University of Southern California Dissertations and Theses

Blood-brain barrier pathophysiology in cognitive impairment and injury

(USC Thesis Other)

Blood-brain barrier pathophysiology in cognitive impairment and injury

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
i

BLOOD-BRAIN BARRIER PATHOPHYSIOLOGY IN
COGNITIVE IMPAIRMENT AND INJURY

by
MELANIE DANIELLE SWEENEY

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA

In Partial Fulfillment of the Requirements for the Degree
DOCTOR OF PHILOSOPHY
(NEUROSCIENCE)

May 2019

Copyright 2019 Sweeney, M.D.

ii
DEDICATION

This dissertation is dedicated to my family for their unwavering love and support.

iii
ACKNOWLEDGEMENTS

First and foremost, I would like to thank my mentor, Dr. Berislav V. Zlokovic, for the
endless support, time, guidance, and opportunities he has given me throughout my Ph.D. training.
His insight and drive have been a source of inspiration over the course of my research training. He
has helped me develop both as a scientist and as a person, and I will forever be grateful for this
experience.
I would also like to thank the members of my dissertation committee, Drs. Christian Pike
(Committee Chair), William Mack, Justin Ichida and Jae Jung, for their guidance throughout
graduate school. Additionally, thank you to my undergraduate research advisor, Dr. Kimberley
Phillips, for always believing in me and continuing to encourage me throughout graduate school.
To current and former Zlokovic laboratory members, thank you for making my Ph.D.
training enjoyable and for all your support along the way. Particularly thanks to Dr. Abhay Sagare,
Dr. Axel Montagne, Divna Lazic, Maricarmen Pachicano, Dr. Kassandra Kisler, Dr. Mariangela
Nikolakopoulou, Ching-Ju (Ruu) Hsu, Edward Zuniga, Sanket Rege, Ashim Ahuja, Dr. Zhen
Zhao, Dr. Zhonghua Dai, Dr. Yaoming Wang, Dr. Ararat Chakhoyan, Dr. Mikko Huuskonen, Sam
Bazzi, Erica Lawson, Alex Armendariz, Krupal Shah, Jacob Prince, and Caroline Murphy.
Thank you to the USC Neuroscience Graduate Program and my Ph.D. cohort – Louise
Menendez, Kirsten Lynch, Talia Nir, Katie Zyuzin, Rorry Brenner, Dan Rinker, Brenton Keller,
and Pan Kong – I could not have navigated the Ph.D. experience without you. Also, many thanks
to the USC Graduate Student Government and Cali Strong Fitness – Wednesday workouts were a
highlight of my week.
On a personal note, a sincere thanks to my entire family – the Sweeneys, McElraths,
Holcombes and O’Briens. Words fall short of how grateful I am for your love, support, and ability
to keep me laughing.

iv
TABLE OF CONTENTS
Page
Dedication ........................................................................................................................... ii
Acknowledgements ............................................................................................................ iii
Table of Contents ............................................................................................................... iv
List of Publications ............................................................................................................ vii
List of Abbreviations ........................................................................................................... x
List of Tables ...................................................................................................................... xi
List of Figures .................................................................................................................. xiii
Chapter 1: Introduction and Background .......................................................................... 16
1.1 The Blood-Brain Barrier (BBB) ...................................................................... 16
1.2 BBB Architecture and Transport Physiology .................................................. 18
1.2.1 Molecular Definition of the BBB Vascular Tree ............................. 20
1.2.2 BBB Junctional Molecules ............................................................... 23
1.2.3 BBB Transport Systems ................................................................... 25
1.2.4 Other Vascular-Mediated Transport ................................................. 34
1.3 Vascular Dysfunction in Alzheimer’s Disease ................................................ 36
1.3.1 Alzheimer’s Disease Pathophysiology ............................................. 36
1.3.2 Genetic Contributions ...................................................................... 38
1.3.3 BBB Breakdown and Dysfunction ................................................... 45
1.4 Summary and Overview of Chapters .............................................................. 52
Chapter 2: BBB Breakdown is an Early Biomarker of Human Cognitive Dysfunction ... 55
2.1 Introduction ..................................................................................................... 55
2.2 Methods ........................................................................................................... 56
2.3 Results ............................................................................................................. 67
2.4 Discussion ....................................................................................................... 82
2.5 Acknowledgements ......................................................................................... 82
2.6 Supplementary Tables ..................................................................................... 84

v
Chapter 3: BBB Breakdown Predicts Early Cognitive Dysfunction in APOE4 Carriers 101
3.1 Abstract ......................................................................................................... 101
3.2 Introduction ................................................................................................... 102
3.3 Methods ......................................................................................................... 103
3.4 Results ........................................................................................................... 111
3.5 Discussion ..................................................................................................... 119
3.6 Current and Future Directions ....................................................................... 121
3.7 Acknowledgements ....................................................................................... 122
Chapter 4: BBB Dysfunction in Autosomal Dominant Alzheimer’s Disease ................ 123
4.1 Abstract ......................................................................................................... 123
4.2 Introduction ................................................................................................... 124
4.3 Methods ......................................................................................................... 125
4.4 Results ........................................................................................................... 127
4.5 Discussion ..................................................................................................... 131
4.6 Acknowledgements ....................................................................................... 133
Chapter 5: A Novel Assay to Validate a Biomarker of Pericyte Injury .......................... 134
5.1 Abstract ......................................................................................................... 134
5.2 Introduction ................................................................................................... 135
5.3 Methods ......................................................................................................... 136
5.4 Results ........................................................................................................... 139
5.5 Discussion ..................................................................................................... 142
5.6 Acknowledgements ....................................................................................... 144
Chapter 6: Understanding the Molecular Mechanisms / Signatures of the
Cerebrovasculature Using a Mouse Model of Hypoxia .................................................. 145
6.1 Abstract ......................................................................................................... 145
6.2 Introduction ................................................................................................... 146
6.3 Methods ......................................................................................................... 147
6.4 Results ........................................................................................................... 155
6.5 Discussion ..................................................................................................... 159

vi
6.6 Future Directions ........................................................................................... 161
6.7 Acknowledgements ....................................................................................... 162
Chapter 7: Conclusions, Synthesis and Future Implications ........................................... 163
7.1 Summary of Findings .................................................................................... 163
7.1.1 Update on Dynamic Biomarkers in Alzheimer’s Disease .............. 165
7.2 Commonalities of Cerebrovascular Dysfunction in Neurodegenerative
Disorders ................................................................................................. 167
7.3 Targeting the BBB for Treatments ................................................................ 170
7.4 Lessons Learned and Future Directions ........................................................ 175
Appendix A: Human Brain Pericytes Shed Soluble PDGFRb ....................................... 178
Appendix B: Regional BBB Permeability by MRI ......................................................... 180
Appendix C: Regional BBB Permeability by MRI in APOE4 Carriers .......................... 182
Appendix D: BBB Breakdown in Neurodegenerative Disorders .................................... 184
Appendix E: BBB-Based Therapeutic Approaches ........................................................ 189
References ....................................................................................................................... 193

vii
LIST OF PUBLICATIONS
* denotes equally contributed first co-authors.
Published:
1. Nation DA*, Sweeney MD*, Montagne A*, Sagare AP, D’Orazio LM, Pachicano M,
Sepehrband F, Nelson AR, Buennagel DP, Harrington MG, Benzinger TLS, Fagan AM,
Ringman JM, Schneider LS, Morris JC, Chiu HC, Law M, Toga AW, Zlokovic BV. Blood-
brain barrier breakdown is an early biomarker of human cognitive dysfunction. Nature
Medicine 2019; 25(2):270-276. PMID: 30643288. PMCID: PMC6367058.
2. Sagare AP*, Sweeney MD*, Nelson AR*, Zhao Z, Zlokovic BV. Prion protein antagonists
rescue Alzheimer’s amyloid-beta-related cognitive deficits. Trends in Molecular Medicine
2019; 25(2):74-76. PMID: 30661727. PMCID: PMC6377285.
3. Sweeney MD, Montagne A, Sagare AP, Nation DA, Schneider LS, Chui HC, Harrington MG,
Pa J, Law M, Wang DJJ, Jacobs RE, Doubal FN, Ramirez J, Black SE, Nedergaard M,
Benveniste H, Dichgans M, Iadecola C, Love S, Bath PM, Markus HS, Salman RA, Allan SM,
Quinn TJ, Kalaria RN, Werring DJ, Carare RO, Touyz RM, Williams SCR, Moskowitz MA,
Katusic ZS, Lutz SE, Lazarov O, Minshall RD, Rehman J, Davis TP, Wellington CL, González
HM, Yuan C, Lockhart SN, Hughes TM, Chen CLH, Sachdev P, O’Brien JT, Skoog I, Pantoni
L, Gustafson DR, Biessels GJ, Wallin A, Smith EE, Mok V, Wong A, Passmore P, Barkof F,
Muller M, Breteler MMB, Román GC, Hamel E, Seshadri S, Gottesman RF, van Buchem MA,
Arvanitakis Z, Schneider JA, Drewes LR, Hachinski V, Finch CE, Toga AW, Wardlaw JM,
Zlokovic BV. Vascular dysfunction – the disregarded partner of Alzheimer’s disease.
Alzheimer’s & Dementia 2019; 15(1):158-167. PMID: 30642436. PMCID: PMC6338083.
4. Sweeney MD*, Zhao Z*, Montagne A, Nelson AR, Zlokovic BV. Blood-brain barrier: From
physiology to disease and back. Physiological Reviews 2019; 99(1):21-78. PMID: 30280653.
PMCID: PMC6335099.
5. Sweeney MD*, Kisler K*, Montagne A*, Toga AW, Zlokovic BV. The role of brain
vasculature in neurodegenerative disorders. Nature Neuroscience 2018; 21(10):1318-1331.
PMID: 30250261. PMCID: PMC6198802.
6. Sweeney MD & Zlokovic BV. A lymphatic waste-disposal system implicated in Alzheimer’s
disease. Nature 2018; 560(7717):172-174. PMID: 30076374. PMCID: PMC6201839.
7. Kisler K, Lazic D, Sweeney MD, Plunkett S, El Khatib M, Vinogradov SA, Boas DA,
Sakadžić S, Zlokovic BV. In vivo imaging and analysis of cerebrovascular hemodynamic
responses and tissue oxygenation in the mouse brain. Nature Protocols 2018; 13(6):1377-
1402. PMID: 29844521.
8. Sweeney MD, Sagare AP, Zlokovic BV. Blood-brain barrier breakdown in Alzheimer’s
disease and other neurodegenerative disorders. Nature Reviews Neurology 2018; 14(3):133-
150. PMID: 29377008. PMCID: PMC5829048.

viii
9. Sweeney MD, Ayyadurai S, Zlokovic BV. Pericytes of the neurovascular unit: key functions
and signaling pathways. Nature Neuroscience 2016; 19(6):771-783. PMID: 27227366.
PMCID: PMC5745011.
10. Montagne A, Nation DA, Pa J, Sweeney MD, Toga AW, Zlokovic BV. Brain imaging of
neurovascular dysfunction in Alzheimer’s disease. Acta Neuropathologica 2016; 131(5):687-
707. PMID: 27038189. PMCID: PMC5283382.
11. Nelson AR, Sweeney MD, Sagare AP, Zlokovic BV. Cerebrovascular contributions to
Alzheimer’s disease pathophysiology. Biochimica et Biophysica Acta 2015; 1862(5):887-900.
PMID: 26705676. PMCID: PMC4821735.
12. Sagare AP, Sweeney MD, Makshanoff J, Zlokovic BV. Shedding of soluble platelet-derived
growth factor receptor-β from human brain pericytes. Neuroscience Letters 2015; 607:97-101.
PMID: 26407747. PMCID: PMC4631673.
13. Sweeney MD, Sagare AP, Zlokovic BV. Cerebrospinal fluid biomarkers of neurovascular
dysfunction in mild dementia and Alzheimer’s disease. Journal of Cerebral Blood Flow and
Metabolism 2015; 35(7):1055-1068. PMID: 25899298. PMCID: PMC4640280.
14. Montagne A, Barnes SR, Sweeney MD, Halliday MR, Sagare AP, Zhao Z, Toga AW, Jacobs
RE, Liu CY, Amezcua L, Harrington MG, Chui HC, Law M, Zlokovic BV. Blood-brain barrier
breakdown in the aging human hippocampus. Neuron 2015; 85(2):296-302. PMID: 25611508.
PMCID: PMC4350773.
15. Dybdal-Hargreaves NF*, Holder

ND*, Ottoson PE*, Sweeney MD*, Williams T.
Mephedrone: Public health risks, mechanisms of action, and behavioral effects. European
Journal of Pharmacology 2013; 714(1-3):32-40. PMID: 23764466.
Under Review or In Preparation:
16. Nikolakopoulou AM*, Montagne A*, Kisler K*, Dai Z*, Wang Y, Huuskonen MT, Sagare
AP, Lazic D, Sweeney MD, Kong P, Wang M, Owens NC, Lawson EJ, Xie X, Zhao Z,
Zlokovic BV. Pericyte loss leads to circulatory failure and pleiotrophin depletion causing
neuron loss. Nature Neuroscience; In Revision.
17. Montagne A, Huuskonen MT, Rajagopal G, Sweeney MD, Nation DA, Sepehrband F,
D’Orazio LM, Harrington MG, Chui HC, Law M, Toga AW, Zlokovic BV. Undetectable
gadolinium brain retention in individuals with an age-dependent blood-brain barrier
breakdown in the hippocampus and mild cognitive impairment. Alzheimer’s & Dementia; In
Revision.
18. Wardlaw J, Benveniste H, Nedergaard M, Zlokovic BV, Mestre H, Lee H, Doubal F, Brown
R, Ramirez J, MacIntosh B, Tannenbaum A, Ballerini L, Rungta R, Boido D, Sweeney MD,
Montagne A, Charpak S, Joutel A, Smith K, Black S. Perivascular spaces in the brain:
Anatomy, physiology, and contributions to pathology of brain diseases. Nature Reviews
Neurology; Under Review.

ix
19. Sweeney MD*, Sagare AP*, Harrington MG, Chui HC, Schneider LS, Ringman JM, Fagan
AM, Morris JC, Toga AW, Zlokovic BV. A novel assay to validate a biomarker of pericyte
injury in human cerebrospinal fluid. Current Pharmaceutical Design. Invited – In Preparation.
20. Sweeney MD*, Montagne A*, Nation DA*, Chakhoyan A, Sagare AP, Pachicano M,
Sepehrband F, Harrington MG, Buennagel DP, Ringman JM, Joe E, Schneider LS, Pa J,
Benzinger TLS, Fagan AM, Morris JC, Law M, Chiu HC, Toga AW, Zlokovic BV. Blood-
brain barrier breakdown predicts early cognitive dysfunction in APOE4 carriers independent
of Alzheimer’s amyloid-β and tau. In Preparation.
21. Sweeney MD, Sagare AP, Coppola G, Chui HC, Zlokovic BV, Ringman JM. Cerebrovascular
dysfunction in autosomal dominant Alzheimer’s disease. In Preparation.
22. Sweeney MD, Huuskonen MT, Montagne A, Murphy C, Zhao Z…Zlokovic BV. Hypoxia-
induced vascular responses and molecular signatures in the adult mouse brain. In Preparation.

x
LIST OF ABBREVIATIONS
AD, Alzheimer’s disease
ADAD, Autosomal dominant Alzheimer’s disease
ADAM10, A disintegrin and metalloproteinase domain-containing protein-10
ADRC, Alzheimer’s Disease Research Center
AJ, Adherens junctions
APOE, Apolipoprotein E
APP, Amyloid precursor protein
Aβ, Amyloid-beta
BBB, Blood-brain barrier
CAA, Cerebral amyloid angiopathy
CBF, Cerebral blood flow
CDR, Clinical Dementia Rating
CNS, Central nervous system
CSF, Cerebrospinal fluid
CypA, Cyclophilin A
DCE-MRI, Dynamic contrast-enhanced magnetic resonance imaging
ELISA, Enzyme-linked immunosorbent assay
HC, Hippocampus
LRP1, Low-density lipoprotein receptor-related protein-1
MC, Mutation carriers
MCI, Mild cognitive impairment
MMP9, Matrix metalloproteinase-9
MSD, Meso Scale Discovery
NCI, No cognitive impairment
NVU, Neurovascular unit
PDGFRβ, Platelet-derived growth factor receptor-beta
PET, Positron emission tomography
PHC, Parahippocampus
PSEN, Presenilin
pTau, Phosphorylated tau
Qalb, Albumin quotient
RAGE, Receptor for advanced glycation endproducts
RBC, Red blood cells
SMC, Smooth muscle cell
sPDGFRβ, Soluble platelet-derived growth factor receptor-beta
TJ, Tight junctions
UDS, Uniform Data Set
VRF, Vascular risk factors

xi
LIST OF TABLES
Table Page

Table 2.1 Total sample by level of Clinical Dementia Rating (CDR) score. ......................... 84
Table 2.2 Total sample by number of cognitive domains impaired. ...................................... 85
Table 2.3 Logistic regression models of CSF sPDGFRβ or regional BBB Ktrans
beyond demographic factors and AD biomarkers in relation to CDR status. ....... 86
Table 2.4 Logistic regression models of CSF sPDGFRβ or regional BBB Ktrans
beyond AD biomarkers in relation to cognitive domains impaired. ..................... 89
Table 2.5 Logistic regression models of CSF sPDGFRβ or regional BBB Ktrans
beyond demographic factors and regional brain volume in relation to
CDR status. ............................................................................................................ 92
Table 2.6 Logistic regression models of CSF sPDGFRβ or regional BBB Ktrans
beyond regional brain volume in relation to cognitive domains impaired. ........... 94
Table 2.7 Logistic regression models of AD biomarkers with demographic factors
and beyond CSF sPDGFRβ in relation to CDR status. ......................................... 95
Table 2.8 Logistic regression models of AD biomarkers with demographic factors
and beyond regional BBB Ktrans in relation to CDR status. .................................. 96
Table 2.9 Logistic regression models of AD biomarkers and predictive value beyond
CSF sPDGFRβ in relation to cognitive domains impaired. .................................. 98
Table 2.10 Logistic regression models of AD biomarkers and predictive value
beyond regional BBB Ktrans in relation to cognitive domains impaired. ............... 99
Table 3.1 Participants’ demographic information for the biofluid cohort. ........................... 111
Table 4.1 Participants’ demographic and clinical information. ............................................ 127
Table 5.1 Summary of reagents used to develop and optimize the sPDGFRβ assay
on the MSD platform. .......................................................................................... 139
Table 5.2 Participants’ demographic information. ............................................................... 141
Table C.1 Participants’ demographic information for the MRI cohort. ............................... 182
Table D.1 BBB disruption by neuroimaging in neurodegenerative disorders. .................... 184

xii
Table D.2 BBB disruption by CSF analysis in neurodegenerative disorders. ...................... 185
Table D.3 BBB disruption by post-mortem tissue analysis in neurodegenerative
disorders. ............................................................................................................. 186
Table E.1 Circumventing, protecting and traversing the BBB for treatments. .................... 189

