Aged microglia promote peripheral T cell infiltration by reprogramming the microenvironment of neurogenic niches

Abstract

Background: The immune cell compartment of the mammalian brain changes dramatically and peripheral T cells infiltrate the brain parenchyma during normal aging. However, the mechanisms underlying age-related T cell infiltration in the central nervous system remain unclear.

Results: Chronic inflammation and peripheral T cell infiltration were observed in the subventricular zone of aged mice. Cell-cell interaction analysis revealed that aged microglia released CCL3 to recruit peripheral CD8⁺ memory T cells. Moreover, the aged microglia shifted towards a pro-inflammation state and released TNF-α to upregulate the expression of VCAM1 and ICAM1 in brain venous endothelial cells, which promoted the transendothelial migration of peripheral T cells. In vitro experiment reveals that human microglia would also transit to a chemotactic phenotype when treated with CSF from the elderly.

Conclusions: Our research demonstrated that microglia play an important role in the aging process of brain by shifting towards a pro-inflammation and chemotactic state. Aged microglia promote T cell infiltration by releasing chemokines and upregulating adhesion molecules on venous brain endothelial cells.

Keywords: Aging; Blood–brain barrier; Endothelial cells; Microglia; Single-cell transcriptome; T cells.

Fig. 1

T cells infiltrate the old neurogenic niches and affect the brain microenvironment. a t-SNE projection of the CD45⁺ cells in the SVZ from three young (3 months old) and three old (28–29 months old) mice. Cell types are color-coded and annotated based on the transcriptomic profiles. b Bar plot showing the fraction of cells associated with each cell type in the young and aged SVZ. *P ≤ 0.05 by one-tailed Wilcoxon rank sum test. c Representative confocal microscopic images of young and old SVZs stained for CD31 and CD8a. Nuclei are labeled with DAPI. Scale bar: 200 μm. d Number of CD8⁺ T cells per coronal section in four young (6–8 weeks old) and four aged (18 months old) male mice. *P = 0.0286, Mann–Whitney test, two sided. Data are shown as mean ± s.e.m. e Violin plots showing expression of various T cell activation- (Cd69, Itk), cytokine release- (Ccl5, Xcl1), and cytotoxicity-related (Gzmb, Gzmk) genes in each of three distinct clusters. f Sankey diagram depicting the interaction between T cells and resident cells in the SVZ based on the Reactome term immunoregulatory interactions between a lymphoid and a non-lymphoid cell (R-HSA-198933). The proportional flow represents the number of gene pairs. Gene pairs are listed in Supplementary Table 2

Fig. 2

Transcriptomic alteration of chemokines and receptors in aged microglia and peripheral T cells. a Volcano plot depicting the differentially expressed receptor genes in aged CD8⁺ T cells compared to young CD8⁺ T cells. Differentially expressed genes (DEGs; |log₂(fold change)| > 0.1, Bonferroni adjusted P-value < 0.05) were colored (red for upregulated DEGs and green for downregulated DEGs). b Bar plot showing the expression levels of Cxcr3, Ccr2, and Ccr5 in CD4⁺ T cells and CD8⁺ T cells from young and old mice. ***Bonferroni adjusted P-value ≤0.001 by Wilcoxon rank sum test. c Expression profiles of Ccr2 and Ccr5 in both young and aged CD8⁺ T cells are shown using the UMAP visualization approach. d Protein expression of CCR2 by splenic CD8⁺ memory T cells of four young (6–8 weeks old) and four aged (18 months old) male mice. *P = 0.0264, two-tailed Student’s t test. Data are shown as mean ± s.e.m. e Violin plots showing expression of Ccl3 and Ccl4 in young and old mouse brains. *** Bonferroni adjusted P-value ≤0.001 by Wilcoxon rank sum test. f Dot plot showing human CCL3 plasma levels. Linear regression is depicted with the colored line. Type II sums of squares were calculated and tested using the F-test. Q-values were estimated using the Benjamini–Hochberg approach. ***P = 4.36 × 10^− 33; q = 1.14 × 10^− 31). g Left: Representative confocal microscopic images of young and old SVZs stained for CCL3 and IBA-1. Nuclei are labeled with DAPI. Scale bars: 100 μm or 25 μm. Right: Quantification of CCL3 fluorescence in IBA1⁺ cells in four young (6–8 weeks old) and four aged (18 months old) male mice. ****P < 0.0001, two-tailed Student’s t test. Data are shown as mean ± s.e.m.

Fig. 3

Age-related changes of BECs from different arteriovenous zone. a t-SNE projection of brain endothelial cells in the SVZ from three young (3 months old) and three old (28–29 months old) (left). Cell types are color-coded and annotated based on the transcriptomic profiles. Bar plot showing the fraction of cells associated with each cell type from both young and aged mice (right). b Dot plot showing the expression of tight junction genes across different vascular segments. Node size represents the magnitude of DEGs (log₂(fold change)) and node color represents adjusted P-values. **Bonferroni adjusted P-value ≤0.01, ***Bonferroni adjusted P-value ≤0.001 by Wilcoxon rank sum test. c Representative confocal microscopic images of young and old mouse brains stained for IgG. Nuclei are labeled with DAPI. Scale bars: 100 μm. d Quantitative analysis of IgG extravascular deposits in four young (6–8 weeks old) and four aged (18 months old) male mice. *P = 0.0117, two-tailed Student’s t test. Data are shown as mean ± s.e.m. e Heatmap of GSEA showing part of the significant (FDR q-value < 0.05) aging-related pathways across vessel segments. Numbers in the legend represent normalized enrichment score (NES); positive values indicate upregulation and negative values indicate downregulation

