Abstract
A subconcussive impact (SCI) is defined as a bump, jolt, or blow that does not result in any apparent symptoms. Many athletes who participate in contact sports suffer from long-lasting effects of neurodegeneration after many years of continued exposure to such SCIs. Since the existing preclinical models do not adequately replicate the delivery of repetitive subconcussive head impacts that are representative of head injuries in a sport, the goal of this research was to develop a model of repetitive subconcussive impacts (rSCIs) based on comparisons with sham rats and rats that received a single or rSCIs using behavioral and histological outcomes. Thirty adult male rats were randomly divided into three groups in which they were treated either with a sham procedure, a single SCI or rSCI. The SCIs were administered using the Marmarou impact acceleration model by dropping a 50-gram weight from a height of 50 cm onto a steel helmet affixed on the rat’s head and delivering 0.25 J of energy with each impact. Surface righting was found to be significantly increased in the rSCI group compared to the sham and single SCI animals. Although no group differences in learned behaviors were observed, the rSCI group showed significantly greater variance in learning. Histological analysis demonstrated a significant presence of beta-amyloid precursor protein–reactive profiles in the rSCI group compared with the sham and SCI groups. This research establishes a foundational subconcussive animal model that is noninjurious in singularity but produces changes in righting reflex and axonal injury with repeated exposures.
Introduction
The behavioral and histological response to traumatic brain injury (TBI) and mild TBI (mTBI) using closed-head injury models in rats and mice have been widely documented in literature,1–6 but there are limited studies on how repetitive subconcussive impacts (rSCIs) affect these responses.7–11 With focus over the last decade on injurious impacts contributing to mTBI and long-term symptoms associated with repetitive exposures in athletes who participate in contact sports, there is a need to investigate impacts that may not immediately appear to be clinically significant, that is., SCIs.
From a real-world perspective, Stojsih et al. reported that more than 99% of the head impacts in boxing are below the threshold of causing a concussion (head injury criterion value 250), 12 and Jansen et al. also reported that more than 90% of the punches in boxing and mixed martial arts (MMA) were below the threshold that had a 25% probability of causing a concussion. 13 Taken together, these SCIs are often overlooked because a single impact does not result in any clinically noticeable symptoms. However, when deceased athletes with severe neurodegenerative diseases were autopsied, the lack of documented concussive events in some of their careers led to the belief that the effects resulting from subconcussive injuries may set the stage for such neurodegeneration.14,15 Due to the persisting effects of routine head injury exposures even after many years, the effects of multiple impacts are said to be cumulative,12,16–18 but these studies do not explain how the progression of the underlying neuropathology correlates with those above the threshold level.
Even with the advancements in human studies of TBI, issues with relying on self-reports of exposure and athlete history persist. Therefore, the establishment of an animal model, especially to study repetitive subconcussive head impacts, is essential. In the last decade, there has been some development in studying the effects of repetitive mTBI and reported subconcussive models in literature;7–11,14,15,19 however, these models have not been adequately validated to show that a single impact at the proposed conditions does not induce any neurodegenerative changes when compared to a sham. If an impact is truly subconcussive, then a single impact should not show any significant differences from a sham animal. This finding needs to be validated prior to labeling any given impact condition as subconcussive. Once this level is determined, then analyzing the cumulative effects of such impacts is possible.
Thus, the goal of the current research is to develop a rSCI model based on a truly single noninjurious impact. The gold standard of histological analysis and neurocognitive assessments was used to determine the level of injury. Once this threshold is determined, then various paradigms, such as duration between impacts, frequency of impacts (to simulate the head impact exposure in different sports), and the latency period after which neurological symptoms become irreversible and apparent, can be evaluated.
Methods
All experiments were approved by the Wayne State University Institutional Animal Care and Use Committee. Relevant guidelines in the Guide for the Care and Use of Laboratory Animals (2011) were also followed. 20
A total of 31 Sprague Dawley rats weighing between 210 and 258 grams were used in this study. Upon arrival, they were pair-housed and allowed to acclimatize in a humid and temperature-controlled vivarium room set to a reverse light/dark cycle (lights off at 6:00 and on at 18:00). The rats were given unlimited access to food and water, except during the head impact procedure and behavioral testing. All rats were handled for 10 days prior to the initiation of the experimental procedures. All rats were randomly assigned to one of the three experimental groups: sham, single SCI, or rSCI. As shown in the experimental timeline (Fig. 1a), the rats underwent an impact or sham procedure (day 1), followed by 6 days of Barnes maze (BM) assessment (days 2 to 7). Following the conclusion of BM testing, they were euthanized by transcardial perfusion to harvest the brain for subsequent histopathological assessment.
