Traumatic brain injury promotes neurogenesis at the cost of astrogliogenesis in the adult hippocampus of male mice

Abstract

Traumatic brain injury (TBI) can result in long-lasting changes in hippocampal function. The changes induced by TBI on the hippocampus contribute to cognitive deficits. The adult hippocampus harbors neural stem cells (NSCs) that generate neurons (neurogenesis), and astrocytes (astrogliogenesis). While deregulation of hippocampal NSCs and neurogenesis have been observed after TBI, it is not known how TBI may affect hippocampal astrogliogenesis. Using a controlled cortical impact model of TBI in male mice, single cell RNA sequencing and spatial transcriptomics, we assessed how TBI affected hippocampal NSCs and the neuronal and astroglial lineages derived from them. We observe an increase in NSC-derived neuronal cells and a concomitant decrease in NSC-derived astrocytic cells, together with changes in gene expression and cell dysplasia within the dentate gyrus. Here, we show that TBI modifies NSC fate to promote neurogenesis at the cost of astrogliogenesis and identify specific cell populations as possible targets to counteract TBI-induced cellular changes in the adult hippocampus.

Fig. 1. Immunohistochemical and behavioral characterization of hippocampal changes following TBI.

Micrograph representative of three independent experiments. Red: GFAP immunohistochemistry showing ipsilateral astrogliosis (a), white box: the location of the controlled cortical impact, dotted line: approximate boundary of the cortical lesion, scale bar, 250 µm. Micrographs representative of three independent experiments showing increased astrogliosis in the hippocampus of Control (b) vs. TBI (c) mice. ML molecular layer, sGCL suprapyramidal granule cell layer, H hilus, iGCL infrapyramidal granule cell layer. Scale bars, 100 µm. Quantification of GFAP surface coverage in the dentate gyrus, n = 3 mice. Data are presented as mean values +/− SEM, ***P < 0.0001 one-way ANOVA with Tukey’s post hoc test (d). Micrographs representative of three independent experiments showing proliferative cells in the DG. Red: BrdU, gray: cell nuclei labeled DAPI, scale bars, 40 µm (e). Quantification of cell proliferation in the DG, n = 4 mice. Data are presented as mean values +/− SEM, **P = 0.0061 two-tailed, unpaired t test (f). Micrographs representative of three independent experiments. Green: Nestin-GFP, red: BrdU (red), magenta: mature astrocyte marker S100B. Cell nuclei were labeled with the DNA marker DAPI (blue). Yellow arrowheads: Nestin-GFP⁺/BrdU⁺ cells, yellow arrows: Nestin-GFP⁺/BrdU⁺ cells in the GCL. Deep pink: S100B⁺/BrdU⁺ astrocytes^, white arrowheads S100B⁺/BrdU⁺ astrocytes in the hilus. Scale bars, 60 µm (g). Quantification of astrocyte proliferation in the DG expressed as the total number of S100B⁺/BrdU+ cells per mm³, n = 4 mice. Data are presented as mean values +/− SEM, **P = 0,0060 two-tailed, unpaired t test (h). Quantification of astrocyte proliferation in the DG expressed as the % of total S100B⁺ cells that incorporated BrdU, n = 4 mice. Data are presented as mean values +/− SEM, ***P = 0,0063 two-tailed, unpaired t test (i). Micrographs representative of three independent experiments showing proliferative NSCs in the DG (Nestin-GFP⁺/GFAP⁺/Mki67⁺ cells). Green: Nestin-GFP, red: GFAP, magenta: Mki67, merge: Nestin-GFP⁺/GFAP⁺/Mki67⁺, scale