The brain-bone marrow axis and its implications for chronic traumatic brain injury

Abstract

Traumatic brain injury (TBI) causes acute and chronic alterations in systemic immune function which contribute to posttraumatic neuroinflammation and neurodegeneration. However, how TBI affects bone marrow (BM) hematopoietic stem/progenitor cells chronically and to what extent such changes may negatively impact innate immunity and neurological function has not been examined. To further understand the role of BM cell derivatives on TBI outcome, we generated BM chimeric mice by transplanting BM from chronically injured or sham congenic donor mice into otherwise healthy, age-matched, irradiated hosts. After 8 weeks of reconstitution, peripheral myeloid cells from TBI→WT mice showed significantly higher oxidative stress levels and reduced phagocytic activity. At eight months after reconstitution, TBI→WT chimeric mice were leukopenic, with continued alterations in phagocytosis and oxidative stress responses, as well as persistent neurological deficits. Gene expression analysis revealed BM-driven changes in neuroinflammation and neuropathology after 8 weeks and 8 months of reconstitution, respectively. Chimeric mice subjected to TBI showed that longer reconstitution periods were associated with increased microgliosis and leukocyte infiltration. Thus, TBI causes chronic activation and progressive dysfunction of the BM stem/progenitor cell pool, which drives long-term deficits in innate immunity and neurological function, as well as altered sensitivity to subsequent brain injury.

Figure 1.. Femur bone marrow responses at acute and chronic stages following TBI.

(A-E) Flow cytometry was performed in whole blood and femur bone marrow at 1, 3, and 7d post-injury. TBI increased the percentage of myeloid cells (% CD11b+ of CD45+ cells) in blood (A) and bone marrow (B). Percentages of lineage-c-Kit+Sca1+ (LSK+) stem/progenitor cells were elevated in bone marrow after TBI (C). A representative dot plot of LSK+ populations at 3d TBI is indicated in the left panel of C. Quantification of LSK+ cells is shown in the right panel of C. The mean fluorescence intensity (MFI) of cell cycle markers Ki-67 (D) and PCNA (E) were seen significantly increased in the LSK+ population at 1d post-injury. n=6–10 mice/group. (F-H) The proliferative status of LSK+ cells was assessed at 120 days post-injury (dpi). n=4 (Sham) and 6 (TBI) mice/group. (I-J) Chronic TBI at 365 dpi caused a significant reduction in circulating (I) and bone marrow (J)-derived white blood (CD45+) cells. n=6–11(blood) and 4–6 (bone marrow) mice/group. For all histograms, light gray=fluorescence minus one (FMO) control. Data were analyzed using one-way ANOVA group analysis with Tukey’s test for multiple comparisons (A-C) or Mann-Whitney for two group comparisons (D-J). **p<0.01, *p<0.05.

Figure 2.. TBI causes long-term alterations in femur bone marrow stem/progenitor cells.

The transcriptomic profile of femur bone marrow lineage-c-Kit+Sca1+ (LSK+) cells at 90 days post-injury (dpi) was assessed using the NanoString nCounter Stem Cell panel. (A) Principal component analysis (PCA) plot indicated that the two main principal components of variation were captured on the x- and y-axis, respectively, showing a clear separation of clusters between the Sham and TBI groups. (B) Pathway enrichment analysis with the MSigDB Hallmark 2020 database revealed that several gene networks related to hematopoietic stem cell proliferation and self-renewal, inflammatory activation, and senescence-associated were enriched after chronic TBI. (C) Pathway analysis based on gene annotations given by NanoString revealed high percentage of genes related to Apoptosis, Rho/ROCK signaling, Senescence & Quiescence, Hypoxia Response, and Cell Cycle pathways being modified by TBI. (D) Heatmap of genes that are uniquely altered in the most differentially expressed genes between Sham and 90 dpi. Color coding was based on z-score scaling. (E-H) Volcano plot analyses for the Apoptosis (E), Senescence and Quiescence (F), Oxidative Stress Response (G), and Epigenetic Modification (H) pathways showed the most differentially and significantly expressed genes within each network. n=4–5 mice/group.

Figure 3.. Blood neutrophil functions are chronically dysregulated during the chronic stages of TBI.