Head impact procedure
SCIs or sham procedures were performed using a modified version of the Marmarou impact acceleration device. After allowing 30 min for habituation, the animals were anesthetized with isoflurane (4%) in oxygen (0.6 L) for 4 min in an induction chamber. Then while maintaining anesthesia (2% isoflurane), they were placed on the polyurethane foam. Similar to the original Marmarou impact acceleration model, 21 a circular steel “helmet” (1 cm wide and 0.5 cm thick) was strapped onto the dorsal side of the rat’s head in between the ears. The rat’s head was positioned on the foam bed contained in a Plexiglas box (12 × 12 × 43 cm), and then the body was secured in place by two Velcro straps while ensuring the rat’s head was directly under the lower end of the Plexiglas tube that guided the impactor. Just prior to the induction of impacts, the nose cone supplying anesthesia was removed. Then single or repetitive SCIs were induced by dropping a 50 gram cylindrical brass weight (18 mm diameter) from a height of 50 cm onto the affixed steel helmet. This imparted 0.25 J of energy on the rat’s head. This drop mass and velocity of impact were determined based on the initial work of Lavender et al. (2020) 11 and our preliminary data 22 in order to create a true SCI.
To avoid a second impact due to the rebound of the drop mass, the Plexiglas box was moved away from underneath the impact tube area immediately after impact. Sham rats were subjected to identical experimental procedures, except for the impact(s). Immediately following the impact or sham procedure in the sham, SCI, and rSCI groups, the nose cone was reattached to maintain continued supply of isoflurane.
Animals in the single SCI group were subjected to one SCI, and animals in the rSCI group were subjected to 10 consecutive SCIs with an interval of 10 sec between each impact (Fig. 1b). In real life, the frequency of each repetitive exposure varies depending on the sports activity. As a baseline, the frequency of head impacts reported in MMA matches was used to estimate the inter-impact interval. According to a previous study, the average number of impacts per athlete for a 5 min session was reported to be 6.9, 13 which results in an impact interval of 43.5 sec. Using the cube scaling law for rats and humans, 23 a 10 sec interimpact interval was used in this rodent model to match the frequency of head impacts in MMA match. Sham animals underwent all procedures without any impact. To minimize variation in anesthesia duration between animals in the three groups, an attempt was made to maintain consistency in the duration of anesthesia. Therefore, animals in the SCI group were maintained for the same anesthesia duration as in the rSCI group, and animals in the SCI group were impacted (1×) at the time point that corresponded with the 10th impact of the rSCI group. After sham or impact procedures, all the animals were placed in a recovery chamber and tested for duration of the surface righting reflex, measured in seconds as a proxy for duration of loss of consciousness.
Behavioral assessment on the BM
A custom-made BM (Formtech Plastics, Oak Park, MI) consisting of a 122 cm diameter circular platform with 20 equally spaced 10 cm diameter holes along its periphery was used to assess cognitive changes (Fig. 2a). 24 The platform was elevated 1 m above the ground to encourage the rats to stay on the platform throughout the duration of the trial. A hidden escape box (25 × 17 × 12.5 cm) was placed beneath one of the holes to allow the rat to enter and hide in a relatively secure, dark space. Visual cues were placed around the room to allow the rat to learn where the escape box was located. Two bright lights placed in each corner of the room, combined with an 80 dB white noise, were used to create an aversive environment to facilitate learning of escape behavior.