bars, 15 µm (j). Quantification of NSC proliferation in the DG expressed as the total number of Nestin-GFP⁺/GFAP⁺/Mki67⁺ per mm³, n = 5 mice. Data are presented as mean values +/− SEM, *P = 0.0374 two-tailed, unpaired t test (k). Micrographs representative of three independent experiments showing cells expressing DCX in the DG of Control (l) or TBI (m) mice, scale bars, 10 µm. Quantification of hippocampal neurogenesis expressed as total numbers of DCX+ cells in the dentate gyrus, Control n = 5 mice, TBI n = 4 mice. Data are presented as mean values +/− SEM, *P < 0.0078 two-tailed unpaired t test (n). Quantification of six DCX+ cell phenotypes according to the presence, shape and orientation of apical dendrites within the DG. Control n = 5 mice, TBI n = 4 mice. Data are presented as mean values +/− SEM, *P = 0.0431 one-way ANOVA with Tukey’s post hoc test (o). Micrographs representative of three independent experiments showing proliferative cells in the DG, Green: DCX antibodies, red: BrdU, blue: cell nuclei labeled DAPI. Yellow somas indicated by yellow arrowheads: DCX⁺/BrdU⁺ cells, red somas indicated by red arrowheads: (DCX^-/BrdU⁺) cells, scale bars, 80 µm (p). Quantification of neuronal cell proliferation in the DG expressed as the total number of DCX⁺/BrdU⁺ cell per mm³, n = 4 mice. Data are presented as mean values +/− SEM, **P = 0.0095 two-tailed, unpaired t test (q). Quantification of neuronal cell proliferation in the DG expressed as the % of total DCX⁺ cells that incorporated BrdU, n = 4 mice. Data are presented as mean values +/− SEM, **P = 0.0068 two-tailed, unpaired t test (r). Micrographs representative of three independent experiments. Magenta: RV-GFP, cyan: DCX. Yellow arrows: RV-GFP⁺/ DCX⁺ cells in the external half of the GCL or the ML, scale bars, 20 µm (s). Quantification of newborn neuron numbers expressed as the % of DCX+ cells that were positive for RV-GFP, n = 4 mice. Data are presented as mean values +/− SEM, *P = 0.0129 two-tailed, unpaired t test (t). Quantification of the number of newborn neurons with two primary dendrites emerging directly from the soma, or a main dendrite bifurcating within the first 10 µm from the soma (atypical dendrites) expressed as the % of total newborn neurons, n = 4 mice. Data are presented as mean values +/− SEM, *P = 0.0146 two-tailed, unpaired t test (u). Quantification of dendritic spines density in secondary and tertiary dendrites of newborn cells expressed as the number of spines per 10 μm long dendritic segment, n = 4 mice. Data are presented as mean values +/− SEM, **P = 0.0014 two-tailed, unpaired t test (v). Escape latency in the Morris Water Maze test Control n = 12 mice, TBI n = 11 mice. Data are presented as mean values +/− SD, #F(1, 109) = 67,0012, P = 5,56 10⁻¹³ time, *F(4, 109) = ¹9,6203, P = 3, 47 10⁻¹² treatment, two-way repeated measures ANOVA with Bonferroni post hoc test. Exact P values were calculated using P = F.DIST.RT(F, DFn, DFd) (w). Percentage of time spent in the target quadrant during the Morris water maze probe trial, Control n = 12 mice, TBI n = 11 mice. Data are presented as mean values +/− SEM, ^#P = 0.011 two-tailed, unpaired t test (x). Source data are provided as a Source Data file.

Fig. 2. Molecular characterization of the NSC-derived neuro- and astrogliogenic lineages within the mouse dentate gyrus and the effects of TBI.

