Repeated mild traumatic brain injury triggers pathology in asymptomatic C9ORF72 transgenic mice

Abstract

Frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS) are fatal neurodegenerative diseases that represent ends of the spectrum of a single disease. The most common genetic cause of FTD and ALS is a hexanucleotide repeat expansion in the C9orf72 gene. Although epidemiological data suggest that traumatic brain injury (TBI) represents a risk factor for FTD and ALS, its role in exacerbating disease onset and course remains unclear. To explore the interplay between traumatic brain injury and genetic risk in the induction of FTD/ALS pathology we combined a mild repetitive traumatic brain injury paradigm with an established bacterial artificial chromosome transgenic C9orf72 (C9BAC) mouse model without an overt motor phenotype or neurodegeneration. We assessed 8-10 week-old littermate C9BACtg/tg (n = 21), C9BACtg/- (n = 20) and non-transgenic (n = 21) mice of both sexes for the presence of behavioural deficits and cerebral histopathology at 12 months after repetitive TBI. Repetitive TBI did not affect body weight gain, general neurological deficit severity, nor survival over the 12-month observation period and there was no difference in rotarod performance, object recognition, social interaction and acoustic characteristics of ultrasonic vocalizations of C9BAC mice subjected to repetitive TBI versus sham injury. However, we found that repetitive TBI increased the time to the return of the righting reflex, reduced grip force, altered sociability behaviours and attenuated ultrasonic call emissions during social interactions in C9BAC mice. Strikingly, we found that repetitive TBI caused widespread microglial activation and reduced neuronal density that was associated with loss of histological markers of axonal and synaptic integrity as well as profound neuronal transactive response DNA binding protein 43 kDa mislocalization in the cerebral cortex of C9BAC mice at 12 months; this was not observed in non-transgenic repetitive TBI and C9BAC sham mice. Our data indicate that repetitive TBI can be an environmental risk factor that is sufficient to trigger FTD/ALS-associated neuropathology and behavioural deficits, but not paralysis, in mice carrying a C9orf72 hexanucleotide repeat expansion.

Keywords: C9orf72; amyotrophic lateral sclerosis; frontotemporal dementia; neurodegeneration; transactive response DNA binding protein 43 kDa (TDP-43).

Figure 1

Repetitive TBI causes mild motor impairment in C9BAC mice. (A) Significant decline in rotarod performance in all studied groups (P = 0.689 for group effect, P < 0.001 for time effect, P = 0.976 for Group × Time interaction). Table indicates significant differences to baseline (ref.). (B) Significant early decline in forelimb grip force after repetitive TBI (rTBI) in C9BAC^tg/− and C9BAC^tg/tg mice (P < 0.001 for group effect, P < 0.001 for time effect, P = 0.967 for Group × Time interaction). (C) Absent increase in the all-limb grip force in C9BAC^tg/− and C9BAC^tg/tg mice after rTBI (P < 0.001 for group effect, P < 0.001 for time effect, P = 0.023 for Group × Time interaction). For clarity in B and C, post hoc pairwise comparisons are only shown for significant differences to Ntg rTBI animals. Data are mean ± SEM. Data of sham and rTBI groups are slightly offset for better visibility. All statistical comparisons were made using mixed effects models. **P < 0.01, ***P < 0.001 (post hoc pairwise comparisons FDR adjusted).

Figure 2

Novel object recognition and social interaction of mice at 12 months after surgery. (A) During the familiarization phase of the novel object recognition test, groups spent an equal percentage of time exploring both objects (dashed line indicates chance performance). (B) During the test phase, both sham operated and rTBI non-transgenic (Ntg) mice spent significantly more time with the novel object. On the contrary, C9BAC hemizygous and homozygous sham and rTBI mice did not discriminate between familiar and novel objects. Between group comparisons were made by one way ANOVA. Data are mean ± SEM. *P < 0.05 versus Ntg sham, ^#  P < 0.05 versus Ntg rTBI. (C) Compared to Ntg sham controls, C9BAC^tg/tg mice displayed significantly increased self-grooming behaviour. (D) Increased rearing behaviour in C9BAC^tg/− and C9BAC^tg/tg mice after rTBI as compared to Ntg sham controls. (E) After rTBI C9BAC^tg/tg mice showed significantly increased number of faecal droppings when compared to all other groups. Data in A–E were compared by one-way ANOVA with post hoc Holm-Šídák test and presented as mean ± SEM. (F) Significantly fewer C9BAC^tg/− and C9BAC^tg/tg mice emitted USV at 52 weeks after rTBI compared to non-transgenic (Ntg) sham controls (P = 0.019, Fisher’s exact test with post hoc Bonferroni adjustment). Hatched bars indicate call presence, open bars indicate call absence. For clarity, only significant results are indicated in the figure.

