Aging disrupts blood-brain and blood-spinal cord barrier homeostasis, but does not increase paracellular permeability

Abstract

Blood-CNS barriers protect the CNS from circulating immune cells and damaging molecules. It is thought barrier integrity becomes disrupted with aging, contributing to impaired CNS function. Using genome-wide and targeted molecular approaches, we found aging affected expression of predominantly immune invasion and pericyte-related genes in CNS regions investigated, especially after middle age, with spinal cord being most impacted. We did not find significant perturbation of endothelial cell junction genes or proteins, nor were vascular density or pericyte coverage affected by aging. We evaluated barrier paracellular permeability using small molecular weight tracers, serum protein extravasation, CNS water content, and iron labelling measures. We found no evidence for age-related increased barrier permeability in any of these tests. We conclude that blood-brain (BBB) and blood-spinal cord barrier (BSCB) paracellular permeability does not increase with normal aging in mouse. Whilst expression changes were not associated with increased permeability, they may represent an age-related primed state whereby additional insults cause increased leakiness.

Keywords: Aging; Blood–brain barrier; Laser-capture; Proteomics; RNA-Seq; scRNA-seq.

Fig. 1

RNAseq of CTX and SC indicates barrier transcriptomes are perturbed by aging. a Barrier-related GO enriched in CTX and SC DEG sets with aging. GO with enrichment > 3 and FDR < 0.01 are shown. b Enrichments plotted for endothelial- (EC), pericyte-, and astrocyte-cell-type-specific genes in the DEG sets for CTX and SC, using cell-type-specific gene sets as described in the text. c Enrichments plotted for ECs located at different parts of the cerebrovasculature tree. d Enrichments plotted for EC junction genes and BBB dysfunction module genes. See Methods for details on enrichment analyses. b-d **p < 0.01 ***p < 0.001

Fig. 2

BSCB is more impacted by aging than BBB. a Heatmap of barrier gene expression. Individual sample expression is relative to average expression of young in matching CNS region. Scale bar is log₂ FC. Superscripted numbers with gene names indicate region(s) reaching statistical significance: CTX¹ CC² HIP³ CB⁴ SCGM⁵ SCWM⁶. b Number of significantly differentially expressed genes (DEGs) of each functional group in each CNS region. c Relative expression of selected barrier genes across mouse adult lifespan (2.5, 4, 8, 14, 26 months) in SC. Expression is relative to 2.5-month group. Dots indicate individual animals. Bars indicate mean. Error bars ± SD. *p < 0.05 corrected. CTX – cortex; CC – corpus callosum; HIP – hippocampus; CB – cerebellum; SC – spinal cord; GM – grey matter; WM – white matter

Fig. 3

Laser microdissection of CNS microvessels. Relative expression of barrier junction genes in a CTX and b SCGM laser microdissected blood vessels. Col-IV-immunolabelled 10 µm SC section c before and d after laser microdissection of microvessels, 40 × objective. Scale bars are 75 µm. e Relative expression of F11r in CTX and SC blood vessels. Bars represent group fold change relative to young, points represent individual animals relative to young group average, error bars are SD for all graphs. n = 5–6/group.* one-tailed Wilcoxon Mann-Whiney U Exact test corrected p-value < 0.05. CTX – cortex; SCGM – spinal cord grey matter

Fig. 4

Aging brain single cell meta-analysis. a UMAP of aging brain cells. Putative cell type clusters: 0—ECs; 1—microglia; 2—astrocytes; 3—microglia; 4—oligodendrocytes; 5—astrocytes; 6—oligodendrocyte precursor cells; 7—ECs; 8—neurons; 9—smooth muscle cells; 10—choroid plexus epithelial cells; 11—perivascular macrophages; 12—ependymal cells; 13—PC; 14—perivascular macrophages; 15—vascular leptomeningeal cells; 16—vascular leptomeningeal cells (ABC); 17—olfactory ensheathing cells; 18—neurons; 19—proliferating; 20—oligodendrocytes. b Number of cells in EC and mural cell clusters. c Venn diagram of EC and mural cell clusters age-related DEGs d Selected enriched GOs of DEGs in cluster 0 (EC1), cluster 7 (EC2), 9 (SMC), and 13 (PC). EC – endothelial cell; SMC – smooth muscle cell; PC – pericyte

Fig. 5

Effects of aging on blood-CNS barrier proteins. a Quantification of claudin-5 protein expression in CTX and SC of young (red columns) and old (blue columns) mice. Columns represent different MW bands of the protein. Western blots from which expression was quantified are shown for b CTX and c SC. d Quantification of occludin protein expression in CTX and SC of young (red columns) and old (blue columns) mice. Columns represent different MW bands of the protein. Western blots from which expression was quantified are shown for e CTX and f SC. g Quantification of tubulin beta-3 protein expression in CTX and SC of young (red columns) and old (blue columns) mice. Western blots from which expression was quantified are shown for h CTX and i SC. Analyses of claudin-5, occludin, and tubulin beta-3, protein expression in j brain ECs and k CTX and l HIP from mice of different ages. Proteomics data was obtained from Todorov-Volgyi et al., 2024 j and Tsumagari et al., 2023 k, l

Fig. 6

Blood vessel density and pericyte coverage in aging mouse CNS. a-e Blood vessel density. a Vessel density was quantified by Col-IV immunolabelling across 6 CNS regions. Points are individual measurements (6–9/animal). Bars are mean and error bars are ± SD. n = 3/gp/region. * Nested t-test p < 0.05. b-e False colour processed images of Col-IV immunolabelling (red) in young b, d and old c, e spinal cord grey b, c and white d, e matter at 40x. f-j Pericyte coverage. f Pericyte coverage was quantified using Col-IV and CD13 double immunolabelling. Points are individual measurements (6–9/animal). Bars are mean and error bars ± SD. n = 8–9/gp/region. g-j False colour processed images of Col-IV (red) and CD13 (green) double (yellow) immunolabelling in young g, i and old h, j spinal cord grey g, h and white i, j matter at 40x. CTX – cortex; CC – corpus callosum; HIP – hippocampus; CB – cerebellum; SC – spinal cord; GM – grey matter; WM – white matter

Fig. 7

Functional assessment of the impact of aging on blood-CNS barrier permeability. a Effects of aging on brain and spinal cord water content expressed as percent of wet weight. n = 6/gp. b Sodium Fluorescein (NaFl) permeability. n = 10–11/gp. c Fluorescein dextran (Fl-Dex) permeability. n = 3–4/gp. d Serum albumin (Ser-Alb) permeability. n = 8–9/gp. e Immunoglobulin G (IgG) permeability. n = 8–9/gp. f Iron labelling for microhaemorrhages. n = 8–9/gp. Graphs: points represent individual animals; bars are mean and error bars are ± SD; * one-tailed Wilcoxon Mann–Whitney U Exact test corrected p-value < 0.05. Legend in a holds for all graphs. CTX – cortex; CC – corpus callosum; HIP – hippocampus; CB – cerebellum; SC – spinal cord; GM – grey matter; WM – white matter