UMAP-based visualization of the major higher-order cell types identified by Seurat in the single-cell dataset. Each dot represents a single cell. Cells with similar molecular profiles group together. Cell types were assigned according to the expression of specific marker genes and are labeled in different colors (a). UMAP-based visualization of single-cell data according to experimental origin (Control or TBI, 5–6 animals per condition) (b). Bar plots showing the relative number of cells per cluster originating from Control or TBI samples against their predicted abundance (62.96% of all cells originating from TBI samples: dashed line), NSC, *P = 4.69 × 10⁻²; Astrocyte linage, *P = 5.99 × 10⁻³⁵; Oligo, *P = 2.92 × 10⁻³; Endo, *P = 5.14 × 10⁻⁵; Mgl1, *P = 2.13 × 10⁻²; Mural, *P = 1.82 × 10⁻⁸ vs. expected abundance, two-sided binomial test. A summary of cell numbers and statistical analyses performed is given in Supplementary Data 1 (c). UMAP-based visualization of NSC-derived neuronal and astrocytic lineages obtained by extraction of neuronal and astrocytic cells followed by reclustering. Each dot represents an individual cell and discrete cell states are identified in specific colors (d). UMAP-based visualization of reclustered data according to experimental origin (Control or TBI) (e). Bar plots showing the relative number of cells per identified cell cluster in Control or TBI samples, against their expected abundance (57.39% of all cells originating from TBI samples: dashed line), N-stage 1, *P = 4.66 × 10⁻²; N-stage 2, *P = 4.07 × 10⁻⁴; N-stage 4, *P = 2.97 × 10⁻⁴; N-stage 5, *P = 2 × 10⁻⁵; A-stage 1, *P = 2.23 × 10⁻⁷; A-stage 2, *P = 3.39 × 10⁻¹⁵; A-stage 4, *P = 2.12 × 10⁻⁶ vs. expected abundance, two-sided binomial test. A summary of cell numbers per population and statistical analyses performed is given in Supplementary Data 2 (f). GO biological pathway matrix for the cell clusters included in the neuronal (g) and astrocytic (h) lineages. Source data are provided as a Supplementary Data file.

Fig. 3. RNA velocity analysis indicates that TBI induces a subtle but functionally significant shift in hippocampal neural stem cell fate.

RNA velocity analysis along both the neuronal and astrocytic lineage in the Control (a) and TBI (b) groups, vectors (arrows) indicate the predicted direction and speed of single-cell transitions in transcriptome space, colors indicate previously characterized cell clusters indicated in the figures. Box blots showing transition probabilities calculated for NSC-stage 1 cells to NSC-stage 1 cells (****P = 4.4 × 10⁻¹⁵), to NSC-stage 2 (****P = 3.7 × 10⁻¹⁵) and to RG-like cells (**P = 1.5 × 10⁻³) (c); NSC-stage 2 cells to NSC-stage 1 (****P = 6.5 × 10⁻⁸), to NSC-stage 2 (*P = 1.7 × 10⁻²), to RG-like (****P = 1.4 × 10⁻⁵) and to N-stage 1 cells (****P = 7.9 × 10⁻⁵) (d); RG-like cells to NCS-stage 1 (**P = 1.2 × 10⁻³), to RG-like (****P = 6.6 × 10⁻¹²), to N-stage 1 (****P = 9.3 × 10⁻⁴) and to A-stage 1 cells (****P = 3.6 × 10⁻⁶) (e); N-stage 1 cells to RG-like (*P = 4.7 × 10⁻²), to N-stage 1 (***P = 7.1 × 10⁻⁴), and to N-stage 2 cells (**P = 1.7 × 10⁻³) (f); and A-stage 1 cells to A-stage 1 (****P = 5.4 × 10⁻⁹) and to A-stage 2 cells (****P = 3.9 × 10⁻⁹) (g). Orange bars: Control group, cyan bars: TBI group, using Control = 5 and TBI = 6 animals per condition. In all panels, independent non-parametric two-sided Wilcoxon test, Control vs TBI. Box plots show the median, first quartile (25%), third quartile (75%), and interquartile range. Whiskers represent data minima and maxima; dots are data points located outside the whiskers, n = number of cells per cell subpopulation Control/TBI are available as Supplementary Data 2 and calculated mean transition probabilities across cell clusters are available in Supplementary Data 14, including exact P values for all comparisons. Source data are provided as Supplementary Data 15.

Fig. 4. TBI induces cell population-specific changes in gene expression in NSCs and NSC-derived cells in the DG.

Heatmaps showing expression of individually upregulated (a) and downregulated (b) genes across NSCs and NSC-derived cell populations. Color bars indicate the relative intensity of expression, indicated as ln fold change in TBI vs. Control groups; SCT normalized data are compared. Violin plots showing changes in Ppp1r14b expression induced by TBI in RG-like cells, ***P = 5 × 10⁻⁴ (c), NSC-stage 2 cells, ***P = 8.28 × 10⁻⁷ (d) astrocytic A-stage 3 cells, ***P = 4.93 × 10⁻⁵ (e), and neurogenic N-stage 1, ***P = 3.21 × 10⁻¹⁵ (f), N-stage 2, ***P = 8.59 × 10⁻¹⁶ (g), N-stage 3, ***P = 1.38 × 10⁻¹³ (h) and N-stage 4 cells, ***P = 6.44 × 10⁻¹² (i), two-sided MAST test with Bonferroni post hoc test on SCT normalized data. Source data are provided at https://zenodo.org/records/10829090/files/AstrocyticAndNeuronalLineage_10x.RData?download=1.

Fig. 5. TBI induces changes in the location of specific cell populations in the dentate gyrus.

Micrograph representative of three independent experiments showing RNAscope visualization of marker genes for A-stage 4 (a, inset iii), A-stage 2 (a, inset ii) and A-stage 1 (a, inset i) cells (individual marker signals shown in Supplementary Fig. 7). ML molecular layer, GCL granule cell layer, SGZ subgranular zone; scale bar, 5 µm. UMAP-based representation of the individual 10X and Molecular Cartography (MC) datasets. scRNA-seq-derived (left) and MC-derived data (right) share a similar distribution within UMAP space. Colors in the 10X dataset indicate neuronal and astrocytic cell lineages (b). UMAP-based representation of the combined 10X-MC dataset (merged data), with colors indicating the ten distinct clusters identified (c). UMAP-based representation of pseudotime for neuronal (d) and astrocytic (e) lineages using the combined 10X-MC dataset. Color bars: relative pseudotime distance from the root population (yellow: minimum, red: maximum). Correlation matrix showing the relative overlap in gene expression data between cell populations identified by scRNA-seq (10X scRNA-seq-defined clusters) and MC (Molecular Cartography signal clusters) (f). Color bar: % of genes in a scRNA-seq cluster present in a MC cluster. DG subdivisions used to compare the location of MC clusters, Hilus, subgranular zone (SGZ), inner granule cell layer (GL1), outer granule cell layer (GL2). ML is indicated as reference (g). Spatial mapping for an example Control DG section (h). Spatial mapping of NSC populations in Control and TBI groups in the DG. Colored dots indicate the location of single-cell populations indicated in the legend (i). Spatial mapping of NSC-derived neuronal cells in Control and TBI groups in the DG. Colored dots indicate the location of single-cell populations indicated in the legend. Pie charts: % localization of N-stage 3 cells to the SGZ (orange) or the GL (yellow) (j). Dot plots showing quantification of N-stage 3 cells in the SGZ (k) or the GL (l), n = 4 Control, n = 5 TBI. Data are presented as mean values +/− SEM, **P = 0.0075 (k), P = 0.0030 (l), two-tailed unpaired t test. Spatial mapping of NSC-derived astrocytic populations in Control and TBI groups. Colored dots indicate the location of single-cell populations indicated in the legend. Pie charts: % localization of astrocytic A-stage 1 cells to the SGZ (green) or the Hilus (blue) (m). Dot plots showing the quantification of A-stage 1 cells in the SGZ (n) or the Hilus (o) n = 4 Control, n = 5 TBI. Data are presented as mean values +/− SEM, **P = 0.0088 (n), P = 0.0043 (o), two-tailed unpaired t test. Spatial mapping of Gfap+ astrocytes in the DG from Control and TBI animals (p); dots indicate the location of single cells, with color encoding the relative intensity of Gfap expression. Spatial representations in (j, m) are representative examples of hippocampal slices analyzed to generate the data in (k, l, n, o). Summary of cell transitions predicted by RNA velocity in the Control and TBI groups. Black arrows indicate the most probable transitions of NSCs and NSC-derived cells along neuronal and astrocytic lineages in uninjured AHN; arrow thickness indicates transition probability. Red +: transition promoted by TBI; Red −: transition inhibited by TBI. Circular arrows: cell populations with predicted self-renewal potential (q). Representations of GCL/SGZ/Hilus areas in the DG indicating the location of NSC-derived neuronal or astrocytic cells determined using MC spatial transcriptomics in both Control and TBI (r). Source data are provided as a Source Data file.