Altered Tau Kinase Activity in rTg4510 Mice after a Single Interfaced CHIMERA Traumatic Brain Injury

Traumatic brain injury (TBI) is an established risk factor for neurodegenerative diseases. In this study, we used the Closed Head Injury Model of Engineered Rotational Acceleration (CHIMERA) to investigate the effects of a single high-energy TBI in rTg4510 mice, a mouse model of tauopathy. Fifteen male rTg4510 mice (4 mo) were impacted at 4.0 J using interfaced CHIMERA and were compared to sham controls. Immediately after injury, the TBI mice showed significant mortality (7/15; 47%) and a prolonged duration of loss of the righting reflex. At 2 mo post-injury, surviving mice displayed significant microgliosis (Iba1) and axonal injury (Neurosilver). Western blotting indicated a reduced p-GSK-3β (S9):GSK-3β ratio in TBI mice, suggesting chronic activation of tau kinase. Although longitudinal analysis of plasma total tau suggested that TBI accelerates the appearance of tau in the circulation, there were no significant differences in brain total or p-tau levels, nor did we observe evidence of enhanced neurodegeneration in TBI mice compared to sham mice. In summary, we showed that a single high-energy head impact induces chronic white matter injury and altered GSK-3β activity without an apparent change in post-injury tauopathy in rTg4510 mice.


Introduction
Traumatic brain injury (TBI) is a leading worldwide cause of death and disability. TBI is clinically classified as mild, moderate, or severe, based on the Glasgow Coma Scale and additional clinical indicators, including loss of consciousness and post-traumatic amnesia. Repetitive exposure to mild TBI (mTBI) is associated with increased risk of developing a neurodegenerative condition called chronic traumatic encephalopathy (CTE) [1][2][3]. The pathognomonic features of CTE include aggregates of perivascular phosphorylated tau (p-tau), located at the depths of the cortical sulci [4]. In addition to repetitive mTBI, several reports and case studies have shown that a single exposure to moderate or severe TBI (msTBI) may also induce p-tau neuropathology.
A previous report comparing fatal TBI cases to controls (n = 5 per age group of <20 years, 20-50 years, >50 years) identified subtle tau immunoreactivity in glial cells during the acute period (<24 h to 1 month) after msTBI [5]. A recent study of 39 postmortem brains of chronic survivors (1-47 years post-injury) of a single msTBI found that, in the young cohort, tau-positive neurofibrillary tangles (NFTs) were more frequently observed in TBI cases than in age-matched controls [6]. Other case studies [7,8] examined post-mortem brains from patients who survived decades (24 to 42 years post-injury) after levels, and ratios of different tau isoforms). In general, TBI appears to have a greater probability of inducing tauopathy in mouse models overexpressing mutant forms of tau.
We developed an impact-acceleration model of TBI known as CHIMERA [21], which allows for unrestrained head motion during impact. We recently published a modification that allows us to deliver a high impact energy injury to mice using an interface to distribute impact energy across the skull, thereby avoiding skull fracture, to mimic more severe TBIs [22,23]. In C57/Bl6 mice, we showed that interfaced TBI at 2.5 J resulted in a mortality of <20%, neurological and memory deficits, elevated plasma total tau and neurofilament light levels, increased brain cytokine levels, and blood-brain barrier disruption [22]. The present study was designed to evaluate the effect of a single high impact-energy TBI on chronic tauopathy in rTg4510 mice. Here we used interfaced CHIMERA to induce a single TBI in 4-mo rTg4510 mice using 4.0 J of impact energy, which produced an overall mortality of 46.7%. Animals were aged for 2 mo before brain tissues were harvested for histological and immunoblotting analyses. Blood samples were collected for plasma biomarker measurements. We observed that a single high-energy TBI was sufficient to induce long-term white matter injury and prolonged activation of GSK-3β. However, tauopathy and neurodegeneration were not significantly different between the sham and TBI animals. These findings suggest that a single high-energy head impact may be sufficient to induce long-term white matter injury and increased activity of tau kinases but without a marked effect on tauopathy or neurodegeneration.