xiii
LIST OF FIGURES
Figure Page

Figure 1.1 The blood-brain barrier. ........................................................................................ 17
Figure 1.2 Molecular definition of the BBB and cerebral blood vessels. .............................. 21
Figure 1.3 BBB endothelial connections. ............................................................................... 24
Figure 1.4 Major BBB transport systems. .............................................................................. 27
Figure 1.5 Brain perivascular and paravascular transport. ..................................................... 35
Figure 1.6 Alzheimer’s disease is a multifactorial and heterogeneous disease. ..................... 37
Figure 1.7 Effects of genetic mutations carrying inheritance or increasing risk for AD
on BBB dysfunction. ............................................................................................. 39
Figure 1.8 BBB breakdown and dysfunction in AD. ............................................................. 46
Figure 2.1 Molecular biomarkers of the neurovascular unit. ................................................. 68
Figure 2.2 Increased CSF sPDGFRβ with CDR impairment is independent of Aβ and
tau status. ............................................................................................................... 69
Figure 2.3 Site-specific analysis confirming increased CSF sPDGFRβ with CDR
impairment, independent of Aβ and tau status. ..................................................... 70
Figure 2.4 Increased CSF sPDGFRb with CDR is independent of VRFs and
confirmed by site-specific analysis. ...................................................................... 71
Figure 2.5 Other CSF biomarkers of the neurovascular unit are not altered with CDR
impairment, confirmed by site-specific analysis. .................................................. 72
Figure 2.6 Increased CSF sPDGFRβ with cognitive domain impairment is
independent of Aβ, tau and VRF status. ................................................................ 74
Figure 2.7 Site-specific analysis confirming increased CSF sPDGFRβ with cognitive
domain impairment, independent of Aβ, tau and VRF status. .............................. 75
Figure 2.8 Other CSF biomarkers of the neurovascular unit are not altered with
cognitive domain impairment, confirmed by site-specific analysis. ..................... 77
Figure 2.9 CSF sPDGFRβ is not related to age but positively associates with CSF
BBB breakdown markers. ..................................................................................... 77

xiv
Figure 2.10 ADAM10 mediates sPDGFRβ shedding in human brain pericytes in vitro. ...... 78
Figure 2.11 CSF sPDGFRβ relates to hippocampal gray matter regions. .............................. 79
Figure 2.12 Increased CSF sPDGFRβ during cognitive impairment is independent of
(para)hippocampal volume. ................................................................................... 81
Figure 3.1 Pericyte injury and BBB breakdown increase in APOE4 carriers with
cognitive impairment. .......................................................................................... 112
Figure 3.2 Other CSF biomarkers of NVU cell and system injury are not differentially
altered during cognitive impairment. .................................................................. 114
Figure 3.3 CSF Aβ1-42 and pTau in APOE carriers during cognitive impairment. ............... 115
Figure 3.4 CSF sPDGFRβ increases in APOE4 carriers with cognitive impairment,
independent of Aβ and tau. ................................................................................. 116
Figure 3.5 Qalb increases in APOE4 carriers with cognitive impairment, independent
of Aβ and tau. ...................................................................................................... 118
Figure 3.6 CSF cyclophilin A increases in APOE4 carriers and with cognitive
impairment. ......................................................................................................... 119
Figure 3.7 High CSF sPDGFRβ at baseline predicts subsequent cognitive decline in
APOE4 carriers. ................................................................................................... 121
Figure 4.1 CSF biomarkers of cerebrovascular dysfunction are increased in ADAD. ........ 129
Figure 4.2 Standard CSF Alzheimer’s biomarkers, Aβ42 and tau, are altered in ADAD. .. 130
Figure 4.3 Cerebrovascular and standard AD biomarkers inform ADAD
pathophysiological progression. .......................................................................... 131
Figure 5.1 Performance summary of the novel sPDGFRβ assay. ........................................ 140
Figure 5.2 Validation of sPDGFRβ as a pericyte injury biomarker in human CSF. ............ 141
Figure 6.1 Cerebral microvascular remodeling during hypoxia. .......................................... 156
Figure 6.2 Transient blood-brain barrier (BBB) breakdown during hypoxia. ..................... 157
Figure 6.3 Chronic mild hypoxia does not affect peripheral physiological functions. ........ 158
Figure 6.4 Molecular regulation of cerebrovascular responses during hypoxia. ................. 159

xv
Figure 7.1 Hypothetical updated Jack model of AD biomarkers to include the role of
brain vasculature. ................................................................................................ 166
Figure 7.2 Commonality of an early involvement of the CNS vasculature in different
neurodegenerative disorders. ............................................................................... 170
Figure 7.3 BBB dysfunction – implications for drug delivery. ............................................ 172
Figure A.1 Abundant PDGFRb expression in human brain pericytes compared to
other vascular cell types. ..................................................................................... 178
Figure A.2 Hypoxia and Ab peptide lead to shedding of sPDGFRb from human brain
pericytes, but not from SMCs. ............................................................................ 179
Figure B.1 BBB breakdown in the (para)hippocampus with increased CDR is
independent of Aβ and tau status and regional volume. ..................................... 180
Figure B.2 BBB breakdown in the (para)hippocampus in individuals with increased
cognitive domain impairment is independent of Aβ and tau status and
regional volume. .................................................................................................. 181
Figure C.1 BBB Ktrans in the (para)hippocampus increases in APOE4 carriers,
independent of Aβ and tau. ................................................................................. 183
Figure D.1 BBB breakdown promotes neurodegeneration. ................................................. 188

16
CHAPTER 1:
INTRODUCTION AND BACKGROUND

Adapted from:
Sweeney MD…Zlokovic BV, Physiological Reviews, 2019
Sweeney MD…Zlokovic BV, Alzheimer’s & Dementia, 2019
Sweeney MD…Zlokovic BV, Nature Neuroscience, 2018
Sweeney MD…Zlokovic BV, Nature Reviews Neurology, 2018
Sweeney MD…Zlokovic BV, Nature Neuroscience, 2016
Sweeney MD…Zlokovic BV, Journal of Cerebral Blood Flow and Metabolism, 2015
1.1 The Blood-Brain Barrier (BBB)
The human brain has approximately 400 miles of blood vessels that supply brain cells with
oxygen, energy metabolites and nutrients, and remove carbon dioxide and other metabolic waste
products from brain into circulation
1,2
. Although 2% of total body mass, the brain consumes
approximately 20% of the body’s glucose and oxygen, and can rapidly increase blood flow and
oxygen delivery to its activated regions, known as neurovascular coupling
2,3
. Capillaries are the
smallest cerebral blood vessels that account for approximately 85% of cerebral vessel length, and
are a major site of the blood-brain barrier (BBB)
1
. In the human brain, capillaries provide
approximately 12 m
2
of endothelial cell surface area available for transport of solutes from the
blood to the brain, and vice versa. The mean inter-capillary distance in the human brain is
approximately 40 μm
4
, which allows almost instantaneous solute equilibration throughout the
brain interstitial space once the molecules cross the BBB. Energy substrates are consumed by the
brain “on the fly” from blood via transport across the BBB, as the brain lacks a reservoir to store
fuel for use when needed
5
.
The BBB is a continuous endothelial barrier with sealed cell-to-cell contacts resulting in
high transendothelial electrical resistance and low para-cellular and trans-cellular permeability
6,7
.
The endothelial monolayer is ensheathed by mural cells (pericytes at capillaries and vascular
smooth muscle cells (SMCs) at arterioles and arteries) and astrocyte endfeet
6,8
. The BBB is present
at all levels of the vascular tree from arteries and arterioles to capillaries to venules and veins

17
(termed arteriovenous axis)
6,8
(Figure 1.1), but capillaries constitute the largest BBB surface area
9
.
In contrast to highly permeable systemic capillaries
10
, brain capillaries exhibit a low rate of trans-
endothelial bulk flow transcytosis which, together with tightly-sealed endothelium, restricts entry
of most blood-derived macromolecules, cells and microbial pathogens into the brain unless they
have specialized carriers and/or receptors in brain endothelium that facilitate their transport across
the BBB. Pericytes, SMCs, and endothelial cells express thousands of transcripts encoding
different transporters, receptors, active efflux pumps, ion channels, and regulatory molecules
11–20
,

Figure 1.1 The blood-brain barrier.
Brain capillaries are a key site of the BBB. The capillary cross section (large inset) shows a tightly
sealed endothelium, which shares a common basement membrane with pericytes and astrocyte
end-feet wrapping around the capillary wall. The arterial cross section (small inset) shows
perivascular flow of interstitial fluid (ISF) through the arterial wall in the opposite direction to
blood flow; paravascular flow might also occur in the same direction as blood flow. Cerebrospinal
fluid (CSF) is produced by the choroid plexus and flows from brain ventricles into subarachnoid
spaces, draining into the meningeal lymphatic system and/or venous blood through the arachnoid
villi. ISF can exchange with CSF in the ventricles (not shown) and subarachnoid spaces.
Abbreviation: ECS, extracellular space.

18
whose expression pattern varies by cell type and location along the vascular tree
18
. The BBB
architecture, junctional molecules and transport systems will be introduced and discussed in
greater detail below in Section 1.2 BBB Architecture and Transport Physiology.
Cerebrovascular integrity at the BBB is maintained by continuous cross-talk between
endothelium, mural cells, glia (astrocytes and microglia) and neurons – cell types that collectively
compose the neurovascular unit (NVU)
1,5,7
. The pattern of cerebral blood vessels follows the major
brain circuits tasked with sensation, memory and motion suggesting that cerebrovascular system
plays an important role in normal central nervous system (CNS) functioning
2,3,5,6
. BBB integrity
is critical for proper synaptic functioning, information processing, and neuronal connectivity. Loss
of BBB integrity results in vascular permeability, transport dysfunction and impaired proteostasis,
and is associated with reduced cerebral blood flow (CBF) and impaired hemodynamic
responses
2,3,5,7,8,21
. BBB breakdown and dysfunction allows toxic blood-derived molecules, cells,
and pathogens to enter the CNS and aberrant transport and clearance of molecules
6,7,9
, which
contribute to neurological deficits. Cerebrovascular dysfunction is increasingly recognized in
Alzheimer’s disease (AD) pathophysiology, which has been studied in animal models of AD, as
recently reviewed
9
. In humans, evidence of BBB breakdown has been determined by
neuroimaging, neuropathological and molecular biofluid findings. This evidence will be presented
and discussed below in Section 1.3 Vascular Dysfunction in Alzheimer’s disease.
Original evidence that BBB breakdown is an early biomarker of cognitive impairment in
humans, and the impact of AD genetic factors on cerebrovascular dysfunction, are presented and
discussed in Chapters 2-4. Sensitive, reliable detection and clinical-relevance of a novel
biomarker of pericyte injury is then validated in Chapter 5. Next, a mouse model of hypoxia that
exhibits BBB breakdown is used to better understand molecular mechanisms and signatures of
cerebrovascular dysfunction in Chapter 6. Finally, Chapter 7 will synthesize the original findings
within the current perspectives in the field of neurodegenerative diseases and discuss gaps in the
field and future directions.
1.2 BBB Architecture and Transport Physiology
In contrast to leaky capillary endothelium in peripheral organs
10
, the BBB endothelium is
sealed by TJs
22,23,6
and has low rate of bulk-flow transcytosis
6,24
. Brain endothelial molecular

19
junctions, transporters, receptors, and channels have been initially discovered by physiological
experiments and ultrastructural studies
1,25
, which was followed by transcriptomic approaches of
endothelial and vascular mural cells. These include suppression subtractive hybridization
13,16,17
,
microarrays
11,12,14
, and RNA-sequencing analysis
15,19,20
.
Earlier studies of the BBB transcriptome were conducted on isolated rat brain capillaries
containing endothelial cells and pericytes together
16,17
. More recent studies have used endothelial-
specific Tie2-eGFP (enhanced green fluorescent protein) transgenic mice to investigate
transcriptomes of eGFP-positive brain capillary endothelial cells purified by fluorescence-
activated cell sorting (FACS)
14
; GFP-positive brain endothelial cells purified by FACS from Tie2-
GFP and Pdgfrb-positive pericytes purified by immunopanning
20
; and microvascular fragments
isolated from brains of pericyte-deficient Pdgfb, Pdgfrb and Pdgfb
ret/ret
mice and controls
11,12
.
RNA-sequencing analysis of mural cell transcriptome has been recently performed in
pericytes purified by FACS for two markers, PDGFRβ and NG2 (neural/glial antigen 2), expressed
in Pdgfrb-eGFP; chondroitin sulfate proteoglycan-4 (Cspg4)-DsRed mice
15
. Single cell RNA-
sequencing analysis of endothelial and pericyte clusters from the mouse brain has been also
reported
18,19
. The recent Nature paper presents a landmark molecular study of cell types and
zonation in the brain vasculature using a clustering approach to identify genes and protein classes
that are enriched along the arteriovenous axis
18
. Briefly, this study was the first to report that, at
the BBB endothelium, transcription factors predominate at arteries and arterioles, transporters
predominate at capillaries and veins, and ribosomal proteins indicative of protein synthesis are
spread along the arteriovenous axis
18
, yielding important insights to biological functions and
endothelial specialization along the brain vascular tree. In contrast to the BBB endothelium’s
gradual phenotypic change along the arteriovenous axis, mural cells formed two distinct groups
comprising pericyte and venous SMCs, and arterial and arteriole SMCs
18
. Although this study did
not examine neurons
18
, an earlier single cell RNA-sequencing study did investigate all NVU cell
types albeit with limited sequencing depth for vascular transcriptomes due to relatively low
abundance of vascular clusters compared to neurons and glial cells
19
. Thus, a more comprehensive
single cell RNA-sequencing study of the entire NVU is needed, in addition to regional
transcriptional analysis of the NVU.
Next, molecular definition of cerebral blood vessels (Section 1.2.1), BBB junctional
molecules (Section 1.2.2), endothelial and pericyte transport systems (Section 1.2.3), and transport

20
of molecules across brain extracellular spaces and by perivascular and paravascular transport
(Section 1.2.4) are discussed.
1.2.1 Molecular Definition of the BBB Vascular Tree
To keep the 86 billion neurons in the human brain working properly requires an adequate
supply of blood, which is accomplished through a vast, well-regulated vascular network of arteries,
arterioles, capillaries, venules and veins reaching approximately 400 miles in length
2,5
. The NVU,
comprised of endothelial cells forming the inner layer of the vessel walls, mural cells along the
vessels that help regulate vascular tone (pericytes and SMCs), astrocytes whose endfeet cover most
of the vasculature, and neurons, work in concert to regulate BBB integrity and CBF
2,5
. The NVU
cellular composition varies along the arteriovenous axis, with rubber band-like SMCs wrapping
vessels at the arterial and arteriole level, pericytes along capillaries, and "stellate" SMCs along
venules
26
(Figure 1.2a).
Endothelial Zonation
The concept of vascular ‘zonation’ landmarks was recently introduced based on single cell
RNA-sequencing of brain vascular cell types isolated from different murine models suggesting
molecular and functional phenotypic differences along the vasculature
18
(Figure 1.2b). Single
endothelial cells isolated from Cldn5-GFP reporter mice with GFP expression driven by claudin-
5, an endothelial tight junction protein, and transcriptional data allowed endothelial cell clustering
into groups based on transcriptomic (dis)similarity. For example, arterial vascular zonation was
identified with arterial-specific endothelial markers, Bmx [encoding BMX non-receptor tyrosine
kinase], Efnb2 [encoding ephrin B2], Vegfc [encoding vascular endothelial growth factor C],
Sema3g [encoding semaphorin 3G], and Gkn3 [encoding gastrokine-3], and venous vascular
zonation with venous-specific endothelial markers, Nr2f2 [encoding nuclear receptor subfamily 2
group F member 2] and Slc38a5 [encoding a sodium-dependent amino acid transporter]
18
. Analysis
of the gradual arteriovenous zonation reveals endothelial genes that peaked in the middle,
representing the capillary enriched genes Mfsd2a [major facilitator superfamily domain-containing
protein 2a (MFSD2a)] and Tfrc [transferrin receptor]
18
. MFSD2a is an essential omega-3 fatty acid
transporter
27
required for BBB formation and omega-3 fatty acid transport function
28
, and
transferrin receptor is important for brain delivery of iron
6
. For the first time, this novel approach

21
enables studies of transcriptional expression with respect to vascular zonation revealing molecular
and functional differences along the brain vascular tree. For example, transcription factors
predominated at the arterial endothelium, whereas transporters predominated at the capillary and
venous endothelium. Some endothelial cells did not fit the arteriovenous zonation pattern
exhibiting high expression of ribosomal protein transcripts, which indicates that protein synthesis
occurs throughout the arteriovenous axis
18
.

Figure 1.2 Molecular definition of the BBB and cerebral blood vessels.
(a) The brain vasculature is a continuum from artery to arteriole to capillary to venule to vein. The
BBB is formed by a continuous endothelium monolayer surrounded by mural cells. Vascular
zonation refers to the molecular and phenotypic changes along the vascular endothelial continuum.
Molecularly, the endothelium is a gradual continuum enriched with cell-specific markers at the
arterial and arteriolar, capillary, and venule and vein levels. Mural cells also cluster at different
vascular segments: SMCs at arterioles and venules and pericytes at capillaries. (b) Representative
curves showing molecular expression patterns of endothelial cells. The arteriole-specific genes
include Bmx, Efnb2, Vegfc, Sema3g, and Gkn3; capillary-specific genes include Mfsd2a and Tfrc;
and venule-specific genes include Nr2f2 and Slc38a5. (c) Representative curves showing
molecular expression patterns of mural cells. Arteriole SMCs enriched genes include Acta2, Myl9,
and Myh11. Arteriole markers represent averaged artery and arteriole expression. Capillary
pericyte enriched genes include Vtn, Pdgfrβ, Kcnj8, and Abcc9. The molecular characterization is
informed from recent single-cell RNA-sequencing studies in multiple murine models
18
.
Abbreviations: aSMCs, arteriole SMCs; vSMCs, venous SMCs.

