mild, it was substantial enough to reliably produce neurodegenerative changes in other parts of the visual system that followed the innervation pathways emanating from the optic tract in the impacted mice. Terminal field argyrophilia was observed in the lateral geniculate (C) and superior colliculi (E) in almost all impacted mice, and sometimes there was more robust silver staining on one side of the brain than the other. No similar silver staining was observed in level-matched brain sections in sham controls (B, D, and F). LGd = lateral geniculate nucleus, dorsal part; LGv = lateral geniculate nucleus, ventral part; opt = optic tract; SC = superior colliculus; sgr = superficial gray layer of the superior colliculus; zo = zonal layer of the superior colliculus. Scale bars: A and B = 50 m C-F = 200 m.

DISCUSSION

There is little doubt that moderate to severe TBI is associated with long-term neurologic sequelae, including cognitive dysfunction.16-18 What is less clear is whether mild TBI can also lead to cognitive impairment, with several studies showing negative results.19-21 More than one million concussions occur annually in this country, with 300,000 resulting from sports-related activities. This is an annual incidence greater than that of all cases of moderate and severe TBI and Alzheimer's disease combined.22

A number of studies performed over the past decade suggest that cognitive performance (using standardized psychometric testing) is significantly poorer in subjects who have had one or more concussions than in age-matched subjects without a history of concussion.1,3-6 Imaging studies have also been conducted in professional boxers with a history of concussion. These studies have shown a higher prevalence of pathologic lesions compared with that observed in age-matched, noninjured controls.23-25

Most people at risk for mild TBI are young, and the health care and societal implications of prolonged learning disability resulting from concussions are profound. A valid animal model of concussion or mild TBI would add useful information concerning several different issues, including the threshold of injury required to induce learning/memory impairments. Quantifying injury thresholds is problematic, however. Weight-drop models have been used recently to produce mild TBI in mice. The description of these models includes the weight of the dropped object and the height from which the object is dropped. This information is not sufficient to quantify the actual force imparted to the mouse head if the hardness of the impacting device and the resilience or deformation characteristics of the head holder are not reported. Just as falls from the same height have different outcomes depending on the "landing surface," the physical characteristics of the impacting device can lead to different results, even with apparently similar initial conditions. Measuring impact force helps directly characterize the potential magnitude for injury, accounting for variations in the hardness of the impactor and holder. Our force data show some variation within and across impact sessions, despite efforts to maintain identical testing conditions. In particular, there was a substantial (marginally nonsignificant) reduction in force from the first to the second impact. Some of this may be due to anatomic differences in mouse heads and slight variations (about 1 mm) in the location of the impact. However, it is also possible that repeated impacts may change physical characteristics of the scalp or skull, perhaps through edema or hematoma of the scalp or microfractures of the skull. This is supported by our observation of a substantial reduction in peak force in the single known case of skull fracture. Although impact force can help specify the potential severity of a closed-head injury, differences in the size and mechanical characteristics of the mouse skull and brain compared with humans make it difficult to scale this mathematically to a comparable impact for people.

Nevertheless, measuring force of impact allows an assessment of the reproducibility of impact parameters across experiments and for studying associations with other important variables such as deformation (strain) and rate of deformation (strain rate) of brain tissue. Ultimately, deformation and rate of deformation of brain tissue are the critical factors that determine brain injury. Force is just one readily measurable parameter related to deformation in our mouse model where deformation is achieved through compression of a somewhat flexible skull. Displacement measurements would provide complementary and more direct measurement of deformation. In humans, or other animals with larger brains and rigid skulls, the prevailing mechanism for concussion is believed to be angular acceleration that leads to deformation of the brain through shearing forces.26

In mice and rats, the small size of the brain precludes the attainment of large deformations by acceleration of the skull. Accordingly, concussions in the weight-drop model most likely arise from brain deformation resulting from compression of the skull. Although the compression model differs from the acceleration model in terms of external mechanical induction, the injury is likely caused by the same mechanical trigger (i.e., abnormally large and fast deformation of brain tissue).

In this study, multiple mild TBI produced deficits in place learning performance as measured in the water navigation task. The data from the one-hour locomotor activity test, sensorimotor battery, and cued trials suggest that the place-learning impairment in the impacted mice was not due to nonassociative factors such as visual or sensorimotor disturbances or altered motivation. However, the slower swimming speeds exhibited by the impacted mice during the probe trial and the last block of place trial training suggest that there may be some subtle sensorimotor or behavioral disturbance not detected by the other tests, such as a relative lack of physical endurance.

Slower swimming speeds should not have affected the path lengths of the impacted mice during the place trials and did not prevent the mice from performing at control-like levels during the probe trial or demonstrating a spatial bias in the target quadrant, suggesting they had learned the task.