Figure 3

Repetitive TBI causes extensive neuronal loss at 12 months after injury in C9BAC mice. (A) Schematic of the mouse skull denoting the approximate location of the impact centre over the intact mouse skull (blue circle) as well as (B) relative to the brain section sampled for histological analysis shown in this figure (dashed line). (C) Loss of NeuN stained neurons in the cerebral cortex of C9BAC^tg/− and C9BAC^tg/tg mice after rTBI (images were taken from an area corresponding to the striped square in C). Scale bar = 120 µm. (C) Cartoon depicting the extent of neuronal loss within the conceptual framework that the combined risk of rTBI (environmental ‘hit’) and C9orf72 gene mutation (genetic ‘hit’) cause neuronal loss [− risk absent (Ntg sham mice), + or ++ risk present (C9BAC^tg/− or C9BAC^tg/tg mice with or without rTBI)]. Data are expressed as % reduction [range 0% (white) to 40% (red)] relative to the corresponding region of interest in Ntg sham mice (lightning bolt denotes the site of impact delivery). Squares indicate the approximate ROI used for quantitative analyses shown in D. (D) Significant reduction in the number of NeuN stained neurons at 12 months after rTBI in C9BAC^tg/− and C9BAC^tg/tg mice, with overall greater neuronal loss in the ipsilateral versus contralateral hemisphere (P < 0.001 for group effect, P = 0.014 for left versus right ROI, P = 0.005 for Group × ROI interaction). For each genotype, data on the right are from the ipsilateral cortex. Data in the bar graph are shown as mean ± SEM. n = 9–12 per group. *P < 0.05, ***P < 0.001 for between group comparisons; ^#  P < 0.05, ^###  P < 0.001 for between hemisphere comparisons (post hoc pairwise comparisons FDR adjusted).

Figure 4

Repetitive TBI accentuates axonal and synaptic loss at 12 months in C9BAC mice. (A) Representative photomicrographs showing SMI-312, PSD-95 and synaptophysin staining signal from the ipsilateral cerebral cortex (square box in inset). Scale bars = 100 µm. Reduced cortical axonal density in C9BAC mice that is accentuated by rTBI as indicated by (B) attenuated SMI-312 staining signal (P < 0.001 for group effect, P < 0.001 for left versus right ROI, P = 0.010 for Group × ROI interaction) as well as (C) reduced number of SMI-312 stained axons profiles (P < 0.001 for group effect, P = 0.003 for left versus right ROI, P = 0.345 for Group × ROI interaction). RTBI accentuates synaptic loss in C9BAC mice as indicated by a reduction in the (D) postsynaptic marker PSD-95 (P < 0.001 for group effect, P = 0.008 for left versus right ROI, P = 0.055 for Group × ROI interaction) and (E) presynaptic marker synaptophysin (P < 0.001 for group effect, P = 0.004 for left versus right ROI, P = 0.940 for Group × ROI interaction). Bar-pairs denote left (contralateral) and right (hatched, ipsilateral) ROI. Data are mean ± SEM. n = 9–12 per group. *P < 0.05, **P < 0.01, ***P < 0.001 for between group comparisons; ^##  P < 0.01, ^###  P < 0.001 for between hemisphere comparisons (all post hoc pairwise comparisons FDR adjusted).

Figure 5

Repetitive TBI causes neuronal TDP-43 mislocalization and accumulation at 12 months in C9BAC mice. (A) Representative photomicrographs showing TDP-43 staining signal from the ipsilateral cerebral cortex. Nuclear localization of TDP-43 in C9BAC sham and non-transgenic (Ntg) rTBI mice (long arrow) at 12 months after rTBI. Nuclear loss and inclusions of cytoplasmatic localized neuronal TDP-43 in C9BAC^tg/− and C9BAC^tg/tg mice (short arrows). Scale bars = 70 µm. (B) Quantified TDP-43 signal from ROIs corresponding to those used for NeuN staining shows a significant reduction in signal intensity of TDP-43 in both the ipsilateral injured (hatched bars) and contralateral non-injured cortex underlying the impact centre at 12 months after rTBI in C9BAC^tg/− and C9BAC^tg/tg mice (P < 0.001 for group effect, P = 0.659 for left versus right ROI, P = 0.976 for Group × ROI interaction). (C) Similar percentage of all NeuN stained neurons expressing TDP-43 between the experimental groups (P = 0.606 for group effect, P = 0.208 for left versus right ROI, P = 0.532 for Group × ROI interaction). (D) Significantly increased proportion of neurons exhibiting TDP-43 mislocalization in C9BAC^tg/− and C9BAC^tg/tg mice that is not observed in Ntg rTBI and sham controls (P < 0.001 for group effect, P = 0.816 for left versus right ROI, P = 0.691 for Group × ROI interaction). Data are mean ± SEM. n = 9–12 per group. *P < 0.05, **P < 0.01, ***P < 0.001 for between group comparisons (post hoc pairwise comparisons FDR adjusted).

Figure 6

Persistent microglial activation in C9BAC mice at 12 months after repetitive TBI. (A) Representative photomicrographs of Iba-1 stained microglia with corresponding outline used for Sholl analysis. (B) At 12 months after rTBI, C9BAC mice exhibited a significant increase in the number of Iba-1 stained cells in both the ipsilateral injured (hatched bars) and contralateral non-injured cortex underlying the impact centre (P < 0.001 for group effect, P = 0.471 for left versus right ROI, P = 0.870 for Group × ROI interaction) without (C) corresponding astroglial activation (P < 0.124 for group effect, P = 0.429 for left versus right ROI, P = 0.404 for Group × ROI interaction). (D) Sholl analysis conducted in the ipsilateral cortex (highlighted by blue square in top diagram) indicated significant enlargement and (E) ramification of microglia of C9BAC mice after rTBI. Scale bar = 10 µm. Data are mean ± SEM. n = 9–12 per group except n = 4 per sham groups and n = 6 per for TBI groups for Sholl analysis. *P < 0.05, **P < 0.01, ***P < 0.001 for between group comparisons in B–D and versus Ntg rTBI in E (post hoc pairwise comparisons were adjusted using Tukey’s test in D and the FDR in B, C and E). For clarity, only significant results are indicated in the figure.