All rats were subjected to BM testing on days 2 through 7. On each day of testing, the rats were brought into the behavioral testing room and allowed to acclimate for at least 30 min prior to testing. The acquisition phase occurred on days 2–5. Each rat underwent two trials per day, with an intertrial interval of 4 min. The testing began by placing the rat at the center of the maze under a start box for 30 sec of acclimation. Each trial began when the start box was lifted, allowing the rat to explore the maze. On the first day of the testing (habituation), the rat was allowed to freely explore the maze with no anxiogenic stimuli for the first trial. After 90 sec of exploration, the rat was gently guided to the escape hole and allowed to stay in the underlying escape box. After 3 min, the rat was removed from the escape box and returned to its home cage. For the second trial, the lights in the room were switched on, and the rat was placed under the start box. Trial 2 began by lifting the start box and allowing the rat to explore the maze. The rat was guided to the escape hole if it did not enter within a minute and was again allowed to stay in the escape box for the remainder of the 3 min trial period. The maze was thoroughly cleaned with 70% ethanol between each trial.
For trials on days 3–5 (acquisition), rats were placed under the start box, and the anxiogenic stimuli were turned on. Trials began when the start box was removed, thereby allowing the rat to find the escape hole and enter the escape box. Trials ended once the rat entered the designated escape hole they were returned to their home cages. All rats were monitored for latency to enter the target zone located in quadrant 3 (Fig. 2a), and the distance travelled each day was calculated as the average of both trials.
On day 6, the rats were tested by a probe trial for assessing spatial learning and memory recall. On this day, all the holes were closed, and the rats underwent two trials for 3 min. After 3 min, the trial ended, and the rats returned to their home cages. The percentage time spent in the escape quadrant (Q3), total distance traveled, and latency to get to the target zone were calculated. On day 7, the rats were tested by a reversal trial to assess their cognitive flexibility. On this day, the target zone, which includes the escape box, was moved directly opposite the original location to quadrant 2 (Q2, Fig. 2a).
The BM behavior was recorded by a ceiling-mounted camera positioned in the behavioral room and processed by Noldus EthoVision XT software (Noldus, Wageningen, Netherlands) to quantify outcome measures. Using this software, the circular platform was divided into 4 quadrants (Q1–Q4) with a starting zone designated. The hole with the escape box was marked as the target hole, and the circular zone around it was marked as the target zone (Fig. 2a). The calculated measures include latency to enter the target zone (in Q3 on days 2–6 or in Q2 on day 7), latency to enter the escape hole, time spent in the quadrants, and distance travelled. Results were shown as an average of two trials on all days except the probe trial on day 6, which reported the results from only the first trial due to methodological differences between trials on that day.
Perfusion and immunohistochemistry
Twenty-four hours after BM testing on day 8, all animals were deeply anesthetized (5% isoflurane) and transcardially perfused with cold 4% paraformaldehyde (PFA) to achieve euthanasia. The brains were harvested and postfixed in PFA for 2 days. The brains were then cryoprotected by immersion in graded sucrose (15% and 30%) solution. Then 40 µm thick sections from bregma −1.5 to −4.5 mm were collected and stored in 1× phosphate-buffered saline filled multiwell plates at 4°C until further processing.
Three representative brain sections from the hippocampus (−2.3, −3.1, and −3.9 mm bregma) of each brain were incubated in citrate buffer (pH 6.0, C9999, Sigma Aldrich, Saint Louis, MO) at 90°C for 30 min to facilitate antigen retrieval. This was followed by immersion in 0.6% hydrogen peroxide solution to quench any endogenous peroxidase activity. The sections were then incubated in a blocking buffer containing 2% normal goat serum (Vector Laboratories, Burlingame, CA) and 2% bovine serum albumin for an additional 1 h at room temperature. Then sets of sections were incubated in blocking buffer solution with anti-ionized calcium-binding adapter molecule 1 (Iba1, cat # 019–19741, Wako Chemicals, Richmond, VA; 1:2000), anti-glial fibrillary acidic protein (GFAP, Thermofisher Scientific, MA; 1:2000), or anti-beta-amyloid precursor protein (β-APP, Zymed, CA; 1:1000) antibodies for 72 h at 4°C. Then the sections were thoroughly washed and incubated in biotinylated goat anti-rabbit or anti-mouse secondary antibodies (Vector Laboratories; 1:500) dilution for 1 h at room temperature. All the sections were incubated in avidin-biotin peroxidase complex solution (Vectastain ABC kit, Vector Laboratories) for another hour, and the peroxidase activity was developed by incubation in diaminobenzidine (DAB kit, Vector Laboratories). Finally, they were mounted onto slides, cover-slipped, and observed under the microscope (EVOS XL Core).