22
Mural Cell Pattern
In contrast to the endothelium with gradual zonation, mural cells exhibited a segregated
zonation pattern. Isolated from Pdgfrb-GFP; Cspg4-DsRed mice, mural cells transcriptionally
separated into two distinct groups including pericytes and venous SMCs (vSMCs), and arterial and
arteriolar SMCs (aSMCs)
18
. Clustering analysis revealed pericyte-enriched genes, Pdgfrb
[encoding platelet-derived growth factor receptor-b], Cspg4 [encoding chondroitin sulfate
proteoglycan neuron-glial antigen 2 (NG2)], Anpep [encoding N-aminopeptidase CD13], Rgs5
[encoding regulator of G protein signaling 5], Abcc9 [encoding SUR2 subunit of K
+
-ATP channel],
Kcnj8 [encoding Kir6.1], Vtn [encoding vitronectin], and Ifitm1 [encoding interferon-induced
transmembrane protein 1], consistent with reported literature
18,19
. aSMCs enriched genes included
Acta2 [encoding alpha smooth muscle actin, α-SMA], Tagln [encoding transgelin], Myh11
[encoding myosin heavy chain 11], Myl9 [encoding myosin light chain 9], Mylk [encoding myosin
light chain kinase], Sncg [encoding synuclein gamma], Cnn1 [encoding calponin-1], and Pln
[encoding phospholamban]
18
(Figure 1.2c).
In contrast to aSMCs, pericytes express barely detectable levels of Acta2 encoding α-
SMA
18
, a protein that plays a key role in cell contractile apparatus, which would argue against
pericyte role in contractility and CBF regulation. However, using strategies that allow rapid
filamentous-actin (F-actin) fixation or prevent F-actin depolymerization, it has been recently
shown that pericytes on mouse retinal capillaries, including those in intermediate and deeper
plexus, express α-SMA
29
. Junctional pericytes were more frequently α-SMA-positive compared
to pericytes on linear capillary segments. Additionally, short interfering α-SMA-siRNA
suppressed α-SMA expression preferentially in high order branch capillary pericytes, confirming
the existence of a smaller pool of α-SMA in distal capillary pericytes that is quickly lost by
depolymerization
29
. Recent RNA-sequencing studies also indicated that pericytes express
moderate-to-robust levels of other contractile proteins including Des [encoding desmin] and Cnn2
[encoding calponin-2]
18
and Myl9 [encoding myosin light chain 9]
19
. How and whether these
contractile proteins contribute to the contractile apparatus in pericytes remains to be determined.
Nevertheless, beyond expression of contractile proteins, more functional studies are needed to
determine how exactly pericytes contribute to CBF regulation. Since capillaries represent the
largest vascular surface area in the brain, even a ‘small’ regulation by pericytes would have an
important impact on CBF.

23
The subsequent sections on BBB junctional molecules and transport systems will primarily
focus on capillaries.
1.2.2 BBB Junctional Molecules
Adherens Junctions (AJ)
Closest to the basolateral membrane, AJ proteins, VE-cadherin and platelet endothelial cell
adhesion molecule-1 (PECAM-1), form homophilic endothelial-to-endothelial contacts roughly
20 nm wide
30,23,31
(Figure 1.3). AJs are connected to cytoskeleton, modulate receptor signaling
32
,
and regulate transendothelial migration of lymphocytes
33
, monocytes
32,34
, and neutrophils
35
.
Tyrosine phosphorylation of VE-cadherin is required for brain transendothelial infiltration of
leucocytes
33,35
.
Gap Junctions
Gap junctions including connexin-37 (CX37), CX40 and CX43 form hemichannels
between endothelial cells
36–38
, albeit with species-dependent differences in distribution, enabling
endothelial intercellular communications
6
. Further, brain endothelial gap junctions also function
to maintain tight junction integrity
37
.
Other Junctional Molecules
These include the endothelial cell adhesion molecule (ESAM) and structurally similar
JAM-A, -B and -C that modulate junctional tightness similar as AJs, and regulate transendothelial
migration of leukocytes
39,40
.
Tight Junctions (TJ)
Closest to the apical membrane, TJ proteins claudin-1, -3, -5, and -12, and occludin limit
paracellular diffusion of solutes and ions across the BBB
23,41
(Figure 1.3). Loss of claudins is
associated with BBB breakdown in human neurodegenerative disorders
42,43
and acute CNS
diseases
1
, as well as in animal models of these diseases
9
. Claudins can be therapeutically targeted
to seal the BBB, as shown by increasing claudin-1 expression at the BBB in murine experimental
autoimmune encephalomyelitis (EAE) model of MS
44
. Ocln knockout mice develop male
infertility, but TJs in both epithelial and CNS endothelial cells appear ultrastructurally normal and

24
maintain normal transendothelial electrical resistance, suggesting that TJs form a functional barrier
in the absence of occludin
45
. Interestingly, Ocln knockout mice develop brain calcifications
45
, and
similarly humans with OCLN mutations develop a severe neurological syndrome known as
pseudo-TORCH-1 that is characterized with bands of gray matter calcification, severe
microcephaly, early-onset seizures, and developmental delay
46
.
TJ proteins are connected to the actin and vinculin-based cytoskeletal filaments via
scaffolding proteins of the membrane-associated guanylate kinase family ZO-1, -2, and -3
47,48

(Figure 1.3). ZO-1 deficiency leads to BBB breakdown in many neurodegenerative and acute CNS
disorders
7
. BBB also expresses the TJ protein LSR, also known as angulin-1, that has been
previously identified at peripheral tricellular junctions
49
. Additionally, dystrophin complex
operates as a scaffold protein to recruit actin and vinculin filaments, which maintains the
endothelial cytoskeletal network
50
. Dystrophin knockout mice exhibit a notable brain
microvascular phenotype with disrupted endothelial TJs, swollen perivascular glial endfeet, and
degenerating microvessels
51
.

Figure 1.3 BBB endothelial connections.
Several types of junctional molecules maintain the endothelial tight structural lining. Closest to
the basolateral membrane, adherens junctions consist of vascular endothelial (VE)-cadherin and
platelet endothelial cell adhesion molecule-1 (PECAM-1). Gap junctions including connexin-30
(CX30) and CX43 form hemichannels between endothelial cells. Other types of junctional
molecules contribute to the tight lining including the endothelial cell adhesion molecule (ESAM)
and junctional adhesion molecule (JAM)-A, -B, and -C. Closest to the apical membrane, tight
junctions consist of lipolysis-stimulated lipoprotein (LSR)/angulin-1; claudin-1, -3, -5, and -12;
and occludin, which limits paracellular diffusion of solutes and ions across endothelial monolayer.
Zonula occludens (ZO)-1, -2, and -3 attach to claudins and occludin and bind to actin and vinculin-
based cytoskeletal filaments. Dystrophin functions as a scaffold to recruit actin and vinculin, which
maintains the endothelial cytoskeletal network.

25
Pericyte-Endothelial Junctions
Pericytes share a common basement membrane with brain capillary endothelial cells
8
.
Direct peg-and-socket contacts between pericytes and endothelial cells are formed by N-
cadherin
52,53
. Gap junction CX43 hemichannels
54–56
enable intercellular communications between
pericytes and endothelial cells
6,8,57
.
Astrocyte Junctions
Astrocytes express gap junction proteins CX30 and CX43
58–60
. Astrocyte-specific CX43
knockout
61
and/or CX43 and CX30 double knockout
58,59
weakens the BBB leading to astrocytic
edema and loss of astrocyte endfeet perivascular polarity
58,59
, and heightened leukocyte
infiltration
61
.
The Basement Membrane
Endothelial cells interact with the extracellular matrix (ECM) proteins in the capillary
basement membrane including collagen, perlecan and laminin via a- and b-integrin receptors,
which form transmembrane heterodimers that functionally link the ECM with the cell
cytoskeleton
23
. Integrins mediate cell signaling by activating ECM ligands, growth factors, and
growth factor receptors, which regulates multiple endothelial cell functions including survival,
migration, differentiation, adhesion and polarity
62
. Integrin dysfunction leads to BBB
abnormalities, as illustrated for example by b1-integrin endothelial knockout mice that develop
aberrant VE-cadherin signaling, loss of caludin-5 and immature BBB
63
. Conditional deletion of
astrocytic laminin g1 and acute knockdown of laminin a2 lead to breakdown of the basement
membrane, loss of astrocyte endfeet polarity, reduced BBB TJs expression and BBB disruption
64
.
Similarly, Lama2 knockout mice lacking laminin a2 have pronounced BBB disruption associated
with reduced pericyte coverage and loss of TJ and AJ proteins
65
. Thus, aberrant astrocyte-capillary
connections compromise BBB integrity and exacerbate microvascular dysfunction.
1.2.3 BBB Transport Systems
The major BBB transporters, receptors and channels in endothelial cells and pericytes have
been validated by transcriptomic studies and/or protein analysis in the rodent brain
11–17,19,20

(Figure 1.4). With the exception of gases (e.g., oxygen and carbon dioxide) and small lipophilic

26
molecules (<400 Da) that freely diffuse across the endothelium
66
, brain endothelial transport
systems regulate molecular exchanges between blood-and-brain and brain-and-blood
1,6,7,67–69
.
Given the close proximity and highly interactive, cooperative signaling between brain vascular
pericytes and endothelial cells, it is relevant and timely to discuss current knowledge of BBB
pericyte transporters from several recent studies
15,18
. While astrocytes also influence BBB integrity
and transporters at astrocytic endfeet are relevant to the BBB since astrocytic endfeet surround
brain vessels
6,64,70
, currently the transcriptome or proteome specifically enriched only in astrocytic
endfeet has not been examined.
Endothelial Solute Carrier-Mediated Transport
Carrier-mediated transport (CMT) enables solutes such as carbohydrates, amino acids
(AA), monocarboxylic acids, hormones, fatty acids, nucleotides, inorganic ions, amines, choline,
and vitamins to cross the BBB via substrate-specific transporters
1,6,7,67–69
.
Carbohydrate Transporters
The GLUT1 uniporter transports glucose, the CNS’s key energy metabolite, down the
concentration gradient
68,69
. GLUT1 has a single binding site that can be accessed by glucose (and
other hexoses) from either side of the luminal or abluminal endothelial membrane extracellularly
or intracellularly
71
. Since glucose concentration is lower in the brain interstitial fluid (ISF)
compared to plasma, GLUT1 favors blood-to-brain transport of circulating glucose
71–73
.
GLUT1 is expressed in endothelial cells, but not in neurons
71,73
. The importance of glucose
transport across the BBB is best illustrated by the fact that Slc2a1 transcript encoding GLUT1 is
one of the most abundant transcripts in brain endothelium
14
. Mutations in human SLC2A gene have
profound effects on brain function, as reviewed in greater detail
6,74
.
Early immunogold electron microscopy studies have shown greater density of GLUT1
transporters on the abluminal endothelial membrane compared to the luminal membrane
75,76
.
Crystallization of human GLUT1 in the inward open conformation
71
, and crystallization of
bacterial GLUT1 homologue have contributed to our understanding of how GLUT1 mediates
glucose transport across the cell membrane
77,78
. Briefly, the u-shaped intracellular helical bundle
of GLUT1 is formed by three helixes and functions as a latch to secure GLUT1 in the outward
open conformation making the sugar binding site accessible extracellularly
71
. After binding to the
extracellular binding site glucose enters GLUT1, which leads to a conformational change causing

27
extracellular transmembrane domains 1, 4, and 7 to function as a latch to secure the inward open
conformation of GLUT1 enabling release of glucose intracellularly
71
.
Endothelial cells also express sodium glucose cotransporter 1 (SGLT1)
79
that is found in
neurons
80
, but its physiological role in glucose transport across the BBB remains elusive.
Myoinositol is transported via sodium/myoinositol transporter (SMIT) and H
+
/myo-inositol
symporter (HMIT) by facilitated diffusion
6
.

Figure 1.4 Major BBB transport systems.
Endothelium: These include solute carrier-mediated transport (CMT), receptor-mediated transport
(RMT), active efflux, and ion transport. CMT systems mediate transport of carbohydrates, amino
acids, monocarboxylates, hormones, fatty acids, nucleotides, organic anions and cations, amines,
choline, and vitamins with precise substrate specificity and directionality, as indicated. RMT
systems transport proteins including transferrin, insulin, leptin, arginine vasopressin, amyloid-β
(Aβ), glycosylated proteins, and apolipoproteins E (APOE) and J (APOJ). Active efflux includes
ATP-binding cassette (ABC) transporters which transport xenobiotics, drugs, drug conjugates, and
nucleosides from endothelium to blood, as indicated. Ion transport underlies the movement of Na
+
,
K
+
, Cl
-
, HCO3
-
, H
+
, and Ca
2+
into and out of the endothelium via ATPases, uniporters, exchangers,
and symporters, as indicated. Pericytes: presently, details about pericyte transporters’ cellular
polarity and precise direction(s) of transport remain elusive. CMT systems transport
carbohydrates, amino acids, carboxylates, organic anions and cations, and folate. RMT system
transports Aβ, APOE, lipophilic molecules, and aminophospholipids. Ion transport of Na
+
, K
+
, Cl
-
, HCO3
-
, H
+
, I
-
, and Ca
2+
occurs via ATPases, uniporters, exchangers, and symporters, as indicated.
All BBB transporters indicated here are validated with RNA-sequencing and/or proteomic analysis
in the rodent brain. See the corresponding text for a more detailed discussion.

28
Amino Acids (AA) Transporters
All essential AA are transported into the brain across the BBB via large neutral endothelial
AA transporter 1 and 2 (LAT1/2) that transport bi-directionally large neutral AA such as
tryptophan and tyrosine
81
, and the cationic AA transporter 1 and 3 (CAT1/3) that transport cationic
AA such as lysine and arginine
1,72,82
. The concentration of essential AA is lower in brain ISF
compared to plasma, which favors blood-to-brain transport
7
.
Glutamine levels are higher in brain ISF
83
. Glutamine is transported into endothelium via
sodium-coupled neutral AA transporter 1, 2, 3, and 5 (SNAT1/2/3/5) and then hydrolyzed in
endothelium to glutamate via glutaminase, and removed into circulation
84
.
On the abluminal endothelial membrane, sodium-dependent transporters for excitatory AA
(EAAT1/2/3) transport glutamate and aspartate out of the brain
1,85
, which limits their excitotoxic
effects on neurons. Sodium-dependent AA transporters ASCT1/2 and GLYT1 at the abluminal
membrane remove non-essential AA alanine, serine and cysteine, and glycine, respectively, from
brain-to-blood
83
. Transporters of neutral and excitatory AA, glycine, taurine, and GABA are
enriched abluminally and with high-affinity transport from brain-to-endothelium in a sodium-
dependent fashion, and then, these AA are transported across the luminal membrane of the BBB
into the blood via low-affinity transporters mediating AA clearance of nitrogen-rich and acidic AA
into the circulation
84,86
.
Monocarboxylate Transporters
Lactate released from skeletal muscles during exercise, and ketone bodies derived from
liver from metabolism of fatty acids are transported from blood into the brain across the BBB by
monocarboxylate transporter-1 (MCT1)
87
, and then utilized as alternative energy metabolites by
the brain
88
.
Hormone Transporters
Hormone endothelial transporters include MCT8 transporter for T3 (triiodothyronine)
thyroid hormone and the organic anion transporting polypeptide 1c1 (OATP1C1) transporter for
T4 (thyroxine) thyroid hormone
89,90
. Mutations in endothelial SLC16A2 gene encoding MCT8
have a profound effect on brain function, causing Allan-Herndon-Dudley syndrome
91,92
that is
characterized by severely impaired neuronal development and functional deficits causing

29
psychomotor retardation and intellectual disability due to deficient transport of T3 from blood-to-
brain
93
, as reviewed
6,74
.
Fatty Acids Transporters
Essential fatty acids are important for brain development and postnatal neural functions.
Brain endothelium expresses luminal transporters for fatty acids, including fatty acid transport
protein 1 and 4 (FATP-1/4)
94
, and the MFSD2A
27
. Previously an orphan major facilitator
superfamily (MFS) transporter, MFSD2A was identified as a BBB transporter for
lysophosphatidylcholine (LPC) esterified docosahexaenoic acid (DHA) supplying the brain with
the essential circulating omega-3 fatty acids
27
. In the brain, MFSD2A is exclusively expressed in
brain endothelium and is required for proper BBB development and functional integrity
28
.
Endothelial MFSD2A mutations lead to development of microcephaly syndrome in humans
associated with loss of neurons and intellectual disability, and some mutations cause lethal
microcephaly
95,96
.
Nucleotide Transporters
Nucleotides and nucleobases, e.g., cytosine, guanine and adenine found in RNA and DNA,
thymine found in DNA and uracil found in RNA, are all transported across the BBB via sodium-
independent concentrative nucleoside transporter-2 (CNT2) and the sodium-independent
equilibrative nucleoside transporter-1 and 2 (ENT1/2)
69,97
. They supply brain with key substrates
for DNA and RNA synthesis.
Organic Anion and Cation Transporters
Organic anions are transported via organic anion transporter-3 (OAT3) and organic anion
transporting polypeptide 1a4 (OATP1A4)
98
and 2b1 (OATP2B1)
99
. OATP1A4 is a known BBB
transporter of statin
98
. Organic cation/carnitine transporter-2 (OCTN2) transports carnitine, an
essential cofactor for fatty acids oxidation in mitochondria
100
. Additionally, organic cations are
transported via organic cation transporters 1-3 (OCT1/2/3). OCT1/2 also transport N-methyl-4-
phenyl-1,2,3,6-tetrahydropyridine (MPTP), a neurotoxin causing Parkinson’s disease-like motor
impairment
101
.

