Traumatic Brain Injury in Mice Generates Early-Stage Alzheimer's Disease Related Protein Pathology that Correlates with Neurobehavioral Deficits

Abstract

Traumatic brain injury (TBI) increases the long-term risk of neurodegenerative diseases, including Alzheimer's disease (AD). Here, we demonstrate that protein variant pathology generated in brain tissue of an experimental TBI mouse model is similar to protein variant pathology observed during early stages of AD, and that subacute accumulation of AD associated variants of amyloid beta (Aβ) and tau in the TBI mouse model correlated with behavioral deficits. Male C57BL/6 mice were subjected to midline fluid percussion injury or to sham injury, after which sensorimotor function (rotarod, neurological severity score), cognitive deficit (novel object recognition), and affective deficits (elevated plus maze, forced swim task) were assessed post-injury (DPI). Protein pathology at 7, 14, and 28 DPI was measured in multiple brain regions using an immunostain panel of reagents selectively targeting different neurodegenerative disease-related variants of Aβ, tau, TDP-43, and alpha-synuclein. Overall, TBI resulted in sensorimotor deficits and accumulation of AD-related protein variant pathology near the impact site, both of which returned to sham levels by 14 DPI. Individual mice, however, showed persistent behavioral deficits and/or accumulation of toxic protein variants at 28 DPI. Behavioral outcomes of each mouse were correlated with levels of seven different protein variants in ten brain regions at specific DPI. Out of 21 significant correlations between protein variant levels and behavioral deficits, 18 were with variants of Aβ or tau. Correlations at 28 DPI were all between a single Aβ or tau variant, both of which are strongly associated with human AD cases. These data provide a direct mechanistic link between protein pathology resulting from TBI and the hallmarks of AD.

Keywords: Behavior; Beta-amyloid; Concussion; Immunohistochemistry; Oligomers; Tau.

Fig. 1

Each brain image was divided into 7–10 generalized regions of interest (ROI), depending on which regions were present in each brain slice. Shown: Upper Cortex CTX (blue); corpus callosum CC (orange); hippocampus HC (yellow); caudoputamen CP (green); fornix FX (purple); corticospinal tract CST (magenta); thalamus TH (orange); hypothalamus HYP (red); amygdala/olfactory/cortical subplate AMY (light blue). Striatum STR is located rostrally and is not clearly visible in this slice. Non-measurable areas including edges, wrinkles, overlaps, tears, and/or contaminants were excluded from the ROI selections. Scale bar 500 µm

Fig. 2

Diffuse TBI resulted in behavioral deficits. A Latency to fall from the rotarod measured at 2, 5, and 7 days post-injury (DPI). Mice subjected to 1 TBI or 2 TBIs had significantly shorter latencies to fall at all time points. B Sensorimotor deficits were assessed with a neurological severity score (NSS). Mice subjected to 1 TBI or 2 TBIs had significantly higher scores at all time points compared to uninjured shams. C-D There were no brain injury-induced deficits on the elevated plus maze or the novel object recognition task. E TBI did not result in anxiety-like behavior assessed with the forced swim task (FST). There were no injury-induced changes to the time spent immobile during the FST. However, there were significant time effects and mice spent more time immobile in subsequent trials compared to the first trial. Data are presented as individual data points with the mean and SEM

Fig. 3

Representative coronal mice brain staining detected with D11C scFv-phage grouped by Sham (craniectomy without TBI; left) or TBI (right). Tissue was collected at (A-B) 7, (C-D) 14, or (E–F) or 28 days post-injury (DPI). Phage were detected with HRP conjugated secondary/DAB (brown) and nuclei detected with hematoxylin (purple). Phage staining was detected throughout the brain, though most intense staining was observed near the site of craniectomy and TBI in tissue collected at 7 DPI. The presence of the target protein variant resolved by 14 and 28 DPI. Scale bar 500 µm

Fig. 4

Combined staining intensity values of each of seven scFv phage in 10 different brain tissue regions. Brain tissue images from each individual mouse were converted to 8-bit black and white images and inverted so phage staining correlated to brightness (0–255). Mean grey value intensity (brightness) of each brain region of interest was measured by ImageJ analysis. Values were normalized to the average grey value intensity of the corresponding brain region in sham treated mice to generate the combined values for each of the seven scFv phage. Brain tissue was collected at 7, 14, or 28 days post-injury (DPI). Mice were subjected to either a control sham injury, 1 TBI, or 2 TBIs. Post hoc analysis of composite staining intensity of all seven scFvs showed statistically significant differences in the composite staining intensity at 7 DPI in TBI versus sham mice in the cortex (1 TBI p < 0.0001, 2 TBIs p < 0.0001), the corpus callosum (1 TBI p < 0.0001, 2 TBIs p = 0.002) and hippocampus (1TBI p < 0.0001). These differences resolved by 14 and 28 DPI. The only composite protein variant staining intensity that did not resolve by 14 DPI was in hypothalamus of mice subjected to 2 TBIs (p = 0.037). We also observed unresolved elevated levels of protein variants in other deeper brain regions among individual outlier cases, but these did not reach statistical significance. Asterisks indicate p < 0.05 significance compared to corresponding sham region. Error bars are ± SD

Fig. 5

Staining with the C6T-phage displayed two distinct patterns, (A) a diffuse staining pattern observed with the other reagents, and (B) a punctate staining pattern unique to C6T phage. Brain tissue was incubated with C6T- or F9T-phage, detected with peroxidase conjugated secondary and stained with DAB. Staining of mouse cortex tissue shown here. At higher magnification, the C6T Aβ variant punctate staining pattern is observed in: (C) cytoplasmic staining of neurons in individual TBI mice, and in human post-mortem (D) Braak stage III, and (E) Braak stage VI AD brain tissue. F9T tau variant staining pattern is observed in: (F) cytoplasmic staining of neurons in individual TBI mice, and in human post-mortem (G) Braak stage V AD brain tissue. White Scale bar 200 µm, black scale bar 20 µm

Fig. 6

Negative correlations between time spent immobile during the forced swim task (FST) and F9T tau variant staining level increased with days post-injury (DPI). At 7 and 14 DPI, there were no statistically significant correlations with staining in the (A) corpus callosum (CC), (B) hypothalamus (HYP), or (C) corticospinal tract (CST). At 28 DPI, a significant correlation with behavioral deficits in the FST was observed. Levels of F9T tau variant showed a positive correlation with FST at 7 DPI, but a strong negative correlation by 28 DPI

Fig. 7

Negative correlations between (A) novel object recognition (NOR) task or (B) forced swim test (FST) task and C6T Aβ variant staining increased with days post-injury (DPI). (A) At 7 and 14 DPI, the discrimination index on the NOR task was not significantly correlated with punctate C6T Aβ variant staining but was significantly correlated with behavioral deficits at 28 DPI in the corpus callosum (CC). (B) At 7 and 14 DPI, time spent immobile during the FST was not significantly correlated with punctate C6T Aβ variant staining, but was significantly correlated at 28 DPI in the hypothalamus (HYP). (C) At 7 and 14 DPI, the NOR task was not significantly correlated with diffuse C6T Aβ variant staining, but was significantly correlated at 28 DPI in the cortex (CTX). (D) At 7 and 14 DPI, FST was not significantly correlated with diffuse C6T Aβ variant staining but was significantly correlated at 28 DPI in the HYP