Fig. 4

Cell–cell interaction between venous BECs and Tcm. a The activation status of biological processes related to leukocyte migration was assessed by calculating their NES using GSEA. Shown are the biological processes associated with leukocyte migration that were predicted to be activated in arterial capillary, capillary, and vein clusters. b Dot plot showing the upregulated genes in the term lymphocyte migration across different vascular segments. Dot size means the power of gene pairs by summing up the log₂(fold change) values of participating genes. Dot color means the aggerated confidence for gene pairs. c Dot plot depicting the interaction between T cells and BECs from different vascular segments. Node size represents the count of interaction pairs and node color represents the power of interaction pairs. d Dot plot showing the gene pairs between venous BECs and central memory T cells. Gene pairs with p-value < 0.05 in permutation test are fetched. Dot color represents the power of gene pairs by multiplying the expression levels of participating genes. e Expression profiles of Itgb1 and Itgb2 in both young and aged CD8⁺ T cells are shown using the UMAP visualization approach. f Violin plots showing expression of Itgb1 and Itgb2 in young and aged CD8⁺ T cells. g–h Representative confocal microscopic images and quantification of VCAM1 (g), ICAM1 h colocalization with CD31 in four young (6–8 weeks old) and four aged (18 months old) male mice. Nuclei are labeled with DAPI. Scale bars: 50 μm. **P = 0.0096 (g), ***P = 0.0008(h), two-tailed Student’s t test. Data are shown as mean ± s.e.m.

Fig. 5

Cell–cell interactions between aged microglia and venous BECs. a Dot plot depicting the interaction of “leukocyte migration” between cells in the SVZ and BECs from different vascular segments. Node size represents count of interaction pairs and node color represents power of interaction pairs. b t-SNE projection of microglia in the SVZ from young and aged mice (left). Cell types were color-coded and annotated based on the transcriptomic profiles. Bar plot showing the fraction of cells associated with each cell type from both young and aged mice (right). c Sankey diagram depicting the interactions between microglia and venous BECs. The proportional flow represents the number of gene pairs. d Dot plot showing the gene pairs between venous BECs (old vs. young) and microglia. Gene pairs with p-value < 0.05 in permutation test are fetched. Dot color represents the power of gene pairs by multiplying the expression levels of participating genes. e Cell plot showing interaction pairs between microglia and vein. The vivid two-cell graph was generated by the InterCellDB package using default settings. f Left: Representative confocal microscopic images of young and old mice brains stained for TNF-α and IBA-1. Nuclei are labeled with DAPI. Scale bars: 20 μm. Right: Quantification of TNF-α fluorescence in IBA1⁺ cells in four young (6–8 weeks old) and four aged (18 months old) male mice. ***P = 0.0008, two-tailed Student’s t test. Data are shown as mean ± s.e.m. g CD8⁺ T cells in inserts were cocultured with young brain slice, aged brain slice, microglia-depleted aged brain slice and aged brain slice with anti-CCL3 for 24 h. Cells in the lower compartment were collected for flow cytometry analysis. Each dot represents one mice. *Adjusted-P = 0.0155(young brain slie vs aged brain slice), # Adjusted-P = 0.0161(aged brain slice vs aged brain slice + anti-CCL3), One-way ANOVA. Data are shown as mean ± s.e.m.

Fig. 6

Transcriptional regulation of chemotaxis-related genes in human microglia after treating with CSF from the elderly. a Sorted primary human microglia were cultured with CSF from the elderly (n = 4) or PBS (n = 2) for 24 h and then collected for RNA sequencing. b mRNA expression levels of microglia (AIF1,TMEM119), neurons (MAP 2, NCAM1), Astrocytes (GFAP, ALDH1L1), Oligodendrocytes (APC, PLP1), Endothelial cells (CLDN5, ITM2A), T cells (CD3E) and B cells (CD79A) marker genes. c Volcano plot depicting the differentially expressed genes in CSF treated microglia compared to PBS treated microglia. Differentially expressed genes were colored (red for upregulated DEGs and green for downregulated DEGs). d Functional enrichment analysis was performed using Metascape on the upregulated DEGs in (c). The significantly overrepresented (P < 0.01) terms were grouped into color-coded clusters based on their membership similarities and rendered as a network plot. Each node represents an enriched term, and one representative term is shown for each cluster. Terms with a similarity > 0.3 are connected by edges. e Expression profiles of DEGs in CSF treated microglia related to leukocyte recruitment pathways

Fig. 7

Microglia promotes T cell infiltration in the aged CNS. Microglia shift towards a pro-inflammatory state in the aged CNS and promote CD8⁺ memory T cell recruitment by releasing CCL3. Aged microglia release TNF-α, which upregulates the expression of VCAM1 and ICAM1 in venous BECs. The activated venous BECs adhere tightly to hematogenous CD8⁺ memory T cells and promote their transmigration into the CNS. The infiltrated T cells interact with brain resident cells by releasing CCL4, CCL5, IFN-γ, XCL1, FASLG