For each section, one image from the CA1 region from the left and right sides was captured. In this manner, six images (2 images/section × 3 sections) of representative microglia/macrophage (Iba1-reactive) and astrocyte (GFAP-reactive) were counted for each animal. For each group, 60 images (6 sections × 10 animals) were considered and analyzed for Iba1 and GFAP-reactive cells. Out of the 180 (60 images/group × 3 groups) images for Iba1 and 180 images for GFAP, four images for the GFAP cells and six images for the Iba1 cells could not be counted due to inconsistent staining. The cell counter function in ImageJ (Bethesda, MD) was used to count the number of Iba1 and GFAP-positive cells from each image encompassing the CA1 region in the hippocampus. For diffuse axonal injury evidenced by β-APP reactive axonal profiles, all three sections of each brain were thoroughly scanned at 20× magnification, which was considered a high-power field (HPF). Then each HPF was scored for the presence or absence of β-APP reactive swollen axons or retraction beads as seen in TBI studies in rats, 25 and the number of HPFs was recorded and averaged based on the number of sections counted from. One section from the rSCI and SCI groups was not considered due to inconsistent staining.
Statistical analysis
Surface righting duration following impacts was analyzed for group-wise differences using GraphPad Prism (v10, CA, USA) statistical software. Several outcomes from the BM testing, including the latency to enter the target zone, distance travelled until it first entered the target zone (days 2–5 and 7), percentage time spent in the target quadrant, and time until it first reached the target zone on day 6, were also analyzed. Additionally, a metric of variance for latency to enter and distance travelled till the target zone during probe trial (day 6) and reversal trial (day 7) for each animal was calculated using the squared deviation from the mean
Histological outcomes, glial cell (Iba1 or GFAP) counts and axonal injury in HPF counts, were also assessed for group differences. Traditional or nonparametric one-way analysis of variance (ANOVA) with appropriate post hoc tests was used. A traditional significance p value threshold of < 0.05 (e.g., 95% confidence interval [CI]) was used.
Results
Repetitive SCIs lead to extended latency to surface righting
On day 1, rats in the rSCI group demonstrated a significantly prolonged duration to surface right compared to those in the sham or SCI group (F(2, 27) = 5.769, p < 0.05). Surface righting duration was not significantly different between rats in the sham and SCI groups (p > 0.05) (Fig. 1c).
Behavioral response
BM Training (Days 2–5)
Rats from all three groups were subjected to BM training between days 2 and 5. As shown (Fig. 2b–c), all the rats learned the task with each passing day (d2–d5), as shown by a reduced latency to enter. The target zone has the escape hole (Fig. 2a). Coinciding with the reduced latency, the rats also showed a reduced distance travelled in their efforts to enter the target zone. However, no significant differences between groups were observed in these metrics.
Probe Trial Day (Day 6)
Following training sessions on days 2–5, the rats in all the groups underwent a probe trial on day 6 to assess the effect of the rSCI on short-term memory recall (i.e., learning). The rats in the rSCI group showed a prolonged duration (18.9 ± 2.3 sec) to reach the target zone, with the escape hole being covered. On the other hand, rats in the sham (9.2 ± 2.1 sec) and SCI (9 ± 2.3 sec) groups demonstrated a relatively shorter duration to reach the target zone. Although the latency to enter the target zone varied greatly, the means were not significantly different (sham and rSCI: p = 0.1256, SCI and rSCI: p = 0.1184) (Fig. 2d). However, when the variance score of the latency was further analyzed, the rSCI group showed a significantly high variance score compared with SCI (p = 0.0096) with no other significant differences between other groups (Fig. 2e). Although the rSCI group travelled longer average distances (157 ± 43 cm) compared with both sham (109 ± 21 cm) and SCI (91 ± 16 cm) groups, the differences between groups were not significantly different (sham and rSCI: p = 0.4917, SCI and rSCI: p = 0.2604) (Fig. 2f). When the variance score of the distance travelled was further analyzed, the rSCI group showed a significantly high variance score compared to SCI (p = 0.0123), with no other significant differences between other groups (Fig. 2g). Additionally, the percentage time spent in the target quadrant also did not yield significant difference between groups (sham and rSCI: p = 0.5632, SCI and rSCI: p = 0.1361, and sham and SCI: p = 0.6150) (Fig. 2h).
