Article | Published:

Aged blood impairs hippocampal neural precursor activity and activates microglia via brain endothelial cell VCAM1

Abstract

An aged circulatory environment can activate microglia, reduce neural precursor cell activity and impair cognition in mice. We hypothesized that brain endothelial cells (BECs) mediate at least some of these effects. We observe that BECs in the aged mouse hippocampus express an inflammatory transcriptional profile with focal upregulation of vascular cell adhesion molecule 1 (VCAM1), a protein that facilitates vascular–immune cell interactions. Concomitantly, levels of the shed, soluble form of VCAM1 are prominently increased in the plasma of aged humans and mice, and their plasma is sufficient to increase VCAM1 expression in cultured BECs and the hippocampi of young mice. Systemic administration of anti-VCAM1 antibody or genetic ablation of Vcam1 in BECs counteracts the detrimental effects of plasma from aged individuals on young brains and reverses aging aspects, including microglial reactivity and cognitive deficits, in the brains of aged mice. Together, these findings establish brain endothelial VCAM1 at the blood–brain barrier as a possible target to treat age-related neurodegeneration.

Access optionsAccess options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Data availability

RNA-seq gene lists with statistics (Figs. 1 and 2, and Extended Data Figs. 1 and 2), full blots (Extended Data Fig. 6) and individual data points graphed for Extended Data Fig. 9n are available as source data files and as Supplementary Tables 16 accompanying this article. Requests for datasets obtained from human research will be subject to additional review steps by the IRB that granted the permit for particular research. Bulk and scRNA-seq datasets that support the findings of this study have been deposited in the Gene Expression Omnibus (GEO) with the series accession number GSE127758 (embedded accession codes GSM3638211 to GSM3638222) and GSE127963. Please contact the corresponding author for additional information.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. 1.

    Wyss-Coray, T. Ageing, neurodegeneration and brain rejuvenation. Nature 539, 180–186 (2016).

  2. 2.

    Harry, G. J. Microglia during development and aging. Pharm. Ther. 139, 313–326 (2013).

  3. 3.

    Mosher, K. I. & Wyss-Coray, T. Microglial dysfunction in brain aging and Alzheimer’s disease. Biochem. Pharmacol. 88, 594–604 (2014).

  4. 4.

    Safaiyan, S. et al. Age-related myelin degradation burdens the clearance function of microglia during aging. Nat. Neurosci. 19, 995–998 (2016).

  5. 5.

    Ming, G. L. & Song, H. Adult neurogenesis in the mammalian brain: significant answers and significant questions. Neuron 70, 687–702 (2011).

  6. 6.

    Lazarov, O. & Marr, R. A. Neurogenesis and Alzheimer’s disease: at the crossroads. Exp. Neurol. 223, 267–281 (2010).

  7. 7.

    Ray, S. et al. Classification and prediction of clinical Alzheimer’s diagnosis based on plasma signaling proteins. Nat. Med. 13, 1359–1362 (2007).

  8. 8.

    Villeda, S. et al. The ageing systemic milieu negatively regulates neurogenesis and cognitive function. Nature 477, 90–94 (2011).

  9. 9.

    Britschgi, M. et al. Modeling of pathological traits in Alzheimer’s disease based on systemic extracellular signaling proteome. Mol. Cell. Proteom. 10, M111.008862 (2011).

  10. 10.

    Villeda, S. A. et al. Young blood reverses age-related impairments in cognitive function and synaptic plasticity in mice. Nat. Med. 20, 659–663 (2014).

  11. 11.

    Castellano, J. M. et al. Human umbilical cord plasma proteins revitalize hippocampal function in aged mice. Nature 544, 488–492 (2017).

  12. 12.

    Katsimpardi, L. et al. Vascular and neurogenic rejuvenation of the aging mouse brain by young systemic factors. Science 344, 630–634 (2014).

  13. 13.

    Rebo, J. et al. A single heterochronic blood exchange reveals rapid inhibition of multiple tissues by old blood. Nat. Commun. 7, 13363 (2016).

  14. 14.

    Engelhardt, B. & Liebner, S. Novel insights into the development and maintenance of the blood–brain barrier. Cell Tissue Res. 355, 687–699 (2014).

  15. 15.

    Fabene, P. F. et al. A role for leukocyte–endothelial adhesion mechanisms in epilepsy. Nat. Med. 14, 1377–1383 (2008).

  16. 16.

    Vivash, L. & OBrien, T. J. Imaging microglial activation with TSPO PET: lighting up neurologic diseases? J. Nucl. Med. 57, 165–168 (2016).

  17. 17.

    Gragnano, F. et al. The role of von Willebrand factor in vascular inflammation: from pathogenesis to targeted therapy. Mediators Inflamm. 2017, 5620314 (2017).

  18. 18.

    Rossi, B., Angiari, S., Zenaro, E., Budui, S. L. & Constantin, G. Vascular inflammation in central nervous system diseases: adhesion receptors controlling leukocyte–endothelial interactions. J. Leukoc. Biol. 89, 539–556 (2011).

  19. 19.

    Berlin, C. et al. α4 integrins mediate lymphocyte attachment and rolling under physiologic flow. Cell 80, 413–422 (1995).

  20. 20.

    Garton, K. J. et al. Stimulated shedding of vascular cell adhesion molecule 1 (VCAM-1) is mediated by tumor necrosis factor-α-converting enzyme (ADAM 17). J. Biol. Chem. 278, 37459–37464 (2003).

  21. 21.

    Pan, J. et al. Patterns of expression of factor VIII and von Willebrand factor by endothelial cell subsets in vivo. Blood 128, 104 –109 (2016).

  22. 22.

    Lee, M. et al. Transcriptional programs of lymphoid tissue capillary and high endothelium reveal control mechanisms for lymphocyte homing. Nat. Immunol. 15, 982–995 (2014).

  23. 23.

    Murakami, M. Signaling required for blood vessel maintenance: molecular basis and pathological manifestations. Int. J. Vasc. Med. 2012, 293641 (2012).

  24. 24.

    Daneman, R. et al. The mouse blood–brain barrier transcriptome: a new resource for understanding the development and function of brain endothelial cells. PLoS One 5, e13741 (2010).

  25. 25.

    Macdonald, J. A., Murugesan, N. & Pachter, J. S. Endothelial cell heterogeneity of blood–brain barrier gene expression along the cerebral microvasculature. J. Neurosci. Res. 88, 1457–1474 (2010).

  26. 26.

    Vanlandewijck, M. et al. A molecular atlas of cell types and zonation in the brain vasculature. Nature 554, 475 (2018).

  27. 27.

    Han, J. et al. Vascular endothelial growth factor receptor 3 controls neural stem cell activation in mice and humans. Cell Rep. 10, 1158–1172 (2015).

  28. 28.

    Fontaine, R. H. et al. Vascular endothelial growth factor receptor 3 directly regulates murine neurogenesis. Genes Dev. 25, 831–844 (2011).