CHIMERA TBI with Interface at 4.0 J Induced High Mortality
We first performed a pilot impact energy titration experiment to establish CHIMERA conditions to mimic a moderate-severe TBI in 4-mo rTg4510 mice, overexpressing the P301L human 4R0N tau. Interfaced CHIMERA TBI impact energies ranging from 2.3 J to 5.8 J were tested. Survival was 100% in mice impacted at 2.3-2.9 J, 100% at 3.5-3.8 J, 53.8% at 4.0 J, 50.0% at 4.2 J, and 0% at 4.5-4.8 J (Supplementary Figure S1A). To produce experimental TBI with an a priori defined overall mortality rate of~50%, we selected an impact energy range of 4.0 J for the full study.

TBI Induced Much Prolonged Loss of Righting Reflex and Chronic Axonal Injury and Microgliosis
For the full study, seven rTg4510 mice were randomized to the sham group, which received all procedures except for impact, and 15 rTg4510 mice were randomized to the TBI group and impacted at 4.0 J (n = 15). Overall, eight TBI mice survived. Non-surviving mice either died immediately after TBI or reached the humane end point soon thereafter, resulting in an overall mortality rate of 46.7% (7/15) ( Figure 1A, Supplementary Figure S1B,C). The duration of loss of the righting reflex (LRR, analogous to loss of consciousness in humans) in all surviving mice is shown in Figure 1B. In addition, one TBI mouse that survived the procedure and regained consciousness but reached the humane end point at 6 h (LRR value = 1680) is also included in the figure. On average, sham-operated mice had an LRR duration of 140 s, whereas TBI mice had a significantly longer LRR duration (2225 s, p = 0.0002).
After sham or TBI procedures, the eight surviving rTg4510 mice were aged for 2 months before brain samples were harvested at 6 months of age. Compared to sham controls, TBI mice had chronic microgliosis at the optic tract, as revealed by Iba1 immunohistochemistry (p = 0.0082), as well as axonal injury in the optic tract, as revealed by silver staining (p = 0.0012) ( Figure 1C). These results are consistent with our previous studies demonstrating chronic white matter injury, particularly in the optic tract, after interfaced CHIMERA impacts in C57Bl/6 mice [22,24]. by silver staining (p = 0.0012) ( Figure 1C). These results are consistent with our p studies demonstrating chronic white matter injury, particularly in the optic tra interfaced CHIMERA impacts in C57Bl/6 mice [22,24].

TBI Induced Chronic Activation of GSK-3β
Since rTg4510 mice express high levels of 4R P310L human tau, we tested w TBI induced changes in the levels of tau kinases. We first performed Western blo RIPA lysates to probe for different forms of GSK-3β, a major Ser/Thr kinase resp for tau phosphorylation. Although TBI did not affect the level of total GSK-3β (p = TBI significantly reduced the ratio of phosphorylated GSK-3β (p-GSK-3β S9) to tot 3β (p = 0.0132) (Figure 2A). Since p-GSK-3β S9 is an inactive form of GSK-3β, this suggests that GSK-3β activity was chronically enhanced in these animals. W investigated AKT (also known as protein kinase B), which is a Ser/Thr kinase and regulator of GSK-3β signaling. When activated, AKT phosphorylates GSK-3β

TBI Induced Chronic Activation of GSK-3β
Since rTg4510 mice express high levels of 4R P310L human tau, we tested whether TBI induced changes in the levels of tau kinases. We first performed Western blotting of RIPA lysates to probe for different forms of GSK-3β, a major Ser/Thr kinase responsible for tau phosphorylation. Although TBI did not affect the level of total GSK-3β (p = 0.6499), TBI significantly reduced the ratio of phosphorylated GSK-3β (p-GSK-3β S9) to total GSK-3β (p = 0.0132) (Figure 2A). Since p-GSK-3β S9 is an inactive form of GSK-3β, this finding suggests that GSK-3β activity was chronically enhanced in these animals. We then investigated AKT (also known as protein kinase B), which is a Ser/Thr kinase and a major regulator of GSK-3β signaling. When activated, AKT phosphorylates GSK-3β at S9, leading to its inhibition. Western blotting of RIPA lysates revealed no significant change in total AKT levels (p = 0.9979). However, there was a strong trend toward a reduced ratio of p-AKT (S473) to total AKT (p = 0.0582) ( Figure 2B). We conclude that high impactenergy TBI in rTg4510 mice chronically activated GSK-3β, potentially due to disrupted AKT-GSK-3β signaling.
in total AKT levels (p = 0.9979). However, there was a strong trend toward a reduced ratio of p-AKT (S473) to total AKT (p = 0.0582) ( Figure 2B). We conclude that high impactenergy TBI in rTg4510 mice chronically activated GSK-3β, potentially due to disrupted AKT-GSK-3β signaling.

