Distinct changes in all major components of the neurovascular unit across different neuropathological stages of Alzheimer's disease

Tunahan Kirabali; Ruslan Rust; Serena Rigotti; Alessandro Siccoli; Roger M Nitsch; Luka Kulic

doi:10.1111/bpa.12895

. 2020 Sep 16;30(6):1056–1070. doi: 10.1111/bpa.12895

Distinct changes in all major components of the neurovascular unit across different neuropathological stages of Alzheimer's disease

Tunahan Kirabali ^1,^✉, Ruslan Rust ¹, Serena Rigotti ^1,², Alessandro Siccoli ^1,³, Roger M Nitsch ^1,⁴, Luka Kulic ^1,⁵

PMCID: PMC8018068 PMID: 32866303

Abstract

In the brain capillaries, endothelial cells, pericytes, astrocytes and microglia form a structural and functional complex called neurovascular unit (NVU) which is critically involved in maintaining neuronal homeostasis. In the present study, we applied a comprehensive immunohistochemical approach to investigate the structural alterations in the NVU across different Alzheimer's disease (AD) neuropathological stages. Post‐mortem human cortical and hippocampal samples derived from AD patients and non‐demented elderly control individuals were immunostained using a panel of markers representing specific components of the NVU including Collagen IV (basement membrane), PDGFR‐β (pericytes), GFAP (astrocytes), Iba1 (microglia), MRC1 (perivascular macrophages) and lectin as an endothelial cell label. Astrocytes (GFAP) and microglia (Iba1) were quantified both in the whole visual‐field and specifically within the NVU, and the sample set was additionally analyzed using anti‐tau (AT8) and three different anti‐Aβ (clones G2‐10, G2‐11, 4G8) antibodies. Analyses of lectin labeled sections showed an altered vascular distribution in AD patients as revealed by a reduced nearest distance between capillaries. Within the NVU, a Braak‐stage dependent reduction in pericyte coverage was identified as the earliest structural alteration during AD progression. In comparison to non‐demented elderly controls, AD patients showed a significantly higher astrocyte coverage within the NVU, which was paralleled by a reduced microglial coverage around capillaries. Assessment of perivascular macrophages moreover demonstrated a relocation of these cells from leptomeningeal arteries to penetrating parenchymal vessels in AD patients. Collectively, the results of our study represent a comprehensive first in‐depth analysis of AD‐related structural changes in the NVU and suggest distinct alterations in all components of the NVU during AD progression.

Keywords: GFAP, Iba1, PDGFR‐β, MRC1, NVU, vascular dysfunction

We applied a comprehensive immunohistochemical approach to investigate the structural alterations in the neurovascular unit (NVU) across different Alzheimer's disease (AD) neuropathological stages. We found alterations in all major components of the NVU during AD progression and further investigated the association of these alterations with the neuropathological hallmarks of the AD.

graphic file with name BPA-30-1056-g004.jpg

Introduction

Alzheimer's disease (AD) is defined by its pathological hallmarks, beta‐amyloid (Aβ) plaques and neurofibrillary tangles composed of tau protein (38). These pathologies alone, however, are not sufficient to explain the complexity of all disease aspects. The two‐hit vascular hypothesis implies that the effects of an impaired cardiovascular system on cognition and vascular changes in disease progression have to be taken into account to fully elucidate the pathophysiology of AD (36, 53). According to this hypothesis, cerebrovascular damage can directly lead to neurodegeneration, while it can also show its effects through the enhanced the accumulation of Aβ in AD (36, 70). Structural and functional deficits in the brain arteries, triggered by the vascular accumulation of Aβ in these vessels, are well‐characterized and have been extensively studied both in transgenic animal models of AD and in humans (11, 21, 23, 28, 29, 57, 60, 64). Although capillaries play a vital role in maintaining brain homeostasis, vascular impairments at the capillary level still require further characterization. Within the capillaries, endothelial cells, pericytes, astrocytic endfeet and processes of microglia around the vessels form the neurovascular unit (NVU) (70). Components of the NVU play important roles in maintaining vital brain functions, such as clearance of waste products, regulation of cerebral blood flow and formation of the blood‐brain barrier (BBB) (70).

Cerebrovascular endothelial cells form the blood‐brain barrier (BBB) through connections of tight and adherent junction proteins which made them the focus of NVU‐related changes in aging, AD and other neurodegenerative diseases (52). In AD, BBB disruption is characterized by the accumulation of blood‐derived molecules (fibrinogen, albumin, IgG) in the brain parenchyma and reduced expression of endothelial cell‐specific junction proteins (claudin‐5, occludin, ZO‐1, VE‐cadherin) (31, 33, 36, 66). In addition, the activation of endothelial cells characterized by an increased expression of cell adhesion molecules (VCAM‐1, ICAM‐1, E‐selectin, P‐selectin) was also shown to induce BBB impairment (67). Pericytes, vascular mural cells that wrap around the capillaries, are also vital for a healthy cerebrovascular system and the integrity of the BBB (65). In contrast to peripheral vessels, brain capillaries are enriched in pericytes and characterized by a higher pericyte to endothelial cell ratio (65). Pericyte depletion was shown to exacerbate AD pathology in transgenic animals and a reduction in pericyte coverage was observed in AD patients in post‐mortem studies (12, 24, 49). Astrocytes within the NVU interact with endothelial cells, vascular mural cells and neurons, and are essential in maintaining the connections between neuronal circuits and vasculature. Astrocytes regulate the water homeostasis and glymphatic clearance with the aquaporin 4 (AQP4) water channels at the astrocytic end‐feet, covering most of the abluminal side of cerebral microvessels (40, 42, 55). Reductions in AQP4 coverage in astrocytic end‐feet of AD patients have recently been identified and associated with impairments in Aβ clearance (68). Although pericytes were also reported to secrete Apolipoprotein E (APOE) (62), astrocytes are known as the main supplier of APOE within the cerebral vasculature (27). Therefore, astrocytes could have an indirect effect on cerebral microvessels and other components of NVU through APOE secretion. Microglia, brain resident immune cells of the myeloid lineage, are also part of the NVU and interact with its other components (22). BBB disruption and extravasation of blood proteins were shown to induce microglial responses and inflammatory processes in the brain (48). Moreover, microglia were suggested to induce detrimental effects by releasing pro‐inflammatory cytokines or matrix‐metalloproteases (MMPs) into the vasculature (69). Nevertheless, the exact mechanisms of microglial effects on vascular impairment and whether microglial activation is the cause or the consequence of BBB disruption and cerebrovascular dysfunction remain elusive. In addition to microglia, perivascular macrophages (PVM) represent another major type of innate immune cells localized around leptomeningeal vessels and in penetrating brain vessels within the perivascular spaces (9). Considering their phagocytic capacity and enriched scavenger receptor expression, these cells were implicated to have a role in Aβ clearance and waste removal in the brain (9, 14). However, they were also shown to promote neurovascular dysfunction via Aβ‐induced reactive oxygen species (ROS) production in AD (37). In addition to Aβ‐associated functions, PVM have been suggested to have a role in BBB regulation and antigen presentation at the central nervous system (CNS) boundaries (22). Further mechanistic and post‐mortem studies are required to understand their exact roles and contributions to AD pathology.

In this study, we aimed to characterize the histopathological changes in the NVU in the course of AD by post‐mortem human cortex and hippocampus samples and analyzed all components of the NVU in the same sample set. Endothelial cells, pericytes, astrocytes, microglia and PVM within the NVU were visualized with immunostainings/labels and quantified across different Braak stages. In addition, we assessed tau and beta‐amyloid pathology in detail and further analyzed the correlations of NVU changes with the main pathological hallmarks of AD. By analyzing the NVU as an entity, here for the first time, we provide a comprehensive characterization of capillary level changes and show alterations in all components of the NVU during AD progression.

Methods

Post‐mortem human brain samples

Paraffin‐embedded human hippocampus (n = 28) and cortex (gyrus frontalis medius) (n = 31) tissue blocks were provided by the Netherlands Brain Bank (NBB), the Netherlands Institute for Neuroscience, Amsterdam, the Netherlands. Written informed consent for a brain autopsy and the use of the material for research were obtained by NBB. Distribution of age, gender, diagnostic status, Braak neurofibrillary tangle stage, APOE4 carrier status and Consortium to Establish a Registry for Alzheimer's Disease (CERAD) score of the sample set is listed in Table 1 (2, 3, 34). Post‐mortem cases included in the study had comparable post‐mortem delays. For the hippocampal samples, the average post‐mortem delay was 6 h 40 minutes [±3 h 48 minutes, standard deviation (SD)]. For the cortex samples, the average post‐mortem delay was 6 h 38 minutes (±2 h 44 minutes, SD).

Table 1.

Paraffin‐embedded post‐mortem sample information.

Braak stage	Number of subjects	Average age at death^*	Gender^†	APOE4 Status^‡	CERAD Score^§	Diagnosis^¶
Hippocampus
1	6	82.5 ± 5.5	F (5) M (1)	APOE4+ (2) APOE4− (4)	O(2), A(3), B(1)	NDCNTRL
2	5	85.6 ± 1.75	F (4) M (1)	APOE4+ (1) APOE4− (3)	A(1), B(3), C(1)	NDCNTRL
3	3	82 ± 5.7	F (1) M (2)	APOE4+ (1) APOE4− (1)	A(1), C(2)	NDCNTRL
4	6	92.3 ± 3.3	F (5) M (1)	APOE4+ (2) APOE4− (4)	C	NDCNTRL(2), AD(4)
5	4	82.7 ± 8.8	F (4)	APOE4+ (3) APOE4− (1)	C	AD
6	4	90 ± 2.9	F (4)	APOE4+ (3) APOE4− (1)	C	AD
Cortex
0	5	69.8 ± 21.4	F (3) M (2)	APOE4+ (0) APOE4− (3)	O(2), A(1), B(1)	NDCNTRL(3), MS(2)
1	4	87.7 ± 2.5	F (2) M (2)	APOE4+ (0) APOE4− (4)	O(1), A(3)	NDCNTRL
2	3	85.3 ± 0.5	F (1) M(2)	APOE4+ (1) APOE4− (1)	A(1), B(3), C(2)	NDCNTRL
3	5	86.6 ± 3.0	F (2) M (3)	APOE4+ (2) APOE4− (2)	A(1), B(2), C(3)	NDCNTRL
4	6	89.3 ± 5.0	F (4) M (2)	APOE4+ (3) APOE4− (3)	B(2), C(4)	NDCNTRL(2), AD(4)
5	3	84.3 ± 10.8	F (3)	APOE4+ (2) APOE4− (1)	C	AD
6	5	89.2 ± 3.0	F (5)	APOE4+ (4) APOE4− (1)	C	AD

Open in a new tab

Mean age ± SD in years.

