Abrogation of atypical neurogenesis and vascular-derived EphA4 prevents repeated mild TBI-induced learning and memory impairments

Abstract

Brain injury resulting from repeated mild traumatic insult is associated with cognitive dysfunction and other chronic co-morbidities. The current study tested the effects of aberrant neurogenesis in a mouse model of repeated mild traumatic brain injury (rmTBI). Using Barnes Maze analysis, we found a significant reduction in spatial learning and memory at 24 days post-rmTBI compared to repeated sham (rSham) injury. Cell fate analysis showed a greater number of BrdU-labeled cells which co-expressed Prox-1 in the DG of rmTBI-injured mice which coincided with enhanced cFos expression for neuronal activity. We then selectively ablated dividing neural progenitor cells using a 7-day continuous infusion of Ara-C prior to rSham or rmTBI. This resulted in attenuation of cFos and BrdU-labeled cell changes and prevented associated learning and memory deficits. We further showed this phenotype was ameliorated in EphA4f.^/f/Tie2-Cre knockout compared to EphA4f.^/f wild type mice, which coincided with altered mRNA transcript levels of MCP-1, Cx43 and TGFβ. These findings demonstrate that cognitive decline is associated with an increased presence of immature neurons and gene expression changes in the DG following rmTBI. Our data also suggests that vascular EphA4-mediated neurogenic remodeling adversely affects learning and memory behavior in response to repeated insult.

Figure 1

rmTBI induces GFAP immunoreactivity, learning and memory impairments, and expansion of neuroblast/immature neurons. (A–C) Representative maximum projection z-stack confocal mosaic image of serial section stained for GFAP (red), CD31 (green) and DAPI (blue) in rSham at 24 dppi. (D–E) Representative confocal image at 24 dppi rmTBI showing increased GFAP immunoreactivity in the hippocampus (D1–F1; insert) and cortex (D2–F2; insert). (G) Experimental timeline showing repeated mild TBI surgeries (*blue), daily BrdU injections (green), behavioral testing (grey), and two time-points for tissue collection (arrows, black). The rmTBI mice show an increase in the # primary errors (H) and latency in seconds (s) (I) made during Barnes Maze testing at 23 dppi. No difference was found in the distance to goal box per trial (J) or day (K). The number of BrdU (L) and Prox1/BrdU-positive (M) cells in the DG was increased in rmTBI mice while no change was seen in Sox/BrdU (N) or NeuN/BrdU (O) at 24 dppi compared to rSham. Scale = 1 mm in (A–F); 500 µm in (D1–F1) and 20 µm in (D2–F2). Data are represented as mean ± SEM. Two-way ANOVA for repeated measures with Bonferroni post-hoc. *p < 0.05, and **p < 0.01. n = 10–15 mice per group for behavior and n = 7–12 for stereology.

Figure 2

AraC rescues rmTBI-induced aberrant neurogenesis. (A) Experimental time line showing 7 days pre-infusion of vehicle or 2% AraC prior to rSham or rmTBI, 50 mg/kg/day BrdU injections, repeated injuries and behavior testing paradigm. (B) Quantified data showing estimated number of BrdU+ cells/mm³ by non-biased stereology in the DG at 24 dppi. A significant attenuation of increased BrdU-labeling was seen in AraC-treated rmTBI mice compared to vehicle rmTBI. (C) AraC also reduced the number of Prox1/BrdU double-positive cells compared to vehicle rmTBI. No significant differences in the estimated number of Sox2/BrdU or NeuN/BrdU positive cells were seen between groups. (F–I) Representative confocal maximum projection z-stack images of the DG immunolabeled with BrdU (red), Prox1 (green) and DAPI (blue) with inset (white boxes). Scale = 250 µm in (F–I) and 50 µm in (F–I) insets. Data are represented as mean ± SEM. One-way ANOVA with Bonferroni post-hoc. *p < 0.05, ****p < 0.000. n = 6–10 mice.

Figure 3

AraC recues cognitive deficits and hyperactivity in the DG following rmTBI. (A) The number of primary errors made by vehicle infused rmTBI mice was significantly reduced in rmTBI mice treated with AraC at 23 dppi. (B) Quantified data showing estimated number of cFos+ cells/mm³ by non-biased stereology in the DG at 24 dppi. Mice receiving pre-treatment with AraC did not display an increase in cFos activity in the DG after rmTBI compared to vehicle treated mice. (C) The Correlation Coefficient plotted for each animal, shows the number of cFos+ cells in the DG is positively correlated with the number of Barnes Maze errors (R² = 0.3; p = 0.01) in vehicle-treated mice. No correlation was observed in AraC mice (R² = 0.0; p = 0.02). (D) The number of Prox1+ cells quantified in the hilus was increased following 24 dppi rmTBI in vehicle but not AraC mice. (E–H) Representative confocal images of cFos (green) and DAPI (blue) staining in the DG of rSham and rmTBI mice pre-infused with vehicle or AraC. Scale = 500 µm in (E–H). *p < 0.05; ***p = 0.001; ****p = 0.0001. n = 9–13 mice per group for behavior and 6–10 for stereology.

Figure 4

EphA4f.^/f/Tie2-Cre mice show attenuation of rmTBI-induced deficits. (A) The number of primary errors in significantly increased in rmTBI EphA4f.^/f (WT) mice at 23 dppi compared rSham. This effect was not observed in EphA4f.^/f/Tie2-Cre (KO) mice. Compared to rSham, rmTBI KO mice did not show an increase in the number of cFos (B), BrdU (C) or Prox1/BrdU (D) positive cells in the DG as was seen in WT mice. No difference in the number of Sox2/BrdU (E) or NeuN/BrdU (F) positive cells were found across the groups of mice. (G–H) Representative confocal images of Prox1/BrdU positive cells in the DG at 24 dppi in WT mice. (I–L) Quantified mRNA expression of total DG tissue using qPCR for FGF2, MCP-1, Gja1 (Cx43), and TGFβ, respectively. All genes were normalized to GAPDH then represented relative to WT rSham levels. *p < 0.05; **p < 0.01, ***p = 0.001. Scale = 100 µm in (G,H); 20 µm in inset. n = 5–10 mice per group for stereology and n = 10–16 for behavior. qPCR was performed using biological triplicates.