TBI Accelerated Increase in Plasma Total Tau in rTg4510 Mice
We collected longitudinal plasma samples from sham and TBI mice starting from baseline (2 weeks before TBI or sham operation), at 6 h post-TBI, and every 2 weeks thereafter. Plasma total tau (t-tau) was assayed using a human t-tau SIMOA assay. Despite loss of some samples due to machine error, we observed the expected time-dependent increase in t-tau from 2 to 8 weeks post-operation ( Figure 3A). We then proceeded to analyze sham vs. TBI differences using four-parameter logistic regression ( Figure 3B). The best fitted models of the sham and TBI groups did not differ greatly in terms of the upper plateau (10 2.567 = 369 pg/mL vs. 10 2.733 = 541 pg/mL). Intriguingly, we noted that the sham group exhibited a point of inflection at 18 days post-sham procedure (10 1.506 = 32 days postbaseline), whereas the TBI group had a point of inflection within 1 day post-TBI (10 1.170 = 14.8 days post-baseline). This finding suggests that interfaced TBI at 4.0 J may accelerate the increase in plasma t-tau in rTg4510 mice by approximately 2 weeks, although it may not affect the final level of plasma tau at 2-mo post-TBI.

TBI Accelerated Increase in Plasma Total Tau in rTg4510 Mice
We collected longitudinal plasma samples from sham and TBI mice starting from baseline (2 weeks before TBI or sham operation), at 6 h post-TBI, and every 2 weeks thereafter. Plasma total tau (t-tau) was assayed using a human t-tau SIMOA assay. Despite loss of some samples due to machine error, we observed the expected time-dependent increase in t-tau from 2 to 8 weeks post-operation ( Figure 3A). We then proceeded to analyze sham vs. TBI differences using four-parameter logistic regression ( Figure 3B). The best fitted models of the sham and TBI groups did not differ greatly in terms of the upper plateau (10 2.567 = 369 pg/mL vs. 10 2.733 = 541 pg/mL). Intriguingly, we noted that the sham group exhibited a point of inflection at 18 days post-sham procedure (10 1.506 = 32 days post-baseline), whereas the TBI group had a point of inflection within 1 day post-TBI (10 1.170 = 14.8 days post-baseline). This finding suggests that interfaced TBI at 4.0 J may accelerate the increase in plasma t-tau in rTg4510 mice by approximately 2 weeks, although it may not affect the final level of plasma tau at 2-mo post-TBI.

TBI Did Not Change Brain Tau Burden in rTg4510 Mice at 2-mo Post-Injury
Since we observed increased brain GSK-3β activity and a possible acceleration of the elevation of t-tau in plasma, we next sought to evaluate whether TBI induced tau burden in rTg4510 mice, using immunohistochemistry, histochemical staining, and Western blotting ( Figure 4A-F). Overall, TBI did not induce significant changes in the level of total tau (DA9, Figure 4A) or neurofibrillary tangles (NFT, by Gallyas Silver stain, Figure 4B). We next performed IHC using antibodies that recognize different phosphorylated tau epitopes: PHF1 (pS396+pS404), AT8 (pS202+pT205), CP13 (pS202), and MC1 (pathological conformation) ( Figure 4C-F, Supplementary Figure S5). When we compared sham vs. all TBI animals, we did not observe a significant increase in any p-tau epitope in any brain region examined (dorsal hippocampus, ventral hippocampus, entorhinal cortex, frontal cortex, amygdala, optic tract, hypothalamus). These findings suggest that, despite increased GSK-3β activity in the brain and accelerated elevation of total tau in plasma, the tau burden in rTg4510 was not increased at 2-mo post-injury.