^†

F = Female; M = Male.

^‡

APOE4+ = APOE4 carrier; APOE4− = APOE4 non‐carrier.

^§

CERAD = Consortium to Establish a Registry for Alzheimer's Disease.

^¶

NDCNTRL = Non‐demented control; AD = Alzheimer's disease patient; MS = Multiple Sclerosis.

Immunohistochemistry/Immunofluorescence

5 μm thick sections from the cortex and hippocampus were mounted on slides and deparaffinized. For deparaffinization, sections were immersed in Xylol, 100% ethanol, 95% ethanol and 70% ethanol in respective order for 3 minutes in each solvent. After deparaffinization, sections were boiled in citrate buffer as previously described (24). After a 20 minutes cooling period, sections were blocked with 10% horse serum in PBS containing 0.2% Triton‐X 100 (PBS‐T). Following blocking, primary antibodies were diluted in 5% horse serum in PBS‐T and sections were incubated with primary antibodies overnight at 4°C. All primary antibodies and dilutions used in the experiments are listed in Table S1. For astrocyte, microglia and perivascular macrophage immunostainings, biotinylated lectin (Ulex europaeus agglutinin I/Vectorlabs) was used as a co‐staining (1:500 dilution). After washing with PBS, secondary antibodies (Jackson Immunoresearch, donkey anti‐mouse/rabbit/goat Cy3, donkey anti‐rabbit Alexa488 and streptavidin‐Alexa 488) were prepared in 1:200 dilution in PBS‐T and sections were incubated with secondary antibodies for 2 h at room temperature. After a brief wash, sections were incubated in DAPI working solution and Sudan Black working solution as previously described (24). Sections were mounted with a water‐based mounting medium and kept at 4°C for further analysis.

For DAB (3,3′‐diaminobenzidine) based anti‐Aβ stainings, sections were deparaffinized as outlined above. Prior to citrate buffer pre‐treatment, sections were incubated in 95% formic acid for 5 minutes and afterward in 2% hydrogen peroxide solution in PBS for 10 minutes at room temperature. After blocking with 10% horse serum in PBS, sections were incubated overnight at 4°C with anti‐Aβ antibodies (listed in Table 1) diluted in 5% horse serum in PBS. After washing with PBS, secondary antibodies (VECTASTAIN anti‐mouse biotinylated, 1:200 dilution) were added and samples were incubated for 2 h at room temperature. Following additional washing steps, sections were incubated with ABC solution from VECTASTAIN DAB kit and incubated for 1 h at room temperature. DAB working solution was prepared by 1:10 dilution of DAB in stable peroxide substrate buffer and sections were incubated with DAB working solution for 10 minutes before they were dehydrated by placing them in distilled water, 70% ethanol, 95% ethanol, 100% ethanol and xylol in respective order. After mounting with xylol‐based mounting medium, sections were kept at room temperature for further analysis.

Quantification

From each subject, two sections per brain region which were at least 50 μm apart from each other were immunostained as outlined above. From each section, 10 random pictures were taken by a Leica DM4000B microscope and were used for further quantification.

Platelet‐derived growth factor receptor‐β (PDGFR‐β) quantification

Previously, with a sample set mostly overlapping with the samples used in this study, we quantified the pericyte coverage with a manual counting strategy (24). In this study, a semi‐automated, area‐based quantification strategy was used. Images were converted to 8‐bit binary and the area covered by PDGFR‐β immunostaining was quantified for each image after applying a constant threshold to remove background and non‐specific signals using Image J software. Vessel density (area covered by Collagen IV immunostaining) in each picture was calculated by the same workflow and pericyte coverage was reported as the ratio between the area covered by PDGFR‐β and Collagen IV. Pericyte coverage is defined as the PDGFR‐β+ area within the Collagen IV+ area and includes both the pericyte processes and cell bodies. In order to exclude smooth muscle cells from the analysis, vessels with more than 10 μm diameter were excluded from the quantification manually. The mean of 20 visual fields was reported as % pericyte coverage for each subject.

Glial fibrillary acidic protein (GFAP) quantification

Images were converted to 8‐bit binary and the area covered by GFAP was quantified for each image after applying constant threshold using Image J software. The means of 20 visual fields were reported as %GFAP+ area for each subject.

Astrocytic endfeet coverage was quantified by analyzing the GFAP immunoreactivity associated with the vasculature, similar to earlier studies (17, 41). Immunoreactivity of GFAP within the NVU was quantified by lectin co‐staining images and a custom‐made Image J macro. Briefly, images were converted to 8‐bit binary and a constant threshold was applied. After applying the threshold, 8‐bit binary images of lectin labeling were enlarged using the “Dilate” function in Image J. After this step, the area within the enlarged vessel walls, immunostained by lectin, was defined as the NVU. In order to specifically measure GFAP immunostaining within this area, enlarged lectin labeling images were converted to mask, images were merged together and only the area co‐localizing within the enlarged lectin labeled vessels was measured. This area was divided by the area covered by enlarged lectin labeling and reported as %GFAP in NVU. The representative workflow for the quantification is shown in Figure S1.

Ionized calcium‐binding adaptor molecule 1 (Iba 1) quantification

Images were converted to 8‐bit binary and the area covered by Iba1 immunostaining was quantified for each image after applying constant threshold using Image J software. For each visual field, the number of microglia was quantified by manually counting Iba1 and DAPI double‐positive cells. By dividing the area covered by Iba1 by the number of microglia, the area covered per microglia was determined for each visual field. The mean of 20 visual fields was reported as %Iba1 per microglia for each subject.

The area covered by microglia in the NVU was quantified by the same Image J macro as explained above for GFAP quantification. Results were reported as %Iba1 in NVU.

Mannose receptor C‐type 1 (MRC1) quantification

MRC1 and DAPI double‐positive cells were manually counted in each visual field separately for brain parenchyma and leptomeningeal vessels. The mean of all visual fields was reported as the number of MRC1+ cells in parenchymal and leptomeningeal vessels.

Lectin quantification

Vessel area fraction, length, branching and nearest distance were quantified using lectin labeling with an automated ImageJ script as previously described for brain and retinal vasculature (44, 45, 46, 47). Briefly, for area fraction, the area covered by lectin labeling was quantified using ImageJ after applying a constant threshold. The vascular length was quantified using the “skeleton length” tool and number of branches was assessed using the “analyze skeleton” tool. The nearest distance between vessels was calculated using the nearest‐neighbor distance tool. For the vascular length and branches, results were normalized per mm² of brain tissue. In the case of lectin quantification, size‐based elimination for the vessels was not applied, therefore, both capillaries and larger vessels were included in the analysis.

Tau quantification

Images were converted to 8‐bit binary and the area covered by phosphorylated tau (AT8+ area) was quantified for each image after applying constant threshold using Image J software. The means of 20 visual fields were reported as %Tau for each subject.

Aβ plaque quantification

Distinct to other markers, for Aβ plaques, adjacent sections from the same subject were immunostained with three different anti‐Aβ antibodies by a DAB‐based immunohistochemistry protocol: C‐terminal‐specific anti‐Aβ40, C‐terminal‐specific anti‐Aβ42 and mid‐domain anti‐Aβ (4G8) antibody. Pictures of full slides were taken using the MIRAX slide scanner. By Panorama viewer software, three images from the same subject were synchronized and 20 random pictures were taken at the same positions in adjacent slides. Images were converted to 8‐bit binary and the area covered by each anti‐Aβ immunostaining was quantified for each image after applying constant threshold with Image J software. The mean of 20 visual fields was reported as %area covered for each immunostaining.

Statistical analysis

For all statistical analyses, R Studio and GraphPad Prism software were used. The assumption of normal distribution was tested with the Shapiro–Wilk test. If the assumption of normality is true, differences between the two groups were tested with a two‐tailed Student's t‐test. Otherwise, differences between the two groups were tested with a non‐parametric two‐tailed Mann–Whitney test. When comparing the differences between Braak stages, the lowest Braak stage (stage 0 in the cortex, stage 1 in the hippocampus) was used as a control and P‐values were calculated compared to the control column. After confirming the normal distribution, P‐values were calculated with one‐way ANOVA followed by Tukey's test for multiple comparisons. For Braak stage correlations, Spearman correlation coefficients (r) and two‐tailed P‐values with 95% confidence intervals were calculated and reported. For correlation matrices, the “corrplot” package in R was used. In the correlation matrices, Spearman correlation coefficients (r) and two‐tailed P‐values with 95% confidence intervals were calculated and reported with heatmaps. Additional statistical parameters for the area‐based quantifications are shown in Table S2. The standard error of the mean (SEM) was calculated for each post‐mortem case by dividing the SD by the square root of the number of measurements. Average SEM ± SD was shown for each immunostaining to illustrate the robustness of the quantification.

Results

Vessel distribution and pericyte coverage are altered in Alzheimer's disease progression

In order to assess the changes in the NVU, different components, that is, endothelial cells, pericytes, astrocytes, microglia and perivascular macrophages, were quantified across different stages of Alzheimer's disease. To begin with, several characteristics of vascular structure such as area fraction, length, branches and nearest distance were investigated in non‐demented controls and AD patients (Figure 1). Endothelial cells were visualized using lectin labeling and the aforementioned parameters were quantified with an automated analysis tool (Figure 1A). Vessel area fraction, length and branches were unaltered in the disease progression and no statistically significant difference was identified between AD patients and non‐demented controls (all P > 0.05) (Figure 1B‐D). However, the nearest distance between the vessels was significantly lower in AD patients compared to non‐demented controls in both the hippocampus (P: 0.02) and the cortex (P: 0.03), suggesting an alteration in vessel distribution in AD patients (Figure 1E).

*Quantification of vascular parameters*. A. Representative pictures of lectin (green) immunostaining from a non‐demented control and an AD patient (Scale bar: 50 µm) B. Diagnosis‐based distribution of vessel area fraction (% Lectin+ area) in hippocampus (n = 28) and cortex (n = 31). C. Diagnosis‐based distribution of vessel length in the hippocampus and cortex. D. Diagnosis‐based distribution of vessel branching in the hippocampus and cortex. E. Diagnosis‐based distribution of nearest‐neighbor distance (µm) in the hippocampus and cortex. In cortex one outlier was removed from the dataset based on the ROUT outlier test (Q = 1). The graphs represent individual values and mean ± SD. P‐values (*P < 0.05) were determined by two‐tailed Student's t‐test.