30
Other Transporters
Amines, choline, and vitamins are also transported across the BBB. Specifically, plasma
membrane monoamine transporter (PMAT) transports organic cations from brain-to-blood,
choline transporter like protein type 1 (CTL1) transports choline bidirectionally across the BBB,
and sodium-dependent multivitamin transporter (SMVT) transports multivitamins from blood-to-
brain
6,67–69
.
Endothelial Receptor-Mediated Transport
Most circulating proteins and large macromolecules (e.g., fibrinogen, immunoglobulins,
albumin, thrombin, plasminogen, growth factors) are not transported across the BBB. However,
some proteins and peptides use receptor-mediated transport (RMT) to traverse the BBB and enter
into the brain. In general, the transport rate of circulating peptides is slower than nutrient transport
across the BBB
102
.
Transferrin and Insulin Receptors
Transferrin receptor (TfR)
103–105
, insulin receptor (IR)
106,107
, and leptin receptor (LEP-
R)
108–110
mediate blood-to-brain transport of transferrin (iron-protein carrier), insulin and leptin
across the BBB, respectively. TfR and IR have been utilized for CNS drug delivery including
therapeutic antibodies and enzymes via Trojan horses’ technology
66
, as discussed in Section 7.3
Targeting the BBB for Treatments. Additionally, the V1 vasopressinergic receptor mediates bi-
directional arginine-vasopressin transport across the endothelium
111,112
.
Lipoprotein Receptors and RAGE
Low density lipoprotein receptor-related protein-1 (LRP1) and LRP2 are expressed in brain
endothelium and co-localize mainly on the abluminal side of the BBB in humans and rodents
6,113–
116
. LRP1 binds Alzheimer’s amyloid-β (Aβ) toxin and mediates its brain-to-blood
clearance
113,115,117,118
. Specifically, LRP1 facilitates clathrin-dependent, receptor-mediated
endothelial endocytosis of Ab at the abluminal membrane of the BBB, which requires
phosphatidylinositol binding clathrin assembly protein (PICALM)
119
. PICALM guides trans-
endothelial trafficking of endocytotic Aβ-containing vesicles to Rab5, and then to Rab11 small
GTPase leading to exocytosis of Ab across the luminal membrane of the BBB into the blood
119
.
LRP1 also binds APOE2 and APOE3, and APOE2-Ab and APOE3-Ab complexes at the

31
abluminal side of the BBB, mediating their efflux from brain-to-blood
120
. LRP1 levels at the BBB
are diminished in AD and AD models contributing to Ab accumulation in the brain
113,115,121,122
.
Additionally, APOJ or clusterin (CLU) binds to LRP2 (or megalin) at the BBB, which mediates
Aβ42 transport from brain to blood
117,123
. Therapeutic strategies based on LRP1-mediated Aβ
clearance are discussed below in Section 7.3 Targeting the BBB for Treatments.
In contrast to lipoprotein receptors, the receptor for advanced glycation endproducts
(RAGE) is expressed mainly at the luminal membrane of the BBB
124
. Its expression at the BBB is
increased in AD and AD models
124,121,122
. Under pathological conditions RAGE mediates re-entry
or influx of circulating Aβ across the BBB into the brain, which is associated with
neuroinflammatory response, CBF reductions and BBB breakdown
124,125
Therapeutic strategies
based on pharmacological blockade of RAGE at the BBB have advanced to Phase 3 clinical trial
in AD patients, as discussed below in Section 7.3 Targeting the BBB for Treatments.
Endothelial Active Efflux
ATP-binding cassette (ABC) transporters utilize ATP as energy source, and are primarily
expressed at the luminal side of the BBB endothelium
6,67,126
. They function to prevent brain
accumulation of drugs, xenobiotics, drug conjugates, and nucleosides via active efflux from
endothelium to blood. Examples include ABCB1 (also known as P-glycoprotein, P-gp), ABCA2,
breast cancer resistance protein (BCRP), and multidrug resistance-associated proteins 1-5
(MRP1/2/3/4/5). ABCB1 contributes to Alzheimer’s Aβ toxin clearance from brain-to-blood
127,128
.
Diminished expression and/or dysfunction of ABCB1 were found in neurodegenerative disorders
including AD and Parkinson’s disease, as discussed below in Section 1.3.3 BBB Breakdown and
Dysfunction for AD and Section 7.2 Commonalities of Cerebrovascular Dysfunction in
Neurodegenerative Disorders for Parkinson’s disease.
Endothelial Ion Transport
The BBB has a major role in controlling concentration of ions in the CNS, which is
important for proper CNS functioning
1,129
.
Sodium Pump
The abluminal sodium pump (Na
+
, K
+
ATPase) is a key regulator of sodium (Na
+
) influx
into the brain and potassium (K
+
) efflux from the brain, which keeps high concentration of Na
+

32
and low levels of K
+
in brain ISF
130,131
. This, in turn, is critical for regulating electrophysiological
activity of neuronal cells including the resting membrane and action potentials, and for maintaining
Na
+
concentration gradient at the BBB (extracellular > intracellular), which drives Na
+
-dependent
transport processes.
Other Ion Transporters
The luminal Na
+
K
+
Cl
-
(chloride) co-transporter (NKCC1) mediates entry of Na
+
, K
+
, and
2Cl
-
ions from blood-to-endothelium
132,133
. The bicarbonate (HCO3
-
)-Cl
-
exchanger mediates entry
of intracellular Cl
-
and the extracellular release of HCO3
-129
. The luminal Na
+
-H
+
(hydrogen)
exchanger transports H
+
protons from the endothelium-to-blood in exchange for intracellular
influx of Na
+
, and is a key regulator of intracellular endothelial pH
134
.
Calcium Transporters
The Na
+
-Ca
2+
(calcium) exchanger co-transporter mediates Ca
2+
efflux from endothelium
into brain ISF, which maintains low intracellular Ca
2+
levels in the microvascular endothelium
2
.
The abluminal transient receptor potential (TRP) channels, also known as non-selective Ca
2+
-
conducting cation channels, are expressed both in arterial endothelium
135
and brain microvascular
endothelial cell lines
136,137
. TRP channels regulate Ca
2+
influx into brain endothelium that releases
soluble factors such as nitric oxide, prostaglandins, and endothelial-derived hyperpolarizing factor
initiating endothelium-dependent vasodilation
135
.
Potassium Channels
Capillary endothelial cells express voltage-gated K
+
channel KV1 and the inward rectifier
K
+
channel KIR2
2,138–140
. During physiological conditions, capillary endothelial K
+
channels
mediate outward K
+
currents causing endothelial cell hyperpolarization that propagates
vasodilatory signals upstream to arterioles contributing to blood flow regulation
138,141
.
Pericyte Transporters
Recent transcriptomic studies suggest that pericytes express multiple transporters,
receptors and ion channels
11,12,14,15,19,20
. Some of these are discussed below.

33
Solute Carrier-Mediated Transport
Pericytes express carbohydrate transporters such as insulin-regulated glucose transporter
GLUT4
15
, facilitative glucose transporter GLUT10
15
, and sodium/myoinositol cotransporter
(SMIT)
11
.
Several AA transporters have been recently identified in pericytes, including the high
affinity excitatory AA transporter EAAT2
11
, sodium-dependent neutral AA transporter
SLC6A17
15
, sodium- and chloride-dependent transporter SLC6A20 for small AA including
glycine and proline
11,14,15
, GABA transporter-1 and 2 (GAT1; GAT2)
11
, and the cationic AA
transporter CAT2
11,15
. These transporters likely contribute to the removal of excitatory and
nitrogen-rich AA from the brain to prevent excitotoxicity, similar to endothelial transporters.
Pericytes also express the monocarboxylic acid transporter-12 (MCT12) that mediates
creatine transport
11,15
and sodium-dependent SLC13A3 for dicarboxylic acids
15
. Organic anions
are transported via the organic anion transporter OATP3A1
11,14,15
. Additionally, pericytes express
the vitamin transporter reduced folate carrier-1 (RFC1)
11,12,14,15,19
.
The precise cellular mechanisms and function of pericyte CMT systems, and whether or
not some are part of serial BBB transport mechanisms supplying brain with energy metabolites
and nutrients as opposed to cell-autonomous role, remains largely unexplored.
Receptor-Mediated Transport
Pericytes express the lipoprotein receptor LRP1
14,142
, which mediates cellular uptake of Aβ
followed by its intracellular degradation and clearance
142,143
. In the case of excessive Aβ load,
accumulation of Aβ can lead to pericyte cell death
142,143
. Additionally, LRP1 on pericytes regulates
cerebrovascular integrity in an APOE-dependent fashion
144
. Studies in transgenic mice expressing
human APOE isoforms have shown that astrocyte-secreted APOE2 and APOE3 bind to LRP1 on
pericytes in vivo, which inhibits the proinflammatory cyclophilin A (CypA) and matrix
metalloproteinase-9 (MMP9) pathway preventing degradation of BBB TJ and basement membrane
proteins
144
. On the other hand, APOE4 has a low affinity for LRP1, which activates the CypA-
MMP9 pathway causing BBB breakdown
144
. Activation of CypA-MMP9 pathway associated with
BBB breakdown has been also shown in human APOE4 carriers by CSF analysis
145
and post-
mortem brain tissue analysis
146,147
. Additionally, pericytes express LRP3
15
which internalizes and
transports lipophilic molecules.

34
Ion Transport
Transcriptomic studies suggest that pericyte express Na
+
K
+
ATPase a- and b-
subunits
11,14,15
, as well as Ca
2+
ATPases
11,14,15
. They also express the Na
+
-K
+
-Ca
2+
exchanger
SLC24A3
11
, the Na
+
-H
+
exchanger SLC9A3R1
12,15
, Cl
-
HCO3
-
exchanger SLC4A4
11
and
SLC4A3
15
, and the Na
+
-Ca
2+
exchanger SLC8A2
15
. Furthermore, pericytes express the ATP-
sensitive K
+
channel ATP-binding cassette sub-family C member 9 (ABCC9), the H
+
-peptide
transporter SLC15A2
11
, the Na
+
-I
-
symporter SLC5A5
15
, and inwardly rectifying potassium (KIR)
channel Kir6.1
11,12,14,15
.
Functionally, adenosine binding to pericyte α1-adrenergic receptors activates ATP-
sensitive K
+
channels causing pericyte hyperpolarization and relaxation
148
. Increases in
intracellular Ca
2+
in response to large increases in extracellular K
+
concentrations activate voltage-
gated Ca
2+
channels in pericytes, which leads to pericyte depolarization and contraction
148
. These
findings coupled by recent physiological experiments on pericyte contractility support that
pericytes play an active role in regulating CBF
2,149–153
.
1.2.4 Other Vascular-Mediated Transport
Besides the major role of trans-vascular transport in clearance of solutes across the BBB
by CMT, RMT, major facilitators, and active efflux transporters
6,7,154
, solutes diffuse across brain
extracellular spaces and are cleared along the basement membranes of the arterial vessel walls by
the perivascular ISF flow, which travels in the reverse direction of blood flow
154–156
(Figure 1.5).
Early studies in rabbits using radiolabeled albumin and in rats using Indian ink, albumin-labelled
with colloidal gold, and Evans blue have suggested that perivascular ISF flow carries solutes and
macromolecules to the subarachnoid space and CSF compartment for drainage into deep cervical
lymph
157,158
. More recent studies in mice using fluorescent solutes have shown that brain has its
own lymphatic vascular system in the dura matter, which drains ISF and macromolecules into the
deep cervical lymph nodes
159–162
. These findings suggest that the brain communicates directly with
the peripheral immune system via meningeal lymphatic vessels. Recent MRI studies in the living
human brain and marmosets utilized a combination of gadolinium-based contrast agent (Gadovist)
and blood-pool contrast agent (Vasovist) to demonstrate existence of the meningeal lymphatic
system
163
.

35

Figure 1.5 Brain perivascular and paravascular transport.
Perivascular interstitial fluid (ISF) flows in the reverse direction of blood flow in the arterial vessel
walls ultimately reaching cerebrospinal fluid (CSF)-filled subarachnoid spaces where ISF-CSF
drains into the meningeal lymphatic vessels and cervical lymph nodes. Paravascular transport of
solutes from subarachnoid spaces flows through Virchow-Robin spaces formed between pia
membrane and glia limitans and is suggested to flow in the same direction as blood flow. At the
capillary level, solutes diffuse across extracellular spaces (ECS) and undergo transvascular
clearance to blood via transport systems as illustrated in Figure 1.4 and discussed in the text.
In addition to the potential role in regulating brain immune responses, the lymphatic system
also plays a role in removing metabolic waste products and proteinaceous toxic accumulates from
the brain. A recent study found that meningeal lymphatic vascular clearance is impaired with
aging, and meningeal lymphatic vessels are a route of Alzheimer’s Ab clearance
164,165
. Drainage
of brain ISF and CSF by the meningeal lymphatics is necessary for proper cognitive function, and
early evidence indicates that promoting local growth of meningeal lymphatic vessels can improve
its clearance function and cognition
164,165
.
To compare clearance routes, transport studies using radiolabeled and unlabeled Ab
peptide have shown that under physiological conditions, the perivascular ISF flow contributes to
15-20% of Ab clearance from the mouse brain
115,117,166
, whereas 80-85% is removed by
transvascular BBB transport. Transvascular Ab clearance fails early in AD and AD models due to
diminished expression of Ab efflux transporters LRP1 (expressed at the abluminal endothelial

36
membrane) and P-gp (expressed at the luminal endothelial membrane) at the BBB
9
. Do damaged
blood vessels and impaired transvascular clearance routes cause an increased burden on the
perivascular Aβ clearance system leading to its disruption? Whether the perivascular lymphatic
system can be therapeutically-targeted to successfully maintain proteostasis, for example to lessen
Ab brain deposition in AD, remains to be determined.
Early studies have suggested that solutes injected into the subarachnoid space can use
paravascular transport from the subarachnoid space to enter the brain through Virchow-Robin
spaces in the same direction to the flow of blood
167
. This concept has been explored by recent
studies. For example, studies using injection of fluorescent tracers into cisterna magna of mice
have suggested that paravascular circulation occurs via CSF convective flow through the
extracellular spaces from the para-arterial to the para-venous spaces, which is regulated by AQP4
water channels on astrocytes, and therefore the system was re-named as ‘glymphatic’ system
168,169
.
Other recent studies
170–173
, however, did not support the proposed ‘glymphatic’ mechanism, nor
the convective, pressure-driven fluid flow of CSF from para-arterial to para-venous spaces
throughout the parenchymal extracellular spaces
174–176,170,171
. A recent report in AQP4 knockout
rodents has shown that loss of AQP4 does not affect transport of fluorescent solutes from
subarachnoid space to brain in rats and mice, suggesting that water production by astrocyte endfeet
does not control transport of solutes across brain extracellular spaces
170
.
Nearly half a century ago, physiologists proposed that CSF acts as a sink for brain-derived
molecules
25,177
. This concept is supported by recent findings showing that under physiological
conditions brain-derived molecules secreted into ISF are present at higher concentrations in the
ISF than in the CSF
154–156,159–161
.
1.3 Vascular Dysfunction in Alzheimer’s Disease
1.3.1 Alzheimer’s Disease Pathophysiology
Alzheimer’s disease (AD) is characterized by the classic hallmark pathology, Aβ plaques,
hyperphosphorylated tau neurofibrillary tangles, and neuron loss, but increasing evidence also
supports that early vascular dysfunction contributes to AD pathophysiology and cognitive
impairment
178–182,43,7,6,183,184
. Some studies have suggested that during preclinical stages, vascular

37
dysfunction is amongst the first detectable biomarker changes reported prior to symptomatic onset
and prior to changes in other standard AD biomarkers, including amyloid deposition and CSF
Aβ42, phosphorylated tau (pTau), and total tau
178,181
.
Neuropathological studies have shown that cerebrovascular pathology is a major risk factor
for clinically diagnosed AD-type dementia with clinical expression associated with low scores in
most cognitive domains
183
. A large autopsy-based neuropathological study importantly revealed
that 80% of patients diagnosed with AD and no evidence of mixed (vascular) dementia had
vascular pathology including cortical infarcts, lacunes, cerebral microbleeds, and multiple
microinfarcts indicative of small vessel disease (SVD), intracranial atherosclerosis,
arteriolosclerosis, enlarged perivascular spaces, and amyloid deposits in the vascular wall
184
,
supporting the concept that cerebrovascular dysfunction is prominent in AD and lowers the
threshold for dementia for a given AD pathology burden.

Figure 1.6 Alzheimer’s disease is a multifactorial and heterogeneous disease.
Alzheimer’s disease (AD) is defined as a unique neurodegenerative disease based on the presence
of Aβ and tau deposits. Additional factors (red), however, contribute to the onset and progression
of AD pathophysiological changes directly affecting brain vascular system (i.e., blood-brain
barrier leakages and cerebral blood flow shortfalls) and innate immune system, and neuronal health
and functioning independently and/or simultaneously with Ab and tau pathologies. This includes,
but is not limited to, genetic risk factors, vascular factors, and environmental factors including
socioeconomic stress, microbiome, and lifestyle. Aging still remains the key risk factor for AD
and also profoundly affects brain vasculature, innate immune responses, and neuronal functions
(blue).

38
AD is a multifactorial and heterogeneous disease with many factors contributing to its onset
and progression. For example, vascular risk factors including hypertension
185,186
, diabetes
187,188
,
hyperlipidemia
189
and cardiovascular disease
190
, as well as environment (e.g., pollution)
191–194
, and
lifestyle (e.g., obesity, sedentary lifestyle)
195,196
are all reported to influence AD onset and
progression, and are also associated with cerebrovascular and BBB dysfunction
180,182
(Figure 1.6).
1.3.2 Genetic Contributions
Autosomal-dominant AD (ADAD) is inherited form of AD caused by mutations in amyloid
precursor protein (APP), and presenilin-1 and 2 (PSEN1; PSEN2) genes
197–201
. ADAD accounts
for ~1% of all AD cases and exhibits early age of onset (<65 years of age)
202
. Several APP and
PSEN1 mutations lead to BBB breakdown and cerebrovascular pathology, as discussed below and
illustrated in Figure 1.7.
The large majority of AD cases, however, are sporadic, late-onset without clear etiology or
inheritance. Nevertheless, several genes are associated with increased or lower risk for sporadic
late-onset AD. Apolipoprotein E-ε4 (APOE4) is the major genetic risk factor for sporadic AD
203–
206
. APOE4 leads to BBB breakdown, vascular pathology and diminished clearance of Ab across
the BBB
207
, as discussed below.
Genome-wide association studies (GWAS) have identified multiple loci associated with
AD including, to name a few, variants in PICALM, CLU, ATP-binding cassette transporter A7
(ABCA7), sortilin related receptor-1 (SORL1), complement receptor 1 (CR1), triggering receptor
expressed on myeloid cells 2 (TREM2), and bridging integrator 1 (BIN1) genes
208–215
. Below, I
examine variants that affect BBB transport and clearance functions associated with PICALM, CLU
and SORL1 genes (Figure 1.7).
APP
Approximately 40 APP mutations have been identified causing ADAD
216
. APP mutations
can lead to cerebrovascular pathology including BBB breakdown and cerebral amyloid angiopathy
(CAA), as shown in humans
197,217–219
and transgenic animal models expressing human APP
mutations
9,113,142,220–231
. CAA is characterized by Aβ deposition in the vascular wall of small and
mid-sized cerebral and leptomeningeal arteries, veins, and cerebral capillaries
232
, and develops as
a result of an imbalance between Aβ production and clearance, particularly faulty trans-vascular

39

Figure 1.7 Effects of genetic mutations carrying inheritance or increasing risk for AD on
BBB dysfunction.
APP: Amyloid precursor protein (APP) vasculotropic mutations E692G (Flemish), E693Q
(Dutch), E693K (Italian), E693G (Arctic), and D694N (Iowa) lead to prominent cerebral amyloid
angiopathy (CAA) causing extensive cerebrovascular pathology and BBB breakdown in humans
(orange red). Dotted boxes denote validation of BBB breakdown in transgenic rodents carrying
the respective human vasculotropic mutations. APP NH2-terminal KM670/671NL (Swedish)
mutations and A673 (Icelandic) mutations lead to a moderate CAA and BBB breakdown in
humans (berry red). Dotted box denotes validation of BBB breakdown in transgenic animals
expressing Swedish mutation. Cerebrovascular function in human carriers of APP COOH-terminal
V715M (French), V715A (German), I716V (Florida), V717I (London), V717F (Indiana), and
L723P (Australian) mutations has not been examined (blue; not studied). However, the BBB
breakdown has been shown in transgenic models carrying Florida, London, and Indiana mutations
(berry red). PSEN1: BBB breakdown and cerebrovascular dysfunction have been reported in
humans carrying different PSEN1 mutations including T113-114 insertion, P117L, M139V,
M146V, L153V, H163R, E184D, G209V, C260V, E280A, L282V, C285T, L420R, and DE9
deletion (orange red). Dotted boxes denote validation of BBB breakdown in transgenic animal
models carrying the respective human PSEN1 mutations. PSEN2: The most common PSEN2
mutation N141I in humans is associated with BBB breakdown (orange red). APOE4:
apolipoprotein E (APOE4), the major genetic risk factor for sporadic AD, leads to BBB breakdown
in humans and transgenic models expressing human APOE4 gene (orange red, dotted box). Others:
phosphatidylinositol binding clathrin assembly protein (PICALM) and clusterin (CLU) regulate
clearance of amyloid-b peptide across the BBB (orange red), while sortilin-related receptor-1
(SORL1) expressed in brain endothelial cells regulates PDGF-BB and LRP1 signaling at the BBB
(berry red), as shown in animal studies. Complement receptor 1 (CR1), triggering receptor
expressed on myeloid cells-2 (TREM2), and bridging integrator 1 (BIN1) have not been studied
for their cerebrovascular effects. The color scale: BBB breakdown is pronounced (orange red),
moderate (berry red), modest (purple), or not studied (blue). See the corresponding text for a more
detailed discussion.