TBI Did Not Change Brain Tau Burden in rTg4510 Mice at 2-mo Post-Injury
Since we observed increased brain GSK-3β activity and a possible acceleration of the elevation of t-tau in plasma, we next sought to evaluate whether TBI induced tau burden in rTg4510 mice, using immunohistochemistry, histochemical staining, and Western blotting ( Figure 4A-F). Overall, TBI did not induce significant changes in the level of total tau (DA9, Figure 4A) or neurofibrillary tangles (NFT, by Gallyas Silver stain, Figure 4B). We next performed IHC using antibodies that recognize different phosphorylated tau epitopes: PHF1 (pS396+pS404), AT8 (pS202+pT205), CP13 (pS202), and MC1 (pathological conformation) ( Figure 4C-F, Supplementary Figure S5). When we compared sham vs. all TBI animals, we did not observe a significant increase in any p-tau epitope in any brain region examined (dorsal hippocampus, ventral hippocampus, entorhinal cortex, frontal cortex, amygdala, optic tract, hypothalamus). These findings suggest that, despite increased GSK-3β activity in the brain and accelerated elevation of total tau in plasma, the tau burden in rTg4510 was not increased at 2-mo post-injury.

TBI Did Not Change Levels of Autophagosomes and Lysosomes, Neurons, Astrocytes, Endothelial Cells, Synapses, or Size of Brain Regions
We then investigated whether the changes in protein degradation pathways may explain the lack of significant change in tau burden in the brain. We probed for autophagosome adaptor protein (p62), autophagosome initiation proteins (LC3B-I and LC3B-II), and lysosomal proteases (cathepsin B and cathepsin D) using IHC or Western blotting of RIPA lysates. No significant differences between sham and TBI mice were observed (Supplementary Figure S3A-D).
We also performed IHC analysis using NeuN, GFAP, CD31, and IgG antibodies to stain for neurons, astrocytes, endothelial cells, and extravasated IgG, respectively. We observed no significant change when comparing sham vs. TBI brains across any region analyzed ( Figure 5A-C). Western blotting analysis also showed no differences in synaptic markers (synaptophysin and PSD95) between sham vs. TBI mice ( Figure 5D).      We also analyzed the size of various brain regions and ventricles in sham vs. TBI animals ( Figure 5E). The sizes of the dorsal hippocampus and amygdala were measured on the coronal plane (approximately −2 mm posterior to the bregma), and the area of the lateral ventricle was measured on the coronal plane of the frontal cortex (approximately 1.2 mm anterior to the bregma). There was no significant difference in the size of these regions in sham vs. TBI animals, suggesting no effect of TBI on neurodegeneration.