For the quantification of pericyte coverage, the well‐established pericyte marker PDGFR‐β was used (Figure 2A). A gradual decrease in pericyte coverage was observed across Braak stages (Figure 2B). A strong negative correlation was found between pericyte coverage and Braak stages both in the hippocampus (r: −0.80, P < 0.0001) and the cortex (r: −0.76, P < 0.0001). A significant reduction in pericyte coverage was observed in Braak stage 4 in both brain regions (Figure 2B). Diagnosis based analysis illustrated a 32% reduction of pericyte coverage in the hippocampus (P < 0.0001) and a 26% reduction of the pericyte coverage in the cortex (P: 0.0004) of AD patients in comparison to non‐demented controls (Figure 2C).

*Quantification of pericyte coverage*. A. Representative pictures of PDGFR‐β (red) and Collagen IV (green) immunostainings from a non‐demented control and an AD patient (Scale bar 50 µm). B. Hippocampal and cortical pericyte coverage in different Braak stages. Pericyte coverage is defined as PDGFR‐β+ area within Collagen IV+ area. P‐values (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001) were determined by one‐way ANOVA and Tukey's multiple comparison test. Comparisons were performed using the lowest Braak stage as a control. Correlation coefficients were calculated using Spearman's correlation (α = 0.05, CI 95%, two‐tailed). C. Diagnosis‐based pericyte coverage in hippocampus (n = 28) and cortex (n = 29). P‐values (***P < 0.001, ****P < 0.0001) were determined by two‐tailed Student's t‐test. The graphs represent individual values and mean ± SD.

Astrocyte coverage within the NVU is higher in AD patients

Astrocytes, another major component of the NVU, were visualized with GFAP immunostaining (Figure 3A). Astrocyte coverage within the NVU was assessed using a custom‐made Image J script. Vessel borders, defined by lectin labeling, were enlarged to cover all the NVU and GFAP+ area within these borders were quantified (Figure 3A, Figure S1). Astrocyte coverage within the NVU was not significantly different between Braak stages, neither in the cortex nor in the hippocampus (Figure 3B). However, NVU‐associated astrocyte immunoreactivity positively correlated with Braak stages in the cortex (r: 0.41, P: 0.02) (Figure 3B). A 20% increase in the GFAP+ area within the NVU was detected in AD patients compared to non‐demented controls in the cortex (P: 0.05) (Figure 3C). In addition to NVU‐associated immunoreactivity, we also assessed the total astrocyte coverage (reported as GFAP+ area) in both brain regions. Similar to NVU‐associated GFAP immunoreactivity, total astrocyte coverage did not show any significant differences between Braak stages (Figure S2A). The GFAP+ area did not correlate with Braak stages in the hippocampus, whereas a significant positive correlation was found in the cortex (r: 0.43, P: 0.01). Astrocyte coverage was 53% higher in AD patients compared to non‐demented controls in the cortex (P: 0.01) (Figure S2B).

*Quantification of astrocytes within the NVU*. A. Representative images of a non‐demented control and an AD patient for GFAP (red) and Lectin (green) immunostainings. Insert shows the magnification of the dashed area. The NVU border (white) in the insert was generated by enlarging the lectin immunostaining as outlined in the methods (Scale bar 50 µm). B. Braak stage distribution of astrocyte coverage within the NVU in the hippocampus (n = 28) and cortex (n = 31). Correlation coefficients were calculated using Spearman's correlation (α = 0.05, CI 95%, two‐tailed). C. Diagnosis‐based distribution of astrocyte coverage within the NVU in the hippocampus and cortex. P‐value (*P < 0.05) was determined by the non‐parametric two‐tailed Mann Whitney test. The graphs represent individual values and mean ± SD.

Changes in both microglia coverage and perivascular macrophage distribution are observed in AD patients

Microglia and perivascular macrophages are the resident immune cells associated with the NVU. In order to quantify the NVU‐associated microglia coverage in the cortex and the hippocampus, we quantified Iba1 immunoreactivity, a well‐established marker for microglia, within the enlarged lectin+ endothelial cell layer in both brain regions (Figure 4A). Although there were no significant differences in microglia coverage between the Braak stages, both in the hippocampus and the cortex, a decreasing trend was observed within NVU as the disease progressed (Figure 4B). A negative correlation between the Braak stages and Iba1+ area within the NVU was identified both in the hippocampus (r: −0.41, P: 0.03) and the cortex (r: −0.41, P: 0.02) (Figure 4B). The microglia coverage within the NVU was significantly lower in AD patients compared to non‐demented controls in both brain regions [a 39% reduction in the hippocampus (P: 0.05) and a 38% reduction in the cortex (P: 0.01)] (Figure 4C). In addition to NVU‐associated microglia coverage, we also assessed the number of microglia and Iba1+ area in the whole visual field to quantify the area covered per microglia in the brain. The area covered per microglia did not show any significant differences between Braak stages, neither in the hippocampus nor in the cortex (Figure S2C). Only in the cortex, a negative correlation between microglia coverage and Braak stages was observed (r: −0.46, P: 0.009). A Diagnosis‐based analysis showed a decreased area covered per microglia in AD compared to non‐demented controls in the cortex (36% reduction, P: 0.02) whereas in the hippocampus no significant differences were found between the groups (Figure S2D).

*Quantification of microglia coverage within the NVU*. A. Representative images of Iba1 (red) and Lectin (green) immunostainings in a non‐demented control and an AD patient. Insert shows the magnification of the dashed area. The NVU border (white) in the insert was generated by enlarging the lectin immunostaining as outlined in the methods (Scale bar 50 µm). B. Braak stage distribution of microglia coverage within the NVU in the hippocampus (n = 28) and cortex (n = 31). Correlation coefficients were calculated using Spearman's correlation (α = 0.05, CI 95%, two‐tailed). C. Diagnosis‐based distribution of microglia coverage within the NVU in the hippocampus and cortex. Student's t‐test was used to test for statistically significant differences between non‐demented controls and AD patients. P‐values (*P < 0.05) were determined by two‐tailed Student's t‐test. The graphs represent individual values and mean ± SD.

Apart from microglia, perivascular macrophages can be found in association with capillaries and larger arteries in the brain. As a next step, we, therefore, quantified perivascular macrophages separately in parenchymal vessels and leptomeningeal arteries using the MRC1 marker (Figure 5A). A significant reduction of leptomeningeal perivascular macrophages was found in Braak stage V in the hippocampus (Figure 5B). Leptomeningeal macrophages were negatively correlated with Braak stages both in the hippocampus (r: −0.53, P: 0.004) and the cortex (r: −0.39, P: 0.03). Significantly higher numbers of leptomeningeal perivascular macrophages were identified in non‐demented controls compared to AD patients in both brain regions [hippocampus (P: 0.03), cortex (P: 0.04)] (Figure 5C). Conversely, in parenchymal vessels, a positive correlation was found between Braak stages and perivascular macrophages in the hippocampus (r: 0.41, P: 0.03) and the cortex (r: 0.39, P: 0.03) (Figure 5D). Higher numbers of perivascular macrophages were detected in AD patients in comparison to non‐demented controls in both brain regions [hippocampus (P: 0.03), cortex (P: 0.06)] (Figure 5E).

*Quantification of perivascular macrophages in leptomeningeal and parenchymal vessels*. A. Representative images of MRC1+ macrophages (red) and lectin+ blood vessels (green) immunostainings in leptomeningeal and parenchymal vessels (Scale bar 50 µm). B. Braak stage distribution for the number of leptomeningeal perivascular macrophages in the hippocampus (n = 28) and cortex (n = 31). P‐value (*P < 0.05) was determined by one‐way ANOVA and Tukey's multiple comparison test using the lowest Braak stage as control. Correlation coefficients were calculated using Spearman's correlation (α = 0.05, CI 95%, two‐tailed). C. Diagnosis‐based distribution of leptomeningeal perivascular macrophages in the hippocampus and cortex. P‐values (*P < 0.05) were determined by Student's t‐test. D. Braak stage distribution of parenchymal perivascular macrophages in the hippocampus and cortex. Correlation coefficients were calculated using Spearman's correlation (α = 0.05, CI 95%, two‐tailed). E. Diagnosis‐based distribution of parenchymal perivascular macrophages in the hippocampus and cortex. P‐values (*P < 0.05) were determined by two‐tailed Student's t‐test. The graphs represent individual values and mean ± SD.

Correlations of NVU components with tau pathology and Aβ plaque load

In order to further characterize our sample set, tau and beta‐amyloid pathologies were analyzed in detail. Phosphorylated tau coverage was quantified by a phospho‐tau‐specific AT8 antibody (Figure S3). In order to assess beta‐amyloid pathology, adjacent sections were stained with three different clones of anti‐Aβ antibodies (Figure S4). The G2‐10 clone recognizes the C‐terminus of Aβ40 whereas the G2‐11 clone specifically targets the C‐terminus of Aβ42. Last, the 4G8 clone binds to the mid‐domain of Aβ is, therefore, expected to detect all major isoforms of Aβ. Aβ plaque density was quantified as the percentage area covered by each antibody. Following the detailed quantification of tau and beta‐amyloid load, the association of the NVU components with the pathological hallmarks of AD was further investigated (Figure 6). Nearest‐neighbor distance negatively correlated with both tau and beta‐amyloid (Aβ40, Aβ42 and 4G8) pathologies in the hippocampus; however, no association was found in the cortex. Pericyte coverage negatively correlated with both tau load and Aβ plaque load (Aβ40, Aβ42 and 4G8) in the hippocampus and the cortex. Astrocyte coverage within the NVU did not show any correlation with tau and beta‐amyloid pathologies; however, it was positively correlated with total GFAP coverage in both brain regions. Iba1 within the NVU did not correlate with tau whereas negative correlations were identified with beta‐amyloid (Aβ42, 4G8 and CERAD score) in the cortex. In both brain regions, microglia coverage within the NVU positively correlated with the area covered per microglia. Parenchymal macrophage counts showed a positive correlation with tau, Aβ42+ and 4G8+ plaque load in the hippocampus whereas in cortex positive correlations were identified only with tau and Aβ40+ plaque load. Moreover, leptomeningeal macrophage counts revealed an opposite trend, namely a negative correlation with tau and all beta‐amyloid parameters in the hippocampus; however, no correlation was identified with these parameters in the cortex. Absolute Spearman r values and P‐values are provided in the Supporting Information (Figure S5).