40
and perivascular clearance of Aβ from the brain
2,9,155,233
. CAA is a major cause of SMCs vascular
degeneration that is associated with BBB breakdown at the arterial and/or arteriolar level, lobar
microbleeds, infarcts, white matter changes and cognitive impairment worsening AD
pathology
154,233
.
Individuals with ‘vasculotropic’ APP mutations within Aβ21-23 residues including Dutch
(E693Q), Arctic (E693G), Flemish (A692G), Iowa (D694N), and Italian (E693K) mutation
197,217–
219,234,235
develop prominent CAA
233,236
causing extensive cerebrovascular pathology. For
example, the Dutch mutation leads to recurrent hemorrhages due to damage of the arterial vessel
wall by the CAA, known as hereditary cerebral hemorrhage with amyloidosis (HCHWA-D), which
is often fatal by mid-life
218,237
. Patients with HCHWA-D rarely develop parenchymal amyloid
plaques and neurofibrillary tangles
238
. CBF reductions have also been reported in humans carrying
vasculotropic APP mutations
197
. Experimental studies have shown that the vasculotropic Aβ
mutant peptides are poorly cleared from brain across the BBB into circulation due to their low
affinity for the BBB clearance receptors including LRP1 compared to wild type Aβ peptides
113,239
;
therefore, Aβ mutant peptides accumulate rapidly along the vessel walls.
In contrast to vasculotropic mutations, APP N-terminal mutations such as Swedish
(KM670/671NL) mutation, and APP C-terminal mutations including A713T, A714I, A714A,
V715M (French), V715A (German), I716V (Florida), I716T, V717I (London), V717F (Indiana),
V717G, V717L, and L723P (Australian) mutations lead aberrant and increased Aβ production by
affecting b-secretase and g-secretase processing activities of APP, respectively
199–201,218
.
Compared to vasculotropic mutations, APP N- and C-terminal mutations are associated with less
pronounced CAA and cerebrovascular pathology
218
. The N-terminal APP A673T (Icelandic)
mutation, however, makes APP a less favorable substrate for b-secretase, resulting in decreased
Aβ production
240
. Interestingly, despite sparse parenchymal amyloid deposition, the Icelandic
mutation leads to mild CAA and vascular pathology including microinfarcts
241
and ischemic
stroke
242
, suggesting vascular vulnerability.
Similar to humans, transgenic mice expressing different human APP mutations exhibit
pronounced vascular pathology including severe BBB permeability changes
223,228,230,231
,
microbleeds
220,222,223,225,230
, BBB leakages of blood-derived molecules
142,224,225,228,243
, impaired Aβ
clearance at the BBB
113,221,227,229
, endothelial cell degeneration
142,225
, loss of SMCs
225,227
and
pericytes
142,225,227
, and CAA
221,226,244–247
.

41
Some studies in transgenic murine models have examined the temporal sequence of
appearance of different pathologies, revealing that progressive BBB breakdown develops early in
APP
Sw/0
mice beginning at 1-3 months of age
142,228,231
prior to Aβ deposition, CAA and behavioral
memory recognition deficits that are observed beginning at 10-12 months of age
225,231,244,248
. These
studies suggest that CAA is not the only cause of BBB breakdown in APP transgenic models.
Although the precise mechanism of early BBB breakdown in APP mice is currently unclear,
oligomeric Aβ toxic species and/or direct APP-mediated vasculotoxicity could play a role in CAA-
independent early BBB breakdown and vascular pathology.
PSEN1
To date, 228 PSEN1 mutations have been identified causing ADAD
197,249,250,200,199,201
.
PSEN1 is the catalytical component of g-secretase
200
. PSEN1 mutations increase the faster release
of long Ab peptide species due to altered carboxypeptidase-like γ-secretase activity that increases
the proportion of Ab42, Ab43, and even longer Ab peptide species (Ab45, Ab46)
251–254
. Moreover,
the ratio of Ab42:Ab40 is altered in most, but not all, PSEN1 mutations carriers
254
and the
significance of the altered ratio is not well understood. PSEN1 mutations lead to faster soluble-to-
fibrillar conversion of Aβ42 promoting amyloid deposition in the brain
199,201,250
. Some PSEN1
mutations lead to very early onset ADAD (<35 years of age)
198,200
.
Human PSEN1 mutation carriers have notable BBB breakdown and cerebrovascular
dysfunction including cerebellar amyloid angiopathy and CAA
255,249,256–265
, disrupted meningeal,
subpial and cortical arterioles
249,258,264
, degeneration of pericytes and SMCs
257
, cerebral
perivascular amyloid deposits
255,257,266,267,258,261
, and diminished FDG transport across the BBB in
an early asymptomatic stage
268,269
. Particularly, cerebrovascular pathology with BBB breakdown
was shown by neuropathological studies in patients with PSEN1 T113-114 insertion
259
,
P117L
257,260
, M139V
262,264
, M146V
264
, L153V
266
, H163R
263
, E184D
261
, G209V
264
, A260V
264
,
C260T
258
, E280A
255
, L282V
256
, C285T
258
, and L420R
249
missense mutations, and De9 deletion
263
.
Similar to humans, mice expressing human PSEN1
M146V
mutations driven by the neuronal
Thy1 promoter develop BBB breakdown with microhemorrhages and basement membrane
degeneration in the absence of Aβ pathology or CAA
270
. Additionally, PSEN1
-/-
mice exhibit
severe microbleeds and endothelial degeneration in the neocortex at embryonic day 18.5
271
,
indicating that PSEN1 loss of function induces BBB damage. Interestingly, hemorrhages and

42
vascular abnormalities in PSEN1
-/-
mice can be corrected by neuron-specific PSEN1 expression
271
,
suggesting that impaired vascular-neuronal cross-talk contributes to vascular
pathophysiology
270,271
.
PSEN2
PSEN2 mutations account for only ~5% of all ADAD cases
272
. The mutations N141I
(Volga German pedigree) and M239V represent ~75% of all PSEN2 mutations
273
. Similar to
PSEN1, PSEN2 mutations may also cause increased production of long Ab peptide species
274,275
,
although this needs more investigation. PSEN2 N141I mutation carriers have severe CAA and
hemorrhagic strokes, but sparse parenchymal amyloid and neurofibrillary tangles
276
. Thus far, few
studies have investigated cerebrovascular dysfunction in individuals with other PSEN2 mutations.
APOE4
APOE4 is the major genetic risk factor for sporadic, late-onset AD
201,204,205,277
. One and
two APOE4 alleles increase risk for AD by ~3.8- and ~12-fold, respectively, compared to
APOE3/APOE3 genotype
74,205,277,278
. The effect of one APOE4 allele on AD risk is stronger in
females than in males. One copy of APOE2 allele decreases risk by ~0.6-fold relative to
APOE3/APOE3 genotype. Additionally, APOE4 increases risk for CAA.
APOE exerts toxic effects on the cerebrovascular system
207
and neurons
279,280
, and
influences Aβ clearance
7,120,144,281–287
and amyloid deposition
203,205,278,283–285,288,289
, and tau-related
neurodegeneration
290
in an allele-dependent manner APOE4>APOE3>APOE2. Human APOE4
carriers compared to noncarriers develop accelerated BBB breakdown and pericyte
degeneration
291,292,145–147,293
, early neurovascular dysfunction
294–296
, impaired cerebrovascular
reactivity
297,298
and diminished regional BBB uptake of glucose
299,300
. Cerebrovascular effects of
APOE4 are associated with AD, stroke and brain hemorrhage
7,278,301–303
.
Studies in animal models support that APOE4 compared to APOE3 and APOE2 diminishes
Aβ clearance across the BBB
117,120
, which has been confirmed in transgenic APOE4 mice
282
.
Transgenic mice expressing human APOE4 gene, but not APOE3 and APOE2, develop an early
BBB breakdown
144,304
, cerebral microhemorrhages
305
, and loss of endothelial GLUT1
expression
306
that is followed by secondary neurodegenerative changes
144
. Transgenic mice
lacking Apoe also develop an early BBB breakdown
144,304,307–314
indicating that APOE is essential

43
for maintaining BBB integrity. Apoe
-/-
mice also develop secondary neurodegenerative changes
after BBB breakdown in the absence of Aβ pathology
144
.
APOE in the brain is primarily synthesized by astrocytes and microglia, as indicated by a
recent single cell RNA-sequencing study in the mouse
18
. APOE3 promotes enzyme-mediated
degradation of Aβ in microglia more efficiently than APOE4
315
, suggesting the APOE and
microglia play a role in the innate immune response in AD
204
. Nevertheless, the role of APOE
secreted by other cell types such as pericytes
316–318
and SMCs
319
should be further explored.
PICALM
The association of PICALM polymorphisms with late onset AD has been reported by the
majority of GWAS studies
209,212,214,320–323
. Although most of the PICALM single nucleotide
variants (SNPs) associated with late onset AD are located outside of the coding regions
119
, the risk
alleles of rs3851179
209
and rs10792832
212
lead to lower expression of PICALM isoform 2 in the
frontal and temporal cortex in the human brain on the expression quantitative trait loci (eQTLs)
324
,
indicating that lower PICALM levels may increase the risk for AD.
The N-terminus of PICALM contains an epsin NH2-terminal homology domain for
phosphatidylinositol-4,5-bisphosphate binding, which allows PICALM to sense membrane
curvature
325
and regulate the size of clathrin-coated vesicles
326
. PICALM controls receptor
internalization and subsequent intracellular trafficking
119
, via R-SNARE-mediated fusion of
clathrin-coated vesicles with endosomes
327
. These functions are central to its role in the clearance
of both tau through autophagy
328
, and Aβ via transvascular transport across the BBB
119
.
PICALM is enriched in the endothelium of human cerebral vessels including capillaries,
but is downregulated in AD patients
119
. As shown in the Picalm
+/-
; APP
Swe
mice, diminished
PICALM levels at the BBB accelerate amyloid pathology and behavioral deficits, which can be
ameliorated by endothelial re-expression of Picalm using an adeno-associated virus (AAV)
119
.
Endothelial cells derived from human iPSCs carrying homozygous protective rs3851179
A
alleles
exhibited increased PICALM expression and improved transvascular clearance of Aβ across
human BBB in vitro, when compared with isogenic cells from iPSCs carrying homozygous risk
rs3851179
G
alleles
119
. Overexpression of PICALM in the primary rat cortical neurons attenuated
the toxicity of soluble Aβ oligomers
329
. However, reducing or overexpressing PICALM levels in
hippocampal neurons of APP
Swe
; PSEN1
L166P
mice with AAV8-mediated delivery of PICALM

44
shRNA or cDNA, respectively, indicated that PICALM might regulate Aβ production
330
.
Moreover, PICALM-guided APP intracellular trafficking to autophagosome for degradation
331

limits Aβ production working in harmony with Aβ transvascular BBB clearance to keep low levels
of Aβ in the brain.
CLU
Several GWAS studies have identified CLU as a significant genetic risk factor for sporadic
AD
209,213,321,320
. The functional impact of CLU polymorphisms is currently elusive
200
. CLU gene
encodes clusterin or APOJ that besides its functions in lipid transport, membrane recycling, cell
adhesion, and apoptosis
332–335
affects transvascular Aβ clearance by promoting Aβ42 efflux across
the BBB
201
. In the brain, astrocytes primarily secrete clusterin which acts as a chaperone molecule
that binds soluble Aβ
333
. LRP2 also known as glycoprotein 330 (gp330) or megalin mediates
transport of Aβ-clusterin complexes across the BBB
117,123
.
SORL1
SORL1 was identified through GWAS as a risk factor for sporadic AD
336,337
. SORL1, a
vacuolar protein sorting-10 (Vps10) domain-containing protein, binds PDGF-BB
338,339
and LRP1
ligands
339
. Proper interactions with PDGF-BB and LRP1 ligands are necessary for functional
downstream signaling of PDGFRβ and LRP1, respectively. SORL1 mutations may impact
PDGFRb signaling in mural cells causing pericyte dysfunction and/or degeneration that is reported
in AD, and may also impair LRP1-mediated transvascular clearance, a key mechanism by which
Ab40 and 42 peptides are cleared from brain-to-blood
340
.
Other genes
AD risk genes affect various biological functions, including vascular, immune, metabolic,
trafficking, transcription and adhesion, or a combination. Some additional highly replicated risk
genes include CR1, BIN1, and TREM2
208,211,212,215,320,321
. CR1 is predominantly expressed by
erythrocytes
212,213
and Ab-C3b-CR1-mediated interactions sequester Ab42 to promote clearance
to periphery
341,342
, suggesting that CR1 variants can increase free blood Ab levels
213
, which in turn
can promote RAGE-mediated Aβ re-entry across the BBB into the brain. BIN1 is broadly
expressed in brain cell types
212
, but the effect of BIN1 variants on cerebrovascular function has not
been explored. Microglial TREM2 variants
211,212
are partial loss of function with impaired transport

45
of APOE-containing lipoproteins
343
. TREM2
208,211,215
and two novel rare coding variants, PLCG2
and ABI3, implicate innate immunity in AD pathophysiology
215
, but whether these variants can
influence vascular-inflammatory cross-talk contributing to AD pathophysiology remains unclear.
1.3.3 BBB Breakdown and Dysfunction
Evidence of BBB breakdown and dysfunction during AD pathophysiology in humans has
come from neuroimaging (MRI and PET) and biofluid findings in living humans, and from post-
mortem neuropathological findings in AD brains. Figure 1.8 illustrates a proposed evidence-based
sequence of BBB breakdown and dysfunction during AD pathophysiological progression based
on clinical findings, neuroimaging, neuropathological and biofluid studies. The current findings
supporting BBB dysfunction in AD is discussed in detail in the following subsections.
Neuroimaging Findings
BBB Permeability
Recent neuroimaging studies in individuals with mild cognitive impairment (MCI) and
early AD reveal BBB breakdown in the hippocampus, including its CA1 and dentate gyrus
subfields
178
, and/or several gray and white matter regions
344–346
, prior to brain atrophy and
dementia. These studies used advanced dynamic contrast-enhanced (DCE)-MRI to quantify the
regional BBB permeability constant, Ktrans, to the contrast tracer gadolinium relative to each
individual’s arterial input tracer function
178
. These studies used Patlak analysis
347–349
, which allows
detection of subtle changes in BBB permeability
178,179,348,350
. Earlier studies using longer resolution
time and semi-quantitative analysis indicated a possible trend of increased BBB permeability in
the hippocampus in MCI compared to controls
351
, and suggested that accumulation of contrast
agent in brains of individuals with probable AD likely occurs via blood-to-brain-to-CSF
pathway
352
, consistent with increased blood vessel permeability.
Cerebral Microbleeds
Microbleeds are RBC-derived iron-containing perivascular hemosiderin deposits that can
be visualized as small hypointense regions on T2*- and/or susceptibility-weighted imaging (SWI)
MRI sequences
353
. CNS microbleeds reflect loss of cerebrovascular integrity and are found in 25%

46

Figure 1.8 BBB breakdown and dysfunction in AD.
Evidence-driven sequence of pathophysiological events, including vascular contributions, during
AD development and progression (left to right). Dynamic contrast-enhanced magnetic resonance
imaging (DCE-MRI) has shown BBB breakdown in the hippocampus in individuals with MCI and
in different gray and white matter regions in early AD, before brain atrophy and dementia occur.
Microbleeds reflecting loss of cerebrovascular integrity and BBB breakdown have been shown by
T2*-MRI and susceptibility-weighted imaging (SWI)-MRI during MCI stage, which progresses
and augments through early stages of AD. Fluorodeoxyglucose positron emission tomography
(FDG-PET) has indicated diminished BBB GLUT1 transporter activity mediating glucose uptake
by the brain before brain atrophy, dementia, or Ab pathology. Similarly, diminished active efflux
ABCB1 (P-gp) BBB transporter activity was shown by verapamil-PET in early AD. Early BBB
breakdown and vascular dysfunction in MCI and AD has been confirmed by some studies by
elevated levels of vascular biomarkers in CSF and blood before Ab and tau pathology, and
dementia. Neuropathological (NP) analysis of mild and advanced AD cases confirmed
accumulation of perivascular blood-derived deposits including, to name a few, fibrinogen,
thrombin, red blood cells (RBC)-derived iron containing products that all are potentially toxic for
the neural tissue. In addition, pericyte degeneration, endothelial degeneration, and brain infiltration
with circulating macrophages and neutrophils were associated with BBB breakdown of AD cases
on NP analysis. Diminished expression levels of BBB GLUT1 and ABCB1 (P-gp) transporters
have been shown by post mortem NP analysis of AD cases, as well as downregulation of Ab BBB
clearance receptors LRP1 and ABCB1, suggesting impaired Ab clearance. Furthermore, APOE4
carriers develop accelerated BBB breakdown associated with activation of proinflammatory
CypA-MMP9 pathway at the BBB, which degrades endothelial tight junction and basement
membrane proteins enhancing BBB damage. How changes in BBB permeability as measured by
advanced neuroimaging techniques in the living human brain relate to disrupted structural and
functional connectivity as measured by diffusion-tensor imaging (DTI)-MRI and functional MRI
(fMRI), and amyloid-PET and tau-PET findings remains unclear at present.