Previous murine models of mild TBI have not shown cognitive impairment resulting from a single impact; however, multiple impacts have been found, under certain conditions, to compromise learning/ memory functions. DeFord et al. reported that repeated impacts with 100-g and 150-g masses dropped from 40 cm resulted in significantly increased escape latencies during place trials compared with mice that received a single impact with a 150-g mass or sham-impacted mice.10 In the present study, subjecting mice to three impacts with a smaller mass (21 g) dropped from a lower height (35 cm) also resulted in impaired performance during the place condition in the MWM. Our mice were of similar age and were tested in the MWM at about the same postimpact time as those used by DeFord et al.10 However, we included a battery of tests (one-hour activity test, sensorimotor battery, cued and probe trials, and calculation of swimming speeds in the MWM) not used by these investigators to determine the presence of nonassociative factors such as sensorimotor disturbances, which are sometimes found following multiple-impact procedures11 and could confound interpretation of learning and memory data. The preponderance of evidence from these tests suggests that the place-learning impairment in the impacted mice was not due to nonassociative factors. However, it may be prudent to consider the possibility that some subtle sensory deficit induced by damage to various parts of the visual system may have contributed to the inferior performance of the impacted mice during the place condition. Histopathologic analysis suggests that there may be damage to the olfactory system, although it is not immediately clear how this might affect spatial navigation. Future studies using our impact model should assess the possibility of visual and olfactory sensory dysfunction and the role this may play in impaired cognitive function.

Studies of the effects of repeated mild TBI in humans have shown that a person who has experienced one concussion is three times as likely to experience a second concussion and eight times as likely to experience three or more concussions.27 Studies investigating the effects of multiple mild TBI have shown that experiencing more than three concussions may result in irreversible cognitive impairment.28 Specifically, Gronwa\ll and Wrightson found that individuals with a history of repeated mild TBI took longer to reach normal information- processing speeds after experiencing a mild TBI.29 Repeated mild TBI is of particular importance for children and adults who participate in collision sports such as football, soccer, ice hockey, and boxing.30 Recent studies among high school and college athletes show prolonged effects resulting from sports-related concussion,5,6 with one recent study showing that high school players had more prolonged effects than collegiate athletes.31 A replicable animal model that adequately reproduces the hallmark symptoms of mild TBI, such as transient loss of consciousness and subsequent disruptions in cognitive processing, would be important in further investigating the effects of repetitive mild TBI.

LIMITATIONS

In humans, acceleration of the head is believed to play a central role in the pathophysiology of TBI. In our weight-drop model, TBI probably occurs from brain deformation resulting from compression of the skull. Although the external mechanical induction of TBI differs, the injury is likely caused by the same event, namely, abnormally large and fast deformation of brain tissue. A second limitation of this model was the rather modest effect on learning. The impacted mice were found to be significantly impaired on the place-learning condition in the MWM relative to sham controls, but group differences may not be large enough to conduct studies involving therapeutic interventions such as neuroprotective agents. Our weight-drop procedure may have to be modified to induce damage in important memory-related brain regions to make it appropriate for these kinds of studies. Future studies will include an evaluation of the role of specific genetic factors in the susceptibility to impact- induced neuropathology and cognitive impairment.

CONCLUSIONS

Multiple impacts to the heads of C57BL/6 mice in the force range of approximately 20 N were associated with prolonged loss of consciousness and impaired spatial learning performance on the MWM compared with sham controls. Cupric silver staining revealed histologic damage mainly in the optic tract and olfactory nuclei and in the terminal fields located in the innervation pathways of these two systems.

References

1. Tysvaer AT, Storli OV, Bachen NI. Soccer injuries to the brain. A neurologic and electroencephalographic study of former players. Acta Neurol Scand. 1989; 80:151-6.

2. Tysvaer AT, Storli OV. Soccer injuries to the brain. A neurologic and electroencephalographic study of active football players. Am J Sports Med. 1989; 17:573-8.

3. Collins MW, Lovell MR, McKeag DB. Current issues in managing sports-related concussion. JAMA. 1999; 282:2283-5.

4. Matser EJ, Kessels AG, Lezak MD, Jordan BD, Troost J. Neuropsychological impairment in amateur soccer players. JAMA. 1999; 282:971-3.

5. Guskiewicz KM, McCrea M, Marshall SW, et al. Cumulative effects associated with recurrent concussion in collegiate football players: the NCAA Concussion Study. JAMA. 2003; 290:2549-55.

6. McCrea M, Guskiewicz KM, Marshall SW, et al. Acute effects and recovery time following concussion in collegiate football players: the NCAA Concussion Study. JAMA. 2003; 290:2556-63.

7. Tang YP, Noda Y, Hasegawa T, Nabeshima T. A concussive-like brain injury model in mice (I): impairment in learning and memory. J Neurotrauma. 1997; 14:851-62.

8. Tang YP, Noda Y, Hasegawa T, Nabeshima T. A concussive-like brain injury model in mice (II): selective neuronal loss in the cortex and hippocampus. J Neurotrauma. 1997; 14:863-73.

9. Uryu K, Laurer H, McIntosh T, et al. Repetitive mild brain trauma accelerates Abeta deposition, lipid peroxidation, and cognitive impairment in a transgenic mouse model of Alzheimer amyloidosis. J Neurosci. 2002; 22:446-54.

10. DeFord SM, Wilson MS, Rice AC, et al. Repeated mild brain injuries result in cognitive impairment in B6C3F1 mice. J Neurotrauma. 2002; 19:427-38.

