Altered Neuroinflammation and Behavior after Traumatic Brain Injury in a Mouse Model of Alzheimer's Disease

Abstract

Traumatic brain injury (TBI) has acute and chronic sequelae, including an increased risk for the development of Alzheimer's disease (AD). TBI-associated neuroinflammation is characterized by activation of brain-resident microglia and infiltration of monocytes; however, recent studies have implicated beta-amyloid as a major manipulator of the inflammatory response. To examine neuroinflammation after TBI and development of AD-like features, these studies examined the effects of TBI in the presence and absence of beta-amyloid. The R1.40 mouse model of cerebral amyloidosis was used, with a focus on time points well before robust AD pathologies. Unexpectedly, in R1.40 mice, the acute neuroinflammatory response to TBI was strikingly muted, with reduced numbers of CNS myeloid cells acquiring a macrophage phenotype and decreased expression of inflammatory cytokines. At chronic time points, macrophage activation substantially declined in non-Tg TBI mice; however, it was relatively unchanged in R1.40 TBI mice. The persistent inflammatory response coincided with significant tissue loss between 3 and 120 days post-injury in R1.40 TBI mice, which was not observed in non-Tg TBI mice. Surprisingly, inflammatory cytokine expression was enhanced in R1.40 mice compared with non-Tg mice, regardless of injury group. Although R1.40 TBI mice demonstrated task-specific deficits in cognition, overall functional recovery was similar to non-Tg TBI mice. These findings suggest that accumulating beta-amyloid leads to an altered post-injury macrophage response at acute and chronic time points. Together, these studies emphasize the role of post-injury neuroinflammation in regulating long-term sequelae after TBI and also support recent studies implicating beta-amyloid as an immunomodulator.

Keywords: Alzheimer's disease; macrophage; neuroinflammation; traumatic brain injury.

FIG. 1.

Lesion and hippocampal volume change over time. (A) Representative three-dimensional (3D) reconstructions show a cortical injury cavity at 3 days post-injury (DPI) in mice with traumatic brain injury (TBI), which is larger at 120 DPI (n = 6-7/group/time point). (B) Volumetric lesion analysis reveals a time-dependent increase in the size of the injury cavity in the R1.40 mice at 120 DPI compared with 3 DPI. (C) Representative 3D reconstructions show the ipsilateral hippocampus in brain and sham injured non-Tg and R1.40 mice at 3 and 120 DPI. (D) Volumetric hippocampal analysis reveals a time-dependent decrease in all groups, which reached statistical significance in R1.40 TBI and sham mice.

FIG. 2.

Traumatic brain injury (TBI) induces morphometric changes in Iba1+ macropahges. (A,C) Representative images of Iba1 immunostaining in sham and TBI mice in ipsilateral cortex (Ipsi-Ctx), ipsilateral thalamus (Ipsi-Thal), and contralateral CA2 (Contra-CA2) brain regions at 3 days post-injury (DPI) and 120 DPI, respectively. (B) Form factor (FF) analysis of Iba1 positive macrophages at 3 DPI revealed higher values in non-Tg and R1.40 TBI mice compared with sham mice in all brain regions analyzed (n = 6/group). (C) FF analysis at 120 DPI also revealed a primary effect of injury in all brain regions (n = 6/ group). Error bars indicate standard error of the mean. Scale bars indicate 20 μm. Main effect of injury, *, p < 0.05, ***, p < 0.001; main effect of genotype, #, p < 0.05.

FIG. 3.

Distinct macrophage phenotype in brain injured R1.40 mice at 3 days post-injury (DPI). (A,C) Representative images of CD45 and F4/80 immunostaining in sham and mice with traumatic brain injury (TBI) in ipsilateral cortex (Ipsi-Ctx), ipsilateral thalamus (Ipsi-Thal), and contralateral CA2 (Contra-CA2) brain regions at 3 DPI. (B) Non-Tg and R1.40 TBI mice show similar levels of increased CD45 immunoreactivity compared with sham animals (n = 6/group) in all brain regions. R1.40 TBI mice showed a decrease in CD45 immunoreactivity compared with non-Tg TBI mice in the Ipsi-Thal. (D) F4/80 immunoreactivity was significantly elevated in non-Tg TBI mice compared with all other groups in each brain region analyzed (n = 6/group). (E,G) Representative images of CD68 and Trem2 immunostaining in sham and TBI mice at 3 DPI. (F) CD68 expression was significantly increased in TBI mice compared with shams with non-Tg TBI mice exhibiting the most CD68 expression (n = 3/group). (H) Trem2 expression was significantly elevated in non-Tg TBI mice compared with all other groups (n = 3/group). Error bars indicate standard error of the mean. Scale bars indicate 20 μm. Main effect of injury: **, p < 0.01; ***, p < 0.001; Main effect of genotype: #, p < 0.05; ##, p < 0.01; ###, p < 0.001; interaction effect: +, p < 0.05; +++, p < 0.001.

