Neurobehavioral, neuropathological and biochemical profiles in a novel mouse model of co-morbid post-traumatic stress disorder and mild traumatic brain injury

Abstract

Co-morbid mild traumatic brain injury (mTBI) and post-traumatic stress disorder (PTSD) has become the signature disorder for returning combat veterans. The clinical heterogeneity and overlapping symptomatology of mTBI and PTSD underscore the need to develop a preclinical model that will enable the characterization of unique and overlapping features and allow discrimination between both disorders. This study details the development and implementation of a novel experimental paradigm for PTSD and combined PTSD-mTBI. The PTSD paradigm involved exposure to a danger-related predator odor under repeated restraint over a 21 day period and a physical trauma (inescapable footshock). We administered this paradigm alone, or in combination with a previously established mTBI model. We report outcomes of behavioral, pathological and biochemical profiles at an acute timepoint. PTSD animals demonstrated recall of traumatic memories, anxiety and an impaired social behavior. In both mTBI and combination groups there was a pattern of disinhibitory like behavior. mTBI abrogated both contextual fear and impairments in social behavior seen in PTSD animals. No major impairment in spatial memory was observed in any group. Examination of neuroendocrine and neuroimmune responses in plasma revealed a trend toward increase in corticosterone in PTSD and combination groups, and an apparent increase in Th1 and Th17 proinflammatory cytokine(s) in the PTSD only and mTBI only groups respectively. In the brain there were no gross neuropathological changes in any groups. We observed that mTBI on a background of repeated trauma exposure resulted in an augmentation of axonal injury and inflammatory markers, neurofilament L and ICAM-1 respectively. Our observations thus far suggest that this novel stress-trauma-related paradigm may be a useful model for investigating further the overlapping and distinct spatio-temporal and behavioral/biochemical relationship between mTBI and PTSD experienced by combat veterans.

Keywords: anxiety and social behavior; cognitive function; mild traumatic brain injury; mouse models; plasma and brain biomarkers; post-traumatic stress disorder.

Figure 1

Timeline and procedures for the experimental mTBI and PTSD paradigm. Mice were pre-trained to obtain their baseline fall latency values on the rotarod test for motor co-ordination for 3 consecutive days, and this was followed by a 7-day acquisition training in the radial arm water maze (RAWM) test for spatial learning and memory. The 21 day experimental paradigm involved an exposure to TMT and restraint on 11 randomly assigned days (blue) at different times of the light-dark cycle. Around the midpoint, on day 12, mice underwent fear conditioning, which involved being placed in a conditioning chamber for 3 min, exposure to a 70 db auditory cue for the last 30 s, culminating with 2 s of a 1 mA foot-shock. Animals in the mTBI group were exposed to a single concussive head injury 1 h after the foot-shock, while under anesthesia. On day 21 animals were tested for their contextual and cued fear memory response. This was followed by a battery of behavioral tests for motor activity/coordination (rotarod), anxiety (elevated plus maze, open field test), social behavior and spatial learning and memory (RAWM). Brain tissue, plasma was collected for further studies.

Figure 2

Contextual, cued fear, and spatial memory. PTSD animal group showed a significant increase in their (%) freezing responses to both the context (A1,A2) and the auditory cue (B1,B2), compared to control animals. mTBI inhibited the retrieval of the contextual fear memory response in the mTBI-PTSD group (A1,A2), there was no effect of mTBI on the cued fear response in the mTBI-PTSD group (B1,B2). No significant effect was seen in the fear memory response from the mTBI only group compared to the control group in both the context and with the auditory cue (P > 0.05; n = 8–13). All animals performed equally in the pre-training session in the RAWM test, reaching the training criteria of one-memory errors by the end of the 7-day acquisition training session (C1). Two-way analysis with repeated measures showed no main effect of exposure (and their interaction with time; P > 0.05) with either PTSD, mTBI or their combinations on spatial learning and memory of a pre-learned task (C1) (n = 8–9). No effect was seen on mean memory error during the last day of the trial (C2). Data are presented as mean ± SEM. X-axis in (A2,B2) represent 30 s and 0.5 min (time) epochs over the entire length of the trial. Data in (A1,C2) were analyzed using One-Way ANOVA. Data in (B1) were analyzed using a regular Two-Way ANOVA. Data in (A2,B2,C1) were analyzed using a repeated measures-Two Way ANOVA. Tukey's multiple comparisons post-hoc test was performed in all cases. Asterisks denote statistical significance as follows: ^**P < 0.01; ^***P < 0.001.