47
of individuals with MCI
354
and 45-78% of individuals with early AD without dementia
355,356,291,357–
361
. This broad range of detection could likely be attributed to the magnet field strength, with
clinical strength 1.5 and 3T magnets likely underestimating the incidence of microbleeds compared
to the research-grade 7T magnet
354,356–358,361,362
. Microbleeds reflect cerebral small vessel disease,
which is observed in approximately 50% of all dementia cases worldwide
3,363
, and is associated
with worse cognitive performance
359
and white matter hyperintensities
354,364
. APOE4 status
accelerates microbleed prevalence in a majority of studies
354,362,364
, but not all
360
.
Microbleed etiology and topography are related, with CAA causing lobar microbleeds in
AD and hypertensive arteriopathy causing deep infratentorial microbleeds
363
. Hypertension
positively associates with microbleed size and prevalence
359,364
particularly in infratentorial
regions, whereas other VRFs such as diabetes and hyperlipidemia do not associate with
microbleeds
362
. Some studies indicate that microbleeds predominate in deep infratentorial regions
during early preclinical AD and MCI stages
360,364
supporting the view that they may precede
amyloid pathology, determined by no difference in CSF Ab42 in MCI subjects with and without
microbleeds
360
, but are later seen in lobar regions
354,356,359,362
reflecting CAA etiology during AD
progression. CAA further contributes to BBB breakdown, infarcts, white matter changes, and
cognitive impairment leading the detection of dementia earlier
154
.
There is a positive association between amyloid deposition in the brain detected by
18
F-
florbetapir positron emission tomography (PET) and the number of microbleeds in individuals
with MCI and AD
365
. Several studies reporting high prevalence of microbleeds in AD
291,355–361
and
MCI
354
, did not perform, however, amyloid-PET imaging
291,355–361
, making it difficult to directly
relate microbleeds to CAA severity.
BBB Endothelial Transporters
18
F-fluoro-2-deoxy-D-glucose (FDG) is a radiolabeled form of the glucose analog, 2-
deoxy-D-glucose (2DG), and is frequently used as a ligand for FDG-PET studies as an “surrogate”
marker for glucose brain uptake
179
. Impaired FDG-PET uptake is often considered an exclusive
biomarker of brain hypometabolism or neurodegeneration
366
. However, since FDG also tracks
BBB transport of glucose, FDG-PET should be considered both a biomarker of glucose (vascular)
transport and cellular uptake, but not a marker of glucose metabolism.

48
Briefly, glucose and its 2DG and FDG analogs are transported from blood-to-brain via
BBB-specific GLUT1, and then taken up by different cell types (e.g., neurons) in the brain via
their respective glucose transporters, which does not include GLUT1
6,73,367
. The ubiquitous
intracellular hexokinase then phosphorylates glucose, 2DG and FDG to their respective 6-
phosphates (6P)
368–371
. However, after this initial hexokinase phosphorylation step there are critical
differences between glucose vs. 2DG/FDG metabolic fates in brain
367–371
. Glucose undergoes
glycolysis followed by pyruvate entry into the Krebs cycle and oxidative phosphorylation, where
2DG and FDG do not enter the glycolytic or Krebs cycle metabolic pathways in the brain
368–372

and are slowly eliminated from the brain. This topic has recently been discussed in greater detail
elsewhere
9,43,373,374
.
Importantly, FDG-PET studies show diminished glucose uptake in several brain regions
prior to any detectable neurodegenerative changes, brain atrophy and/or conversion to AD
375
. In
AD, reduced regional FDG brain uptake was not due to brain atrophy
376
. Longitudinal FDG-PET
findings have suggested that reductions in hippocampal glucose uptake during normal aging can
predict cognitive decline years in advance of clinical AD diagnosis
377
. Diminished regional
glucose uptake has been repeatedly shown by FDG-PET in early AD
378
, and also in individuals at
genetic risk for AD
299,300
, with a positive family history of AD
379
, and/or MCI or no cognitive
impairment prior to progression to AD
380
. The patterns of FDG brain uptake can also discriminate
individuals with normal cognition from MCI and AD patients
378
, suggesting region-specific
insufficiency in brain delivery and uptake of glucose to the brain. FDG-PET changes preceding
neurodegeneration are not only found in humans
375–377,381
, but also in transgenic AD models
382
.
P-glycoprotein (P-gp) encoded by the ABCB1 gene, mediates active efflux of drugs and
xenobiotics from endothelium to blood, which prevents their accumulation in brain
6,67
.
Experimental studies have shown that P-gp (luminally expressed) clears Aβ across the BBB, which
requires the abluminal lipoprotein receptor low-density lipoprotein receptor-related protein 1
(LRP1)
127,128,383
. P-gp function is clinically assessed with radiolabeled ligand (R)-
11
C-verapamil
and PET. Verapamil-PET studies in AD have demonstrated increased uptake of verapamil in
frontal, parietal, temporal and occipital cortices, and posterior and anterior cingulate
384
. Similarly,
verapamil-PET studies in patients with mild AD found significantly lower P-gp activity in the
parietotemporal, frontal, and posterior cingulate cortices and hippocampus
385
. Collectively, these
studies suggest that decreased P-gp function is involved in the pathogenesis of AD either by

49
allowing higher levels of xenobiotics to accumulate in brain, which can injure neurons and promote
inflammation, and/or by reducing Aβ clearance across the BBB. Thus, P-gp together with LRP1
could be an important therapeutic target in AD.
Neuropathological Findings
BBB Breakdown
Post-mortem studies show vascular capillary leakages of blood-derived proteins in the
prefrontal and entorhinal cortex and hippocampus of AD patients including perivascular
accumulation of fibrinogen, thrombin, albumin, immunoglobulin G (IgG), and iron-containing
proteins such as hemosiderin
146,147,293,386–390
. These blood-derived proteins are often found co-
localized with Aβ
147,387,389
and are more pronounced in APOE4 carriers compared to
noncarriers
146,147,292,293
.
Pericyte Degeneration
Ultrastructural studies using electron microscopy reveal accumulations of osmiophilic
materials in the capillary mural cells in AD cortex suggestive of pericyte loss
391,392
. Pericyte loss
has been confirmed by decreased levels of pericyte marker PDGFRβ in the precuneus
390
and loss
of pericytes in the subcortical white matter
393
. Similarly, PDGFRβ immunostaining revealed
significantly reduced pericyte coverage of brain capillaries as well as reduced pericyte numbers in
AD cortex and hippocampus compared to control brains
389
, which is accelerated by APOE4
gene
146
.
Endothelial Degeneration
Reductions in capillary length suggestive of brain endothelial degeneration, reduced
expression of tight junction proteins, and capillary basement membrane changes have been
reported in AD
7,146,292,389,392,394,395
. These changes might reflect aberrant brain angiogenesis caused
by homeobox gene MEOX2, a regulator of vascular differentiation, whose expression in brain
endothelium is low in AD
395
. Pro-angiogenic factors are expressed in AD brains
396
, which in the
presence of reduced MEOX2 expression leads to cell death and microvascular reductions via AFX1
transcription factor that regulates apoptosis
395
. Pericyte-derived soluble factors that maintain

50
healthy endothelium might also be lacking in AD due to pericyte degeneration, which could
potentially contribute to endothelial degeneration, as shown in animal models
397
.
Cell Extravasation
RBCs extravasation
386
and brain infiltration by peripheral macrophages
398,147
and
neutrophils
399
is reported in AD post-mortem studies, suggesting that the brain’s innate immune
system is activated, which can contribute to pathophysiological changes.
Dysregulated Molecular Transport
The levels of endothelial-specific GLUT1 transporter at the BBB are greatly reduced in
AD
400–403
. AD brain microvessels also exhibit reduced levels of LRP1, a major Aβ clearance
receptor at the BBB
113,115,121,122,146
. Besides oxidative stress and Aβ causing faulty LRP1 folding
leading to its proteosomal degradation
113
, reduced GLUT1 levels may also inhibit LRP1
transcriptionally via sterol regulatory element binding protein 2 (SREBP2)
73
. In contrast to LRP1,
expression of RAGE, a major Aβ influx receptor, is increased in AD brain endothelium and mural
cells
121,122,124
, likely contributing to circulating Aβ influx, neuroinflammation, and reduced
CBF
124,125
.
Brain pericytes and endothelial cells in AD APOE4 carriers exhibit increased activity of
BBB-degrading CypA-MMP9 pathway compared to noncarriers
146
. CypA mRNA is also increased
in AD brains
404
.
Angiogenesis
Increased levels of several angiogenic factors in response to reduced CBF and regional
hypoxic changes have been reported in AD brains
396
. Despite a pronounced pro-angiogenic
response, AD brains fail to mount an adequate angiogenic response to renew lost capillary
networks likely due to the loss of the homeobox gene MEOX2 from endothelium
395
, and on-going
pericyte degeneration
8
.
Molecular Findings in Biofluids
The most common biofluid marker of BBB breakdown is the albumin quotient (Qalb),
which is the ratio of CSF albumin levels to serum albumin levels. Several studies report increased
Qalb in individuals with preclinical AD
145
, MCI
178
, and AD
405–407
. However, other studies did not

51
find a change in Qalb in AD
408
unless AD was associated with VRFs
189,408–410
including mild
arterial hypertension, diabetes mellitus, ischemic heart disease
408,410
or dyslipidemia
189
, which
contribute to 65% to 80% of all AD cases at age 65 and 85, respectively
3,411,412
. Although some
studies have not specifically examined the relationship between VRFs and Qalb
145,405
, VRFs did
not impact the magnitude of BBB breakdown to gadolinium as found by DCE-MRI Ktrans
permeability analysis
178
supporting the view that BBB breakdown is associated with AD
independently from VRFs. Future studies should examine more carefully whether ischemic
vascular damage from comorbidities and VRFs
3,180,413
can additionally augment BBB breakdown
in AD.
It is important to recognize, however, that CSF albumin levels could be influenced by
albumin uptake by brain macrophages, microglia, astrocytes, neurons, and neuron-glial antigen 2-
positive cells, and proteolytic cleavage
414–416
. Therefore, Qalb may underestimate the degree of
BBB breakdown in some cases. More sensitive tests of BBB integrity including Ktrans permeability
(by DCE-MRI)
178
, microbleeds (by T2* MRI)
361
, and/or alternative CSF blood-derived biomarkers
such as fibrinogen
417
or plasminogen
418
, previously used to indicate BBB breakdown in MCI and
early AD, respectively, should be used alongside Qalb to accurately assess the degree of BBB
breakdown.
BBB Breakdown and Dysfunction in Animal Studies
Consistent with human studies, studies in different models relevant to AD pathophysiology
also indicate the presence of BBB breakdown and dysfunction. AD transgenic animal models are
generated from ADAD human mutations in APP and PSEN1, as well as APOE4 and mutations in
tau (MAPT). This topic has been comprehensively reviewed recently
9
, and will not be examined
in great detail here.
Briefly, APP transgenic models exhibit perivascular accumulation of blood-derived
proteins (e.g., fibrinogen, IgG, albumin), vascular leakage of circulating exogenous tracers, loss
of tight junction proteins, loss of pericyte coverage, pericyte and endothelial degeneration, and
microbleeds, altogether indicating BBB breakdown
9,142
. Although most studies did not examine
the time course of BBB changes in relation to other brain pathologies, those that did indicated that
BBB breakdown occurs early prior to amyloid accumulation, behavioral deficits or brain
degenerative changes
9,142,228
. Simultaneously, BBB transporter dysfunction occurs, including

52
decreased luminal P-glycoprotein function, decreased LRP1-mediated Aβ clearance, increased
RAGE-mediated Aβ influx, which all accelerates Aβ accumulation in APP models
127,124,113,125
.
Decreased GLUT1 BBB expression also contributes to BBB breakdown and Aβ pathology by
transcriptionally downregulating LRP1
73
, thus corroborating evidence from human BBB studies.
Different PSEN1 models (e.g., PSEN1 knockouts and PSEN1 mutations driven by neuronal
promoters) exhibit microbleeds and endothelial degeneration indicative of BBB breakdown
270,271
,
occurring prior to brain Aβ and CAA pathology
270
. AD models with tau (MAPT) mutations also
show BBB leakage of exogenous tracers, IgG deposits, microbleeds, and leukocyte infiltration
despite no evidence of brain Aβ or CAA accumulation
419
. Finally, APOE knockout mice and mice
with targeted replacement of human APOE4 gene show BBB breakdown and dysfunction (e.g.,
loss of BBB GLUT1 and increased RAGE)
305,420
prior to development of behavioral deficits,
synaptic changes and neuronal dysfunction
144
. Altogether these studies support that
cerebrovascular dysfunction contributes to AD progression.
1.4 Summary and Overview of Chapters
The BBB is a complex, dynamic structure that maintains cerebrovascular integrity and
brain health by functioning as a gatekeeper. The BBB sanctions entry of oxygen and energy
substrates while denying entry of macromolecules, cells and pathogens under proper physiological
conditions. This is accomplished by the expression of a myriad of transporters, receptors, active
efflux pumps, ion channels, regulatory molecules, junctional molecules, etc. and interactive
signaling amongst BBB cell types. Moreover, CBF regulation within a heavily vascularized CNS
is essential to meet the energy demands of neurons. Important physiological data on the molecular
and cellular mechanisms regulating BBB integrity and CBF have been generated from animal
models during both normal physiology and pathophysiology
9,74,374
, providing a foundation to
investigate cerebrovascular dysfunction in disease states in humans.
Existing evidence of BBB dysfunction in humans with AD has been shown by
neuroimaging and biofluid findings and post-mortem brain tissue analysis, as presented
above
43,74,182,374
. Consistently, neuropathological studies have shown that cerebrovascular
pathology is a major risk factor for cognitive impairment in clinically diagnosed AD-type
dementia
183
. A large autopsy-based neuropathological study importantly revealed vascular

53
pathology in 80% of patients diagnosed with AD and no evidence of mixed dementia (i.e.,
concomitant vascular dementia, etc.)
184
, supporting the concept that cerebrovascular dysfunction
is prominent in AD and lowers the threshold for dementia for a given AD pathology burden. While
evidence of cerebrovascular dysfunction is increasingly reported in AD and associated with early
stages of AD (as illustrated in Figure 1.8), cerebrovascular/BBB dysfunction is not yet widely
accepted as an important contributor to AD pathophysiology according to the recent update by the
National Institute on Aging (NIA) and Alzheimer’s Association (AA) in their article NIA-AA
Research Framework: Toward a biological definition of Alzheimer’s disease
366
. The NIA-AA
Research Framework acknowledges that “a vascular biomarker group could be added, that is,
ATV(N), when a clear definition of what constitutes V+ is developed”
366
, although numerous
imaging-based biomarkers of cerebrovascular dysfunction already exist and are routinely used in
the pathological criteria for other neurological disorders (e.g., small vessel disease, etc.)
373
.
The goal of this dissertation is to evaluate and validate biomarkers of BBB dysfunction that
are clinically relevant to cognitive impairment and dementia including AD, and also to explore
basic mechanisms underlying cerebrovascular dysfunction as related to pathophysiology that will
ultimately aid in vascular-directed therapeutic efforts. First, this dissertation presents original
evidence that BBB breakdown is an early biomarker of cognitive impairment in humans (Chapter
2). BBB breakdown is determined by a capillary pericyte injury marker, soluble platelet-derived
growth factor receptor-b (sPDGFRb) in CSF and regional BBB permeability to a contrast agent
by DCE-MRI, which are independent yet related measures of BBB breakdown. These markers of
BBB breakdown are further impacted by genetic factors including APOE4, the major genetic risk
factor for sporadic AD (Chapter 3). Additionally, CSF evidence of cerebrovascular dysfunction
is also shown in ADAD individuals carrying PSEN1 or APP mutations (Chapter 4). Next, a novel
assay to detect CSF sPDGFRb was developed and validated as a reliable and clinically-relevant
biomarker of pericyte injury during cognitive impairment (Chapter 5). Then, in order to elucidate
and better understand molecular mechanisms and signatures of cerebrovascular dysfunction, I turn
to a mouse model of hypoxia that exhibits cerebrovascular dysfunction including BBB breakdown
(Chapter 6). The final chapter will synthesize the original findings from Chapters 2-6 with the
current perspective in the AD field and also relate the observed cerebrovascular dysfunction to
what is similarly seen in other dementias and/or neurodegenerative disorders including Parkinson’s
disease (PD), Huntington’s disease (HD), amyotrophic lateral sclerosis (ALS), multiple sclerosis

54
(MS), human immunodeficiency virus-1 (HIV-1)-associated dementia (HAD) and chronic
traumatic encephalopathy (CTE) (Chapter 7). In conclusion, opportunities to target the BBB for
treatments will be presented, and finally gaps in the field and future directions will be discussed.

55
CHAPTER 2:
BBB BREAKDOWN IS AN EARLY BIOMARKER OF HUMAN COGNITIVE
DYSFUNCTION

Adapted from:
Nation DA*, Sweeney MD*, Montagne A*…Zlokovic BV, Nature Medicine, 2019
* denotes equally contributed first co-authors.
2.1 Introduction
Vascular contributions to cognitive impairment are increasingly recognized
5,7,43,412,421
as
shown by neuropathological
183,184
, neuroimaging
2,43,178,181,344,345
, and cerebrospinal fluid (CSF)
biomarker
43,181
studies (presented in Chapter 1 above). Moreover, small vessel disease of the brain
has been estimated to contribute to approximately 50% of all dementias worldwide, including
those caused by Alzheimer’s disease (AD)
5,43,363
. Vascular changes in AD have been typically
attributed to vasoactive and/or vasculotoxic effects of amyloid-b (Aβ)
2,5,9
, and more recently
tau
422
. Animal studies suggest that Ab and tau lead to blood vessel abnormalities and blood-brain
barrier (BBB) breakdown
9,419,422
. Although neurovascular dysfunction
2,5
and BBB breakdown
develop early in AD
7,43,178,181,344,345,363,412
, how they relate to changes in AD classical biomarkers
Ab and tau, which also develop prior to dementia
366
, remains unknown. To address this question,
here I studied brain capillary damage using a novel CSF biomarker of BBB-associated capillary
mural cell pericyte, namely soluble platelet-derived growth factor receptor-β (sPDGFRβ)
178,423
.
We previously reported that human pericytes, compared to human SMCs, predominantly express
PDGFRβ, and pericyte injury by treatment with Ab peptide or hypoxia causes increased
sPDGFRβ
423
(Appendix A). Additionally, in collaboration with colleagues in our group, regional
BBB permeability was assessed using dynamic contrast-enhanced (DCE)-magnetic resonance
imaging (MRI)
178,344,345
. Altogether our data show that individuals with early cognitive
dysfunction develop brain capillary damage and BBB breakdown in the hippocampus irrespective
of Alzheimer’s Ab and/or tau biomarker changes, suggesting that BBB breakdown is an early
biomarker of human cognitive dysfunction independent of Ab and tau.