Discussion
This study was designed to investigate the outcomes of a single high-energy head impact induced by interfaced CHIMERA in a mouse model of tauopathy. At an impact energy of 4.0 J, designed to mimic a moderate-severe TBI, the overall mortality rate in TBI mice was 46.7%. This mortality rate is vastly different from previous CHIMERA studies (0 mortality at 0.1-J to 0.7-J impacts without interface [21,[24][25][26][27][28]; 9-20% morality at 2.5-J impacts with interface [22,23]). The mortality rate in the current study is well aligned with clinical observations of msTBI, in which mortality rate can be as high as 43-46% [29][30][31]. In addition, the high-energy head impacts in this study induced a much longer duration of LRR in surviving mice (median >30 min in this study, compared to~5 min in non-interface 0.1-to 0.7-J CHIMERA TBI [21,[24][25][26][27][28] and~15 min in 2.5-J interfaced TBI [22,23]).
The most important finding in this study is that TBI chronically activates GSK-3β, which is one of the most well-studied tau kinases, and it phosphorylates tau at multiple sites [32][33][34]. Its activation has been found in NFTs of AD patients [34], and its inhibition is an active area of therapeutic research for neurodegenerative diseases [35]. In rodent models of TBI [36][37][38][39], increased GSK-3β phosphorylation at serine 9 (which inactivates GSK-3β activity [40]) has been observed during the acute post-injury phase and may be important to neuronal survival in the early period. However, the role of GSK-3β in the chronic post-injury phase has not been reported. Our study suggests that TBI chronically disrupts AKT-GSK-3β signaling, leading to increased GSK-3β activity. This finding highlights the potential benefits of inhibiting or competing against GSK-3β activity after TBI in chronic management.
At 2 mo post-injury, we observed no significant change in tau burden. Although we confirmed that tau expression remained unchanged (Supplementary Figure S2E-G), we noted that the TBI group showed increased variability across several measures. Specifically, the coefficient of variation (CoV) of tau and p-tau across multiple brain areas in sham mice is~50%, but it was 90% in the TBI group. Similar observations were found in ventricle sizes. Therefore, it is possible that variations in post-injury tau clearance efficiency may have masked the effects of tau kinase activity. Using Iba1 as microglia marker, we observed no significant differences in the cortical microglial response. Future studies could use phagocytic markers, such as CD68 and Trem2 [41,42], to determine whether microglia showed potential phagocytic differences. In addition, we used p62, LC3B, cathepsin B, and cathepsin D to investigate autophagolysosomes, a major cellular degradation pathway of tau [43][44][45], and we did not observe any significant differences between sham and TBI animals (Supplementary Figure S3). Notably, p62 and CTSD levels were significantly and positively correlated with cortical tau levels, with a subset of TBI samples with very high p62 levels (Supplementary Figure S4). This observation is aligned with previous reports suggesting that TBI may impair autophagy flux [46]. Future studies could investigate whether variations in post-injury autophagy responses contribute to tau clearance.
To our knowledge, we are the first to report longitudinal plasma total tau changes in a mouse model of tauopathy, and we report that the temporal elevation of plasma tau can be accelerated by TBI. However, we did not observe a significant change in brain tau burden at 2 mo post injury, which could be due to the following limitations. First, for brain tauopathy, we had only one time point of observation. Second, our study had a small N and high mortality, which could have induced potential survivor bias in the TBI group. Third, it is unclear how plasma total tau reflects brain tau levels. Nevertheless, the observation that TBI appears to accelerate a shift to high plasma tau levels is potentially interesting, and future studies could incorporate additional blood biomarkers.
In conclusion, a single high impact-energy TBI delivered to rTg4510 mice using the CHIMERA platform resulted in extended loss of the righting reflex, chronic white matter injury, increased GSK-3β activity, and potential acceleration of the elevation of plasma total tau levels.

Animals
All experiments were approved by the University of British Columbia Animal Care Committee and were compliant with the Canadian Council of Animal Care (A15-0096). Male rTg4510 mice (Jackson Laboratory #024854) were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). These mice express a tetracycline-regulatable tetO-MAPT*P301L transgene under the control of the murine prion protein (PrP) promoter, leading to overexpression of human four-repeat (4R0N) mutant P301L tau. These mice also harbor the CaMK2a-tTA transgene, which suppresses tau expression upon exposure to tetracycline or its analogs. Tetracycline was not used in this study. The mice were housed with environmental enrichment on a 12-h/12-h reversed light cycle and received the 2918 Teklad Global 18% Protein rodent diet (Inotiv, Madison, WI, USA) and autoclaved reverse osmosis water ad libitum.

Traumatic Brain Injury
At 4 mo (133.4 ± 0.5 days) of age, male rTg4510 mice received a single interfaced TBI using the CHIMERA device, as previously described [22]. Immediately prior to TBI, the mice received 0.5 mL of NaCl for fluid supplementation and 1 mg/kg of meloxicam for analgesia. Anesthesia was induced using 5% isoflurane at 2.5 L/min oxygen and thereafter maintained at 3-4%. Anesthetized mice were restrained by abdominal straps in the supine position on the CHIMERA device such that their heads were free to move and rested at an angle of approximately 145 • relative to the body. A polylactic acid (PLA)-silicone interface fitted to the contour of the mouse skull was placed under the animal's head to protect it from skull fracture and distribute impact energy evenly across the skull. Head acceleration and rotational motion were on the sagittal plane. For sham controls, mice received fluid supplementation, analgesia, anesthesia, and positioning in the CHIMERA device but no impact. During the TBI procedures, the duration of isoflurane exposure and loss of consciousness were recorded. Impact energies ranged from 2.3 J to 5.8 J in a pilot experiment designed to define the maximum tolerable impact energy, and an impact energy of 4.0 J was used for the remainder of this study. Chest compressions with oxygen supplementation were performed on all mice that experienced cardiac/respiratory arrest immediately after the TBI procedure. Most of the animals that did not survive the procedure died within seconds to minutes post-injury. A summary of mortality and causes of death is provided in Supplementary Figure S1A-C.