*Correlation of NVU components with beta‐amyloid, tau and histopathological assessments*. Rows indicate the NVU‐associated parameters: vessel area, vessel length, vessel branch, nearest distance between vessels (Nearest dist.), PDGFR‐β, GFAP in NVU, Iba1 in NVU, MRC1+ cells in parenchymal vessels (MRC1 para.) and MRC1+ cells in leptomeningeal vessels (MRC1 lepto.). Columns indicate the pathophysiological hallmarks and parameters provided by the brain bank: GFAP, Iba1, tau, Aβ42, Aβ40, 4G8, Braak, CERAD score and diagnosis. Heatmap represents the Spearman correlation coefficient for each correlation. Significant correlations were marked with stars on the heatmap (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). Correlation coefficients were calculated using Spearman's correlation (α = 0.05, CI 95%, two‐tailed) with corrplot package in R. Absolute numbers for correlation coefficients and P‐values are listed in Figure S5.

Discussion

Due to the high glucose and oxygen demand of the brain, a healthy cerebrovascular system is vital for the maintenance of brain homeostasis (6). Impairments of cerebral vessels were identified in aging as well as in various neurodegenerative disorders (52). In AD, vascular dysfunction is believed to be a major contributor to disease pathophysiology and may well account for the well‐established heterogeneity and complexity of the disease (53). In this study, we aimed to identify the capillary level changes in AD pathology by applying a comprehensive immunohistological analysis of the NVU in a well‐characterized sample set of human post‐mortem brains. Using widely distributed non‐demented control subjects across Braak stages 0 to IV and AD patients from Braak stage IV to VI, we investigated the NVU associated changes during disease progression.

We applied a well‐established image processing pipeline in order to assess structural changes in the brain vasculature. This pipeline was previously used to quantify vascular growth in retinal development (44) and vascular remodeling following stroke (45, 47). Our analysis indicated that vascular area fraction, vessel length and branching do not significantly change in AD progression whereas the nearest distance between vessels is significantly reduced in AD patients. Ongoing angiogenesis and vascular remodeling were previously reported in transgenic AD mouse models and post‐mortem human samples (8, 16, 58, 59). In line with these findings, our results suggest an alteration in vessel distribution in AD patients. The negative correlation between the nearest distance between vessels and pathophysiological hallmarks of AD might imply that changes in tau and beta‐amyloid could influence vascular distribution in the brain. Eventually, the neuronal loss may also lead to vascular alterations and re‐distribution. However, we could not fully address the factors leading to these alterations with our post‐mortem characterization.

Reduced pericyte coverage in AD patients was previously shown in post‐mortem studies by both paraffin‐embedded sections and brain homogenates (12, 30, 49). Our results confirm these earlier findings and additionally demonstrate a Braak‐stage dependent decrease in pericyte coverage starting already in early Braak stages. Within the NVU, the reduction in pericyte coverage showed the strongest correlation with Braak stages. Moreover, pericyte loss was among the earliest changes, with a significant reduction in pericyte coverage detected already at Braak stage IV. In agreement with our earlier assessments using a manual vessel counting‐based quantitative method (24), automated area‐based quantification of brain pericytes also revealed a decrease as early as Braak stage II for some subjects, thus confirming an early involvement of pericyte‐associated pathophysiology in AD. Furthermore, in this study, pericyte coverage was also found to be well‐correlated with both beta‐amyloid and tau pathology in our sample set further supporting their role and importance in disease progression. It is crucial to point out that our quantification strategy is solely based on the area covered by PDGFR‐β immunoreactivity but not on absolute cell counts, similar to earlier studies (12, 49). Therefore, the reduction in pericyte coverage does not necessarily reflect a reduction in cell density but may, for example, also be explained by a retraction of pericyte processes or—alternatively increased shedding of membrane‐anchored PDGFR‐β upon injury and pericyte dysfunction.

Astrogliosis and increased the activation of astrocytes upon CNS damage and disease are commonly used as markers of neuroinflammation (50). Astroglial activation, as revealed by increased GFAP immunoreactivity, was frequently reported in the AD brain and around Aβ plaques (39, 61). A Braak stage‐dependent increase in GFAP immunoreactivity and a positive correlation of activated astrocytes with Aβ plaques was previously observed (19). Supporting the earlier findings, we found increased GFAP immunoreactivity in the cortex of AD patients, positively correlated with both tau and beta‐amyloid pathology. We did not observe this association in the hippocampus, possibly due to lack of Braak stage 0 subjects and given the low but detectable levels of tau and beta‐amyloid pathology already in the early Braak stages in the hippocampus. Further analysis of NVU‐specific GFAP immunoreactivity showed an increased GFAP coverage within the NVU. This observation suggests that astrocyte coverage of small capillaries is not reduced and structurally, at least, the glial barrier formed by astrocytic end‐feet appears to be intact in AD patients. However, increased GFAP immunoreactivity within the NVU could also reflect an increased the activation of astrocytes associated with small capillaries. An enhanced GFAP immunoreactivity within the NVU could trigger vascular dysfunction since activated astrocytes were previously shown to release pro‐inflammatory cytokines leading to the extravasation of peripheral immune cells into the brain and BBB disruptions (51).

Microglia, the brain resident innate immune cells, were reported to display distinct morphologies in the brain (1). Possibly due to these dynamic morphological changes in the microglial phenotype, contradictory results have been published with regards to changes in microglial coverage in AD using well‐established microglial markers such as Iba1 (15). Therefore, instead of simply quantifying the Iba1+ area, we additionally quantified the number of microglia in each visual field, normalized our results accordingly and reported the average Iba1 coverage per microglial cell in different Braak stages. Although we were not able to detect any significant changes between the Braak stages, our results showed a decreased Iba1 coverage per microglial cell in AD patients in the cortex. A reduced Iba1 coverage per microglial cell likely illustrates a shift from a ramified to a more amoeboid or less ramified dystrophic morphology along the AD continuum (5, 7, 20). Furthermore, we identified a reduction in Iba1+ coverage within the NVU in AD patients compared to non‐demented controls. In AD patients, microglia were found in the vicinity of Aβ plaques, where they are suggested to be responsible for the clearance and phagocytosis of aggregated Aβ (18). The observed negative correlation between the Aβ plaque load and vascular Iba1 coverage in the cortex might suggest a redistribution of microglial cells from the vessels to predominantly parenchymal areas around aggregated Aβ species.

In addition to the microglia, we quantified the perivascular macrophages both in penetrating vessels in the brain parenchyma and leptomeningeal vessels by MRC1 as a marker; this marker was previously reported to be expressed only by perivascular macrophages but not by brain‐resident microglia (10). Interestingly, our quantitative analysis revealed that in non‐demented controls, perivascular macrophages were more abundant in leptomeningeal arteries, outside the brain parenchyma, while in AD patients, more perivascular macrophages were detected in parenchymal vessels within the brain. The role of perivascular macrophages in AD has mainly been investigated by studies in transgenic animals. Perivascular macrophages were reported to be associated with Aβ plaques and vascular Aβ deposits in AD mouse models (35, 56). Considering their phagocytic capabilities, perivascular macrophages were considered to contribute to Aβ clearance (9) and supporting this hypothesis, depletion of perivascular macrophages was shown to exacerbate Aβ accumulation in transgenic animals (14). In line with these in vivo studies, our data also suggest that Aβ accumulation might trigger the infiltration of perivascular macrophages from leptomeningeal vessels to parenchymal vessels penetrating into the brain. Perivascular macrophage accumulation around parenchymal vessels positively correlated with Aβ plaque load both in the cortex and hippocampus. Interestingly, in the hippocampus, stronger associations were identified with 4G8+ plaques whereas in the cortex the number of perivascular macrophages was correlated with Aβ40+ plaques. Suggesting different mechanisms of infiltration in different brain regions, the number of leptomeningeal macrophages was negatively correlated with both tau and beta‐amyloid pathologies in the hippocampus; however, no association was found in the cortex. To our knowledge, our study represents the first comprehensive analysis of perivascular macrophages in the AD continuum. Nevertheless, further in vitro and in vivo mechanistic studies are required to better understand the role of perivascular macrophages in AD pathophysiology and to delineate the signaling pathways resulting in the observed the relocation of these cells in AD patients.

The major limitations of our study are the relatively low number of post‐mortem samples and inconsistencies regarding the subject heterogeneity between the two brain regions used in our analysis. Although we included subjects from different Braak stages to cover the whole spectrum of the disease from preclinical phases to established AD dementia, in certain stages (ie, Braak stage V cortex or Braak stage III hippocampus) we were able to include only three subjects. Considering the natural variations observed between human subjects, including a larger number of samples from each Braak stage would have improved the power of our study to differentiate more subtle changes between the different stages. Unfortunately, no hippocampal samples from Braak stage 0 individuals were available for our analysis, whereas cortical tissue from Braak stage 0 subjects was included in the study. The lack of availability of hippocampal samples derived from Braak stage 0 subjects may potentially account for some of the observed differences in specific immunostainings detected between the hippocampus and the cortex. Our sample set was age‐matched and no statistically significant age differences were detected between non‐demented controls and AD subjects nor between the different Braak stage groups; however, we had a gender imbalance in our sample set. In general, more female subjects were included in the study and all post‐mortem AD brains were derived from female donors. The high female to male ratio in our sample set is most likely due to higher availability of female donors, especially in the diseased state, since more females are diagnosed with AD compared to males in every age group (63). Gender effects in non‐demented controls were controlled using simple logistic regressions and no statistical difference was found between males and females in any of the quantified parameters. The APOE4 allele was previously shown to affect blood‐brain barrier integrity and was associated with reduced pericyte coverage (12, 32). In this study, we were not able to evaluate the effect of the APOE genotype due to a marked imbalance in the number of APOE4 carriers between non‐demented controls and AD cases in our sample set. The cerebrovascular system maintains dynamic and regionally varying oxygen and nutrient demand of the brain by modulating cerebral blood flow (25). However, changes in the vasculature were shown to affect neurovascular coupling and brain perfusion in AD (25, 26, 54). Alterations within the NVU, such as pericyte dysfunction or changes in the astrocytic end‐feet coverage might ultimately lead to functional impairments in neurovascular coupling. Since our quantification strategy was based on the area coverage but did not include any morphological parameters (ie, vessel diameter) nor parameters reflecting cerebral blood flow, we were not able to assess the functional consequences of the observed NVU‐related changes potentially contributing in neurovascular dysfunction.