Figure 3

Elevated plus maze, open-field, and social behavior. Elevated plus maze; PTSD only group spent 50% less time on average in the open arms compared to control, however this trend was not significant (P > 0.05) following multiple comparisons test (A). Animals that received mTBI spent relatively more time and number of entries into the open arm compared to control and PTSD only groups (A,B). This comparison was shown to be statistically significant between the PTSD and mTBI-PTSD groups (P < 0.05). Data in (A,B) were analyzed using non-parametric Kruskal Wallis-One Way ANOVA followed by Dunn's correction for multiple testing. Open field test demonstrated a statistically significant reduction in mean center zone entries over the 15 min trial (P < 0.01) in both mTBI and PTSD animals compared to control (C). However there was a noticeable trend in the behavior of mTBI-PTSD animals, compared to PTSD only and mTBI only animals. The latter showed an increased mean number of entries into the center zone, although this was not statistically significant (P > 0.05) compared to the other groups (C). Only the PTSD group showed a significant reduction in the mean total time spent in the center zone over the 15-min trial compared to control (P < 0.05). These data were evident despite the fact that all groups demonstrated a comparable motor activity in the open field test (see Figure 4A). Data in (C) were analyzed using repeated measure-Two Way ANOVA with Tukey's post-hoc test, and data in (D) by One-Way ANOVA followed by Tukey's post-hoc test. Social behavior was measured in the social interaction/recognition test (E,F). mTBI and PTSD only groups showed no significant distinction between the time spent in the “empty” and “stranger I” chamber (E). Control and mTBI-PTSD animals performed similarly in the social interaction test (E). In the social recognition test PTSD only animal group also did not show any distinction between “stranger I” and “stanger II” chambers, indicating a dysfunction in social memory and preference (F). mTBI only and mTBI-PTSD animal groups performed similarly as the control group (F). Data are presented as mean ± SEM. Data in (E,F) were analyzed using non-parametric Mann Whitney U-test for each group. Asterisks denote statistical significance as follows: ns; not significant; ^*P < 0.05; ^**P < 0.01; ^***P < 0.001. N = 11–13 in all behavioral tests, except in the social interaction test (n = 6–8).

Figure 4

Motor activity/coordination and growth rate. No main effect was observed in the motor activity during the open field test, and motor coordination in the rotarod test in any of the groups (A,B). Data in (A,B) were analyzed using repeated—Two Way ANOVA followed by post-hoc Tukey's test (n = 11–13). Mice in the control, mTBI, and combination groups showed a 2–5% increase in body weight over the 21 day stress experimental paradigm, while PTSD only mice showed a 5% reduction in their body weight (C). Data are presented as mean ± SEM.

Figure 5

Neuroendocrine response to mTBI and PTSD. Plasma baseline levels of corticosterone showed a trend toward increase in PTSD and mTBI-PTSD only animals (> +70 ng/ml) compared to control. This trend did not reach statistical significance after analyses with Krukal Wallis-One Way ANOVA followed by Dunn's correction for multiple testing. mTBI animals showed a similar baseline level of corticosterone as controls. Data are presented as mean ± SEM. (N = 4–5).

Figure 6

Immune/inflammatory systemic response at acute time-point post-mTBI-PTSD. Levels of 6 different cytokines (IL-1β, IL-10, TNFα, IL-6, IL-17A, IFNγ) were measured in the plasma to detect systemic immune/inflammatory responses to mTBI and/or PTSD. There was an obvious trend toward increase in IL-1B, IL-10, TNFα, and IL-6 in PTSD only animals compared to control and mTBI groups, however this was only statistically significant with TNFα (P = 0.036) and a marginal trend with IL-1β (P = 0.06) (A–D). Intriguingly we observed a rise (P = 0.049) in IL-17A in mTBI only animals compared to controls (E). No significant effect was seen in the levels of IFNγ amongst the different groups (F). Data are presented as mean ± SEM. Data in (A,B,D–F) were analyzed with Kruskal Wallis-One-Way ANOVA followed multiple testing. Asterisk denote statistical significance as follows: ^*P < 0.05. (N = 4–5).

Figure 7

Global brain-related changes observed at acute post-mTBI-PTSD time-point. Hallmark potential biomarkers of mTBI (in humans) were analyzed in brain hemispheres. Western blot analysis showed a significant increase in axonal marker, neurofilament L (NFL) and inflammatory marker CD54 (ICAM)-1 levels in mTBI-PTSD groups compared to control (P = 0.022 and P = 0.016 respectively—A,C). No statistically significant changes were observed in astrocyte cytoskeletal protein—GFAP and phospho-tau protein (detected by s202 antibody—CP13) in all groups analyzed (B,D). Data are presented as mean ± SEM. Arbitrary units were calculated by normalizing with beta-actin levels [beta-actin blot depicted in (A), was used to calculate GFAP and ICAM-1 arbitrary values]. Data were analyzed with Kruskal Wallis-One-Way ANOVA followed by Dunn's correction for multiple testing. Asterisk denotes statistical significance as follows: ^*P < 0.05. (N = 4–5).

