Age dictates brain functional connectivity and axonal integrity following repetitive mild traumatic brain injuries in mice

Abstract

Traumatic brain injuries (TBI) present a major public health challenge, demanding an in-depth understanding of age-specific symptoms and risk factors. Aging not only significantly influences brain function and plasticity but also elevates the risk of hospitalizations and death following TBIs. Repetitive mild TBIs (rmTBI) compound these issues, resulting in cumulative and long-term brain damage in the brain. In this study, we investigate the impact of age on brain network changes and white matter properties following rmTBI by employing a multi-modal approach that integrates resting-state functional magnetic resonance imaging (rsfMRI), graph theory analysis, diffusion tensor imaging (DTI), and neurite orientation dispersion and density imaging (NODDI). Our hypothesis is that the effects of rmTBI are worsened in aged animals, with this group showing more pronounced alterations in brain connectivity and white matter structure. Utilizing the closed-head impact model of engineered rotational acceleration (CHIMERA) model, we conducted rmTBIs or sham (control) procedures on young (2.5-3-months-old) and aged (22-months-old) male and female mice to model high-risk groups. Functional and structural imaging unveiled age-related reductions in communication efficiency between brain regions, while injuries induced opposhigh-risking effects on the small-world index across age groups, influencing network segregation. Functional connectivity analysis also identified alterations in 79 out of 148 brain regions by age, treatment (sham vs. rmTBI), or their interaction. Injuries exerted pronounced effects on sensory integration areas, including insular and motor cortices. Age-related disruptions in white matter integrity were observed, indicating alterations in various diffusion directions (mean diffusivity, radial diffusivity, axial diffusivity, and fractional anisotropy) and density neurite properties (dispersion index, intracellular and isotropic volume fraction). Neuroinflammation, assessed through Iba-1 and GFAP markers, correlated with higher dispersion in the optic tract, suggesting a neuroinflammatory response in injured aged animals compared to sham aged. These findings offer insight into the interplay between age, injuries, and brain connectivity, shedding light on the long-term consequences of rmTBI.

Keywords: Aging; CHIMERA; Diffusion tensor imaging; Microglia; Repetitive mild TBI; Resting state functional MRI.

Fig. 1.

Experimental design and neuroimaging analysis. (A) Schematic representation showing the number of repetitive mild traumatic brain injuries (rmTBI), rsfMRI and tissue collection. (B) Resting state functional MRI processing pipeline. (C) Diffusion tensor imaging (DTI) and neurite orientation dispersion and density imaging (NODDI) workflow. ROI, region of interest; rsfMRI, resting-state functional magnetic resonance imaging; ICA, independent component analysis; FA, fractional anisotropy; MD, mean diffusivity; AD, axial diffusivity; RD, radial diffusivity; ISOVF, isotropic volume fraction; ICVF, intracellular volume fraction; OD, orientation dispersion index.

Fig. 2.

Global network changes in young and aged mice. TBI-related physiological symptoms were assessed immediately after injury. (A) In rmTBI groups, there was a higher loss of righting reflex (LRR), serving as a rodent analog to loss of consciousness in humans. Aged mice exhibited a more severe phenotype compared to young mice. (B) rmTBI-induced effects in LRR were independent of weight. Brain network was assessed by using (C) network strength (measures overall network connectivity), (D) characteristic path length (quantifies network communication efficiency), (E) global efficiency (assesses information transmission efficiency), and (F) small-world index (SWI, evaluates local and global information balance). Two-way ANOVA was conducted in LRR and weight measurements while repeated measures two-way ANOVA were used in network analyses examining a range from 2 to 40 edge density thresholds. Statistically significant difference is indicated as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. rmTBI, repetitive mild trauma brain injury; Y, young; A, Aged; sec, seconds; g, grams.

Fig. 3.

Age and treatment effects on rsfMRI metrics in brain. A threshold of 10 % edge density was applied for local analyses, which is within the most robust and least dense connections. (A) Main effects and interaction of age and injury (injury = sham vs rmTBI) on the following rsfMRI metrics: node degree (ND), node strength (NS), node efficiency (NE), betweenness centrality (BC), and eigenvector centrality (EC). Out of the 148 regions of interest evaluated, only those with 2 or more significant values are represented in the heatplot. Statistically significant (p < 0.05) effects for each region of interest (ROI) resulting from Type III two-way ANOVA with Benjamini false discovery rate (FDR-BH) correction are represented in color. White squares denote non-significant values. (B) Percentage of significant ROI affected by age, injury, interaction age and injury for each metric. (C) Significant main effects distribution in right and left-brain hemisphere. rsfMRI = resting-state functional magnetic resonance imaging. ROI names are detailed in Supplemental Table 1.

Fig. 4.