Blood Collection and Euthanasia
Longitudinal blood (EDTA-plasma) samples were collected from the saphenous vein 1 week before TBI (defined as baseline), at 6 h after TBI, and every 2 weeks thereafter, using capillary collection tubes (ThermoFisher, Waltham, MA, USA). At 2 months post-TBI (6 mo of age), animals were euthanized with 150 mg/kg ketamine (Zoetis, Parsippanny, NJ, USA) and 20 mg/kg xylazine (Bayer, Creve Coeur, MO, USA). Cardiac puncture was performed to collect terminal blood samples. The mice were then perfused with 50 mL of ice-cold heparinized PBS (5 USP unit/mL). The mouse brain was dissected longitudinally, and 1 hemibrain was frozen for protein homogenization, while the other half was fixed in 4% paraformaldehyde (PFA) for histology. All blood samples were centrifuged at 1000× g for 10 min and the supernatant was stored at −80 • C as EDTA-plasma.

Plasma Total Tau Analysis
Plasma samples from 17 mice were analyzed with the Quanterix ® Simoa HD-1 ® analyzer. Samples were analyzed with the Simoa Tau Advantage Kit (101552) using the manufacturer's protocol. Plasma samples were diluted off board at a 50-fold dilution using the sample diluent provided. Longitudinal samples were collected from all mice over 6 time points (baseline, 6 h, 2 wk, 4 wk, 6 wk, 8 wk). However, some samples were lost due to machine error. All remaining samples (n = 64) were randomized and analyzed on a single plate using the provided 8-point calibrator curve and 2 controls. The curve had an average percentage error of 11% and an average recovery of 100%. Both controls were within acceptable ranges as specified by the manufacturer. n = 52 samples were run in duplicate, with an average CV of 11%. n = 11 samples from 8 mice were run singly due to volume constraints. n = 7 samples were greater than the upper limit of quantification but within the limits of detection. No samples were less than the lower limits of quantification. reagents (Roche 06402712001) on a LightCycler ® 96 system (Roche, Basel, Switzerland). Each sample was assayed in duplicate and normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The following primer sequences were used: mouse Gapdh forward 5 -AAG GTC ATC CCA GAG CTG AA-3 , reverse 5 -CTG CTT CAC CAC CTT CTT GA-3 ; mouse Actb (Actin) forward 5 -ACG GCC AGG TCA TCA CTA TTG-3 , reverse 5 -CAA GAA GGA AGG CTG GAA AAG-3 ; human MAPT (Tau) forward 5 -CCC AAT CAC TGC CTA TAC CC-3 , reverse 5 -CCA CGA GAA TGC GAA GGA-3 ; and tTa forward 5 -GGA CGA GCT CCA CTT AGA CG-3 , reverse 5 -CAA CAT GTC CAG ATC GAA ATC-3 .

Statistics
In this study, most analyses (loss of righting reflex, histology, immunohistochemistry, and Western blot) were comparisons between 2 groups (sham and TBI). Data were analyzed by Student's t-test (if normally distributed) or the Mann-Whitney U test (if not normally distributed). The statistical method used for each analysis is stated in the figure legend. Longitudinal measurements of plasma total tau were analyzed by log tau level vs. log time, using a 4-parameter logistic curve model. The effect of impact energy on mortality was modeled using a logistic regression model. Correlational analyses of tau vs. autophagolysosomal markers were performed using Pearson's correlation.