Figure 8

Doublecortin (DCX) cell counts in the dentate gyrus (DG). Numbers of DCX+ cells were counted in the subgranular zone (SGZ) of the DG as depicted in (A). No significant change was observed in the number of DCX+ cells in the subgranular zone (SGZ) of DG. Data are presented as mean ± SEM. Data were analyzed using Kruskal Wallis-One Way ANOVA followed by Dunn's correction for multiple testing. (N = 4–5). Scale bar represents 95 μm.

Figure 9

Neuroglial response in the corpus callosum (CC), hippocampus (CA1), basolateral amygdala (BLA) and hypothalamus (HypoT), Parietal cortex- Injury site (P-Cortex-IS). (A–D) depicts microglia IBA+ immunoreactivity in the corpus callosum (CC). There was a significant increase in mTBI animals compared to controls (P = 0.010); this was confirmed by quantitative densitometry (optical segmentation) (V). (E–T) shows astrocyte GFAP+ immunoreactivity in the aforementioned brain regions. There was a significant increase in astroglial activation in the CC (E–H), hippocampus (I–L), and P-Cortex-IS of mTBI animals compared to controls (P = 0.04; P = 0.022, and P = 0.034 respectively—U). Likewise there was also a significant increase in astroglial activation in the CC (E–H), and a marginal trend in the hippocampus (I–L) of mTBI-PTSD animals (P = 0.009; P = 0.09 respectively—U). No significant change was observed in the P-Cortex-IS of mTBI-PTSD animals compared to controls (U). Notably PTSD and mTBI-PTSD animals showed an increase in GFAP immunoreactivity in the BLA (M–P) and HypoT by qualitative assessment (Q–T). This was partly confirmed by densitometric analysis showing a significant increase in the BLA in mTBI-PTSD animals compared to controls (P = 0.006), and a marginal trend in PTSD mice compared to controls (P = 0.071). Data are presented as mean ± SEM. Data were analyzed using One-Way ANOVA followed by multiple comparisons post-hoc test for each brain region. A 1 in 10 series of >five average sections were analyzed per region for each cellular marker per animal. Asterisks denote statistical significance as follows: ^*P < 0.05; ^**P < 0.01. (N = 4–5). Images show segmented profiles of representative positive immunostaining. Scale bar represents 95 μm.

Figure 10

Distinct periventricular and perivascular astroglia response to mTBI. GFAP immunostaining revealed a layer of GFAP+ activated cells in the most superficial layer of the cortex (See white arrows in representative image—A), beneath the injury site of head injured animals (in both mTBI and mTBI-PTSD). There was some evidence of multifocal perivascular fibrillary astrogliosis (B). In some cases these perivascular astroglial cells appeared hypertrophic with a large cell soma and thick cellular processes (B–D). A distinct remarkable astrogliosis was observed in the walls of the ventricles (gray arrows—E) and the neighboring adjacent brain parenchyma (E,F). Abbreviation; blood vessel (B.V). Scale bar represents 120 μm in all images.

Figure 11

Distinct white matter pathology and microglia response to mTBI. There was an apparent reduction in the immunoreactivity of extracellular matrix protein—Laminin in the corpus callosum (CC) of mTBI and mTBI-PTSD animals (A–D). Microglia response to injury is shown with IBA-1 and CD45 antibodies. IBA-1 staining depicted an increase in microglial activation in white matter/neighboring regions of injured animals (E–J). These regions included the ventral lateral geniculate nucleus (VLG) in (E), the optic tract (Op.T) projections surrounding the dorsal lateral geniculate nucleus (DLG), lateral post-thalamic nuclei (LPTN), and the internal capsule (I.C) in (F–H) respectively. CD45+ cell (a marker of activated microglia/macrophages) was also positive in these white matter brain regions (1, J). Notably ICAM-1/CD54+ cells were also observed in white matter regions of mTBI-injured animals (both mTBI and combination), they resembled oligodendroglia-like (and in some cases microglia-like cells—inset K) with a small round cell body and short sparse processes (see K,L). Scale bar in (L) represents 95 μm in (A–D); 55 μm in (E–G); 65 μm in (H) and 45 μm in (I–L).