Despite the aforementioned limitations, our study, to our knowledge, represents the first comprehensive quantification of different NVU components using the same sample set. Hence, our results give an overview of changes in more than one cell type as revealed by a detailed immunohistochemical characterization. Our quantification strategy provides quantitative information about the immunoreactivity of different components specifically within the NVU. For endothelial cells and pericytes, for example, a simple colocalization analysis would be sufficient to investigate NVU associated changes since these cells are solely part of the vasculature. Moreover, astrocytes and microglia are found abundantly outside the vasculature; our analysis method, therefore, provides a unique opportunity to investigate NVU‐related changes in these cells. With this method, we showed for the first time increased GFAP immunoreactivity and a reduced Iba1 coverage around capillaries in AD. By using a marker specific for perivascular macrophages, we also showed the infiltration of these cells into parenchymal penetrating vessels in AD patients. Consequently, our comprehensive post‐mortem characterization illustrates distinct alterations in the NVU in AD and outlines the relationship between these changes and AD pathological hallmarks. In addition to previously reported morphometric changes in the brain microvasculature, our study further contributes to the characterization of the cerebrovascular alterations in AD (4, 13, 43). Our findings indicate that capillary level vascular dysfunction in AD pathology incorporates more than one cell type and highlights the need for advancing our understanding of the NVU in AD pathogenesis.

Conflict of Interest

The authors declare that they have no conflict of interest.

Supporting information

Figure S1. Workflow for the quantification of immunoreactivity associated with the neurovascular unit. Representative workflow for the quantification of GFAP immunoreactivity within the NVU. Briefly, Lectin and GFAP images for the same visual fields were converted to 8‐bit and pre‐defined threshold was applied using Image J. Subsequently, Lectin+ area was enlarged using “dilate” function in Image J. Area of enlarged lectin image was measured using “Measure” function. Afterwards, GFAP and enlarged lectin images were merged and colocalized area was quantified. Colocalized area was divided to enlarged lectin area (defined as NVU) and results were reported as %GFAP+ area within the NVU.

Figure S2. Quantification of astrocyte and microglia coverage. A. Braak stage distribution of astrocyte coverage in the hippocampus and cortex. Astrocyte coverage was defined as GFAP+ area. Correlation coefficients were calculated using Spearman's correlation (α = 0.05, CI 95%, two‐tailed). B. Diagnosis‐based distribution of astrocyte coverage in hippocampus and cortex. P value (*P < 0.05) was determined by student's t‐test. C. Braak stage distribution of microglia coverage in the hippocampus and cortex. Microglia coverage (Iba1+ area) was normalized to cell count in each visual field. Correlation coefficients were calculated using Spearman's correlation (α = 0.05, CI 95%, two‐tailed). D. Diagnosis‐based distribution of astrocyte coverage in hippocampus and cortex. P value (*P < 0.05) was determined by student's t‐test. The graphs represent individual values and mean ± standard deviation.

Figure S3. Quantification of phosphorylated tau in the human brain. A. Representative images of tau (AT8) (red) and Lectin (green) immunostainings in a non‐demented control and an AD patient. B. Braak stage dependent analysis of tau tangles in the hippocampus and cortex. In hippocampus P values (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001) were determined by Kruskal–Wallis test followed by Dunn's multiple comparison test using the lowest Braak stage as a control. In cortex, P values (***P < 0.001, ****P < 0.0001) were determined by one‐way ANOVA followed by Tukey's multiple comparison test using the lowest Braak stage as a control. Correlation coefficients were calculated using Spearman's correlation (α = 0.05, CI 95%, two‐tailed). C. Diagnosis‐based analysis of tau tangles in the hippocampus and cortex. P values (****P < 0.0001) were determined by Mann–Whitney test. The graphs represent individual values and mean ± standard deviation.

Figure S4. Quantification of the beta‐amyloid plaque load with different anti‐Aβ antibodies. A. Representative images of serial sections derived from a non‐demented control and an AD patient which were immunostained with anti‐Aβ40, anti‐Aβ42 and pan‐Aβ (4G8) antibodies. B. Braak stage distribution of Aβ40+ plaques in the hippocampus and cortex. P‐values (*P < 0.05, ***P < 0.001) were determined by one‐way ANOVA and Tukey's multiple comparison test using the lowest Braak stage as control. Correlation coefficients were calculated using Spearman's correlation (α = 0.05, CI 95%, two‐tailed). C. Diagnosis‐based distribution of Aβ40+ plaques in the hippocampus and cortex. P‐values (****P < 0.0001) were determined by Mann–Whitney test. D. Braak stage distribution of Aβ42+ plaques in the hippocampus and cortex. P‐values (*P < 0.05, **P < 0.01) were determined by Kruskal–Wallis test followed by Dunn's multiple comparison test using the lowest Braak stage as control. Correlation coefficients were calculated using Spearman's correlation (α = 0.05, CI 95%, two‐tailed). E. Diagnosis‐based distribution of Aβ42+ plaques in the hippocampus and cortex. P‐values (****P < 0.0001) were determined by Mann–Whitney test. F. Braak stage distribution of 4G8+ plaques in the hippocampus and cortex. P‐values (*P < 0.05, **P < 0.01, ***P < 0.001) were determined by one‐way ANOVA and Tukey's multiple comparison test in the hippocampus and by Kruskal–Wallis test followed by Dunn's multiple comparison test in the cortex using the lowest Braak stage as control. Correlation coefficients were calculated using Spearman's correlation (α = 0.05, CI 95%, two‐tailed). G. Diagnosis‐based distribution of 4G8+ plaques in the hippocampus and cortex. P‐values (****P < 0.0001) were determined by Mann–Whitney test. The graphs represent individual values and mean ± standard deviation.

Figure S5. Spearman r and P values for the correlation matrices. Absolute numbers for Spearman r correlation coefficients and P values shown in correlation matrices.

Table S1. List of primary antibodies.

Table S2. Additional statistical parameters of quantifications.

Click here for additional data file.^{(5.5MB, docx)}

Data Availability Statement

The datasets used in this study are available from the corresponding author upon reasonable request.