  29. 29.

    Hosokawa, Y., Hosokawa, I., Ozaki, K., Nakae, H. & Matsuo, T. Cytokines differentially regulate ICAM-1 and VCAM-1 expression on human gingival fibroblasts. Clin. Exp. Immunol. 144, 494–502 (2006).

  30. 30.

    Zhang, J. et al. Regulation of endothelial cell adhesion molecule expression by mast cells, macrophages, and neutrophils. PLoS One 6, e14525 (2011).

  31. 31.

    Bruunsgaard, H., Pedersen, M. & Pedersen, B. K. Aging and proinflammatory cytokines. Curr. Opin. Hematol. 8, 131–136 (2001).

  32. 32.

    Sun, B. B. et al. Consequences of natural perturbations In the human plasma proteome. Preprint at bioRxiv https://www.biorxiv.org/content/10.1101/134551v1 (2017).

  33. 33.

    Shultz, L. D., Ishikawa, F. & Greiner, D. L. Humanized mice in translational biomedical research. Nat. Rev. Immunol. 7, 118–130 (2007).

  34. 34.

    Ridder, D. A. et al. TAK1 in brain endothelial cells mediates fever and lethargy. J. Exp. Med. 208, 2615–2623 (2011).

  35. 35.

    Kokovay, E. et al. VCAM1 is essential to maintain the structure of the SVZ niche and acts as an environmental sensor to regulate SVZ lineage progression. Cell Stem Cell 11, 220–230 (2012).

  36. 36.

    Elices, M. J. et al. VCAM-1 on activated endothelium interacts with the leukocyte integrin VLA-4 at a site distinct from the VLA-4/fibronectin binding site. Cell 60, 577–584 (1990).

  37. 37.

    Miller, D. H. et al. A randomized, placebo-controlled trial of natalizumab for relapsing multiple sclerosis. N. Engl. J. Med. 354, 899–910 (2006).

  38. 38.

    Chin, J. E. et al. Airway recruitment of leukocytes in mice is dependent on α4-integrins and vascular cell adhesion molecule-1. Am. J. Physiol. Cell. Mol. Physiol. 272, L219–L229 (1997).

  39. 39.

    Montagne, A., Zhao, Z. & Zlokovic, B. V. Alzheimer’s disease: a matter of blood–brain barrier dysfunction? J. Exp. Med. 214, 3151–3169 (2017).

  40. 40.

    Obermeier, B., Daneman, R. & Ransohoff, R. M. Development, maintenance and disruption of the blood–brain barrier. Nat. Med. 19, 1584–1596 (2013).

  41. 41.

    Singh, R. J. R. et al. Cytokine stimulated vascular cell adhesion molecule-1 (VCAM-1) ectodomain release is regulated by TIMP-3. Cardiovasc. Res. 67, 39–49 (2005).

  42. 42.

    Tchalla, A. E. et al. Elevated soluble vascular cell adhesion molecule-1 is ssociated with cerebrovascular resistance and cognitive function. J. Gerontol. Ser. A 72, 560–566 (2017).

  43. 43.

    Merat, S., Fruebis, J., Sutphin, M., Silvestre, M. & Reaven, P. D. Effect of aging on aortic expression of the vascular cell adhesion molecule-1 and atherosclerosis in murine models of atherosclerosis. J. Gerontol. 55, B85–B94 (2000).

  44. 44.

    Richter, V. et al. Circulating vascular cell adhesion molecules VCAM-1, ICAM-1, and E-selectin in dependence on aging. Gerontology 49, 293–300 (2003).

  45. 45.

    Ballantyne, C. M. & Entman, M. L. Soluble adhesion molecules and the search for biomarkers for atherosclerosis. Circulation 106, 766–767 (2002).

  46. 46.

    Okugawa, Y. et al. Soluble VCAM-1 and its relation to disease progression in colorectal carcinoma. Exp. Ther. Med. 1, 463–469 (2010).

  47. 47.

    Schlesinger, M. & Bendas, G. Vascular cell adhesion molecule-1 (VCAM-1)—an increasing insight into its role in tumorigenicity and metastasis. Int. J. Cancer 1, 2504–2514 (2015).

  48. 48.

    Ewers, M., Mielke, M. M. & Hampel, H. Blood-based biomarkers of microvascular pathology in Alzheimer’s disease. Exp. Gerontol. 45, 75 (2010).

  49. 49.

    Matsuda, M., Tsukada, N., Miyagi, K. & Yanagisawa, N. Increased levels of soluble vascular cell adhesion molecule-1 (VCAM-1) in the cerebrospinal fluid and sera of patients with multiple sclerosis and human T lymphotropic virus type-1-associated myelopathy. J. Neuroimmunol. 59, 35–40 (1995).

  50. 50.

    Elahy, M. et al. Blood–brain barrier dysfunction developed during normal aging is associated with inflammation and loss of tight junctions but not with leukocyte recruitment. Immun. Ageing 12, 2 (2015).

  51. 51.

    Cook-Mills, J. M., Marchese, M. E. & Abdala-Valencia, H. Vascular cell adhesion molecule-1 expression and signaling during disease: regulation by reactive oxygen species and antioxidants. Antioxid. Redox Signal. 15, 1607–1638 (2011).

  52. 52.

    Vestweber, D. How leukocytes cross the vascular endothelium. Nat. Rev. Immunol. 15, 692–704 (2015).

  53. 53.

    Ritzel, R. M. et al. Age-associated resident memory CD8 T cells in the central nervous system are primed to potentiate inflammation after ischemic brain injury. J. Immunol. 196, 3318–3330 (2016).

  54. 54.

    Yousef, H. et al. Systemic attenuation of the TGF-β pathway by a single drug simultaneously rejuvenates hippocampal neurogenesis and myogenesis in the same old mammal. Oncotarget 6, 11959–11978 (2015).

  55. 55.

    Hu, X.-L. et al. Persistent expression of VCAM1 in radial glial cells is required for the embryonic origin of postnatal neural stem cells. Neuron 95, 309–325 (2017).

  56. 56.

    Erdő, F., Denes, L. & de Lange, E. Age-associated physiological and pathological changes at the blood–brain barrier: a review. J. Cereb. Blood Flow Metab. 37, 4–24 (2017).

  57. 57.

    Montagne, A. et al. Blood–brain barrier breakdown in the aging human hippocampus. Neuron 85, 296–302 (2015).

  58. 58.

    Bien-Ly, N. et al. Lack of widespread BBB disruption in Alzheimer’s disease models: focus on therapeutic antibodies. Neuron 88, 289–297 (2015).

  59. 59.

    Steffen, B. J., Breier, G., Butcher, E. C., Schulz, M. & Engelhardt, B. ICAM-1, VCAM-1, and MAdCAM-1 are expressed on choroid plexus epithelium but not endothelium and mediate binding of lymphocytes in vitro. Am. J. Pathol. 148, 1819–1838 (1996).