Multiple comparison and effect sizes on top affected brain regions in young and aged mice after rmTBI. (A) Multiple comparisons tests were conducted on the regions of interest (ROI) showing significant main effects in the two-way ANOVA. Adjusted p-values resulting from pairwise comparisons with Benjamini-Hochberg (FDR-BH) correction are color-coded in the heatplot. The magnitude and direction of the effect were determined by Cohen’s d effect size. Red represents an effect size equal to or greater (red) or lower (blue) than +/−0.8 (large effect). ROIs are arranged from left to right based on the highest number of significant p-values across metrics. Diagrams on the right illustrate scores assigned to a region of interest (ROI, blue circle) based on its connectivity to other regions (white circles), ranging from low to high values. (B) In young mice, the secondary motor area (left smot1) exhibited the greatest impact by rmTBI, while (C) in the aged group, the ventral agranular insula (left vains) emerged as the most affected region. NS = node strength, EC = eigenvector centrality. Bar graphs represent the mean with standard deviation (SD). Statistically significant p-values from two-way ANOVA are represented as follows: *p < 0.05, ***p < 0.001. Gr1, group1 is first from left to right in panel; Young-Sham (Y-Sh), n = 8; Young-rmTBI (Y-TBI), n = 10; Aged-Sham (A-Sh), n = 6; Aged-rmTBI (A-TBI), n = 8.

Fig. 5.

Functional brain connectivity by subdivisions. (A) Connectome maps illustrating the average functional connectivity, eigenvector centrality, and total edge densities among all ROIs within the groups. (B) A basic diagram depicting node degree and eigenvector centrality relationship. (C) Brain areas were further dissected into subdivisions showing age differences in eigenvector centrality in the (D) hippocampal formation, (E) isocortex, (F) striatum, and (G) thalamus. Statistically significant p-values from two-way ANOVA are represented as follows: **p < 0.01, ****p < 0.0001. EC = eigenvector centrality, HPC, hippocampus; ROI, regions of interest.

Fig. 6.

Microstructural changes in ROIs: DTI and NODDI metrics. (A) Main effects and interactions on each region of interest (ROI)’s microstructure. Significant adjusted p-values resulting from two-way ANOVA (type III) with Benjamini-Hochberg (FDR-BH) correction are color-coded in the heatplot. Diffusion tensor imaging (DTI) metrics include fractional anisotropy (FA, measures directional water diffusion), mean diffusivity (MD, reflects overall water diffusion.), axial diffusivity (AD, measures diffusion along the main voxel axis), and radial diffusivity (RD, analyzes diffusion perpendicular to the main voxel axis). The three compartments measured in neurite orientation dispersion and density imaging (NODDI) analysis were: isotropic volume fraction (ISOVF, quantifies non-neurite components), intracellular volume fraction (ICVF, measures volume taken by neurite), and orientation dispersion index (OD, indicate neurite alignment and organization). (B) The percentage of significant ROIs affected by age (young vs aged), injury (sham vs rmTBI) and interaction between these factors. (C) Shared regions affected between groups demonstrating the unique effect by age and injury in white matter integrity. Young-Sham (n = 5), Young-rmTBI (n = 4), Aged-Sham (n = 5), and Aged-rmTBI groups (n = 7). Supplemental Table 2 describe ROI names in DTI and NODDI analyses.

Fig. 7.

Optic tract and anterior cingulate: key regions in DTI and NODDI analysis comparison. (A) Multiple comparison tests were conducted on the regions of interest (ROI) displaying significant main effects in the two-way ANOVA for both DTI and NODDI analyses. DTI measures include fractional anisotropy (FA, directional water diffusion), mean diffusivity (MD, overall water diffusion), axial diffusivity (AD, diffusion along the main voxel axis), and radial diffusivity (RD, diffusion perpendicular to the main voxel axis). NODDI measures comprise isotropic volume fraction (ISOVF, non-neurite components), intracellular volume fraction (ICVF, volume occupied by neurites), and orientation dispersion index (OD, neurite alignment and organization). Significant effects with a large effect size (Cohen’s d) are indicated in color, while non-effects are shown in white. Adjusted p-values were corrected using Benjamini-Hochberg (FDR) correction. Red signifies an effect size equal to or greater than 0.8, and blue indicates an effect size equal to or less than —0.8. ROIs are ordered based on the highest number of significant p-values, highlighting the optic tract (OT) and anterior cingulate (ACC) as the most affected regions. Representative images correspond to their respective metrics. (B-D) Inverse Pearson correlations between FA and RD for the motor area (MOs), corticospinal tract (CST), and hippocampal commissure (HPComm) indicates an overall reduction of RD by aging. (E-F) Interaction between aging and brain injury augments inverse correlations between FA vs RD signal in retrosplenial area (RSP). Gr1, group1 is first from left to right in panel; Young-Sham (Y-Sh), n = 7; Young-rmTBI (Y-TBI), n = 10; Aged-Sham (A-Sh), n = 6; Aged-rmTBI (A-TBI), n = 8.

Fig. 8.

Gliosis is induced by rmTBI in optic tract. (A). Following rmTBI, aged mice displayed an increase in (A-B) Iba-1 and (C-D) GFAP expression in the optic tract. Microglia are indicated by the blue arrowhead, and astrocytes by the red arrowhead. Brain coronal view: 800 μm; Scale bar for inset (enlarged view of the optic tract) = 100 μm. Statistical analysis performed using unpaired t-test with Welch’s correction; Aged-Sham (n = 10), Aged-rmTBI (n = 9). SD: standard deviation; Iba-1, Ionized calcium-binding adaptor molecule 1; GFAP, Glial fibrillary acidic protein; μm, micrometers; OT, optic tract.