References

1. Boche D, Perry VH, Nicoll JAR(2013) Review: activation patterns of microglia and their identification in the human brain. Neuropathol Appl Neurobiol 39:3–18. [DOI] [PubMed] [Google Scholar]
2. Braak H, Alafuzoff I, Arzberger T, Kretzschmar H, Del Tredici K (2006) Staging of Alzheimer disease‐associated neurofibrillary pathology using paraffin sections and immunocytochemistry. Acta Neuropathol 112:389–404. [DOI] [PMC free article] [PubMed] [Google Scholar]
3. Braak H, Braak E (1995) Staging of Alzheimer's disease-related neurofibrillary changes. Neurobiol Aging 16:271–278. [DOI] [PubMed] [Google Scholar]
4. Burke MJC, Nelson L, Slade JY, Oakley AE, Khundakar AA, Kalaria RN (2014) Morphometry of the hippocampal microvasculature in post‐stroke and age‐related dementias. Neuropathol Appl Neurobiol 40:284–295. [DOI] [PMC free article] [PubMed] [Google Scholar]
5. Cai Z, Hussain MD, Yan L-J (2014) Microglia, neuroinflammation, and beta-amyloid protein in Alzheimer's disease. Int J Neurosci 124:307–321. [DOI] [PubMed] [Google Scholar]
6. Camandola S, Mattson MP (2017) Brain metabolism in health, aging, and neurodegeneration. EMBO J 36:1474–1492. [DOI] [PMC free article] [PubMed] [Google Scholar]
7. Clayton KA, Van Enoo AA, Ikezu T (2017) Alzheimer's disease: the role of microglia in brain homeostasis and proteopathy. Front Neurosci. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5733046/ (accessed 23 April 2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
8. Desai BS, Schneider JA, Li J-L, Carvey PM, Hendey B (2009) Evidence of angiogenic vessels in Alzheimer's disease. J Neural Transm 116:587–597. [DOI] [PMC free article] [PubMed] [Google Scholar]
9. Faraco G, Park L, Anrather J, Iadecola C (2017) Brain perivascular macrophages: characterization and functional roles in health and disease. J Mol Med 95:1143–1152. [DOI] [PMC free article] [PubMed] [Google Scholar]
10. Goldmann T, Wieghofer P, Jordão MJC, Prutek F, Hagemeyer N, Frenzel K et al (2016) Origin, fate and dynamics of macrophages at central nervous system interfaces. Nat Immunol 17:797–805. [DOI] [PMC free article] [PubMed] [Google Scholar]
11. Guidoux C, Hauw J-J, Klein IF, Labreuche J, Berr C, Duyckaerts C et al (2018) Amyloid angiopathy in brain hemorrhage: a postmortem neuropathological-magnetic resonance imaging study. Cerebrovasc Dis 45:124–131. [DOI] [PubMed] [Google Scholar]
12. Halliday MR, Rege SV, Ma Q, Zhao Z, Miller CA, Winkler EA et al (2016) Accelerated pericyte degeneration and blood–brain barrier breakdown in apolipoprotein E4 carriers with Alzheimer's disease. J Cereb Blood Flow Metab 36:216–227. [DOI] [PMC free article] [PubMed] [Google Scholar]
13. Hase Y, Ding R, Harrison G, Hawthorne E, King A, Gettings S et al (2019) White matter capillaries in vascular and neurodegenerative dementias. Acta Neuropathol Commun 7:16. [DOI] [PMC free article] [PubMed] [Google Scholar]
14. Hawkes CA, McLaurin J (2009) Selective targeting of perivascular macrophages for clearance of β-amyloid in cerebral amyloid angiopathy. PNAS 106:1261–1266. [DOI] [PMC free article] [PubMed] [Google Scholar]
15. Hopperton KE, Mohammad D, Trépanier MO, Giuliano V, Bazinet RP (2018) Markers of microglia in post‐mortem brain samples from patients with Alzheimer's disease: a systematic review. Mol Psychiatry 23:177–198. [DOI] [PMC free article] [PubMed] [Google Scholar]
16. Jefferies WA, Price KA, Biron KE, Fenninger F, Pfeifer CG, Dickstein DL (2013) Adjusting the compass: new insights into the role of angiogenesis in Alzheimer's disease. Alzheimer's Res Ther 5:64. [DOI] [PMC free article] [PubMed] [Google Scholar]
17. Kacem K, Lacombe P, Seylaz J, Bonvento G (1998) Structural organization of the perivascular astrocyte endfeet and their relationship with the endothelial glucose transporter: a confocal microscopy study. Glia 23:1–10. [PubMed] [Google Scholar]
18. Kamphuis W, Kooijman L, Schetters S, Orre M, Hol EM (2016) Transcriptional profiling of CD11c-positive microglia accumulating around amyloid plaques in a mouse model for Alzheimer's disease. Biochim Biophys Acta 1862:1847–1860. [DOI] [PubMed] [Google Scholar]
19. Kamphuis W, Middeldorp J, Kooijman L, Sluijs JA, Kooi E-J, Moeton M et al (2014) Glial fibrillary acidic protein isoform expression in plaque related astrogliosis in Alzheimer's disease. Neurobiol Aging 35:492–510. [DOI] [PubMed] [Google Scholar]
20. Karperien A, Ahammer H, Jelinek H (2013) Quantitating the subtleties of microglial morphology with fractal analysis. Front Cell Neurosci 7:3. Available at: https://www.frontiersin.org/articles/10.3389/fncel.2013.00003/full (accessed 23 April 2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
21. Keable A, Fenna K, Yuen HM, Johnston DA, Smyth NR, Smith C et al (2016) Deposition of amyloid β in the walls of human leptomeningeal arteries in relation to perivascular drainage pathways in cerebral amyloid angiopathy. Biochim Biophys Acta 1862:1037–1046. [DOI] [PMC free article] [PubMed] [Google Scholar]
22. Kierdorf K, Masuda T, Jordão MJC, Prinz M (2019) Macrophages at CNS interfaces: ontogeny and function in health and disease. Nat Rev Neurosci 20, 547–562. [DOI] [PubMed] [Google Scholar]
23. Kim SJ, Seo Y, Kim HJ, Na DL, Seo SW, Kim Y et al (2018) Pathologically confirmed cerebral amyloid angiopathy with no radiological sign in a patient with early onset Alzheimer's disease. Yonsei Med J 59:801–805. [DOI] [PMC free article] [PubMed] [Google Scholar]
24. Kirabali T, Rigotti S, Siccoli A, Liebsch F, Shobo A, Hock C et al (2019) The amyloid-β degradation intermediate Aβ34 is pericyte‐associated and reduced in brain capillaries of patients with Alzheimer's disease. Acta Neuropathol Commun 7:194. [DOI] [PMC free article] [PubMed] [Google Scholar]
25. Kisler K, Nelson AR, Montagne A, Zlokovic BV (2017) Cerebral blood flow regulation and neurovascular dysfunction in Alzheimer disease. Nat Rev Neurosci 18:419. Available at: https://www.nature.com/nrn/journal/vaop/ncurrent/full/nrn.2017.48.html (accessed 2 June 2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
26. Kisler K, Nelson AR, Rege SV, Ramanathan A, Wang Y, Ahuja A et al (2017) Pericyte degeneration leads to neurovascular uncoupling and limits oxygen supply to brain. Nat Neurosci 20:406–416. [DOI] [PMC free article] [PubMed] [Google Scholar]
27. Koistinaho M, Lin S, Wu X, Esterman M, Koger D, Hanson J et al (2004) Apolipoprotein E promotes astrocyte colocalization and degradation of deposited amyloid-β peptides. Nat Med 10:719–726. [DOI] [PubMed] [Google Scholar]
28. Kulic L, McAfoose J, Welt T, Tackenberg C, Späni C, Wirth F et al (2012) Early accumulation of intracellular fibrillar oligomers and late congophilic amyloid angiopathy in mice expressing the Osaka intra-Aβ APP mutation. Transl Psychiatry 13:e183. [DOI] [PMC free article] [PubMed] [Google Scholar]
29. Magaki S, Tang Z, Tung S, Williams CK, Lo D, Yong WH et al (2018) The effects of cerebral amyloid angiopathy on integrity of the blood‐brain barrier. Neurobiol Aging 70:70–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
30. Miners JS, Schulz I, Love S (2017) Differing associations between Aβ accumulation, hypoperfusion, blood–brain barrier dysfunction and loss of PDGFRB pericyte marker in the precuneus and parietal white matter in Alzheimer's disease. J Cereb Blood Flow Metab 38:103–115. [DOI] [PMC free article] [PubMed] [Google Scholar]
31. Montagne A, Barnes SR, Sweeney MD, Halliday MR, Sagare AP, Zhao Z et al (2015) blood‐brain barrier breakdown in the aging human hippocampus. Neuron 85:296–302. [DOI] [PMC free article] [PubMed] [Google Scholar]
32. Montagne A, Nation DA, Sagare AP, Barisano G, Sweeney MD, Chakhoyan A et al (2020) APOE4 leads to blood‐brain barrier dysfunction predicting cognitive decline. Nature 581:71–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
33. Montagne A, Zhao Z, Zlokovic BV (2017) Alzheimer's disease: a matter of blood‐brain barrier dysfunction? J Exp Med 214:3151–3169. [DOI] [PMC free article] [PubMed] [Google Scholar]
34. Montine TJ, Phelps CH, Beach TG, Bigio EH, Cairns NJ, Dickson DW et al (2012) National Institute on Aging-Alzheimer's Association guidelines for the neuropathologic assessment of Alzheimer's disease: a practical approach. Acta Neuropathol 123:1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
35. Mucke L, Masliah E, Yu GQ, Mallory M, Rockenstein EM, Tatsuno G et al (2000) High-level neuronal expression of abeta 1–42 in wild-type human amyloid protein precursor transgenic mice: synaptotoxicity without plaque formation. J. Neurosci 20:4050–4058. [DOI] [PMC free article] [PubMed] [Google Scholar]
36. Nelson AR, Sweeney MD, Sagare AP, Zlokovic BV (2016) Neurovascular dysfunction and neurodegeneration in dementia and Alzheimer's disease. Biochim Biophys Acta 1862:887–900. [DOI] [PMC free article] [PubMed] [Google Scholar]
37. Park L, Uekawa K, Garcia-Bonilla L, Koizumi K, Murphy M, Pistik R et al (2017) Brain perivascular macrophages initiate the neurovascular dysfunction of Alzheimer Aβ peptides. Circ Res 121:258–269. [DOI] [PMC free article] [PubMed] [Google Scholar]
38. Perl DP (2010) Neuropathology of Alzheimer's disease. Mt Sinai J Med 77:32–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
39. Pike CJ, Cummings BJ, Cotman CW (1995) Early association of reactive astrocytes with senile plaques in Alzheimer's disease. Exp Neurol 132:172–179. [DOI] [PubMed] [Google Scholar]
40. Plog BA, Nedergaard M (2018) The glymphatic system in central nervous system health and disease: past, present, and future. Annu Rev Pathol 24:379–394. [DOI] [PMC free article] [PubMed] [Google Scholar]
41. Rajkowska G, Hughes J, Stockmeier CA, Javier Miguel-Hidalgo J, Maciag D (2013) Coverage of blood vessels by astrocytic endfeet is reduced in major depressive disorder. Biol Psychiat 73:613–621. [DOI] [PMC free article] [PubMed] [Google Scholar]
42. Rasmussen MK, Mestre H, Nedergaard M (2018) The glymphatic pathway in neurological disorders. Lancet Neurol 17:1016–1024. [DOI] [PMC free article] [PubMed] [Google Scholar]
43. Richard E, van Gool WA, Hoozemans JJM, van Haastert ES, Eikelenboom P, Rozemuller AJM et al (2010) Morphometric changes in the cortical microvascular network in Alzheimer's disease. J Alzheimers Dis 22:811–818. [DOI] [PubMed] [Google Scholar]
44. Rust R, Grönnert L, Dogançay B, Schwab ME (2019) A revised view on growth and remodeling in the retinal vasculature. Sci Rep 9:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
45. Rust R, Grönnert L, Gantner C, Enzler A, Mulders G, Weber RZ et al (2019) Nogo-A targeted therapy promotes vascular repair and functional recovery following stroke. PNAS 116:14270–14279. [DOI] [PMC free article] [PubMed] [Google Scholar]
46. Rust R, Kirabali T, Grönnert L, Dogancay B, Limasale YDP, Meinhardt A et al (2020) A practical guide to the automated analysis of vascular growth, maturation and injury in the brain. Front Neurosci 14:244. [DOI] [PMC free article] [PubMed] [Google Scholar]
47. Rust R, Weber RZ, Grönnert L, Mulders G, Maurer MA, Hofer A-S et al (2019) Anti-Nogo-A antibodies prevent vascular leakage and act as pro-angiogenic factors following stroke. Sci Rep 9:20040. [DOI] [PMC free article] [PubMed] [Google Scholar]
48. Ryu JK, McLarnon JG (2009) A leaky blood‐brain barrier, fibrinogen infiltration and microglial reactivity in inflamed Alzheimer's disease brain. J. Cell Mol Med 13:2911–2925. [DOI] [PMC free article] [PubMed] [Google Scholar]
49. Sengillo JD, Winkler EA, Walker CT, Sullivan JS, Johnson M, Zlokovic BV (2013) Deficiency in mural vascular cells coincides with blood‐brain barrier disruption in Alzheimer's disease. Brain Pathol 23:303–310. [DOI] [PMC free article] [PubMed] [Google Scholar]
50. Sofroniew MV (2015) Astrogliosis. Cold Spring Harb Perspect Biol 7:a020420. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4315924/ (accessed 7 November 2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
51. Sofroniew MV (2015) Astrocyte barriers to neurotoxic inflammation. Nat Rev Neurosci 16:249–263. [DOI] [PMC free article] [PubMed] [Google Scholar]
52. Sweeney MD, Kisler K, Montagne A, Toga AW, Zlokovic BV (2018) The role of brain vasculature in neurodegenerative disorders. Nat Neurosci 21:1318–1331. [DOI] [PMC free article] [PubMed] [Google Scholar]
53. Sweeney MD, Montagne A, Sagare AP, Nation DA, Schneider LS, Chui HC et al (2019) Vascular dysfunction-the disregarded partner of Alzheimer's disease. Alzheimers Dement 15:158–167. [DOI] [PMC free article] [PubMed] [Google Scholar]
54. Tarantini S, Tran CHT, Gordon GR, Ungvari Z, Csiszar A (2017) Impaired neurovascular coupling in aging and Alzheimer's disease: contribution of astrocyte dysfunction and endothelial impairment to cognitive decline. Exp Gerontol 94:52–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
55. Tarasoff-Conway JM, Carare RO, Osorio RS, Glodzik L, Butler T, Fieremans E et al (2015) Clearance systems in the brain—implications for Alzheimer disease. Nat Rev Neurol 11:457–470. [DOI] [PMC free article] [PubMed] [Google Scholar]
56. Thanopoulou K, Fragkouli A, Stylianopoulou F, Georgopoulos S (2010) Scavenger receptor class B type I (SR-BI) regulates perivascular macrophages and modifies amyloid pathology in an Alzheimer mouse model. Proc Natl Acad Sci USA 107:20816–20821. [DOI] [PMC free article] [PubMed] [Google Scholar]
57. Tian J, Shi J, Mann DMA (2004) Cerebral amyloid angiopathy and dementia. Panminerva Med 46:253–264. [PubMed] [Google Scholar]
58. Tong X-K, Nicolakakis N, Kocharyan A, Hamel E (2005) Vascular remodeling versus amyloid β-induced oxidative stress in the cerebrovascular dysfunctions associated with Alzheimer's disease. J Neurosci 25:11165–11174. [DOI] [PMC free article] [PubMed] [Google Scholar]
59. Vagnucci AH, Li WW (2003) Alzheimer's disease and angiogenesis. Lancet 361:605–608. [DOI] [PubMed] [Google Scholar]
60. Van Dorpe J, Smeijers L, Dewachter I, Nuyens D, Spittaels K, Van den Haute C et al (2000) Prominent cerebral amyloid angiopathy in transgenic mice overexpressing the london mutant of human APP in neurons. Am J Pathol 157:1283–1298. [DOI] [PMC free article] [PubMed] [Google Scholar]
61. Vehmas AK, Kawas CH, Stewart WF, Troncoso JC (2003) Immune reactive cells in senile plaques and cognitive decline in Alzheimer's disease. Neurobiol Aging 24:321–331. [DOI] [PubMed] [Google Scholar]
62. Verbeek MM, Van Nostrand WE, Otte-Höller I, Wesseling P, De Waal RMW (2000) Amyloid-β-induced degeneration of human brain pericytes is dependent on the apolipoprotein E genotype. Ann N Y Acad Sci 903:187–199. [DOI] [PubMed] [Google Scholar]
63. Viña J, Lloret A (2010) Why women have more Alzheimer's disease than men: gender and mitochondrial toxicity of amyloid-β peptide. J Alzheimer's Dis 20:S527–S533. [DOI] [PubMed] [Google Scholar]
64. Weller RO, Subash M, Preston SD, Mazanti I, Carare RO (2008) Perivascular drainage of amyloid-beta peptides from the brain and its failure in cerebral amyloid angiopathy and Alzheimer's disease. Brain Pathol 18:253–266. [DOI] [PMC free article] [PubMed] [Google Scholar]
65. Winkler EA, Sagare AP, Zlokovic BV (2014) The pericyte: a forgotten cell type with important implications for Alzheimer's disease? Brain Pathol 24:371–386. [DOI] [PMC free article] [PubMed] [Google Scholar]
66. Yamazaki Y, Kanekiyo T. blood‐brain barrier dysfunction and the pathogenesis of Alzheimer's disease. Int J Mol Sci 2017;18:1965. [DOI] [PMC free article] [PubMed] [Google Scholar]
67. Zenaro E, Piacentino G, Constantin G. The blood‐brain barrier in Alzheimer's disease. Neurobiol Dis. Available at: http://www.sciencedirect.com/science/article/pii/S0969996116301656 (accessed 2 December 2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
68. Zeppenfeld DM, Simon M, Haswell JD, D’Abreo D, Murchison C, Quinn JF et al (2017) Association of perivascular localization of aquaporin-4 with cognition and Alzheimer disease in aging brains. JAMA Neurol 74:91–99. [DOI] [PubMed] [Google Scholar]
69. Zhao X, Eyo UB, Murugan M, Wu L-J (2018) Microglial interactions with the neurovascular system in physiology and pathology. Dev Neurobiol 78:604–617. [DOI] [PMC free article] [PubMed] [Google Scholar]
70. Zlokovic BV (2011) Neurovascular pathways to neurodegeneration in Alzheimer's disease and other disorders. Nat Rev Neurosci 12:723–738. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S5. Spearman r and P values for the correlation matrices. Absolute numbers for Spearman r correlation coefficients and P values shown in correlation matrices.