  60. 60.

    Vukovic, J., Colditz, M. J., Blackmore, D. G., Ruitenberg, M. J. & Bartlett, P. F. Microglia modulate hippocampal neural precursor activity in response to exercise and aging. J. Neurosci. 32, 6435–6443 (2012).

  61. 61.

    Muggeo, V. M. R. Estimating regression models with unknown break-points. Stat. Med. 22, 3055–3071 (2003).

  62. 62.

    Muggeo, V. M. R. Segmented: an R package to fit regression models with broken-line relationships. R News 8, 20–25 (2008).

  63. 63.

    Yousef, H., Czupalla, C. J., Lee, D., Butcher, E. C. & Wyss-Coray, T. Papain-based single cell isolation of primary murine brain endothelial cells using flow cytometry. Bio-Protoc. 8, e3091 (2018).

  64. 64.

    Xaio, H., Banks, W. A., Niehoff, M. L. & Morley, J. E. Effect of LPS on the permeability of the blood–brain barrier to insulin. Brain Res. 896, 36–42 (2001).

  65. 65.

    Trapnell, C. et al. Differential analysis of gene regulation at transcript resolution with RNA-seq. Nat. Biotechnol. 31, 46–53 (2013).

  66. 66.

    Picelli, S. et al. Full-length RNA-seq from single cells using Smart-seq2. Nat. Protoc. 9, 171 (2014).

  67. 67.

    Darmanis, S. et al. A survey of human brain transcriptome diversity at the single cell level. Proc. Natl Acad. Sci. USA 112, 7285–7290 (2015).

  68. 68.

    Montesano, R. et al. Increased proteolytic activity is responsible for the aberrant morphogenetic behavior of endothelial cells expressing the middle T oncogene. Cell 62, 435–445 (1990).

  69. 69.

    Czupalla, C. J., Liebner, S. & Devraj, K. In vitro models of the blood–brain barrier. Methods Mol. Biol. 1135, 415–437 (2014).

  70. 70.

    Calabria, A. R., Weidenfeller, C., Jones, A. R., de Vries, H. E. & Shusta, E. V. Puromycin-purified rat brain microvascular endothelial cell cultures exhibit improved barrier properties in response to glucocorticoid induction. J. Neurochem. 97, 922–933 (2006).

  71. 71.

    Luo, J. et al. Glia-dependent TGF-β signaling, acting independently of the TH17 pathway, is critical for initiation of murine autoimmune encephalomyelitis. J. Clin. Invest. 117, 3306–3315 (2007).

  72. 72.

    Czupalla, C. J., Yousef, H., Wyss-Coray, T. & Butcher, E. C. Collagenase-based single cell isolation of primary murine brain endothelial cells using flow cytometry. Bio-Protoc. 8, e3092 (2018).

  73. 73.

    Luo, J. et al. Long-term cognitive impairments and pathological alterations in a mouse model of repetitive mild traumatic brain injury. Front. Neurol. 5, 12 (2014).

  74. 74.

    Hoffmann, A. et al. High and low molecular weight fluorescein isothiocyanate (FITC)-dextrans to assess blood–brain barrier disruption: technical considerations. Transl. Stroke Res. 2, 106–111 (2011).

  75. 75.

    Kijanka, G., Prokopowicz, M., Schellekens, H. & Brinks, V. Influence of aggregation and route of injection on the biodistribution of mouse serum albumin. PLoS One 9, e85281 (2014).

  76. 76.

    van Meer, P. & Raber, J. Mouse behavioural analysis in systems biology. Biochem. J. 389, 593–610 (2005).

  77. 77.

    Wolf, A., Bauer, B., Abner, E. L., Ashkenazy-Frolinger, T. & Hartz, A. M. S. A comprehensive behavioral test battery to assess learning and memory in 129S6/Tg2576 mice. PLoS One 11, e0147733 (2016).

  78. 78.

    Leger, M. et al. Object recognition test in mice. Nat. Protoc. 8, 2531 (2013).

  79. 79.

    Engelhardt, B., Vajkoczy, P. & Weller, R. O. The movers and shapers in immune privilege of the CNS. Nat. Immunol. 18, 123–131 (2017).

  80. 80.

    Banks, W. A. From blood–brain barrier to blood–brain interface: new opportunities for CNS drug delivery. Nat. Rev. Drug Discov. 15, 275–292 (2016).

Download references

Acknowledgements

We thank L. Sutherland and C. Cain for managing the core flow cytometry facility at the VA in Palo Alto and providing H.Y. and C.J.C. training on the instruments, as well as C. Cain for his experimental advice, assistance with flow cytometry and analysis of PBMCs, and thoughtful discussion. We thank O. Leyva for assistance in staining and microscopy analysis for the experiment shown in Extended Data Fig. 4a–d. We would also like to thank R. Watts and N. Bien-Ly for sharing a BEC isolation protocol used in modified form for RNA-seq. This work was funded by the Department of Veterans Affairs (T.W.-C.), the National Institute on Aging (F32-AG051330 to H.Y., R01-AG045034 and DP1-AG053015 to T.W.-C.), the NOMIS Foundation (T.W.-C.), the D.H. Chen Foundation (T.W.-C.), the Glenn Foundation for Aging Research (T.W.-C.), a SPARK grant to H.Y. through the Stanford Clinical and Translational Science Award (CTSA) to Spectrum (UL1-TR001085), the National Institutes of Health (R01-GM37734 and R37-AI047822 to E.C.B., RO1-AI109452 to H.H.), the Wu Tsai Neurosciences Institute (T.W.-C.), the Stanford Institute for Immunity, Transplantation and Infection (C.J.C.) and the Edinger Institute (C.J.C.). The CTSA program is led by the National Center for Advancing Translational Sciences (NCATS) at the National Institutes of Health.

Author information

H.Y. and T.W.-C. designed research. H.Y., C.J.C. and D.L. isolated BECs. H.Y. and C.J.C. performed and analyzed flow cytometry. H.Y. and A.N.B. performed in vitro experiments. H.Y., A.N.B., J.Z., D.L. and T.M. performed staining and microscopy analysis and cell counts. H.Y., A.N.B. and J.Z. performed ELISAs and western blots. H.Y. and C.J.C. performed tissue dissections. H.Y., C.J.C. and D.L. performed plasma injections. H.Y. and L.N.B. performed parabiosis. M.S. provided an Slco1c1-CreERT2 breeding pair. B.L. analyzed human proteomics data. H.H. helped with the design of the immune phenotyping panel and flow cytometry staining and analysis of PBMCs. K.A.Z., H.Y., D.L. and C.J.C. performed behavioral studies. M.S.B., T.P. and K.A.Z. carried out Barnes maze and novel object recognition studies. H.Y. and D.L. sectioned, stained and analyzed neuroinflammation in the peri-infarct region. A.C.Y., D.L., V.M. and H.Y. performed and analyzed BBB permeability with dextran experiments. V.M., A.N.B., D.L. and M.B.C. edited the manuscript. C.J.C. and E.C.B. helped with experimental design and edited the manuscript. C.J.C. developed protocols for BEC isolation, cultivation and flow cytometry following LPS stimulation. H.Y. and E.B. performed the bulk RNA-seq experiment. H.Y., M.B.C. and D.L. performed the scRNA-seq experiment. S.R.Q. supervised scRNA-seq data collection and analysis and reviewed the manuscript. H.Y., R.V.N., B.L. and E.B. analyzed bulk transcriptomic data. M.B.C. and H.Y. analyzed single-cell transcriptomic data. J.K. provided the plasmid for a BBB-specific AAV vector. H.Y. analyzed data and generated the figures. H.Y. and T.W.-C. wrote the manuscript. T.W.-C. supervised the study.