11. Laurer HL, Bareyre FM, Lee VM, et al. Mild head injury increasing the brain's vulnerability to a second concussive impact. J Neurosurg. 2001; 95:859-70.

12. Kanayama G, Takeda M, Niigawa H, et al. The effects of repetitive mild brain injury on cytoskeletal protein and behavior. Methods Find Exp Clin Pharmacol. 1996; 18:105-15.

13. Wozniak DF, Brosnan-Watters G, Nardi A, et al. MK-801 neurotoxicity in male mice: histologic effects and chronic impairment in spatial learning. Brain Res. 1996; 707:165-79.

14. Ho N, Liauw JA, Blaeser F, et al. Impaired synaptic plasticity and cAMP response element-binding protein activation in Ca2+/calmodulin-dependent protein kinase type IV/Gr-deficient mice. J Neurosci. 2000; 20:6459-72.

15. Ikonomidou C, Qin Y, Labruyere J, Kirby C, Olney JW. Prevention of trauma-induced neurodegeneration in infant rat brain. Pediatr Res. 1996; 39:1020-7.

16. Nemetz PN, Leibson C, Naessens JM, et al. Traumatic brain injury and time to onset of Alzheimer's disease: a population-based study. Am J Epidemiol. 1999; 149:32-40.

17. Guo Z, Cupples LA, Kurz A, et al. Head injury and the risk of AD in the MIRAGE study. Neurology. 2000; 54:1316-23.

18. Plassman BL, Havlik RJ, Steffens DC, et al. Documented head injury in early adulthood and risk of Alzheimer's disease and other dementias. Neurology. 2000; 55:1158-66.

19. Chandra V, Kokmen E, Schoenberg BS, Beard CM. Head trauma with loss of consciousness as a risk factor for Alzheimer's disease. Neurology. 9 199 89; 39:1576 -8.

20. Fratiglioni L. Epidemiology of Alzheimer's disease. Issues of etiology and validity. Acta Neurol Scand Suppl. 1993; 145:1 -70.

21. Launer LJ, Andersen K, Dewey ME, et al. Rates and risk factors for dementia and Alzheimer's disease: results from EURODEM pooled analyses. EURODEM Incidence Research Group and Work Groups. European Studies of Dementia. Neurology. 9 1999; 52:78 -84.

22. Sosin DM, Sniezek JE, Thurman DJ. Incidence of mild and moderate brain injury in the United States, 1991. Brain Inj. 1996; 10:47 -54.

23. Casson IR, Siegel O, Sham R, Campbell EA, Tarlau M, DiDomenico A. Brain damage in modern boxers. JAMA. 1984; 251:2663 - 7.

24. Ross RJ, Cole M, Thompson JS, Kim KH. Boxers -computed tomography, EEG, and neurological evaluation. JAMA. 1983; 249:211 - 3.

25. Ross RJ, Casson IR, Siegel O, Cole M. Boxing injuries: neurologic, radiologie, and neuropsychologic [iota]valuation. Clin Sports Med. 1987; 6:41 -51.

26. Shaw NA. The neurophysiology of concussion. Prog Neurobiol. 2002; 67:281 -344.

27. Annegers JF, Grabow JD, Kurland LT, Laws ER Jr. The incidence, causes, and secular trends of head trauma in Olmsted County, Minnesota, 1935 -1974. Neurology. 1980; 30: 912 -9.

28. Macciocchi SN, Earth JT, Littlefield LM. Outcome after mild head injury. Clin Sports Med. 1998; 17:27 -36.

29. Gronwall D, Wrightson P. Cumulative effect of concussion. Lancet. 1975; 2:995 -7.

30. Sturmi JE, Smith C, Lombardo JA. Mild brain trauma in sports. Diagnosis and treatment guidelines. Sports Med. 1998; 25:351 -8.

31. Field M, Collins MW, Lovell MR, Maroon J. Does age play a role in recovery from sports -related concussion? A comparison of high school and collegiate athletes. J Pediatr. 2003; 142:546 -53.

Catherine E. Creeley, MA, David F. Wozniak, PhD, Philip V. Bayly, PhD, John W. Olney, MD, Lawrence M. Lewis, MD

From the Department of Psychiatry (CEC, DFW, JWO), Department of Mechanical Engineering (PVB), and Division of Emergency Medicine (LML), Washington University School of Medicine, St. Louis, MO.

Received November 20, 2003; revision received February 10, 2004; accepted March 10, 2004.

Presented in part at the SAEM annual meeting, Boston, MA, May 2003.

Supported in part by National Institutes of Health grant AG11355 (to DFW/JWO) and funding from the Alzheimer's Disease Research Center (P50AG05681) (to LML).

Address for correspondence and reprints: Lawrence M. Lewis, MD, Division of Emergency Medicine, Washington University School of Medicine, Box 8072, 600 South Euclid Avenue, St. Louis, MO 63110. Fax: 314-362-0478; e-mail: lewisl@msnotes.wustl.edu.

doi:10.1197/j.aem.2004.03.006

Copyright Hanley & Belfus, Inc. Aug 2004

More News in this Category