Table S1. List of primary antibodies.

Table S2. Additional statistical parameters of quantifications.

Click here for additional data file.^{(5.5MB, docx)}

Data Availability Statement

The datasets used in this study are available from the corresponding author upon reasonable request.

[bpa12895-bib-0001] 1. Boche D, Perry VH, Nicoll JAR(2013) Review: activation patterns of microglia and their identification in the human brain. Neuropathol Appl Neurobiol 39:3–18. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0002] 2. Braak H, Alafuzoff I, Arzberger T, Kretzschmar H, Del Tredici K (2006) Staging of Alzheimer disease‐associated neurofibrillary pathology using paraffin sections and immunocytochemistry. Acta Neuropathol 112:389–404. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0003] 3. Braak H, Braak E (1995) Staging of Alzheimer's disease-related neurofibrillary changes. Neurobiol Aging 16:271–278. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0004] 4. Burke MJC, Nelson L, Slade JY, Oakley AE, Khundakar AA, Kalaria RN (2014) Morphometry of the hippocampal microvasculature in post‐stroke and age‐related dementias. Neuropathol Appl Neurobiol 40:284–295. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0005] 5. Cai Z, Hussain MD, Yan L-J (2014) Microglia, neuroinflammation, and beta-amyloid protein in Alzheimer's disease. Int J Neurosci 124:307–321. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0006] 6. Camandola S, Mattson MP (2017) Brain metabolism in health, aging, and neurodegeneration. EMBO J 36:1474–1492. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0007] 7. Clayton KA, Van Enoo AA, Ikezu T (2017) Alzheimer's disease: the role of microglia in brain homeostasis and proteopathy. Front Neurosci. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5733046/ (accessed 23 April 2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0008] 8. Desai BS, Schneider JA, Li J-L, Carvey PM, Hendey B (2009) Evidence of angiogenic vessels in Alzheimer's disease. J Neural Transm 116:587–597. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0009] 9. Faraco G, Park L, Anrather J, Iadecola C (2017) Brain perivascular macrophages: characterization and functional roles in health and disease. J Mol Med 95:1143–1152. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0010] 10. Goldmann T, Wieghofer P, Jordão MJC, Prutek F, Hagemeyer N, Frenzel K et al (2016) Origin, fate and dynamics of macrophages at central nervous system interfaces. Nat Immunol 17:797–805. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0011] 11. Guidoux C, Hauw J-J, Klein IF, Labreuche J, Berr C, Duyckaerts C et al (2018) Amyloid angiopathy in brain hemorrhage: a postmortem neuropathological-magnetic resonance imaging study. Cerebrovasc Dis 45:124–131. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0012] 12. Halliday MR, Rege SV, Ma Q, Zhao Z, Miller CA, Winkler EA et al (2016) Accelerated pericyte degeneration and blood–brain barrier breakdown in apolipoprotein E4 carriers with Alzheimer's disease. J Cereb Blood Flow Metab 36:216–227. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0013] 13. Hase Y, Ding R, Harrison G, Hawthorne E, King A, Gettings S et al (2019) White matter capillaries in vascular and neurodegenerative dementias. Acta Neuropathol Commun 7:16. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0014] 14. Hawkes CA, McLaurin J (2009) Selective targeting of perivascular macrophages for clearance of β-amyloid in cerebral amyloid angiopathy. PNAS 106:1261–1266. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0015] 15. Hopperton KE, Mohammad D, Trépanier MO, Giuliano V, Bazinet RP (2018) Markers of microglia in post‐mortem brain samples from patients with Alzheimer's disease: a systematic review. Mol Psychiatry 23:177–198. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0016] 16. Jefferies WA, Price KA, Biron KE, Fenninger F, Pfeifer CG, Dickstein DL (2013) Adjusting the compass: new insights into the role of angiogenesis in Alzheimer's disease. Alzheimer's Res Ther 5:64. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0017] 17. Kacem K, Lacombe P, Seylaz J, Bonvento G (1998) Structural organization of the perivascular astrocyte endfeet and their relationship with the endothelial glucose transporter: a confocal microscopy study. Glia 23:1–10. [PubMed] [Google Scholar]