Competing interests

The authors declare no competing interests.

Correspondence to Tony Wyss-Coray.

Extended data

  1. Extended Data Fig. 1 Bulk and single-cell transcriptome and proteome profiling of young and aged BECs reveals an increased inflammatory signature with aging.

    a, Schematic of flow sorting of CD31+CD45 BECs from mouse cortex and hippocampus. n = 6 young and 6 aged biologically independent samples; each sample pools the cortices and hippocampi from two biologically independent mice. There were 1,006 significantly differentially expressed genes (*q < 0.05, Cuffdiff statistical package65). b, FACS gating strategy to isolate single BECs. PI+ dead cells were excluded. CD11b+ and CD45+ cells were gated to exclude monocytes/macrophages and microglia. CD31+Cd11bCD45 cells were defined as the BEC population. c, FPKM values of CNS cell-type-specific markers. n = 6 young and 6 aged biologically independent samples. Data are shown as the mean ± s.e.m. d, FPKM values of leukocyte binding adhesion molecules, including Vcam1. n = 6 young and 6 aged biologically independent samples. Bars show the mean; error bars are derived from s.e.m. Specific q values shown were derived from the Cuffdiff statistical package (*q = 0.0015). See the Methods for details; source data are available online. e, FPKM values of tight junction genes. n = 6 young and 6 aged biologically independent samples. Data are shown as the mean ± s.e.m. q = 0.16, *q = 0.0013, **q = 0.0015, Cuffdiff statistical package. See the Methods for details; source data are available online. f, FPKM values of transcripts in young and aged mouse CD31+ BECs corresponding to human plasma proteins that change with age (see Supplementary Table 2 for a list of human plasma proteins expressed in mouse BECs). n = 6 young and 6 aged biologically independent samples. Data are shown as the mean ± s.e.m. *q = 0.0015, **q = 0.021, Cuffdiff statistical package. See the Methods for details; source data are available online. g, C57BL/6 mice were injected with anti-VCAM1-DL488 or IgG-DL488 isotype control (r.o.) 2 h before perfusion to label BECs in vivo followed by brain dissociation, staining and FACS. h, Flow gating and histogram plots of pooled (n = 4 mice per age group) young or aged hippocampi isolated from healthy mice injected with anti-VCAM1-DL488 or IgG-DL488 isotype control as depicted in g. i, Quantification of CD31+VCAM1+ cells isolated from healthy cortices (n = 4 mice per age group, individually measured) (left) and four technical replicates of hippocampi that were pooled from four mice per age group (right). Data are shown as the mean ± s.e.m. *P = 0.0015, two-tailed Student’s t test. j, sVCAM1 ELISA in plasma from young isochronic or heterochronic parabionts following 5 weeks of parabiosis. n = 11 mice per group pooled from two independent experiments. **P = 0.0031, two-tailed Student’s t test. Data are shown as the mean ± s.e.m. k, Confocal images in the DG of VCAM1, lectin and AQP4 of young isochronic or heterochronic parabionts 5 weeks after surgery. Quantification is shown in Fig. 1j. Hoechst labels cell nuclei. Scale bar, 100 µm. n = 8 mice in the young isochronic group and 13 mice in the young heterochronic group from two independent experiments; representative images are shown. l, Box plot of expression levels of classical pan-endothelial and BBB-specific transcripts (n = 272 BECs in total). Minima, maxima, medians and percentiles are listed in Supplementary Table 3 (n = 146 capillary BECs, n = 59 venous BECs, n = 67 arterial BECs pooled from the hippocampi of eight mice). m, Overlay of Vcam1 mRNA levels on corresponding coordinates on the CD31 versus VCAM1 fluorescence intensity plots obtained during FACS sorting. n, Validation of the correlation (Spearman’s rho = 0.704) between protein and mRNA levels for 77 single BECs sorted from both the VCAM1+ and VCAM1 gates. The scatterplot shows VCAM1 fluorescence intensity as measured by FACS and corresponding transcript counts (per million). o, tSNE visualization colored by cell identity (aged versus young) (n = 160 young BECs and 112 aged BECs pooled from the hippocampi of four mice per age group). p, Comparison of VCAM1 expression levels in young and aged hippocampal CD31+ BECs collected from the VCAM1+ gate during FACS sorting (bars, mean; error bars, s.d.; n = 160 young BECs and 112 aged BECs pooled from the hippocampi of four mice per age group). *P = 0.017, two-tailed Mann–Whitney test. q, Violin plots of mRNA levels of Icam1 in all isolated BECs (bottom) and specifically in VCAM1+-enriched BECs (top). Other adhesion molecules, namely Psele and Sele, were not found to be expressed in isolated CD31+ BECs. All BECs: n = 160 young and 112 aged BECs pooled from the hippocampi of four mice per age group; VCAM1+-enriched BECs: n = 56 VCAM1+ young BECs and 44 VCAM1+ aged BECs pooled from the hippocampi of four mice per age group. Minima, maxima, medians and percentiles are listed in Supplementary Table 3. r, Violin plots of tight junction markers in all isolated young and aged BECs. n = 160 young BECs and 112 aged BECs pooled from the hippocampi of four mice per age group. Minima, maxima, medians and percentiles are listed in Supplementary Table 3.