56
2.2 Methods
2.2.1 Study Participants
Participants were recruited from two sites, including the University of Southern California
(USC), Los Angeles, CA, and Washington University, St. Louis, MO. At the USC site, participants
were recruited through the USC Alzheimer’s Disease Research Center (ADRC): combined USC
and the Huntington Medical Research Institutes (HMRI), Pasadena, CA. At the Washington
University site, participants were recruited through the Washington University Knight ADRC. The
study and procedures were approved by the Institutional Review Board of USC ADRC and
Washington University Knight ADRC indicating compliance with all ethical regulations, and
informed consent was obtained from all participants prior to study enrollment. Participants from
both sites were included in cerebrospinal fluid (CSF) biomarker studies. All participants underwent
neurological and neuropsychological evaluations performed using the Uniform Data Set (UDS),
and additional neuropsychological tests, as described below. Participants from the USC ADRC
were included in dynamic contrast-enhanced (DCE)-MRI studies for assessment of BBB
permeability if they had no contraindications for contrast injection or MRI.
This study included 161 participants for CSF biomarker studies (74 from USC/HMRI and
87 from Washington University). A group of 35 participants from the Washington University
Knight ADRC underwent Pittsburgh compound B (PiB)-positron emission tomography (PET)
imaging for amyloid-b. A group of 73 participants recruited from the USC ADRC underwent
DCE-MRI. All biomarker assays and quantitative MRI scans were conducted by investigators
blinded to the clinical status of the participant.
2.2.2 Participant Inclusion and Exclusion Criteria
Included participants (≥45 years of age) with neuropsychologically-confirmed no cognitive
dysfunction and/or early cognitive dysfunction had no current or prior history of any neurological
or psychiatric conditions that might better account for any observed cognitive impairment,
including organ failure, brain tumors, epilepsy, hydrocephalus, schizophrenia, major depression.
Participants were stratified based on CSF analysis as either Ab1-42-positive (Ab1-42+, <190 pg/mL)
or Ab1-42-negative (Ab-, >190 pg/mL), or pTau181-positive (pTau+, >78 pg/mL) or pTau181-
negative (pTau-, <78 pg/mL), using the accepted cutoff values
424–426
. Participants were excluded
if they were diagnosed with vascular cognitive impairment or vascular dementia. These clinical

57
diagnoses were conducted by neurologists and the criteria whether the patient 1) had a known
vascular brain injury and 2) the clinician judged that the vascular brain injury played a role in their
cognitive impairment, and/or pattern and course of symptoms. In addition to clinical diagnosis,
presence of vascular lesions was confirmed by moderate-to-severe white matter changes and
lacunar infarcts by fluid-attenuated inversion recovery (FLAIR) MRI and/or subcortical
microbleeds by T2*-weighted MRI
363
. Participants were also excluded if they were diagnosed with
Parkinson’s disease, Lewy body dementia or frontotemporal dementia. History of a single stroke
or transient ischemic attack was not an exclusion unless it was related to symptomatic onset of
cognitive impairment. Participants also did not have current contraindications to MRI and were
not currently using medications that might better account for any observed cognitive impairment.
2.2.3 Clinical Dementia Rating (CDR)
Clinical Dementia Rating (CDR) assessments followed the standardized UDS procedures.
Participants underwent clinical interview, including health history, and a physical exam.
Knowledgeable informants were also interviewed. Given the lack of scientific consensus regarding
the categorization of older adults along the aging-to-MCI-to-AD dementia spectrum and the time
course and sequence of biomarker abnormalities, we did not use clinical diagnosis in our biomarker
comparisons but rather stratified participants along objective neuropsychological metrics of
cognitive impairment and biological metrics of AD biomarker status using established
cutoffs
424,425
. Participant CDR score was obtained through standardized interview and assessment
with the participant and a knowledgeable informant.
2.2.4 Neuropsychological Evaluation and Domains of Impairment
Neuropsychological performance was used to identify domain impairment. All participants
underwent neuropsychological testing using the UDS battery (version 2.0 or 3.0) plus
supplemental neuropsychological tests at each site. Test impairment for UDS tests was determined
using age-, sex- and education-corrected scores from the National Alzheimer’s Coordinating
Center (NACC) (www.alz.washington.edu). Normalized scores from a total of 10
neuropsychological tests were used in determining domain impairment, including three tests per
cognitive domain (memory, attention/executive function and language) and one test of global
cognition. Domain impairment was determined using previously described neuropsychological
criteria
426
, and was defined as a score >1 standard deviation (SD) below norm-referenced values

58
on two or more tests within a domain
427
. Multiple domain impairment (2+) was assigned when
more than one domain fit the impairment criteria, or three or more tests were impaired across
domains
426,427
. Prior studies have established improved sensitivity and specificity of these criteria
relative to those employing a single test score, as well as adaptability of this diagnostic approach
to various neuropsychological batteries
426–428
. Cognition was presumed normal unless multiple
impaired tests were identified as specified by the criteria. Individuals with low Mini Mental State
Exam (MMSE) or Montreal Cognitive Assessment (MOCA) scores (<25) who had multiple
missing neuropsychological test scores due to difficulty completing testing were considered to
have domain impairment. Test battery specifics for each UDS version and recruitment site are
listed below.
Global Cognition
Mini Mental State Exam (MMSE) for UDS version 2 and Montreal Cognitive Assessment
(MOCA) for UDS version 3.
Memory
The Logical Memory Story A Immediate and Delayed free recall tests [modified from the
original Wechsler Memory Scales – Third Edition (WMS-III)] for UDS version 2 and the Craft
Stories Immediate and Delayed free recall for UDS version 3. For supplemental tests the USC
participants underwent the California Verbal Learning Test – Second Edition (CVLT-II) and the
Selective Reminding Test (SRT) sum of free recall trials. Norm-referenced scores for these
supplemental test scores were derived from a nationally representative sample published with the
test manual (CVLT-II)
429
and in studies of normally aging adults (SRT).
Attention and Executive Function
The Trails A, Trails B, and Wechsler Adult Intelligence Scale - Revised (WAIS-R) Digit
Span Backwards tests for UDS version 2 and the Trails A, Trails B and Digit Span Backwards
tests for UDS version 3.
Language
The Animal Fluency, Vegetable Fluency, and Boston Naming Tests for UDS version 2 and
Animal Fluency, Vegetable Fluency, and Multilingual Naming Test (MINT) for UDS version 3.

59
2.2.5 Vascular Risk Factors
Participant vascular risk factor (VRF) burden was evaluated through physical exam,
clinical blood tests and interviews with the participant and informant, and included history of
cardiovascular disease (heart failure, angina, stent placement, coronary artery bypass graft,
intermittent claudication), hypertension, hyperlipidemia, type 2 diabetes, atrial fibrillation, and
transient ischemic attack (TIA) or minor stroke, and total VRF burden was defined by the sum of
these risk factors. We have previously shown that older adults with AD exhibiting two or more
VRFs are more likely to exhibit occult cerebrovascular disease at autopsy, whereas a single VRF
is common and not necessarily associated with increased cerebrovascular disease in this
population
430,431
. Thus, elevated VRFs burden was defined as having two or more VRFs.
2.2.6 Lumbar Puncture and Venipuncture
Participants underwent a lumbar puncture and venipuncture in the morning after an
overnight fast. The CSF was collected in polypropylene tubes, processed (centrifuged at 2000 g,
10 minutes, 4°C), aliquoted into polypropylene tubes and stored at -80°C until assay. Blood was
collected into ethylenediaminetetraacetic acid (EDTA) tubes and processed (centrifuged at 2000
g, 10 minutes, 4°C). Plasma and buffy coat were aliquoted in polypropylene tubes and stored at -
80°C; buffy coat was used for DNA extraction and APOE genotyping.
2.2.7 APOE Genotyping
DNA was extracted from buffy coat using the Quick-gDNA Blood MiniPrep (Cat. No.
D3024, Zymo Research, Irvine, CA). APOE genotyping was performed using polymerase chain
reaction restriction fragment length polymorphism approach (PCR-RFLP). The DNA was
amplified in a 50 µL reaction with Qiagen reagents (Cat. Nos. 201203 and 201900, Qiagen). Two
primers were used to amplify a 318 base pair fragment: upstream sequence (5’
ACTGACCCCGGTGGCGGAGGAGACGCGTGC) and downstream sequence (5’
TGTTCCACCAGGGGCCCCAGGCGCTCGCGG). The upstream primer introduces an AflIII
site in the amplified product, yielding a unique RFLP pattern for each APOE allele following
enzymatic digestion. The PCR reaction mixture was incubated at 94°C for 3 min, then 40 cycles
of amplification (94°C, 10 sec; 65°C, 30 sec; 72°C, 30 sec), and finally elongation at 72°C for 7
min. Restriction digests containing 10 μl amplicons and either 2.5 U AflIII or 1.5 U HaeII were
incubated at 37°C overnight. The digested products were analyzed on a 4% agarose gel. APOE

60
genotype was determined from the unique digestion pattern: APOE2/2 [A: 231; H: 267], APOE2/3
[A: 231; H: 231 and 267], APOE2/4 [A: 231 and 295; H: 231 and 267], APOE3/3 [A: 231; H:
231], APOE3/4 [A: 231 and 295; H: 231], and APOE4/4 [A: 295; H: 231]; the brackets denote
base pairs of amplicons following the AflIII (A) and HaeII (H) digestions.
2.2.8 Molecular Biofluid Assays
Quantitative Western Blotting of sPDGFRβ
The quantitative Western blot analysis was used to detect sPDGFRβ in human CSF
(ng/mL), as we previously reported
178
. Briefly, standard curves were generated using recombinant
human PDGFRβ (0.25 – 5 ng/well) (Catalog no. 385-PR-100/CF, R&D Systems, Minneapolis,
MN). Human CSF was mixed with reducing NuPAGE
®
LDS sample buffer, boiled, and subjected
to SDS-polyacrylamide gel electrophoresis (4-12% Bis-Tris gel, Life technologies, Carlsbad, CA).
Gel transfer was performed using iBlot2 (Thermo Fisher Scientific) at 20 V for 9 minutes. The
nitrocellulose membrane was incubated with SuperBlock-TBST (Thermo Fisher Scientific) for 1
hour at room temperature and incubated overnight with PDGFRβ primary antibody (Catalog no.
AF1042, R&D Systems). The following day, the membrane was incubated with secondary anti-
goat HRP-conjugated antibody (1:5000) prepared in 5% nonfat dry milk for 1 hour at room
temperature with shaking. Membranes were incubated with SuperSignal West Pico PLUS (Thermo
Fisher Scientific), exposed to CL-XPosure film (Thermo Scientific) and developed in a X-OMAT
3000 RA film processor (Kodak, Rochester, NY). The films were scanned, band intensities were
analyzed by densitometry, and CSF sPDGFRβ concentrations were expressed in ng/mL.
BBB Breakdown Markers
Albumin quotient (Qalb, the ratio of CSF-to-plasma albumin levels) was determined using
enzyme-linked immunosorbent assay (ELISA) (Cat. No., E-80AL, Immunology Consultants
Laboratory, Inc., Portland, OR). CSF levels of fibrinogen were determined by ELISA (Cat. No. E-
80FIB, Immunology Consultants Laboratory, Inc., Portland, OR).
Astrocyte Marker
CSF levels of the astrocytic cytokine, S100 calcium-binding protein B (S100B), were
determined using ELISA (Cat. No. EZHS100B-33K, EMD Millipore, Billerica, MA).

61
Inflammatory Markers
Meso Scale Discovery (MSD) multiplex assay was used to determine CSF levels of
interleukin-2 (IL-2), IL-4, IL-6, IL-8, IL-10, IL-12p70, IL-13, IL-1β, tumor necrosis factor α
(TNF-α), and interferon γ (IFN-γ) (Cat. No. K15049G, MSD, Rockville, MD).
Amyloid-β Peptides
MSD multiplex assay (Cat. No. K15200E, MSD, Rockville, MD) was used to determine
CSF levels of Aβ38, Aβ40 and Aβ42. The CSF Aβ42 cutoff level of 190 pg/mL was applied as
previously reported for the MSD Aβ peptide assay
424
.
Amyloid-β Oligomers
CSF Aβ oligomers were measured by ELISA (protocol modified from IBL Cat. No. 27725
and Holtta et al
432
). Aβ1-16 peptide dimer was used as the standard protein prepared at 0, 1, 2.5, 5,
7.5, 10, 15, 20 pM, and 100 µL of prepared standards and neat CSF were added to each well on an
uncoated 96-well plate along with 20 µL/well of HRP-conjugated anti-human Aβ (N) (82E1)
mouse IgG monoclonal antibody; the plates were incubated overnight at 4°C on an orbital plate
shaker at 600 rpm. 100 µL/well was transferred to a 96-well plate precoated with anti-human Aβ
(N) (82E1) mouse IgG monoclonal antibody and incubated for 1 hour at 4°C with shaking. Plates
were washed (Tris buffered saline with 0.1% Tween-20, TBST) and ELAST ELISA amplification
was performed (Perkin Elmer). Briefly, 100 µL/well of biotinyl-tyramide (1:100 dilution) was
incubated for 15 minutes at room temperature with shaking. Plates were washed and 100 µL/well
of streptavidin-HRP (1:500 dilution) was incubated for 30 minutes at room temperature with
shaking. Plates were washed and 100 µL/well of tetramethylbenzidine (TMB) substrate
(Kirkegaard & Perry Laboratories Cat. No. 53-00-01) was incubated in the dark for 60 minutes,
then 100 µL/well of 2N HCl was added, and the plates were read at 450 nm.
Tau
MSD assay was used to determine CSF levels of total tau (Cat. No. K15121G, MSD,
Rockville, MD). Phosphorylated tau (pT181) was determined by ELISA (Cat. No. 81581, Innotest,
Belgium). The CSF pTau181 cutoff level of 78 pg/mL was applied as previously reported
425
.
Tau Oligomers

62
CSF tau oligomers were measured by direct ELISA using tau oligomer-specific antibody
(T22)
433
. Briefly, 12 µL CSF was diluted in a total volume of 50 µL 0.05 M carbonate-bicarbonate
buffer, added to a 96-well MaxiSorp plate (Nunc) and incubated overnight at 4°C on an orbital
plate shaker at 600 rpm. Plates were washed (TBST) and blocked with 300 µL of 10% nonfat dry
milk (BioRad) for 2 hours room temperature with shaking. Plates were washed and incubated with
100 µL/well of T22 antibody (1:250 diluted in 5% nonfat milk) and incubated for 1 hour at room
temperature with shaking. Plates were washed and incubated with HRP-conjugated anti-rabbit IgG
antibody (1:3000 diluted in 5% nonfat dry milk) and incubated for 1 hour at room temperature
with shaking. Plates were washed and incubated in the dark with 100 µL/well TMB substrate for
14 minutes, then 100 µL/well 2N HCl was added, and the plates were read at 450 nm.
Neuronal Marker
CSF levels of neuron specific enolase (NSE) were determined using ELISA (Cat. No. E-
80NEN, Immunology Consultant Laboratories, Portland, OR).
2.2.9 In vitro Analysis of sPDGFRb Shedding
Primary Human Brain Mural Cell Isolation and Culture
Primary human brain vascular smooth muscle cells (SMCs) were isolated from
leptomeningeal arteries (>100 µm diameter) as described and characterized as reported
423
. SMCs
were >98% positive for a-smooth muscle actin (SMA), myosin heavy chain, calponin and SM22
and negative for von Willebrand factor (endothelial cells), GFAP (astrocytes) and CD11b
(microglia). Cells were cultured in smooth muscle cell medium (Cat. No. 1101, ScienCell) in 5%
CO2 at 37°C. Early passage (P5-P6) cultures were used in the present study.
Primary human brain microvascular pericytes were isolated from cortical brain tissue after
removal of leptomeninges as previously described
423,434
. Pericytes were derived from
intraparenchymal microvessels that were completely free from leptomeningeal vessel
contamination. Purified microvessels were largely brain capillaries (>97%) with diameter <6 µm.
Cells were cultured in human pericyte medium (Cat. No. 1201, ScienCell) in 5% CO2 at 37°C and
were then characterized. Pericytes were positive for the pericyte markers PDGFRb, NG2 and
CD13 and negative for von Willebrand factor (endothelial cells), GFAP (astrocytes) and CD11b
(microglia). Early passage (P5-P6) cultures were used in the present study.

63
Treatment Conditions
Primary human brain SMCs and pericytes were plated in equal cell number for all
conditions. For ADAM10 knockdown, Accell ADAM10 siRNA (Cat. No. E-004503-00-0010,
Dharmacon) at a final concentration of 1 µM in Accell Delivery Media (Cat. No. B-005000-500,
Dharmacon) was added into 90% confluent cultured pericytes in twelve-well tissue culture plates
and after 96 hours as recommended by the manufacturer, cells underwent treatment conditions.
Specifically, cells were subjected to treatment with ionomycin (2.5 µM) and/or marimastat (4 µM)
prepared in reduced serum OptiMEM (Gibco) or media only (control condition) for 40 minutes at
37°C, as previously described
435
. After the 40-minute treatment, the media and cell lysates were
collected for additional analysis described below.
Immunoprecipitation of sPDGFRb
Immunoprecipitation was performed on the pericyte and SMCs media as described by the
manufacturer with optimizations (all wash steps performed as described). For antibody-bead
coupling, 50 µL protein G Dynabeads (Cat. No. 10004D, Invitrogen) and 2 µg of PDGFRb
antibody (goat anti-human, Cat. No. AF385, R&D Systems) were incubated with rotating for 10
minutes at room temperature. Conditioned media and equal volume lysis buffer were added to the
Dynabeads-coupled PDGFRb antibody and incubated with rotating for 30 minutes at room
temperature. Target antigen was eluted in denaturing conditions and quantitative Western
immunoblot was performed as described below.
Western Immunoblot Analysis
Quantitative Western immunoblot on immunoprecipitated media was performed using
carrier-free human recombinant PDGFRb as a protein standard (Cat. No. 385-PR/CF, R&D
Systems). Gel transfer was performed using iBlot2 (Thermo Fisher Scientific) at 20 V for 9
minutes. The nitrocellulose membrane was incubated with SuperBlock-TBST (Thermo Fisher
Scientific) for 1 hour at room temperature and primary antibody (Cat. No. AF385, R&D Systems,
1 µg/mL) prepared in SuperBlock was incubated overnight at room temperature with shaking.
Secondary anti-goat antibody (1:5000) prepared in 5% nonfat dry milk was incubated for 1 hour
at room temperature with shaking. SuperSignal West Pico PLUS (Thermo Fisher Scientific) and
film was used to develop the membrane, and sample protein concentration (ng/mL) was calculated.