Bulk RNA-seq was performed on blood neutrophils at 120 dpi. (A) PLSDA of all normalized gene counts revealed a clear separation of aged sham and aged TBI samples into individual groups across the first two principal components. n=3 mice/group. (B) Pathways involved in biosynthetic processes and translation were down-regulated after TBI, whereas those involved in stress and viral defense responses were up-regulated. (C) Heatmap of unsupervised clustering of the top 30 differentially expressed genes. Color coding was based on z-score scaling. (D-E) Sample-level enrichment analysis (SLEA) for chromatin remodeling pathway (D) and epigenetic regulation (E) of gene expression. (F) Venn Diagram of chromatin remodeling pathway, epigenetic regulation of gene expression pathway, and differentially expressed (DE) epigenetic pathway genes genes between AT vs. AS. (G) Six out of 268 genes from chromatin remodeling pathway were differentially expressed after TBI, including Smarcad1, Tspyl4 and Hgfl3 up-regulated after TBI and Bcl7c, Chd3 and Wdhd1 down-regulated after TBI. (H-I) Neutrophil function including phagocytosis (MFI of pHrodo E. coli., H) and oxidative stress levels [mean fluorescence intensity (MFI) of DCF, I] from the spleen was evaluated by flow cytometry. n=4–6 mice/group. For all histograms, light gray=fluorescence minus one (FMO) control. Data were analyzed using Mann-Whitney for two group comparisons. **p<0.01, *p<0.05. AS: aged sham; AT: aged TBI.

Figure 4.. Bone marrow cells transplanted from chronic TBI mice reprogram host innate immune function in the absence of the brain injury environment.

(A) Using a novel bone marrow transplantation approach, sham and 90 days post-injury (dpi) donor cells were transplanted into irradiated, congenic WT recipients. Femur bone marrow cells were harvested from 90 dpi and sham congenic Pepboy (CD45.1) donor mice, and 100 ml of BM cells (1 × 10⁶ cells/mouse) were intravenously injected by retroorbital injection in recipient WT (CD45.1) C57BL/6 mice. Mice were allowed to reconstitute for 8 weeks following transplantation. (B-D) Blood donor cells (CD45.1, B-C) and leukocyte (CD45+, D) were counted using flow cytometry. n=5–6 mice/group. (E-G) Circulating and bone marrow-derived Ly6G+ neutrophils from TBI→WT chimeric mice exhibited reductions in 0.5–1.0 mm Red Beads+ cells and increase in the mean fluorescence intensity (MFI) of DHR123+ ROS. Left panels of E-F are representative dot plots. n=5–6 mice/group. (H-I) Bone marrow-derived Ly6G+ neutrophils from TBI→WT chimeric mice showed significant reductions in TNF and IL-1b production. Left panels of H are representative dot plots of TNF+ cells. n=7 mice/group. For all histograms, light gray=fluorescence minus one (FMO) control. Data were analyzed using one-way ANOVA group analysis with Tukey’s test for multiple comparisons (E-G) or Student’s T-test for two group comparisons (H-I). ****P<0.0001, **p<0.01, *p<0.05. SH: Sham, WT wildtype.

Figure 5.. Short-term reconstitution of bone marrow cells transplanted from chronic TBI mice affect host neurological function and neuroinflammatory signaling.

(A) Chimera paradigm is illustrated. Mice were allowed to reconstitute for 8 weeks following transplantation. n=11–16 mice/group. (B) Body weight of animal was monitored before and at 4 and 8 weeks (wks) after irradiation (IR). (C-F) Graphs showed parameters (body speed, stride length, swing speed, and swing durations) of CatWalk gait analysis tested at 8 weeks after IR. (G-J) Locomotor function and depressive-like behavior were assessed in the accelerating rotarod (G), Grip strength (H), and Open field (I), and tail suspension (J) tests. (K) Short-term spatial working memory was evaluated in Y-maze test. (L-N) Plasma cytokine levels were examined using multiplex Enzyme-Linked Immunosorbent Assay (ELISA). The pro-aging/neurodegenerative cytokines IP-10, CCL11, and KC were significantly increased in TBI→WT mice compared to controls. n=10–13 mice/group. (O-Q) Brain microglia and CD45^hi leukocytes were detected using flow cytometry. A representative dot plot of leukocyte populations in the brain is illustrated in O. Quantification of CD45^intCD11b⁺ microglia (P) and CD45^hiCD11b⁺ myeloid cells (Q) are shown. (R-V) The transcriptomic profile of brain tissue at 8 weeks after IR was assessed using the NanoString nCounter Neuroinflammation panel. n=4 mice/group. The Partial least squares-discriminant analysis (PLSDA) is illustrated (R). Heatmap of genes that are associated with Microglial Function (S), Inflammatory Signaling (T), and the Innate Immune Response (U) are shown. Color coding was based on Average Transcription Counts (normalized by housekeeping genes). Volcano plot analysis is indicated (V). Data were analyzed using one-way ANOVA group analysis with Tukey’s test for multiple comparisons. W: WT Naive; S: SH→WT; T: TBI→WT. ****P<0.0001, ***p<0.001, **p<0.01, *p<0.05. SH: Sham.