  2. Extended Data Fig. 2 Single-cell transcriptome profiling of VCAM1+-enriched BECs reveals specialized subclusters and plasma from aged individuals upregulates VCAM1 on cultured BECs.

    a, Violin plots of classical arterial (top) and venous (bottom) markers in each cluster. Putative neurogenic secreted factors include Jag1 and Efnb2. Minima, maxima, medians and percentiles are listed in Supplementary Table 3. n = 146 capillary BECs, n = 59 venous BECs, n = 67 arterial BECs pooled from the hippocampi of eight mice. b, Violin plots of various genes related to angiogenesis and Notch signaling in each of the three distinct clusters. Putative neurogenic secreted factors include Vegfc. Minima, maxima, medians and percentiles are listed in Supplementary Table 3. n = 146 capillary BECs, n = 59 venous BECs, n = 67 arterial BECs pooled from the hippocampi of eight mice. c, Representative images of Bend.3 cells immunostained for BBB-specific markers of adherens junctions (AJ) and tight junctions (TJ), specifically β-catenin, claudin-5 and VE-cadherin. All Bend.3 cells and primary BECs were validated with these markers before experimentation; confirmed independently in >10 experiments. Hoechst labels cell nuclei. Scale bar, 100 µm. d, Dose–response graph depicting cultured Bend.3 cells stimulated overnight with increasing concentrations of recombinant mouse TNF-α followed by flow cytometry to quantify the percentage of CD31+VCAM1+ cells. n = 2 pooled samples per condition. e, CD31 and VCAM1 quantification (left) and histogram (right) of Bend.3 cells stimulated overnight with recombinant mouse TNF-α, IL-1β or IL-6 followed by flow cytometry to measure VCAM1. n = 3 biologically independent samples per condition. ****P < 0.0001, one-way ANOVA with Tukey’s post hoc test for group comparisons. Data are shown as the mean ± s.e.m; experiment repeated four times independently with similar results. f, Primary BECs and Bend.3 cells cultured in 10% plasma from young or aged mice (YMP, 3 months old; AMP, 18 months old) or young or aged humans (<25 years or >65 years, YHP/AHP) for 16 h and then stained for VCAM1. Representative images are shown. Scale bar, 100 µm. Each plasma treatment experiment in primary BECs or Bend.3 cells with mouse or human plasma was repeated at least three times independently with similar results. g, Quantification of the percentage of area with VCAM1 staining. Primary BECs treated with plasma from young or aged mice: n = 7 YMP and 9 AMP biologically independent replicates pooled from two experiments. *P = 0.0343. Bend.3 cells with YMP or AMP: n = 4 biologically independent replicates per group derived from different cell flasks. ***P = 0.0003. Bend.3 cells with YHP or AHP: n = 6 biologically independent replicates derived from different cell flasks per group. ****P < 0.0001. P values were determined by two-tailed Student’s t test. Data are shown as the mean ± s.e.m. h, Bend.3 cells cultured in 10% plasma from young or aged mice for 16 h followed by flow cytometry for CD31 and VCAM1. n = 5 biologically independent replicates per group. A graph of CD31 and VCAM1 quantification is shown with a histogram of Bend.3 cells. **P = 0.0082, two-tailed Student’s t test. Data are shown as the mean ± s.e.m. i, Quantification of the percentage of CD31+ Bend.3 cells treated with plasma from young or aged mice and co-stained for CD31 and ICAM1, E-selectin or P-selectin. n = 5 biologically independent replicates per group for ICAM1; n = 6 biologically independent replicates per group for E-selectin and P-selectin. Data are shown as the mean ± s.e.m. Histogram plots are shown to the right of quantification. Not significant; P = 0.2355 (ICAM1), P = 0.1959 (E-selectin), P = 0.0825 (P-selectin), two-tailed Student’s t test. j, Representative images of ICAM1, MECA-99, lectin and Hoechst (to label cell nuclei) of young (3-month-old) mice that received seven retro-orbital injections of pooled plasma from young (3-month-old) or aged (18-month-old) mice over 4 d as described in the schematic in Fig. 3a. n = 10 mice treated with YMP and 11 mice treated with AMP. Scale bar, 100 µm. k, Quantification on the right using n = 4 mice per group. Data are shown as the mean ± s.e.m. Not significant; P = 0.5222, two-tailed Student’s t test. l,m, Quantification in the DG of total BrdU+SOX2+ NPCs in young (3-month-old) mice injected retro-orbitally daily over 5 d (2 µg per injection) with TNF-α (n = 4 mice per group) or with three LPS injections (0.5 mg per kg intraperitoneally) at 28 h, 22 h and 2 h before perfusion (n = 8 mice per group). In each experiment, mice were pulsed with BrdU every 8 h for three injections before perfusion. *P = 0.0194 (TNF-α), *P = 0.0122 (LPS), two-tailed Student’s t test. Data are shown as the mean ± s.e.m.

  3. Extended Data Fig. 3 Assessment of Slco1c1-CreERT2+/–; Vcam1fl/fl young mice, Sudan Black B treatment quenches autofluorescence caused by lipofuscin revealing VCAM1 cerebrovascular specificity and immunodeficient mice exposed to plasma from aged humans over 3 weeks have increased hallmarks of brain aging.

    a, Schematic. n = 5 mice per group. b, Quantification in the DG of VCAM1 from immunostained confocal images. n = 5 mice per group. Data are shown as the mean ± s.e.m. *P = 0.0451, unpaired two-tailed Student’s t test. c, Quantification in the DG of BrdU+ and SOX2+ NPCs and triple-labeled GFAP+ neural stem cells from confocal images of immunostained sections. Scale bar, 100 µm. n = 5 mice per group. Data are shown as the mean ± s.e.m. ***P = 0.007, **P = 0.0227, *P = 0.0038, unpaired two-tailed Student’s t test. d, Quantification in the DG of IBA1 and CD68 from confocal images of immunostained sections. n = 5 mice per group. Data are shown as the mean ± s.e.m. *P = 0.0454, unpaired two-tailed Student’s t test. e, Experimental design. n = 6 Cre and 7 Cre+ mice per group. f, Quantification of the percentage of VCAM1+ area in the lectin+ vasculature of immunostained sections from six Cre and five Cre+ mice per group. ****P < 0.0001, unpaired two-tailed Student’s t test. Data are shown as the mean ± s.e.m. g,h, Quantification of the total number of BrdU+ cells and BrdU+SOX2+ NPCs (g) and average number of DCX+ immature neurons (h) per section in the DG of immunostained sections. n = 6 Cre and 7 Cre+ mice per group. *P = 0.0012, **P = 0021, ***P = 0.0028, unpaired two-tailed Student’s t test. Data are shown as the mean ± s.e.m. i, Quantification of IBA1 and CD68 in the DG of immunostained sections. n = 6 Cre and 7 Cre+ mice per group. Bars, mean; error bars, s.e.m. Staining experiments were repeated twice with similar results; similar mouse experiments using these validated transgenic mice were repeated four times with similar results (Supplementary Table 4). j, Confocal images of brain sections of aged Cre+ or Cre mice treated with tamoxifen in young adulthood (aged 2 months) and aged to 18 months stained for VCAM1 (or IgG isotype control), AQP4 and GFAP. Hoechst labels cell nuclei. Brain sections from aged (18-month-old) mice were treated with Sudan Black B to remove lipofuscin background in the granular and hilus layers of the DG. Sudan Black B treatment removes the majority of lipid-based artifacts typically seen in aged tissues without suppressing immunofluorescent labeling. Scale bar, 100 µm. Experiment repeated three times with similar results. k, Brain sections from aged (18-month-old) Cre+ and Cre mice were immunostained with the regular protocol, without Sudan Black B treatment. Heavy lipofuscin background is present in the Cy3 fluorescence channel. Experiment repeated three times with similar results.

  4. Extended Data Fig. 4 VCAM1 is not expressed in CNS cell types other than BECs in the hippocampus, is increased during neurodegeneration and is deleted in brain endothelium for Slco1c1-CreERT2+/–; Vcam1fl/fl transgenic mice.

    a, Representative 2D and 3D z-stacked high-magnification confocal images (51 slices with an interval of 0.4 µm) of VCAM1 in the granular layer of the DG of the hippocampus of a young (3-month-old) NSG mouse acutely treated with plasma from aged humans. Brain sections were co-stained for DCX and NeuN to label immature and mature granule neurons, respectively. VCAM1 is not expressed in these cell types. Light blue lines outline the granule layer. Experiment repeated three times independently with similar results. 2D scale bar, 50 µm. Two 3D renderings of the 2D images are displayed. 3D scale bar, 20 µm. b, Quantification of VCAM1, AQP4 and lectin, with Hoechst labeling of cell nuclei, in the hippocampus and cerebellum of EAE (multiple sclerosis), Npc1–/– (Niemann Pick disease type C) and Grn–/– (frontotemporal dementia) disease models. EAE: n = 4 naive and 8 EAE induced mice, *P = 0.006, **P = 0.0125; Npc1–/–: n = 6 mice per group, ***P = 0.0274, ****P < 0.0001; Grn–/–: n = 4 mice per group, *****P = 0.0004. P values were determined by unpaired two-tailed Student’s t test. Data are shown as the mean ± s.e.m. c, Representative 2D and 3D z-stacked high-magnification confocal images (51 slices with an interval of 0.4 µm) of VCAM1 in the granular layer of the DG of the hippocampus co-stained with SOX2 and GFAP to label neural stem and progenitor cells (SOX2+GFAP+) and hilus GFAP+ astrocytes. VCAM1 is not expressed in these cell types in the DG. Light blue lines outline the granule layer. Experiment repeated three times independently with similar results. 2D scale bar, 50 µm. Two 3D renderings of the 2D images are displayed. 3D scale bar, 20 µm. d, Cre+ or Cre littermates (3 months old) were treated daily with tamoxifen (i.p., 150 mg per kg) for 5 d followed by 4 d of rest. Mice received three LPS injections (0.5 mg per kg i.p.) at 28, 22 and 2 h before perfusion. Mice also received a retro-orbital injection of fluorescently conjugated mouse anti-VCAM1 monoclonal antibody (100 µg) 2 h before perfusion. FACS gating strategy to analyze single BECs. PI+ dead cells were excluded. Cells negative for CD11a, CD11b, CD45 and Ter-119 were gated to exclude erythrocytes, monocytes/macrophages and microglia. CD13 and ACSA-2 staining were applied to exclude pericytes and astrocytes, respectively. CD31+MECA-99+ cells were defined as the BEC population. e,f, Quantification (e) of flow cytometry (f) that was performed on primary BECs isolated from Cre+ or Cre mice treated as described in d. n = 3 Cre+ or Cre mice received LPS, while one Cre mouse was given PBS vehicle control instead. The VCAM1 gate was set on the basis of a Cre mouse injected with fluorescently conjugated IgG. **P = 0.0011, unpaired two-tailed Student’s t test. Data are shown as the mean ± s.e.m. g, Representative confocal images of cortex and DG for VCAM1 and Hoechst (to label cell nuclei) in LPS-stimulated mice as described in d. Loss of Vcam1 in Cre+ mice, but not Cre mice, in BBB endothelium but not meninges is shown. Experiment repeated three times independently with similar results. Scale bar, 100 µm.

  5. Extended Data Fig. 5 Brain endothelial- and epithelial-specific Vcam1 deletion in young mice mitigates the negative effects of administration of plasma from aged individuals and anti-VCAM1 antibody reduces hallmarks of brain aging in female mice.

    a, Experimental design. n = 8 mice per group. b,c, Representative confocal images (b) and quantification (c) in the DG of VCAM1, MECA-99 and AQP4. Hoechst labels cell nuclei. Scale bar, 100 µm. Arrows indicate VCAM1+ vessels. n = 4 mice per group analyzed. Data are shown as the mean ± s.e.m. ***P = 0.0002, two-way ANOVA with Tukey’s multiple-comparisons test. df, Quantification of the total number of BrdU+ cells, BrdU+SOX2+ NPCs and DCX+ immature neurons in the DG of immunostained sections. n = 8 mice per group. Data are shown as the mean ± s.e.m. *P = 0.0193, **P = 0.0283, ***P = 0.0015, two-way ANOVA with Tukey’s multiple-comparisons test. g,h, Quantification of the total number of surviving EdU+DCX+ immature neurons and EdU+NeuN+ neurons in the DG of immunostained sections. n = 8 mice per group. Data are shown as the mean ± s.e.m. *P = 0.0181, two-way ANOVA with Tukey’s multiple-comparisons test. i,j, Quantification of IBA1 and CD68 in the DG of immunostained sections. n = 8 mice per group. Data are shown as the mean ± s.e.m. ****P < 0.0001 for both, two-way ANOVA with Tukey’s post hoc test. k, Schematic. Aged (18-month-old) C57BL/6J female mice received intraperitoneal injections of a mouse-specific anti-VCAM1 monoclonal antibody or IgG isotype control (9 mg per kg) every 3 d for a total of seven injections. Mice also received BrdU daily (100 mg per kg i.p.) for six consecutive days followed by perfusion 2 d after the last injection. n = 9 IgG-treated and 10 anti-VCAM1 monoclonal-antibody-treated mice per group. l, Quantification of VCAM1 and lectin staining from confocal images in the DG. n = 3 mouse brain sections stained and quantified per group. Data are shown as the mean ± s.e.m. *P = 0.0128, one-way ANOVA with Tukey’s multiple-comparisons post hoc test. m, Quantification of BrdU+ and BrdU+SOX2+ staining from confocal images in the DG. n = 9 IgG-treated and 10 anti-VCAM1 monoclonal-antibody-treated mice per group. Data are shown as the mean ± s.e.m. *P = 0.0325, ***P = 0.0003, unpaired two-tailed Student’s t test. n, Quantification of IBA1 and CD68 staining from confocal images in the DG. n = 9 IgG-treated and 10 anti-VCAM1 monoclonal-antibody-treated mice per group. **P = 0.0008, *P = 0.0427, unpaired two-tailed Student’s t test. Data are shown as the mean ± s.e.m. o, sVCAM1 ELISA of the plasma of young (4-month-old) and aged (18-month-old) female mice. n = 6 mice per group. ****P < 0.0001, unpaired two-tailed Student’s t test. Data are shown as the mean ± s.e.m.

  6. Extended Data Fig. 6 Circulating sVCAM1 does not contribute to inhibitory effects of administration of plasma from aged individuals while anti-VCAM1 antibody prevents the inhibitory effects of plasma from aged humans.

    a, Experimental design. n = 7 mice per group. b, Ponceau S staining showing total protein pulldown from plasma by both IgG and anti-VCAM1 monoclonal antibody conjugated to beads. Experiment repeated three times with similar results. c, Western blot showing human sVCAM1 (93 kDa) pulled down during immunodepletion by anti-human-VCAM1 antibody but not IgG. Experiment repeated three times with similar results. Source data are available online. d, Human sVCAM1 ELISA of depleted plasma. n = 4 mice per group. Data are shown as the mean ± s.e.m. e,f, Representative confocal images (e) and quantification (f) in the DG of VCAM1, MECA-99 and AQP4. Hoechst labels cell nuclei. Scale bar, 50 µm for merged images and 20 µm for the ×4 single-channel VCAM1 images outlined with white boxes. Arrows indicate VCAM1+ vessels. n = 5 mice per group analyzed. Data are shown as the mean ± s.e.m. ***P = 0.0004, **P = 0.0025, one-way ANOVA with Tukey’s multiple-comparisons post hoc test. g, Quantification of the total number of BrdU+ and BrdU+SOX2+ NPCs in the DG of immunostained sections. n = 7 mice per group. Data are shown as the mean ± s.e.m. *P = 0.0237, **P = 0.0123, ***P = 0.0320, ****P = 0.0094, one-way ANOVA with Tukey’s multiple-comparisons post hoc test. h,i, Quantification (h) and representative confocal images (i) of the DG for DCX and Hoechst (to label cell nuclei). Scale bar, 100 µm. n = 5 mice per group analyzed. Data are shown as the mean ± s.e.m. **P = 0.0017, *P = 0.0385, one-way ANOVA with Tukey’s multiple-comparisons post hoc test. j, Quantification of the IBA1 and CD68 staining from confocal images in the DG. n = 7 mice per group. *P = 0.0156, **P = 0.0242, ***P = 0.0034, ****P = 0.0237, P = 0.0546 (PBS as compared to anti-VCAM1 activated microglia counts), one-way ANOVA with Tukey’s multiple-comparisons post hoc test. Data are shown as the mean ± s.e.m. k, Experiment schematic. n = 9 PBS-treated, 8 AHP + IgG-treated and 8 AHP + anti-VCAM1 monoclonal-antibody-treated mice. l, Quantification in the DG of VCAM1 in lectin+ blood vessels from immunostained confocal images. n = 9 PBS-treated, 8 AHP + IgG-treated and 7 AHP + anti-VCAM1 monoclonal-antibody-treated mice analyzed. Data are shown as the mean ± s.e.m. **P = 0.006, one-way ANOVA with Tukey’s multiple-comparisons post hoc test. m, Quantification in the DG of CD68 in IBA1+ stained microglia from immunostained confocal images. n = 9 PBS-treated, 8 AHP + IgG-treated and 8 AHP + anti-VCAM1 monoclonal-antibody-treated mice. Data are shown as the mean ± s.e.m. ***P = 0.0006, **P = 0.0067, one-way ANOVA with Tukey’s multiple-comparisons post hoc test. n,o, Quantification of BrdU+SOX2+ progenitor cells (n) and DCX+ immature neurons (o) from confocal images. n = 9 PBS-treated, 8 AHP + IgG-treated and 8 AHP + anti-VCAM1 monoclonal-antibody-treated mice. Data are shown as the mean ± s.e.m. *P = 0.018, **P = 0.0386, ***P = 0.0344, ****P = 0.0167, one-way ANOVA with Tukey’s multiple-comparisons post hoc test. p, Quantification of the total numbers of EdU+ surviving cells in the DG of immunostained sections. n = 9 PBS-treated, 8 AHP + IgG-treated and 8 AHP + anti-VCAM1 monoclonal-antibody-treated mice. Data are shown as the mean ± s.e.m. ***P = 0.0009, ****P = 0.0002, one-way ANOVA with Tukey’s multiple-comparisons post hoc test. Source Data

  7. Extended Data Fig. 7 BBB integrity is not compromised with aging or conditional deletion of Vcam1.

    a, Quantification of fluorescent signal measured with a microplate reader from homogenized brain tissue samples from mice that were injected with Texas Red-labeled 70-kDa dextran retro-orbitally and perfused with FITC-labeled 2-MDa dextran 3 h after injection. Cre and Cre+ mice were used. n = 3 young (5-month-old) Cre mice, 5 aged (19-month-old) Cre mice, 2 young Cre mice that underwent TBI as a positive control, 3 young Cre control mice not injected with dextran and 3 aged Cre control mice not injected with dextran. Data are shown as the mean ± s.e.m. b, Quantification of the fluorescent signal from homogenized brain tissue samples measured with a microplate reader. Cre and Cre+ mice were used as described in a. n = 3 young Cre mice, 5 aged Cre mice and 5 ‘aged Vcam1-ST’ (19-month-old) mice, which are Cre+ mice that were treated with tamoxifen for 4 d, 2 months before they were killed, and that were infused with dextran before being killed as described in a. Data are shown as the mean ± s.e.m. c,d, Quantification of mean fluorescence intensity from confocal images of tissue sections from mice injected as in a and b. Cre and Cre+ mice were used as described in a. n = 4 young Cre mice, 5 aged Cre mice, 5 aged Vcam1-ST Cre+ mice that were treated with tamoxifen and that were infused with dextran as described in a and b, and 1 young and 1 aged Cre control mice not infused with dextran. Data are shown as the mean ± s.e.m. P = 0.895 (young versus aged), 0.9097 (aged versus aged Vcam1-ST), one-way ANOVA with Tukey’s multiple-comparisons post hoc test. e, Schematic of flow cytometry analysis of various immune cell populations from mouse cortex and hippocampus. f, Flow cytometry gating strategy for individual hippocampal immune cell populations labeled for various immune cell markers, including α4β1 integrins (VLA-4). n-1 was used to gate for VLA-4+ cell populations.

  8. Extended Data Fig. 8 Brain-resident leukocyte composition does not change with aging or conditional deletion of Vcam1.

    a, Mouse model and experimental groups. n = 5 young Cre mice, 9 aged Cre mice, 4 ‘Vcam1-LT’ mice and 4 Vcam1-ST mice. Cre and Cre+ mice were used. The tamoxifen treatment paradigm is described in the schematic. b, Gating plots of CD31+VCAM1+ hippocampal and cortex cells isolated from one LPS-stimulated aged (19-month-old) Cre+ (Vcam1 deletion short term) mouse and one Cre mouse injected with anti-VCAM1-DL488 (r.o.) 2 h before killing to confirm VCAM1 levels on BECs were reduced. One additional Cre mouse was treated with LPS and injected with IgG-DL488 isotype control before killing to serve as a control for VCAM1 gating. cj, Quantification of various cell populations present in young Cre (n = 5), aged Cre (n = 9), aged Vcam1-LT (n = 4) and aged Vcam1-ST (n = 4) mice per group. Data are shown as the mean ± s.e.m. *P = 0.0413, **P = 0.0245, ***P = 0.0429, ****P = 0.0023, one-way ANOVA with Tukey’s multiple-comparisons post hoc test.

  9. Extended Data Fig. 9 VCAM1 and VLA-4 perturbations reduce hallmarks of brain aging.

    a, Experimental design for anti-VLA-4. n = 7 mice per group. b,d, Representative confocal images (b) and quantification (d) of VCAM1, lectin and Hoechst (to label cell nuclei). Scale bar, 100 µm. n = 3 mice per group analyzed. Data are shown as the mean ± s.e.m. c,f, Representative confocal images (c) and quantification (f) in the DG of CD68, IBA1 and Hoechst. Scale bar, 100 µm. n = 7 mice per group. Data are shown as the mean ± s.e.m. *P = 0.0436, **P = 0.0175, two-tailed Student’s t test. e, Quantification of confocal images of the DG of NPCs co-labeled with BrdU and SOX2. n = 7 mice per group. Data are shown as the mean ± s.e.m. g,h, Quantification of IBA1+ and IBA1+CD68+ microglia in DG from the experiment described in Fig. 6f. n = 8 mice per group. Data are shown as the mean ± s.e.m. i–k, Thirteen-month-old NSG mice were injected with anti-VCAM1 monoclonal antibody or IgG every 3 d for 1 month and underwent novel object recognition or fear conditioning tasks during the last week (n = 11 mice per group). i, Quantification of the percentage of time spent exploring objects in the novel object placement task. j, Percentage freezing observed during the training phase. k, The average of trials 3–5 for contextual freezing. Data are shown as the mean ± s.e.m. *P = 0.0485, two-way ANOVA with Sidak’s multiple-comparisons test. There were no significant differences between groups for contextual freezing (P = 0.2722, two-tailed Student’s t test). l–n, Twenty-three-month-old C57BL/6 mice were injected with anti-VCAM1 or IgG every 3 d for 1 month and underwent fear conditioning during the last week (n = 7 PBS-, 12 IgG- and 13 anti-VCAM1-treated mice per group). l, Percentage freezing observed during training. Cued (m) and contextual (n) tests. Data are shown as the mean ± s.e.m. *P = 0.0493, two-way ANOVA with Tukey’s multiple-comparisons test between groups at each time point. Source data are available online. Source Data

  10. Extended Data Fig. 10 Blood from aged individuals inhibits hippocampal NPC activity and activates microglia through VCAM1 at the BBB.

    In young, healthy mice, neurovascular homeostasis is maintained with low expression levels of systemic sVCAM1 and BBB-specific VCAM1, active neurogenesis with neural stem cells (NSCs) differentiating into NPCs, immature neurons and mature neurons, and nonreactive microglia in a low-inflammation environment. During aging or exposure to plasma from aged individuals, we propose that (1) inflammatory factors in plasma from aged individuals (IL-1β and TNF-α, among others) induce arterial and venous BEC activation and upregulation of VCAM1 through their cytokine receptors TNFRSF1A and IL1R1. (2) Venous VCAM1 facilitates tethering, but not transmigration, of leukocytes, which sustain BEC inflammation. (3) Inflamed and activated venous and arterial VCAM1+ brain endothelium relays (unknown) signals to the parenchyma, leading to loss of homeostasis, decline in NPC activity and chronic activation of microglia. (4) Anti-VCAM1 monoclonal antibody protects young brains from the detrimental effects of plasma from aged individuals by reducing BEC-mediated inflammation. (5) Anti-VCAM1 monoclonal antibody rejuvenates aged brains by reducing BEC-mediated inflammation and the reduction mediated by VCAM1+ BECs in NPC proliferation.

Supplementary information

  1. Supplementary Information

    Supplementary Table 4

  2. Reporting Summary

  3. Supplementary Tables

    Supplementary Table 1–3, 5 and 6

Source data

  1. Source Data Fig. 1 and Extended Data Fig. 1

    FPKM values and statistics for differential gene expression analysis of Young and Aged BECs

  2. Source Data Extended Data Fig. 6

    Unprocessed western blots

  3. Source Data Extended Data Fig. 9

    Individual points graphed for Extended Data Fig. 9n

Rights and permissions

To obtain permission to re-use content from this article visit RightsLink.

About this article

Fig. 1: BECs are activated with age, and systemic and cerebrovascular VCAM1 levels increase with age and heterochronic parabiosis.
Fig. 2: Single-cell RNA-seq of VCAM1+-enriched BECs from young and aged mice.
Fig. 3: Administration of blood from aged mice into young mice activates brain vasculature and microglia and reduces hippocampal NPC activity.
Fig. 4: Brain endothelial- and epithelial-specific Vcam1 deletion in young mice mitigates the effects of administration of plasma from aged mice.
Fig. 5: Anti-VCAM1 antibody prevents inhibitory effects from administration of plasma from aged humans and mice in young mice.
Fig. 6: VCAM1 perturbation reverses age-related impairments and improves hippocampal-dependent learning and memory.
Extended Data Fig. 1: Bulk and single-cell transcriptome and proteome profiling of young and aged BECs reveals an increased inflammatory signature with aging.
Extended Data Fig. 2: Single-cell transcriptome profiling of VCAM1+-enriched BECs reveals specialized subclusters and plasma from aged individuals upregulates VCAM1 on cultured BECs.
Extended Data Fig. 3: Assessment of Slco1c1-CreERT2+/–; Vcam1fl/fl young mice, Sudan Black B treatment quenches autofluorescence caused by lipofuscin revealing VCAM1 cerebrovascular specificity and immunodeficient mice exposed to plasma from aged humans over 3 weeks have increased hallmarks of brain aging.
Extended Data Fig. 4: VCAM1 is not expressed in CNS cell types other than BECs in the hippocampus, is increased during neurodegeneration and is deleted in brain endothelium for Slco1c1-CreERT2+/–; Vcam1fl/fl transgenic mice.
Extended Data Fig. 5: Brain endothelial- and epithelial-specific Vcam1 deletion in young mice mitigates the negative effects of administration of plasma from aged individuals and anti-VCAM1 antibody reduces hallmarks of brain aging in female mice.
Extended Data Fig. 6: Circulating sVCAM1 does not contribute to inhibitory effects of administration of plasma from aged individuals while anti-VCAM1 antibody prevents the inhibitory effects of plasma from aged humans.
Extended Data Fig. 7: BBB integrity is not compromised with aging or conditional deletion of Vcam1.
Extended Data Fig. 8: Brain-resident leukocyte composition does not change with aging or conditional deletion of Vcam1.
Extended Data Fig. 9: VCAM1 and VLA-4 perturbations reduce hallmarks of brain aging.
Extended Data Fig. 10: Blood from aged individuals inhibits hippocampal NPC activity and activates microglia through VCAM1 at the BBB.