[bpa12895-bib-0018] 18. Kamphuis W, Kooijman L, Schetters S, Orre M, Hol EM (2016) Transcriptional profiling of CD11c-positive microglia accumulating around amyloid plaques in a mouse model for Alzheimer's disease. Biochim Biophys Acta 1862:1847–1860. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0019] 19. Kamphuis W, Middeldorp J, Kooijman L, Sluijs JA, Kooi E-J, Moeton M et al (2014) Glial fibrillary acidic protein isoform expression in plaque related astrogliosis in Alzheimer's disease. Neurobiol Aging 35:492–510. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0020] 20. Karperien A, Ahammer H, Jelinek H (2013) Quantitating the subtleties of microglial morphology with fractal analysis. Front Cell Neurosci 7:3. Available at: https://www.frontiersin.org/articles/10.3389/fncel.2013.00003/full (accessed 23 April 2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0021] 21. Keable A, Fenna K, Yuen HM, Johnston DA, Smyth NR, Smith C et al (2016) Deposition of amyloid β in the walls of human leptomeningeal arteries in relation to perivascular drainage pathways in cerebral amyloid angiopathy. Biochim Biophys Acta 1862:1037–1046. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0022] 22. Kierdorf K, Masuda T, Jordão MJC, Prinz M (2019) Macrophages at CNS interfaces: ontogeny and function in health and disease. Nat Rev Neurosci 20, 547–562. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0023] 23. Kim SJ, Seo Y, Kim HJ, Na DL, Seo SW, Kim Y et al (2018) Pathologically confirmed cerebral amyloid angiopathy with no radiological sign in a patient with early onset Alzheimer's disease. Yonsei Med J 59:801–805. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0024] 24. Kirabali T, Rigotti S, Siccoli A, Liebsch F, Shobo A, Hock C et al (2019) The amyloid-β degradation intermediate Aβ34 is pericyte‐associated and reduced in brain capillaries of patients with Alzheimer's disease. Acta Neuropathol Commun 7:194. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0025] 25. Kisler K, Nelson AR, Montagne A, Zlokovic BV (2017) Cerebral blood flow regulation and neurovascular dysfunction in Alzheimer disease. Nat Rev Neurosci 18:419. Available at: https://www.nature.com/nrn/journal/vaop/ncurrent/full/nrn.2017.48.html (accessed 2 June 2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0026] 26. Kisler K, Nelson AR, Rege SV, Ramanathan A, Wang Y, Ahuja A et al (2017) Pericyte degeneration leads to neurovascular uncoupling and limits oxygen supply to brain. Nat Neurosci 20:406–416. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0027] 27. Koistinaho M, Lin S, Wu X, Esterman M, Koger D, Hanson J et al (2004) Apolipoprotein E promotes astrocyte colocalization and degradation of deposited amyloid-β peptides. Nat Med 10:719–726. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0028] 28. Kulic L, McAfoose J, Welt T, Tackenberg C, Späni C, Wirth F et al (2012) Early accumulation of intracellular fibrillar oligomers and late congophilic amyloid angiopathy in mice expressing the Osaka intra-Aβ APP mutation. Transl Psychiatry 13:e183. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0029] 29. Magaki S, Tang Z, Tung S, Williams CK, Lo D, Yong WH et al (2018) The effects of cerebral amyloid angiopathy on integrity of the blood‐brain barrier. Neurobiol Aging 70:70–77. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0030] 30. Miners JS, Schulz I, Love S (2017) Differing associations between Aβ accumulation, hypoperfusion, blood–brain barrier dysfunction and loss of PDGFRB pericyte marker in the precuneus and parietal white matter in Alzheimer's disease. J Cereb Blood Flow Metab 38:103–115. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0031] 31. Montagne A, Barnes SR, Sweeney MD, Halliday MR, Sagare AP, Zhao Z et al (2015) blood‐brain barrier breakdown in the aging human hippocampus. Neuron 85:296–302. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0032] 32. Montagne A, Nation DA, Sagare AP, Barisano G, Sweeney MD, Chakhoyan A et al (2020) APOE4 leads to blood‐brain barrier dysfunction predicting cognitive decline. Nature 581:71–76. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0033] 33. Montagne A, Zhao Z, Zlokovic BV (2017) Alzheimer's disease: a matter of blood‐brain barrier dysfunction? J Exp Med 214:3151–3169. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0034] 34. Montine TJ, Phelps CH, Beach TG, Bigio EH, Cairns NJ, Dickson DW et al (2012) National Institute on Aging-Alzheimer's Association guidelines for the neuropathologic assessment of Alzheimer's disease: a practical approach. Acta Neuropathol 123:1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0035] 35. Mucke L, Masliah E, Yu GQ, Mallory M, Rockenstein EM, Tatsuno G et al (2000) High-level neuronal expression of abeta 1–42 in wild-type human amyloid protein precursor transgenic mice: synaptotoxicity without plaque formation. J. Neurosci 20:4050–4058. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0036] 36. Nelson AR, Sweeney MD, Sagare AP, Zlokovic BV (2016) Neurovascular dysfunction and neurodegeneration in dementia and Alzheimer's disease. Biochim Biophys Acta 1862:887–900. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0037] 37. Park L, Uekawa K, Garcia-Bonilla L, Koizumi K, Murphy M, Pistik R et al (2017) Brain perivascular macrophages initiate the neurovascular dysfunction of Alzheimer Aβ peptides. Circ Res 121:258–269. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0038] 38. Perl DP (2010) Neuropathology of Alzheimer's disease. Mt Sinai J Med 77:32–42. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0039] 39. Pike CJ, Cummings BJ, Cotman CW (1995) Early association of reactive astrocytes with senile plaques in Alzheimer's disease. Exp Neurol 132:172–179. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0040] 40. Plog BA, Nedergaard M (2018) The glymphatic system in central nervous system health and disease: past, present, and future. Annu Rev Pathol 24:379–394. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0041] 41. Rajkowska G, Hughes J, Stockmeier CA, Javier Miguel-Hidalgo J, Maciag D (2013) Coverage of blood vessels by astrocytic endfeet is reduced in major depressive disorder. Biol Psychiat 73:613–621. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0042] 42. Rasmussen MK, Mestre H, Nedergaard M (2018) The glymphatic pathway in neurological disorders. Lancet Neurol 17:1016–1024. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0043] 43. Richard E, van Gool WA, Hoozemans JJM, van Haastert ES, Eikelenboom P, Rozemuller AJM et al (2010) Morphometric changes in the cortical microvascular network in Alzheimer's disease. J Alzheimers Dis 22:811–818. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0044] 44. Rust R, Grönnert L, Dogançay B, Schwab ME (2019) A revised view on growth and remodeling in the retinal vasculature. Sci Rep 9:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0045] 45. Rust R, Grönnert L, Gantner C, Enzler A, Mulders G, Weber RZ et al (2019) Nogo-A targeted therapy promotes vascular repair and functional recovery following stroke. PNAS 116:14270–14279. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0046] 46. Rust R, Kirabali T, Grönnert L, Dogancay B, Limasale YDP, Meinhardt A et al (2020) A practical guide to the automated analysis of vascular growth, maturation and injury in the brain. Front Neurosci 14:244. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0047] 47. Rust R, Weber RZ, Grönnert L, Mulders G, Maurer MA, Hofer A-S et al (2019) Anti-Nogo-A antibodies prevent vascular leakage and act as pro-angiogenic factors following stroke. Sci Rep 9:20040. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0048] 48. Ryu JK, McLarnon JG (2009) A leaky blood‐brain barrier, fibrinogen infiltration and microglial reactivity in inflamed Alzheimer's disease brain. J. Cell Mol Med 13:2911–2925. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0049] 49. Sengillo JD, Winkler EA, Walker CT, Sullivan JS, Johnson M, Zlokovic BV (2013) Deficiency in mural vascular cells coincides with blood‐brain barrier disruption in Alzheimer's disease. Brain Pathol 23:303–310. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0050] 50. Sofroniew MV (2015) Astrogliosis. Cold Spring Harb Perspect Biol 7:a020420. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4315924/ (accessed 7 November 2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0051] 51. Sofroniew MV (2015) Astrocyte barriers to neurotoxic inflammation. Nat Rev Neurosci 16:249–263. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0052] 52. Sweeney MD, Kisler K, Montagne A, Toga AW, Zlokovic BV (2018) The role of brain vasculature in neurodegenerative disorders. Nat Neurosci 21:1318–1331. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0053] 53. Sweeney MD, Montagne A, Sagare AP, Nation DA, Schneider LS, Chui HC et al (2019) Vascular dysfunction-the disregarded partner of Alzheimer's disease. Alzheimers Dement 15:158–167. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0054] 54. Tarantini S, Tran CHT, Gordon GR, Ungvari Z, Csiszar A (2017) Impaired neurovascular coupling in aging and Alzheimer's disease: contribution of astrocyte dysfunction and endothelial impairment to cognitive decline. Exp Gerontol 94:52–58. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0055] 55. Tarasoff-Conway JM, Carare RO, Osorio RS, Glodzik L, Butler T, Fieremans E et al (2015) Clearance systems in the brain—implications for Alzheimer disease. Nat Rev Neurol 11:457–470. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0056] 56. Thanopoulou K, Fragkouli A, Stylianopoulou F, Georgopoulos S (2010) Scavenger receptor class B type I (SR-BI) regulates perivascular macrophages and modifies amyloid pathology in an Alzheimer mouse model. Proc Natl Acad Sci USA 107:20816–20821. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0057] 57. Tian J, Shi J, Mann DMA (2004) Cerebral amyloid angiopathy and dementia. Panminerva Med 46:253–264. [PubMed] [Google Scholar]

[bpa12895-bib-0058] 58. Tong X-K, Nicolakakis N, Kocharyan A, Hamel E (2005) Vascular remodeling versus amyloid β-induced oxidative stress in the cerebrovascular dysfunctions associated with Alzheimer's disease. J Neurosci 25:11165–11174. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0059] 59. Vagnucci AH, Li WW (2003) Alzheimer's disease and angiogenesis. Lancet 361:605–608. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0060] 60. Van Dorpe J, Smeijers L, Dewachter I, Nuyens D, Spittaels K, Van den Haute C et al (2000) Prominent cerebral amyloid angiopathy in transgenic mice overexpressing the london mutant of human APP in neurons. Am J Pathol 157:1283–1298. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0061] 61. Vehmas AK, Kawas CH, Stewart WF, Troncoso JC (2003) Immune reactive cells in senile plaques and cognitive decline in Alzheimer's disease. Neurobiol Aging 24:321–331. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0062] 62. Verbeek MM, Van Nostrand WE, Otte-Höller I, Wesseling P, De Waal RMW (2000) Amyloid-β-induced degeneration of human brain pericytes is dependent on the apolipoprotein E genotype. Ann N Y Acad Sci 903:187–199. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0063] 63. Viña J, Lloret A (2010) Why women have more Alzheimer's disease than men: gender and mitochondrial toxicity of amyloid-β peptide. J Alzheimer's Dis 20:S527–S533. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0064] 64. Weller RO, Subash M, Preston SD, Mazanti I, Carare RO (2008) Perivascular drainage of amyloid-beta peptides from the brain and its failure in cerebral amyloid angiopathy and Alzheimer's disease. Brain Pathol 18:253–266. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0065] 65. Winkler EA, Sagare AP, Zlokovic BV (2014) The pericyte: a forgotten cell type with important implications for Alzheimer's disease? Brain Pathol 24:371–386. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0066] 66. Yamazaki Y, Kanekiyo T. blood‐brain barrier dysfunction and the pathogenesis of Alzheimer's disease. Int J Mol Sci 2017;18:1965. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0067] 67. Zenaro E, Piacentino G, Constantin G. The blood‐brain barrier in Alzheimer's disease. Neurobiol Dis. Available at: http://www.sciencedirect.com/science/article/pii/S0969996116301656 (accessed 2 December 2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0068] 68. Zeppenfeld DM, Simon M, Haswell JD, D’Abreo D, Murchison C, Quinn JF et al (2017) Association of perivascular localization of aquaporin-4 with cognition and Alzheimer disease in aging brains. JAMA Neurol 74:91–99. [DOI] [PubMed] [Google Scholar]

[bpa12895-bib-0069] 69. Zhao X, Eyo UB, Murugan M, Wu L-J (2018) Microglial interactions with the neurovascular system in physiology and pathology. Dev Neurobiol 78:604–617. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bpa12895-bib-0070] 70. Zlokovic BV (2011) Neurovascular pathways to neurodegeneration in Alzheimer's disease and other disorders. Nat Rev Neurosci 12:723–738. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Distinct changes in all major components of the neurovascular unit across different neuropathological stages of Alzheimer's disease

Tunahan Kirabali

Ruslan Rust

Serena Rigotti

Alessandro Siccoli

Roger M Nitsch

Luka Kulic

Abstract

Introduction

Methods

Post‐mortem human brain samples

Table 1.

Immunohistochemistry/Immunofluorescence

Quantification

Platelet‐derived growth factor receptor‐β (PDGFR‐β) quantification

Glial fibrillary acidic protein (GFAP) quantification

Ionized calcium‐binding adaptor molecule 1 (Iba 1) quantification

Mannose receptor C‐type 1 (MRC1) quantification

Lectin quantification

Tau quantification

Aβ plaque quantification

Statistical analysis

Results

Vessel distribution and pericyte coverage are altered in Alzheimer's disease progression

Figure 1.

Figure 2.

Astrocyte coverage within the NVU is higher in AD patients

Figure 3.

Changes in both microglia coverage and perivascular macrophage distribution are observed in AD patients

Figure 4.

Figure 5.

Correlations of NVU components with tau pathology and Aβ plaque load

Figure 6.

Discussion

Conflict of Interest

Supporting information

Data Availability Statement

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases