. Endothelial Cells are Heterogeneous in Different Brain Regions and are Dramatically Altered in Alzheimer's Disease. 2023 Feb 16 10.1101/2023.02.16.528825 (version 1) bioRxiv.

Recommends

Please login to recommend the paper.

Comments

  1. This is an important and comprehensive study from the Bennett Lab that illustrates the diversity of endothelial cells across human brain regions and the complexity of their response to AD pathology. It is one of the first single-cell studies to take a broader view of the human brain’s complexity—here, spatially by analyzing different cortical regions and defining the molecular adaptations in human brain endothelial cells that meet the needs of each region. Thus, it goes beyond most published AD single-cell studies, which analyze one, well-defined cortical region (commonly the dorsolateral prefrontal cortex).

    This work raises intriguing questions about brain vascular dysfunction and whether it tracks with Braak staging, or follows its own pattern. In general, this data could also inform brain region-specific drug delivery targets, or better biomarkers across diseases exhibiting distinct brain region staging patterns.

    It is also one of the first studies to connect molecular changes in human brain endothelial cells with pathology, including amyloid plaques, cerebral amyloid angiopathy, and tau pathology. This provides insightful context for what these changes could mean.

    It was interesting to see fewer AD-associated changes in endothelial cells in vulnerable brain regions, such as the entorhinal cortex, as compared to other regions. There are various interpretations, and it’ll be important to investigate them if we are to understand the role of brain endothelial cells in brain region vulnerability. Likewise, the modest APOE4-associated endothelial changes should be further studied, given prior reports linking APOE4 to endothelial dysfunction (Bell et al., 2012; Rieker et al., 2019; Blanchard et al., 2020; Barisano et al., 2022; Yang et al., 2022). Why there would be almost no transcriptomic changes in arterial and venous endothelial cells with pathology would also be worth looking into.

    Future studies could extend beyond cortical brain regions, include other diverse vascular cell types—such as pericytes and macrophages—and evaluate vascular cell vulnerability in addition to gene dysregulation.

    As with most sequencing studies, it’s still difficult to know what gene expression changes are protective versus pathological—for example, the commonly upregulated proteostasis genes. Still, this rich dataset sets the foundation for such functional studies.

    References:

    . Apolipoprotein E controls cerebrovascular integrity via cyclophilin A. Nature. 2012 May 24;485(7399):512-6. PubMed. Correction.

    . Apolipoprotein E4 Expression Causes Gain of Toxic Function in Isogenic Human Induced Pluripotent Stem Cell-Derived Endothelial Cells. Arterioscler Thromb Vasc Biol. 2019 Sep;39(9):e195-e207. Epub 2019 Jul 18 PubMed.

    . Reconstruction of the human blood-brain barrier in vitro reveals a pathogenic mechanism of APOE4 in pericytes. Nat Med. 2020 Jun;26(6):952-963. Epub 2020 Jun 8 PubMed. Correction.

    . A "multi-omics" analysis of blood-brain barrier and synaptic dysfunction in APOE4 mice. J Exp Med. 2022 Nov 7;219(11) Epub 2022 Aug 30 PubMed.

    . A human brain vascular atlas reveals diverse mediators of Alzheimer's risk. Nature. 2022 Mar;603(7903):885-892. Epub 2022 Feb 14 PubMed.

    View all comments by Andrew C. Yang
  2. This work by the Bennett lab explored region-specific differences in the endothelial cell transcriptome in normal healthy aging and Alzheimer’s brains using snRNA-Sequencing of various cortical regions. In recent years, several studies looked at endothelial genetic changes utilizing the mouse brain (He et al., 2018; Vanlandewijck et al., 2018) and also the human brain (Lau et al., 2020; Yang et al., 2022); however, the regional differences across multiple cortical areas have not been investigated. Interestingly, the authors highlight that 80.5 percent of the isolated endothelial nuclei came from capillaries, mirroring the findings that capillary endothelial cells comprise about 85 percent of the blood-brain barrier (Montagne et al., 2017). Furthermore, when comparing AD to non-AD brains, endothelial cell differentially expressed genes (DEGs) among the cortical areas were identified. Genes related to angiogenesis and vasculogenesis were upregulated in the visual cortices, whereas oxidative stress-related genes were more expressed in the entorhinal cortex. Interestingly, genes relating to the heat-shock protein family were uniformly increased across all examined cortical areas, irrespective of Aβ or tau protein aggregates, suggesting a brain-wide susceptibility to proteostatic stress. Given the importance of the vascular contribution and of region-specific pathology manifestation in AD, determining transcriptomic changes of endothelial cells is of the utmost importance to expand our understanding of the underlying mechanisms.

    This publication identified the endothelial cells of AD brains to show upregulation of protein-folding genes and specific endothelial cell transcriptomic variations in response to cerebral amyloid angiography (CAA) and Aβ plaques. The aim of the study was to focus on cortical areas to follow the Braak staging scheme of tau accumulation. However, investigation of the endothelial cell transcriptome in white matter (WM) areas would also be advantageous due to early AD pathological hallmarks associated with these brain regions, including WM lesions and cerebral blood flow alterations (Gaubert et al., 2021; Korte et al., 2020). Ideally, single-cell RNA-Seq would have been utilized to access the cytoplasmic transcripts as well, which would provide more information regarding other biological pathways involved in AD progression, such as endoplasmic reticulum stress (Ajoolabady et al., 2022; Lindholm et al., 2006). Lastly, the spatial relationship between endothelial cell molecular changes and Aβ plaques/CAA was not explored, limiting our understanding of the distinct endothelial genetic expression to pathology accumulation along the vascular tree.

    This paper adds to the data pool generated by recent publications of human brain vascular atlases (Yang et al., 2022; Vanlandewijck et al., 2018). As opposed to Yang et al. (2022) this study is looking specifically at the distinct brain regions rather than pooling tissue, providing more information on the spatial aspect of the translatome, which has not been achieved so far. It appears that the paper’s sole focus is on the endothelial cell’s genetic signature, brushing slightly on the genes shared by other cell types including the astrocytes, microglia, neurons, and oligodendrocytes. However, endothelial cells are also in a close relationship with pericytes, which did not get much attention here, as opposed to by Yang et al., who identified two distinct types of pericytes, the transport and the matrix subtypes, based on genetic signatures. Furthermore, loss of brain vascular nuclei across various cell types, including endothelial cells, smooth muscle cells, and pericytes, was previously observed, with the matrix pericytes being particular vulnerable, confirming a molecular basis for the structural blood-brain barrier (BBB) breakdown seen in dementia (Montagne et al., 2015; Nation et al., 2019; Montagne et al., 2020). There is growing evidence of early BBB and pericyte dysfunction in AD and other dementias advocating the importance of looking at the pericytome alongside endothelial cell transcriptome (Barisano et al., 2022; Montagne et al, 2015; Procter et al., 2021). Interestingly, the current paper did not observe/mention any changes to the nuclei numbers isolated from healthy aged and AD brains, in contrast to Yang and collaborators. More discussion linking the regional endothelial cell gene expression changes, especially in the entorhinal cortex, and actual BBB functions would have been appreciated. For instance, it was shown using contrast MRI that the medial temporal lobe seems to suffer from BBB dysfunction first with aging (Montagne et al., 2015) and is accelerated with dementia (Nation et al., 2019Montagne et al., 2020). 

    Recently, a molecular atlas of cell types and brain vasculature zonation in mice identified new brain arterio-venous markers of endothelial cells, but also of pericyte and of fibroblast-like cells (Vanlandewijck et al., 2018). The current publication identifies the proportion of arteries, capillaries, and veins within the various cortical areas, but unfortunately, it does not tell us if this ratio changes with AD, which is necessary for the identification of more targeted therapeutics. It would also be advantageous to see if there is any overlap between the human and mouse endothelial as well as mural cell transcriptome in relation to the arterio-venous segmentation among different brain regions.

    Overall, this study confirms the importance of investigating vascular zonation. The vasculature is not one entity but rather multiple vascular beds throughout the arterio-venous axis; brain endothelial cells do have different molecular signatures if we look at arterioles versus capillaries versus venules, and also the transitional zone in between the main beds. This is also true for the very closely associated pericytes, which present with varying subtypes based on their location along the arterio-venous tree, including the ensheathing, thin-strand, and mesh types which have different roles, functions, and also molecular signatures (He et al., 2016; Vanlandewijck et al., 2018). 

    Even though endothelial cells and pericytes are very closely associated and a pericyte role in neurodegeneration is evident, it seems that this component of the neurogliovascular unit is still somewhat overlooked. It becomes clear that it is of the utmost importance to focus in depth on the pericyte transcriptome, i.e., investigating across different vascular beds as well as brain regions. Furthermore, as stated above, we need to start assessing the transcriptome of endothelial cells in other brain regions, including healthy and damaged WM regions, which are a prevalent imaging marker of AD and neurodegeneration. And we need to relate the varying gene signatures to spatial amyloid distribution and immune responses. Lastly, in agreement with the authors, studies looking at the transcriptomic changes of endothelial and associated cells of the vasculature are required in people who have an elevated probability of developing AD (e.g., APOE4 carriers) to uncover the effect of this risk allele on disease development and progression.

    References:

    . ER stress and UPR in Alzheimer's disease: mechanisms, pathogenesis, treatments. Cell Death Dis. 2022 Aug 15;13(8):706. PubMed.

    . Blood-brain barrier link to human cognitive impairment and Alzheimer's Disease. Nat Cardiovasc Res. 2022 Feb;1(2):108-115. Epub 2022 Feb 7 PubMed.

    . Topographic patterns of white matter hyperintensities are associated with multimodal neuroimaging biomarkers of Alzheimer's disease. Alzheimers Res Ther. 2021 Jan 18;13(1):29. PubMed.

    . Analysis of the brain mural cell transcriptome. Sci Rep. 2016 Oct 11;6:35108. PubMed.

    . Single-cell RNA sequencing of mouse brain and lung vascular and vessel-associated cell types. Sci Data. 2018 Aug 21;5:180160. PubMed.

    . Cerebral blood flow decrease as an early pathological mechanism in Alzheimer's disease. Acta Neuropathol. 2020 Dec;140(6):793-810. Epub 2020 Aug 31 PubMed.

    . Single-nucleus transcriptome analysis reveals dysregulation of angiogenic endothelial cells and neuroprotective glia in Alzheimer's disease. Proc Natl Acad Sci U S A. 2020 Oct 13;117(41):25800-25809. Epub 2020 Sep 28 PubMed.

    . ER stress and neurodegenerative diseases. Cell Death Differ. 2006 Mar;13(3):385-92. PubMed.

    . Blood-brain barrier breakdown in the aging human hippocampus. Neuron. 2015 Jan 21;85(2):296-302. PubMed.

    . Alzheimer's disease: A matter of blood-brain barrier dysfunction?. J Exp Med. 2017 Nov 6;214(11):3151-3169. Epub 2017 Oct 23 PubMed.

    . APOE4 leads to blood-brain barrier dysfunction predicting cognitive decline. Nature. 2020 May;581(7806):71-76. Epub 2020 Apr 29 PubMed.

    . Blood-brain barrier breakdown is an early biomarker of human cognitive dysfunction. Nat Med. 2019 Feb;25(2):270-276. Epub 2019 Jan 14 PubMed.

    . Interplay between Brain Pericytes and Endothelial Cells in Dementia. Am J Pathol. 2021 Nov;191(11):1917-1931. Epub 2021 Jul 27 PubMed.

    . Pericytes of the neurovascular unit: key functions and signaling pathways. Nat Neurosci. 2016 May 26;19(6):771-83. PubMed.

    . A molecular atlas of cell types and zonation in the brain vasculature. Nature. 2018 Feb 14; PubMed.

    . A molecular atlas of cell types and zonation in the brain vasculature. Nature. 2018 Feb 14; PubMed.

    View all comments by Axel Montagne
  3. The human brain is vascularized by over 400 miles of specialized blood vessels to meet the vast metabolic demands of the central nervous system, including delivery of oxygen and nutrients, removal of carbon dioxide and waste products, and immune cell trafficking. As these metabolic demands vary over space and time according to neuronal activity, the cerebrovasculature is endowed with enormous structural and functional diversity that remains far less understood than are peripheral vascular systems. Indeed, concepts such as the “neurovasculome,” which refers to the arteries, capillaries, veins, and lymphatics that make up the entire vascular supply to the brain parenchyma and meninges, and the “neurovascular complex,” which refers to a diverse set of functional units comprising vascular, perivascular, and neural cells across extracranial, intracranial, and intraparenchymal branches, have been introduced to reflect the exceptional functional, structural, and molecular diversity of the cerebrovasculature. Dysfunction of the cerebrovasculature is intimately associated with several neurological disorders, especially Alzheimer’s disease (AD); autopsy studies have revealed cerebrovascular alterations in over half of patients diagnosed with AD. Thus, considerable efforts are being made to characterize the human neurovasculome. Advances in sequencing methods can provide new insights into the complexity of transcriptional heterogeneity across the cerebrovascular tree. For example, endothelial cells display a transcriptomic continuum that shifts across zones of the artery-arteriole-capillary-venule-vein network.

    As reported in this elegant paper, Bryant et al. used single-nucleus RNA sequencing (snRNA-Seq) to demonstrate an additional axis of heterogeneity in endothelial cell transcriptomic profiles, namely, that of regional variation. From frozen postmortem tissue, the authors generated snRNA-Seq data from a total of 32 donors (16 high-pathology AD donors at Braak stage V/VI, five intermediate-pathology AD donors at Braak III/IV, and 11 low-pathology AD donors at Braak 0-I) across five cortical regions—entorhinal cortex, inferior temporal gyrus, prefrontal cortex, visual association cortex, and primary visual cortex—to produce a dataset of 19,271 genes measured in 51,586 endothelial cell nuclei.

    Analysis of low-pathology brains shows that although there were no clear differences in endothelial cell profiles across regions, there were region-specific differences in the relative proportions of vascular zones as well as in potential region-specific marker genes of endothelial cells. For example, APOE was identified as a candidate endothelial cell gene in the entorhinal cortex but not in the other regions examined.

    Comparison of high-pathology with low-pathology profiles revealed 936 upregulated and 962 downregulated genes in high-pathology cases, with heat-shock family proteins being the most common class. Pathway analysis showed that high-pathology endothelial cells were more engaged in protein folding, maintenance of the blood-brain barrier, and leukocyte adhesion, whereas lipid and glycoprotein metabolism pathways were downregulated. Interestingly, the ATP binding cassette (ABC)A1, which is responsible for ApoE and ApoA-I lipidation, was reduced in high-pathology AD. Further, protein folding, oxidative phosphorylation, cellular response to heat, apoptosis signalling, vasculogenesis, aging, and cytokine production pathways were enriched across all five regions in high-pathology cases. Blood-brain barrier maintenance was enriched only in the prefrontal cortex, inferior temporal gyrus, and somewhat in primary visual cortex, but not in entorhinal cortex or secondary visual cortex.