Reversal Trial (Day 7)
On day 7, all the rats were subjected to a reversal trial assessment to measure behavioral-cognitive flexibility. During this trial, the target zone that includes the escape hole was moved directly opposite the original location to a new quadrant (Q2, Fig. 2a). As shown (Fig. 2i–l), sham rats demonstrated comparably shorter latency (sham = 30.8 ± 4.7 sec) to enter the new escape hole compared to rats in the SCI (38.8 ± 6.7 sec) and rSCI (40.9 ± 5.4 sec) groups. Similarly, sham rats also traveled a shorter distance (280 ± 39 cm) compared with those in SCI (359 ± 49 cm) and rSCI (374 ± 60 cm) in their efforts to enter the new target zone. However, these data were not significantly different. The calculated variance scores for the latency to zone and distance travelled also did not report significant differences between groups.
Astroglia and microglia/macrophage changes
The average number of cells per section was higher for the rSCI group for both microglia/macrophages (rSCI = 51.3 ± 1.4, sham = 48.6 ± 1.1, SCI = 49 ± 1.3) as well as astrocytes (rSCI = 109.6 ± 6.3, sham = 106.6 ± 6.2, SCI = 99.9 ± 4.2). However, there were no statistically significant differences between the groups (Fig. 3).
Representative photomicrographs of microglia and astrocytes in sham, SCI, and rSCI groups respectively. Number of microglia in sham = 46, SCI = 47, and rSCI = 58, and number of astrocytes in sham = 97, SCI = 96, rSCI = 121. Scalebar = 100 µm (left). Top right are representative plates used to identify sections for quantifying glial cells counted from the CA1 region from sections depicted in −2.28 mm, −3.12 mm, and −3.84 mm bregma (adapted from George Paxinos and Charles Watson, The Rat Brain in Stereotaxic Coordinates6th edition, 2007). Bottom right chart shows microglial and astrocytic counts in sections from three groups. Scalebar = 100 µm. All data were reported as mean + SEM.
Analysis of brain sections from all the three groups showed a significantly higher number of β-APP-reactive profiles in the rSCI group compared with the sham and SCI groups (F(2,27) = 14.74, p < 0.05) (Fig. 4). The rSCI group had an average of 25.9(±3.2) β-APP-reactive profiles compared with the sham 10.7(±1.7) and SCI groups 10.7(±1.6). β-APP-reactive profiles appeared to be beaded and swollen, as shown in a representative section from the rSCI group. It should be noted that there were no skull fractures or subdural hematomas noted in any group.
A representative image from rSCI group showing nine beta-amyloid precursor protein (β-APP)-positive axonal profiles (left). Arrows
Discussion
The aim of this study was to validate a model of rSCI to provide the ability to further assess SCIs in a controlled manner. The establishment of an SCI was validated using comparisons between sham rats, rats receiving a single SCI, and rats receiving repetitive SCI based on histological and behavioral outcomes.
The prolonged surface righting duration in the rSCI group suggests a more severe injury outcome compared to the SCI and sham groups. Loss of righting reflex in rat models of TBI is analogous to loss of consciousness in a human following a head injury. It is typically associated with loss of neuromotor control, motor impairments, memory and cognitive impairments, and a greater extent of neuropathology. 27 Longer righting reflex duration is not a measure of injury itself but a predictor of head injury symptom outcomes. 28 Previous studies have demonstrated the sensitivity of the righting reflex when comparing severe TBI to less severe impacts. 25 Bree et al. (2020) reported that 8–9-week-old rats had righting reflex times of less than 2 min when subjected to three SCIs using the impact acceleration model. 29 This is in line with the rats in the sham and SCI impact groups in the current study, with righting reflex durations of 114 (±16) and 130 (±8) sec, respectively.
DeWitt et al. (2013) stated that righting reflex duration of around 15–20 min was considered moderate TBI, and greater than 20 min could be considered severe TBI. 30 The mean righting reflex duration of the rSCI group was about 3 min, with only one rat in this group having a righting reflex duration of 300 sec. These data indicated that the injury severity was significantly higher in the rSCI compared to the sham and single subconcussive test conditions, however lower than the values previously reported for more severe brain injuries.
The effect of repetitive subconcussive events on the rats’ learning and memory was evaluated using the BM methodology. Previous studies have shown that at 6 weeks post repeated mTBI, rats were reported to have taken a longer time to enter the target zone on the BM during the acquisition trial (day 1) compared to a single mTBI group; and both groups reported significantly longer time compared to the sham group. 31 In the current study, the time taken and distance travelled to enter the target zone for all three groups during acquisition (days 2 to 5; Fig. 2b–c), probe trial day 6 (Fig. 2d,f), and reversal trial day 7 (Fig. 2i–k) were not significantly different. These differences between mTBI and rSCI rats in our study may be related to the difference in injury severity produced by the two models, with the current study being less severe. This is in line with other studies involving awake closed head injury using the controlled cortical impactor, where no differences between injured and noninjured groups during the learning, probe, and reversal phase, 32 , and after 1 h and 1 day of injury were reported. 33
Prior research using rat TBI models has shown that all rats spent a significantly longer time in the target quadrant compared to the other quadrants.32,34 The rats in the current study followed this trend, staying in the target quadrant between 38% to 47% of the time. This suggests the utilization of spatial cues to navigate to the position of the escape hole. While the latency and distance travelled to first reach the target zone (Fig. 2d,f) on the probe day showed the rSCI group taking longer average times (rSCI = 18.9 ± 2.3, sham = 9.2 ± 2.1, SCI = 9 ± 2.3 sec) and travelling longer distances on average (rSCI = 157 ± 43, sham = 109 ± 21, SCI = 91 ± 16 cm), there was a lack of statistically significant differences in these parameters. Therefore, it could not be concluded that there were any cognitive deficits in terms of learning on the BM (on days 2–5), memory retention (probe trial on day 6), or cognitive flexibility (reversal trial on day 7) in the rSCI rats compared with the sham and SCI impact groups.34,35
To further understand the effects of the repetitive impacts on the behavior in the rSCI group, the variance scores for latency and distance travelled were calculated for all groups. 26 ANOVA results of the variance scores for the latency to enter the target zone on probe day indicated that not all rats in the rSCI group learnt the task uniformly as evident by wide variations in their latency to the zone (Fig. 2e). The variance score for the rSCI group was significantly different than the variance score of the SCI group (p = 0.0096), but not significantly different than the sham group (p = 0.1701). Similarly to the variance scores for the latency, ANOVA of the variance scores for the distance travelled till the rat entered the new target zone also revealed that the rSCI group showed significantly more variability while completing the task compared to the SCI group (p = 0.0123), but not the sham group. In this study, the energy delivered in each impact (0.25 J) is almost 20 times less than the energy delivered in a traditional mTBI model (4.4 J). For that matter, people exposed to similar brain injury severity may express differential functional outcomes, and individual characteristics such as enhanced diversity may contribute to cognitive resilience, 36 and such genetic variations could be a factor in the observed variations in the current study, albeit they need to be explored further. Additionally, genetic polymorphism influencing recovery from TBI has also been reviewed. 37 Moreover, other robust behavioral testing protocols may also need to be explored that would be more sensitive and help elucidate such subtle individual SCI-related changes.
Other studies that explored the effects of repeated subconcussion have also reported similar findings regarding memory deficits. Hoogenboom et al. (2023) reported that for repetitive SCI, the rats’ short-term memory was not affected in a novel object placement test compared to the sham group when tested 24 h and 4 weeks after rSCI, 7 whereas Wilson et al. (2023) reported that male rats subject to rSCIs showed memory deficits in a novel object recognition test compared to sham rats as indicated by longer percentage times spent with the novel object 30 days after injury. 9 The differences in agreement could be attributed to a different injury paradigm where higher impact intensity could have resulted in memory deficits. Previous studies have imparted a higher energy level to create the injury. Hoogenboom et al. impacted the awake rats under a CCI with 2.5 m/s impactor velocity, 5 mm depth, while Wilson et al. impacted anesthetized rats with 6.5 m/s impactor velocity, 8 mm depth.
For the histological assessment, the average number of glial cells was counted in all groups from the CA1 region in the hippocampus, considering that the learning and memory assessments performed using the BM testing involved the hippocampus.35,38 For the three groups (sham, SCI, and rSCI), there were no significant differences in the number of Iba1 and GFAP reactive cells (p > 0.05, Fig. 3). This is similar to the findings by Clay et al. (2024) 39 and Hoogenboom et al. (2023), 7 where no significant differences in astrocytes and microglia were seen after 7 days and 1 month, respectively, after the induction of a repetitive subconcussive injury. Stemper et al. (2022) also shared similar results where the sham rats did not exhibit significantly different gliosis compared to the moderate injury rats, but were significantly different from the high exposure group. 8 Since their study modeled moderate and high impact exposures, their results were difficult to compare with the rSCI groups in this study. Microglia and astrocytes are generally present in the brain parenchyma in a passive state to phagocytize apoptotic neurons and debris, and they get activated acutely in the event of any pathology in the brain. 40 However, the SCI and rSCI did not seem to be severe enough to elicit a higher immune response in the brain that persisted beyond the acute period. Future studies would also benefit from analyzing the morphological changes, such as ramification, complexity, and shape in glial cells, as demonstrated by Young et al. (2018). 41
Similar to the current study, Hoogenboom et al. (2023) found evidence of increased β-APP reactivity in the reported rSCIs group. Amyloid deposition is considered the beginning of the pathophysiological cascade, followed by tau deposition and eventually toward neurodegenerative diseases such as Alzheimer’s disease or chronic traumatic encephalopathy. 42 This is in line with our results and hypothesis that a single SCI does not cause pathological changes, but rSCI exposure over time is additive.
Conclusion
This study used a single energy level to develop a subconcussive and rSCIs rat model evaluated at 1-week post exposure. Varying the impact parameters, such as impact energy, duration between impacts, and number of days of impacts, to simulate the head impacts in a season in sports and increasing the survival time would greatly increase our understanding of the rat’s behavior as well as neuropathology resulting from rSCI. Second, the use of rats in this study has economical and practical benefits, but translation of these injuries to a human may be difficult due to anatomical features and head impact biomechanics. However, this research provides the foundation to further study the previously mentioned variables in a controlled manner.
This novel SCI induction protocol using Marmarou’s impact acceleration model and neuropathological assessment using the BM and histological analyses demonstrated that a single impact could not be differentiated from the sham group, while repetitive exposures resulted in marked responses. Righting reflex reported in the rSCI group indicated increased injury severity due to the repetitive impacts. An investigation into the cognitive deficits of repetitive impacts yielded no significant differences between impact groups. Additionally, there were no differences in glial (microglia and astrocytes) proliferation in the hippocampus in the repetitive impacts group as assessed at a 1-week time point; however, the significantly higher number of β-APP-positive profiles highlighted the potential axonal injury effect of repetitive impacts. This is a promising study that showcases the neuropathology of rSCIs resulting from one match in contact sports in humans. While no behavioral deficits were reported acutely, there are well-documented research reports in athletes suffering from cognitive deficits due to repeated impacts over time.43–47 Establishing the threshold for subconcussive injuries in terms of number of impacts, impact energy delivered, frequency of impacts over an individual’s lifetime, and duration between impacts would be instrumental in understanding the progression of pathology in contact sports.
Transparency, Rigor, and Reproducibility Statement
The authors affirm that no AI tools were used in the conception, design, analysis, writing, or editing of this article. All content was produced solely by the authors.
Authors’ Contributions
J.V.: Conceptualization, methodology, formal analysis, investigation, and writing—original draft; S.K.: Methodology, formal analysis, writing—original draft, and supervision; K.O.: Investigation; S.A.P.: Methodology, resources, and writing—review and editing; C.B.: Conceptualization, methodology, resources, and writing—review and editing.
Footnotes
Acknowledgments
The authors would like to thank Cameron Davidson, PhD, Assistant Professor, Oakland University, MI, for his help with the interpretation of data and statistical analysis.
Author Disclosure Statement
The author(s) have no competing interests to disclose.
Funding Information
Funds were provided by the Department of Biomedical Engineering, Wayne State University.