Figure 6.. Long-term reconstitution of bone marrow cells transplanted from chronic TBI mice causes innate immune dysfunction and neurological decline in host mice.

(A) Chimera paradigm is illustrated. Mice were allowed to reconstitute for 8 months following transplantation. (B) Body weight was monitored at 8 months after irradiation (IR). (C-F) A battery of neurobehavioral tests was performed to assess forelimb grip strength (C), time on rotarod (D), time immobility in the tail suspension test (E), and cognitive function in Y-maze test (F). (G-H) Blood donor cells (CD45.1, G) and leukocyte (CD45+, H) were counted using flow cytometry. (I) Percentages of bone marrow LSK+ frequencies were detected. (J) Spleen Ly6G+ neutrophils were examining using flow cytometry, showing reduced mean fluorescence intensity (MFI) of IL-1b+ cells in TBI→WT mice compared to SH→WT animals. Light gray=fluorescence minus one (FMO) control. (K-M) Phagocytic engulfment of IgG-coated beads (K), latex beads (L), and pHrodo-labeled E.coli particles (M) were detected in spleen Ly6G+ neutrophils. Left panels of K-M are representative dot plots. Data were analyzed using one-way ANOVA group analysis with Tukey’s test for multiple comparisons (B-I) or Student’s T-test for two group comparisons (J-M). n=7 mice/group. ****P<0.0001, **p<0.01, *p<0.05. SH: Sham.

Figure 7.. Long-term reconstitution of bone marrow cells transplanted from chronic TBI mice worsens ongoing neurodegeneration at the transcriptomic level in host brains.

The transcriptomic profile of whole brain hemisphere tissue at 8 months after irradiation was assessed using the NanoString nCounter Neuropathology panel. (A) Chimera paradigm is illustrated. (B) Partial least squares-discriminant analysis (PLSDA) plot showed a separation of clusters between the TBI→WT (T) and SH→WT (S) groups. (C) Pathway enrichment analysis with the MSigDB Hallmark 2020 database revealed that several gene networks related to apoptosis, oxidative phosphorylation, UV response, and complement pathways were enriched in TBI→WT compared to SH→WT mice. (D) Volcano plot analyses are shown between two groups. (E) Heat map analysis of the top 20 most differentially expressed genes are indicated. Color coding was based on z-score scaling. (F) Normalized transcription count of the top 9 genes presented as Log10 abundance. n=4 mice/group.

Figure 8.. The acute neuroinflammatory response to a subsequent TBI is sensitized with increasing time post-injury.

(A) Chimera paradigm is illustrated. Mice reconstituted for 8 weeks following transplantation were subjected to TBI. (B) Representative dot plots of immune populations in the ipsilateral brain hemisphere at 48 h after TBI. (C-D) Quantification of CD45^intCD11b⁺ microglia and CD45^hiCD11b⁺-infiltrating myeloid cell counts per hemisphere are shown for Sham and TBI in chimeric mice. n=4–8 mice/group. (E) Chimeric mice reconstituted for 8 months following transplantation were subjected to TBI. (F) A representative dot plot of leukocyte populations in the ipsilateral brain at 48 h after TBI. (G-H) Quantification of CD45^intCD11b⁺ microglia (G) and CD45^hiCD11b⁺ myeloid cells (H) are shown. n=6–8 mice/group. Data were analyzed using two-way ANOVA group analysis with Tukey’s test for multiple comparisons. **p<0.01, *p<0.05. SH: Sham.

Figure 9.. The senolytic drug, ABT-263, has beneficial effects on normal age-related motor function but does not confer robust protection to acute TBI.

(A) Experimental design. Naïve aged mice (18-month-old) were treated with ABT-263 (50 mg/kg) or vehicle (Veh) once daily for 2 weeks by oral gavage. After two-weeks on and one week off drug, mice were subjected to TBI up to 72h post-injury. (B) Body weight was monitored before and at 2 weeks (wks) after the treatment. (C-H) A battery of neurobehavioral tests was performed to assess forelimb grip strength (C), time on rotarod (D), and CatWalk gait analysis for body speed (E), stride length (F), swing speed (G), and cadence (H). (I) Representative dot plots of immune populations in the ipsilateral (Ipsi) and contralateral (Contra) brain hemisphere at 72 h after TBI. (J-L) Quantification of CD45^intCD11b⁺ microglia, CD45^hiCD11b⁺-infiltrating myeloid cell, CD45^hiCD11b⁻-infiltrating lymphocytes counts per hemisphere are shown for Sham, TBI, treatment in aged mice. n=9–10 mice/group. Data were analyzed using two-way ANOVA group analysis with Tukey’s test for multiple comparisons. ***p<0.001, **p<0.01, *p<0.05. SH: Sham.