    AD-related endothelial transcriptomic changes also increased along the Braak stage continuum but showed limited overlap with region-specific genes. Analysis of how endothelial cell genes associate with AD pathological burden revealed six overall trends suggestive of distinct temporal relationships with AD disease progression. Intriguingly, CAA-related changes were found to be distinct from changes associated with parenchymal amyloid plaques, and amyloid had a greater effect on gene expression of endothelial cells than did either phosphorylated tau or presence of the APOE4 allele, although the number of APOE4 individuals was too low to fully address the role of APOE genotype. Genes associated with other cerebrovascular insults, such as stroke, white-matter hyperintensities, or cerebral small vessel disease, were not enriched in the present dataset. Finally, they found that 32 endothelial-specific genes, coding for secreted proteins, were elevated in high-pathology cases, which, if confirmed, might lead to novel biomarkers for AD diagnosis.

    Overall, Bryant et al. found that brain endothelial cells show distinct profiles across different cortical regions and that AD pathology affects these profiles, especially in capillaries. These regional differences may help to explain regional vulnerability to AD pathology and will form a useful guide for future studies that focus on cerebral endothelial cell pathophysiology.

    The pivotal role of cerebrovascular endothelial cells in neurological health is beginning to receive appropriate attention, which may be driven in part by vascular adverse events associated with anti-amyloid immunotherapies. In addition to the distinct changes induced by parenchymal and vascular amyloid pathology, it will be important to evaluate how endothelial cell transcriptomic profiles differ in the presence of arteriolosclerosis or venous collagenosis, which are common pathologies observed in cerebral small vessel disease. Understanding how cerebral endothelial cells respond to factors that increase risk of cognitive decline, such as hypertension, Type 2 diabetes, and ischemic-reperfusion injury, may also yield significant insights into the cerebrovascular pathophysiological pathways that affect blood-brain barrier dysregulation, vascular inflammation, and drainage of interstitial and cerebrospinal fluids.

    Considering that women are nearly twice as likely as men to develop AD, future studies should be designed to have sufficient power to address how biological sex influences the endothelial cell transcriptome. Finally, how cerebral endothelial cell transcriptomic profiles may vary by age in humans is not yet known, and may provide additional insights into potential targetable pathways of interest.

    View all comments by Jerome Robert
  4. This paper provides an in-depth analysis of vascular endothelial cell transcriptional state with regional resolution in aged non-AD and AD brains. Bryant et al. importantly find heterogeneity across five different brain regions in aged non-AD, regional changes in AD, and distinct endothelial cell responses to cerebral amyloid angiopathy (CAA) versus amyloid plaque.

    This dataset is highly relevant as we consider the regional heterogeneity of disease pathology and how this affects the function of endothelial cells in maintaining the blood-brain barrier (BBB), regulating blood flow, as well as crosstalk between neighboring pericytes and astrocytes. Notably, high pathology donors showed upregulated heat shock family proteins and inflammatory cytokines, as well as dysregulated metabolic pathways based on gene-set enrichment analysis. These findings suggest endothelial cell stress responses may contribute to neuroinflammation and metabolic distress. Interestingly, a subset of transcriptional changes was not unique to endothelial cells, being similarly observed in microglia.

    As our understanding of the cell-type specific biological responses that occur in AD and in aging becomes deeper with application of single-cell and single-nucleus RNA-Sequencing technology, this opens new avenues for potential therapeutic approaches. This study also provides further insights into the biological relevance of AD genome-wide association study risk variants as we learn of their altered expression in endothelial cells (i.e. PICALM and CLU). Since we know relatively little about the biology of vascular endothelial cells and their response to aging and disease, it is critical to delve deeper, given the presence of vascular deposited plaque and the challenges in its safe removal with anti-amyloid therapies. This study also highlights a need for researchers to exert caution when selecting brain samples for analysis, because a whole-brain assessment does not provide the full picture that assessing multiple brain regions does.

    A next big question this study raises is whether any druggable pathways or targets arise from these findings that could modulate endothelial cell biology. Additionally, given that numerous neurological diseases have been associated with pathological changes in the BBB that affect both function and integrity of the neurovascular unit, these current findings provide a basis to help generate new hypotheses around pathology-specific biological pathways that may be implicated. Finally, a key question to address is if regional heterogeneity influences transport of therapeutics differently across brain regions, which could affect efficacy as the AD therapeutic landscape broadens.