FIG. 4.

Proportion of microglia and monocytes are similar between non-Tg and R1.40 mice with traumatic brain injury (TBI). (A,B) Representative contour plots of microglia (CD45^low/F4/80 cells) and monocytes (CD45^high/F4/80 cells) in non-Tg and R1.40 mice with TBI at 3 days post-injury (DPI). (C) R1.40 TBI mice show a significant decrease in CD45⁺/F4/80⁺ compared with non-Tg TBI mice at 3 DPI (n = 3–7/group). (D,E) TBI induced a significant reduction in microglia (CD45^low/F4/80 cells) at 3 DPI; however, no significant differences in monocytes were identified at 3 DPI (CD45^high/F4/80 cells). Error bars indicate standard error of the mean; main effect of injury: *, p < 0.05; interaction effect: +, p < 0.05.

FIG. 5.

Macrophage activation persists in subcortical structures at 120 days post-injury (DPI). (A,C) Representative images of CD45 and F4/80 immunostaining of both non-Tg and R1.40 sham and mice with traumatic brain injury (TBI) at 120 DPI. (B) CD45 immunoreactivity remained elevated in the ipsilateral thalamus (Ipsi-Thal) in both non-Tg and R1.40 TBI mice (n = 6/group). (D) A main effect of injury was apparent in F4/80 immunoreactivity in the Ipsi-Thal and contralateral CA2 (Contra-CA2) region. Error bars indicate standard error of the mean. Scale bars indicate 20 μm. Main effect of injury: *, p < 0.05, **; p < 0.01. Ipsilateral cortex, Ipsi-Ctx.

FIG. 6.

Traumatic brain injury (TBI) causes time dependent changes in APP expression. (A,D) Western blot of cortical extracts were probed with antibodies CT15, 6E10, and GAPDH at 3 and 120 days post-injury (DPI) respectively. (B) There was a significant increase in APP in non-Tg TBI mice compared with non-Tg shams at 3 DPI. R1.40 TBI mice displayed higher levels of APP than non-Tg TBI mice, but no differences in APP expression were identified between R1.40 TBI and R1.40 sham mice at 3 DPI. (C) No difference in human APP expression was found between R1.40 TBI and R1.40 sham mice at 3 DPI. As expected, human APP was undetectable in non-Tg mice. (E) There was a significant reduction in APP in non-Tg TBI mice compared with non-Tg shams at 120 DPI. APP levels were reduced in non-Tg mice compared with R1.40 mice regardless of injury group. (F) R1.40 TBI mice displayed a significant reduction in human APP expression compared with R1.40 sham mice at 120 DPI (3 DPI, n = 5–6/group; 120 DPI, n = 6/group). Human APP was undetectable in non-Tg mice. (G,H) Levels of Aβ 1-40 in ipsilateral cortex homogenates were higher in R1.40 mice compared with non-Tg mice at 3 and 120 DPI, respectively; however these differences were not statistically significant (n = 5–6/group). (I) Representative images of CT15 immunostaining in sham and TBI mice at 3 DPI (top two rows) and 120 DPI (bottom row). APP accumulation was apparent in the ipsilateral external capsule and cortex at 3 DPI in mice with TBI. By 120 DPI, only R1.40 mice showed enhanced APP immunostaining in the ipsi-cortex (n = 3/group/time point). Error bars indicate standard error of the mean. Scale bars indicate 20 μm. Main effect of injury: *, p < 0.05; main effect of genotype: #, p < 0.05; ###, p < 0.001.

FIG. 7.

Traumatic brain injury (TBI) induces time dependent changes in motor and cognitive function. (A) Mean latency to fall from the rotating rod was shorter in TBI mice at all post-injury time points compared to shams regardless of genotype. No between genotype differences in rotarod performance were detected in sham and TBI mice (n = 12–14/group). (B) No significant differences between groups in average spontaneous alternation in the y maze. R1.40 TBI mice showed a significant reduction in percent spontaneous alternation between arms from 7 days post-injury (DPI) to 90 DPI. (C) Average total arm entries were similar between TBI groups; however, sham mice demonstrated a significant reduction in arm entries over time (n = 12–14/group). (D) A significant interaction effect between injury and genotype was detected in latency to reach the hidden platform at 120 DPI (n = 12–14/group). (E) Average acquisition index shows that latency to reach the platform significantly decreases within a testing day in R1.40 compared with non-Tg mice. (F) Average savings index shows that latency to reach the platform significantly increases between testing days in R1.40 compared with Non-Tg mice. Error bars indicate standard error of the mean. Main effect of injury: *, p < 0.05; ***, p < 0.001; Interaction effect: ++, p < 0.01; +++, p < 0.001.