    View all comments by Joy Zuchero
  5. In this study, the authors conducted a high-resolution analysis to investigate the transcriptomic changes in endothelial cells across different brain regions in Alzheimer’s disease at the single-cell level. Their results provide valuable insights into the biological changes that occur in the brain vasculature in AD, and potentially offer a molecular basis for the causal relationship between brain vasculature dysregulation and AD (Fisher et al., 2022; Kelleher and Soiza, 2013). The authors present a well-designed and informative dataset that delineates the overall transformation of endothelial cells in AD, which enables a higher-resolution examination of this cell type.

    This study identified conventional endothelial markers shared across multiple brain regions, such as FLT1, CLDN5, and ABCB1. The discovery of region-specific genes clarifies the subtle differences in the biological processes executed by endothelial cells in various brain regions, such as their tendency to positively regulate cytokine production in the entorhinal cortex and organize the microtube cytoskeleton during mitosis in the prefrontal cortex.

    Regarding AD, our previous research on single-nucleotide RNA sequencing of the AD brain revealed the induction of a subpopulation of angiogenic endothelial cells characterized by increased expression of genes associated with angiogenesis and immune response pathways in AD pathogenesis (Lau et al. 2020). With their expanded data collection, the authors demonstrate that while some biological processes, such as protein folding and vasculogenesis, are generally strengthened in endothelial cells in different brain regions in AD, they maintain specific patterns in activated pathways within each group—for example, blood–brain barrier maintenance in the inferior temporal gyrus and prefrontal cortex as well as the regulation of vascular permeability in the prefrontal and primary visual cortices. The authors also observed a heterogeneous changing pattern in the expression of genes that are related to diverse biological processes during AD progression, revealing another layer of complexity in the dynamic changes in gene regulation in AD.

    In addition, the authors identified differentially expressed genes associated with risk factors such as Aβ and cerebral amyloid angiopathy, thereby demonstrating the impacts of different neuropathological factors on gene regulation. Specifically, they showed that Aβ has the greatest effect on endothelial gene expression, regardless of the brain region. Thus, these disease factors might function via various pathways, which should be validated in future work.

    This comprehensive study emphasizes the heterogeneity of brain endothelial cells in both normal aged individuals and patients with AD, demonstrating diverse patterns of alteration during AD progression. It also sheds light on the importance of considering the spatiotemporal deviation of the brain when examining brain-related malfunctions. Despite this study’s valuable findings—namely the alterations of the brain vasculature and identification of capillaries as the most affected vessels in AD—there is still a need for more detailed quantification of endothelial heterogeneity and investigation of the dynamic changes in vasculature transformation during AD progression. Meanwhile, this study’s findings and dataset enable the exploration of whether the level of endothelial heterogeneity is linked to the severity of AD and how risk factors (e.g., Aβ and cerebral amyloid angiopathy) are involved in this process. Furthermore, this study provides valuable information on the cell types other than endothelial cells, which could further inform future investigations into the interactions between endothelial cells and other brain components.

    References:

    . Pathological changes within the cerebral vasculature in Alzheimer's disease: New perspectives. Brain Pathol. 2022 Mar 14;:e13061. PubMed.

    . Evidence of endothelial dysfunction in the development of Alzheimer's disease: Is Alzheimer's a vascular disorder?. Am J Cardiovasc Dis. 2013 Nov 1;3(4):197-226. PubMed.

    . Single-nucleus transcriptome analysis reveals dysregulation of angiogenic endothelial cells and neuroprotective glia in Alzheimer's disease. Proc Natl Acad Sci U S A. 2020 Oct 13;117(41):25800-25809. Epub 2020 Sep 28 PubMed.

    View all comments by Nancy Ip

Make a Comment

To make a comment you must login or register.

This paper appears in the following:

News

  1. Brain Endothelial Cells Are Diverse, Perturbed by Amyloid