64
Western immunoblot analysis of primary human brain pericyte cell lysates was performed
with primary antibodies ADAM10 (Cat. No. ab124695, Abcam) and GAPDH (Cat. No. 2118L,
Cell Signaling).
2.2.10 Neuroimaging and Analysis
Magnetic Resonance Imaging
The MRI data sets were obtained at Keck Medical Center of USC. All participants
underwent a blood draw to ensure appropriate kidney function for contrast agent administration
prior to imaging. The imaging protocol performed was developed to detect subtle BBB changes in
patients with cognitive impairment and is detailed in Montagne et al. 2015
178
. Briefly, all images
were obtained on a GE 3 T HDXT MR scanner with a standard eight-channel array head coil.
Anatomical coronal spin echo T2-weighted scans were first obtained through the hippocampi
(TR/TE 1550/97.15 ms, NEX = 1, slice thickness 5 mm with no gap, FOV = 188 x 180 mm, matrix
size = 384 x 384). Baseline coronal T1-weighted maps were then acquired using a T1-weighted
3D spoiled gradient echo (SPGR) pulse sequence and variable flip angle method using flip angles
of 2°, 5° and 10°. Coronal dynamic contrast-enhanced (DCE)-MRI covering the hippocampi and
temporal lobes were acquired using a T1-weighted 3D SPGR pulse sequence (FA = 15°, TR/TE =
8.29/3.09 ms, NEX = 1, slice thickness 5 mm with no gap, FOV 188 x 180 mm, matrix size 160 x
160, voxel size was 0.625 x 0.625 x 5 mm
3
). This sequence was repeated for a total of 16 min with
an approximate time resolution of 15.4 s. Gadolinium-based contrast agent, Gadobenate
dimeglumine (MultiHance®, Bracco, Princeton, New Jersey) or Gadoterate meglumine
(Dotarem®, Guerbet, France) (0.05 mmol/kg) was administered intravenously into the antecubital
vein using a power injector, at a rate of 3 mL/s followed by a 25 mL saline flush, 30 s into the
DCE scan.
Quantification of the Subtle Blood-Brain Barrier Permeability
Post-processing analysis was performed using Rocketship
347
running with Matlab. The
arterial input function (AIF), which was extracted from a region-of-interest (ROI) positioned at
the internal carotid artery, was fitted with a bi-exponential function prior to fitting with Patlak
model
349
. The Patlak linearized regression mathematical analysis was used to generate the BBB
permeability Ktrans maps
178,347,349
with high spatial and temporal resolutions allowing not only

65
simultaneous measurements of the regional BBB permeability in different white (WM) and gray
matter (GM) regions, but also accurate calculations of the Ktrans values in anatomical regions as
small as the subdivisions of the hippocampus. We determined in each individual AIF from the
internal carotid artery. In a few cases when the common carotid artery was not clearly visible a
nearby large vessel was used. Individual AIF measurements are important particularly if the
studied population diverges by age as changes in blood volume and flow may affect AIF and the
Ktrans measurements.
The present analysis requires that the tracer’s diffusion across the BBB remains
unidirectional during the acquisition time. The total tracer concentration in the tissue, Ctissue (t),
can be described as a function of the blood concentration, CAIF (t), the intravascular blood volume,
vp, and a blood-to-brain transfer constant, Ktrans, that represents the flow from the intravascular to
the extravascular extracellular space using equation below:
𝐶
"#$$%&
(𝑡)=𝐾
",-.$
/ 𝐶
012
(𝑢)
"
4
𝑑𝑢 + 𝑣
9
𝐴𝐼𝐹(𝑡)
A statistically significant intersubject variability in the measurement of vp was not
observed. ROI-averaged analysis of DCE-MRI output maps was performed by an experienced
neuroradiologist who manually drew ROIs for each participant based on their own anatomy since
a substantial variability between individuals is seen at a macroscopic level (e.g., enlarged
ventricles, cortical atrophy, hippocampal shrinkage, etc.). Thus, the regional BBB Ktrans
permeability were measured in 13 different GM ROIs including the hippocampus (HC), HC
subfields (i.e., dentate gyrus [DG], CA3, CA1), parahippocampus (PHC), caudate nucleus,
superior frontal cortical gyri, inferior temporal cortical gyri, thalamus, and striatum, and white
matter ROIs including subcortical white matter fibers, corpus callosum, and internal capsule.
Quantification of Regional Brain Volumes
HC and PHC morphometry were performed using the FreeSurfer (v5.3.0) software
package
436
, which is documented and freely available online (http://surfer.nmr.mgh.harvard.edu/).
In brief, HC and PHC gyri were segmented using the included FreeSurfer Desikan-Killiany and
subcortical atlases
437,438
. Then, regional volumes (mm
3
) were derived accordingly. The technical
details of this procedure are described in previous publications
439,440
. Data processing was
performed using the Laboratory of Neuro Imaging (LONI) pipeline system
(http://pipeline.loni.usc.edu)
441,442
.

66
Positron Emission Tomography
Pittsburgh compound B (PiB)-positron emission tomography (PET) imaging was
conducted at Washington University Knight ADRC using procedures and analysis as previously
described
443,444
.
2.2.11 Statistical Analyses
All continuous variables were screened for outliers (+/- 3 SDs from mean) and evaluated
for departures from normality through quantitative examination of skewness and kurtosis, as well
as visual inspection of frequency distributions. Where departures of normality were identified,
log10-transformations were applied, and distribution normalization was confirmed prior to
parametric analyses. Participant demographics and clinical characteristics were initially compared
across both CDR and domain impairment stratifications using chi-square tests and one-way
ANOVAs, with post-hoc Tukey tests.
All CSF biomarkers were compared in parallel analyses applied across the entire sample
stratified by the global CDR score and the number of impaired cognitive domains using ANCOVA,
with post-hoc Bonferroni corrected comparisons. For CDR analyses, model covariates included
age, sex, education and APOE4 carrier status. For domain impairment analyses, age, sex and
education-corrected values were used to identify impairment groups and APOE4 carrier status was
used as a covariate. Site-specific analyses and interaction effect analyses did not include APOE4
carrier status as a covariate to conserve statistical power. For analysis of interactions by Aβ1-42,
pTau and VRF burden, statistical interactions and main effects were examined in similar
ANCOVA models.
The same approach described above was used in all analyses of other CSF glial,
inflammatory and neuronal markers, and for DCE-MRI data. With regard to missing data, all
participants had complete data for primary outcomes (CSF sPDGFRβ and DCE-MRI), and the
extent of missing data was capped at < 10% for all other CSF biomarkers and clinical measures
(i.e., >90% of participants had complete data).
Given the large number of analyses, false discovery rate (FDR)-correction was applied to
all ANCOVA omnibus p-values using the Benjamini-Hochberg method
445
.
Where significant CSF sPDGFRβ and BBB Ktrans findings were identified (CDR 0.5 vs. 0
and domain impairment 1+ vs. 0), separate post-hoc analyses of CSF sPDGFRβ and BBB Ktrans

67
differences controlling for CSF Aβ1-42 and pTau, PiB-PET amyloid deposition, pTau oligomers,
and HC and PHC volumes also utilized ANCOVA models. In addition, separate hierarchical
logistic regression analyses evaluated whether CSF sPDGFRβ and BBB Ktrans predicted cognitive
impairment (CDR 0.5 vs. 0 and domain impairment 1+ vs. 0) after controlling for CSF Aβ1-42 and
pTau, PiB-PET amyloid deposition, pTau oligomers, and HC and PHC volumes. For both
ANCOVA and logistic regression analyses, covariates were entered into the model in the first
block and in the second block either CSF sPDGFRβ or specific regional BBB Ktrans values were
entered. Additional demographics and APOE4 carrier status were included in overall models
correcting for CSF Aβ1-42 and pTau, and models correcting for HC and PHC volumes.
2.3 Results
Different cell types and systems of the neurovascular unit (NVU) are altered in AD
pathophysiology, but the relative relationship between detectable biomarker changes and the
progression of cognitive impairment and/or AD is not well understood. Here, I developed a novel
battery of ~40 biomarkers that reflect different NVU cells and systems including pericyte injury,
BBB disruption/breakdown, endothelial cell injury, vascular growth factors, astrocyte injury,
inflammatory response, white-matter damage, Ab, and neuronal injury (including tau) (Figure
2.1). With the use of multiplexing and ultrasensitive platforms, the entire panel of ~40 biomarkers
are measurable using only 0.5 mL of CSF. In this study, I measured the NVU biomarkers in
individuals with normal cognition or early cognitive impairment from two clinical sites (USC and
Washington University, see Section 2.2.1 Study Participants) and related biomarker changes to
participants’ cognitive status. Tables 2.1 and 2.2 show demographics, clinical data and prevalence
of vascular risk factors (VRFs) by level of Clinical Dementia Rating (CDR) score and number of
cognitive domains impaired, respectively. Individuals diagnosed with vascular dementia and
vascular cognitive impairment, and other disorders that might account for cognitive impairment
were excluded (see Section 2.2.2 Participant Inclusion and Exclusion Criteria).

68

Figure 2.1 Molecular biomarkers of the neurovascular unit.
Biofluid-based biomarkers of the neurovascular unit categorized by cell and system-specific injury
measured in cerebrospinal fluid (CSF) and/or blood. Biomarker categories include: 1) Pericyte
injury, 2) BBB tight/adherens junction proteins*, 3) BBB disruption/breakdown, 4) endothelial
cell injury, 5) vascular growth factors, 6) astrocyte injury, 7) inflammatory response, 8) white-
matter damage, 9) amyloid and 10) neuronal injury. Abbreviations: sPDGFRb, soluble platelet-
derived growth factor receptor-b; VE-cadherin, vascular endothelial cadherin; ZO-1, zonula
occludens-1; CypA, cyclophilin A; MMP9, matrix metalloproteinase-9; sVCAM-1, soluble
vascular cell adhesion molecule-1; sICAM-1, soluble intercellular adhesion molecule-1; PDGF-
BB, platelet-derived growth factor-BB; CRP, C reactive protein; SAA, serum amyloid A; VEGF,
vascular endothelial growth factor; VEGFR1, vascular endothelial growth factor receptor-1; PlGF,
placental growth factor; FGF, fibroblast growth factor; S100B, S100 calcium-binding protein B;
TNFa, tumor necrosis factor-a; IFN-g, interferon-g; IL, interleukin; MBP, myelin basic protein;
MAG, myelin-associated glycoprotein; Ab, amyloid-b; pTau, phosphorylated tau; NSE, neuron-
specific enolase. *Currently in development.
It is known that PDGFRb is primarily expressed in brain by vascular mural cells – capillary
pericytes and vascular smooth muscle cells (SMCs), but not by neurons, astrocytes, endothelial
cells, oligodendrocytes and/or microglia
18,397,446,447
. PDGFRb expression in pericytes is noticeably
higher than in SMCs
18,423,448
. Human brain pericytes, but not SMCs, shed sPDGFRb in the culture
media, which is increased by hypoxia or injurious stimuli
178,423
. Additionally, loss of pericytes in
mice elevates CSF sPDGFRb
178
, supporting that CSF sPDGFRb is a biomarker of brain capillary
pericyte injury. I studied individuals who were cognitively normal as well as those with early
cognitive dysfunction and found increased CSF sPDGFRβ with more advanced CDR impairment
(CDR 1 > 0.5 > 0) (Figure 2.2a), suggesting progressive damage of pericytes
178,423
with cognitive

69

Figure 2.2 Increased CSF sPDGFRβ with CDR impairment is independent of Aβ and tau
status.
(a-c) CSF soluble platelet-derived growth factor receptor-β (sPDGFRβ) (a), Aβ1-42 (b) and pTau
(c) levels in individuals with clinical dementia rating (CDR) score 0 (n=82), 0.5 (n=65) and 1
(n=17). (d) CSF sPDGFRβ in individuals with no cognitive impairment (CDR 0) that are CSF Aβ1-
42 negative (Aβ-; n=53) or positive (Aβ+; n=29), and with cognitive dysfunction (CDR 0.5) that
are Aβ- (n=38) or Aβ+ (n=38). (e) CSF sPDGFRβ in CDR 0 participants that are CSF pTau
negative (pTau-; n=60) or positive (pTau+; n=21) and CDR 0.5 participants that are pTau- (n=33)
or pTau+ (n=29). (f) CSF sPDGFRβ controlled for CSF Aβ1-42 and pTau levels in CDR 0 (n=80)
and CDR 0.5 (n=61) participants. (g) CSF sPDGFRβ is increased with CDR (CDR 0, n=26; CDR
0.5, n=7), independent of amyloid positivity by (11)C-Pittsburgh compound B positron emission
tomography (PiB-PET). (h) No differences were observed in CSF Ab oligomer levels in
individuals with CDR 0 (n=17) versus CDR 0.5 (n=18). (i) No differences were observed in CSF
tau oligomer levels in individuals with CDR 0 (n=63) versus CDR 0.5 (n=63). (j) Increases in CSF
sPDGFRb of individuals with CDR 0.5 vs. CDR 0 remain significant after statistically controlling
for the impact of CSF tau oligomers. Panels a-e, g-i: Box-and-whisker plot lines indicate median
values, boxes indicate interquartile range and whiskers indicate minimum and maximum values.
Panels a-e, g: Significance tests after FDR correction from ANCOVAs with post-hoc Bonferroni
comparisons. Panels f and j: Estimated marginal means ± SEM from ANCOVA models. Panels
h,i: Significance by two-tailed Student’s t-test at a=0.05.
dysfunction. There were no significant differences in CSF Aβ1-42 or phosphorylated tau (pTau)
levels between CDR 0.5 and CDR 0 individuals, although I saw reduced CSF Aβ1-42 in CDR 1
relative to CDR 0.5 participants (Figure 2.2b,c).
Since Ab and tau are both vasculotoxic and Ab1-42 and pTau are hallmark pathological
markers of AD, I wondered whether Ab and tau were driving the increase in CSF sPDGFRβ.
Participants were stratified upon CSF analysis as either Ab-positive (Ab1-42+, <190 pg/mL) or Ab-
negative (Ab1-42-, >190 pg/mL), or pTau-positive (pTau+, >78 pg/mL) or pTau-negative (pTau-,

70
<78 pg/mL), using accepted cutoff values
424–426
. CSF sPDGFRβ was increased in participants with
CDR 0.5 relative to CDR 0 regardless of CSF Aβ1-42 (Figure 2.2d) or pTau (Figure 2.2e) status,
i.e., irrespective of Ab+ or Ab-, or pTau+ or pTau-. Moreover, higher CSF sPDGFRβ remained a
significant predictor of cognitive impairment after statistically controlling for CSF Aβ1-42 and
pTau, as shown by estimated marginal means from ANCOVA models (Figure 2.2f) indicating
medium-to-large incremental effect sizes with η
2
partial range = .10-.12, which has been confirmed
by logistic regression models (Table 2.3a-c). Analyses from the combined cohorts (Figure 2.2a-
e) are also confirmed by site-specific analyses (Figure 2.3).
To further test whether Aβ and/or tau oligomers may mediate the increase in CSF
sPDGFRβ, I measured CSF levels of Aβ and tau oligomers and we investigated Aβ plaques
measured by Pittsburgh compound B (PiB)-positron emission tomography (PET) (PET was
performed at Washington University) in a subset of participants. I found that individuals with CDR
0.5 still exhibited increased CSF sPDGFRβ relative to those with CDR 0 even after

Figure 2.3 Site-specific analysis confirming increased CSF sPDGFRβ with CDR impairment,
independent of Aβ and tau status.
(a,b) Site-specific analysis of CSF sPDGFRβ and standard AD biomarkers, Aβ1-42 and pTau,
indicates an early increase in sPDGFRb with increasing CDR in both independent clinical sites,
USC (a) and Washington University (b). There were no changes in Ab1-42 and pTau at USC site
(a), whereas Ab1-42, but not pTau, was altered at Washington University site; supports Figure 2.2a-
c. (c,d) Site-specific analysis of CSF sPDGFRβ increases with CDR, independent of CSF Aβ1-42
and pTau status in two independent sites, USC (c) and Washington University (d); supports Figure
2.2d-e. Box-and-whisker plot lines indicate median values, boxes indicate interquartile range and
whiskers indicate minimum and maximum values. Significance tests from ANCOVAs. Brackets
denote sample size (n) in each analysis.

71

Figure 2.4 Increased CSF sPDGFRb with CDR is independent of VRFs and confirmed by
site-specific analysis.
(a-c) CSF sPDGFRβ is increased with CDR, independent of vascular risk factors (VRFs) burden
in the combined site analysis (a) and in two independent clinical sites from USC (b) and
Washington University (c). VRFs 0-1: no or 1 vascular risk factor. VRFs 2+: 2 or more vascular
risk factors. See Table 2.1 for the list of VRFs. Box-and-whisker plot lines indicate median values,
boxes indicate interquartile range and whiskers indicate minimum and maximum values.
Significance tests from ANCOVAs. Brackets denote sample size (n) in each analysis.
statistically controlling for PiB-PET amyloid levels (Figure 2.2g), consistent with CSF Aβ1-42
findings (Figure 2.2d). Additionally, I found no differences in CSF Ab and tau oligomer levels
between CDR 0 and CDR 0.5 groups (Figure 2.2h,i). CSF sPDGFRb remained significantly
elevated in CDR analysis after statistically controlling for CSF tau oligomers in ANCOVA models
(Figure 2.2j), suggesting that sPDGFRb increase is not dependent on oligomer levels.
Next, I wondered whether vascular risk factor (VRF) burden influenced the observed
increase in CSF sPDGFRβ. Participants’ vascular risk factor burden was evaluated by the presence
or absence of cardiovascular disease, hypertension, hyperlipidemia, type 2 diabetes, atrial
fibrillation, and transient ischemic attack or minor stroke (see Section 2.2.5 Vascular Risk
Factors). Elevated VRFs burden was defined as having two or more VRFs, as previously
reported
430,431
. I found that increased CSF sPDGFRβ in impaired individuals was independent of
VRFs, as indicated by VRF burden analysis for the entire sample and confirmed by site-specific
analysis (Figure 2.4). Since the present study sample excluded participants with vascular dementia
and vascular cognitive impairment and substantial cerebrovascular pathology, it is probably not
surprising that BBB dysfunction in the present analysis was independent of traditional systemic
VRFs. The possibility of interactive or synergistic effects of traditional VRFs and BBB
dysfunction in populations with more severe vascular lesions, vascular dementia and vascular
cognitive impairment is not ruled out, however, by the present findings. Nevertheless, the fact that

72
brain capillary mural cell damage is independent of traditional VRFs, as I show, is critical
information that underscores the heterogeneity of vascular pathologies in the aging brain.
Moreover, beyond the prominent increase observed in CSF sPDGFRβ reflecting capillary
pericyte injury with increasing cognitive impairment, I then evaluated changes to other NVU
biomarkers of glial and inflammatory responses and neuronal degeneration
182,409
. Out of the 20
biomarkers studied in these categories (Figure 2.1; see Section 2.2.8 Molecular Biofluid Assays),
I did not observe any differences in CSF biomarkers of glial and inflammatory responses or
neuronal degeneration between impaired and unimpaired individuals on CDR exams. This is
illustrated by no changes in a few representative biomarkers including the astrocyte injury

Figure 2.5 Other CSF biomarkers of the neurovascular unit are not altered with CDR
impairment, confirmed by site-specific analysis.
(a-c) CSF markers of glial, inflammatory, or neuronal injury exhibited no significant differences
between unimpaired and impaired individuals on CDR, including S100 calcium-binding protein B
(S100B), interleukin-6 (IL-6), tumor necrosis factor-a (TNFa), or neuron-specific enolase (NSE)
in the combined site analysis (a) and similarly in site-specific analysis of individuals from USC
(b) and from Washington University (c). Box-and-whisker plot lines indicate median values, boxes
indicate interquartile range and whiskers indicate minimum and maximum values. Significance
tests from ANCOVAs. Brackets denote sample size (n) in each analysis.

73
marker S100 calcium-binding protein B (S100B), cytokines interleukin-6 (IL-6) and tumor
necrosis factor-a (TNFa), and neuronal injury marker neuron-specific enolase (NSE) (Figure
2.5a) in the entire sample and also confirmed by site-specific analysis (Figure 2.5b,c). So far,
collectively these findings indicate that increased CSF sPDGFRβ develops early in older adults
with cognitive dysfunction, which is independent of Ab and tau biomarker changes, is not
influenced by VRFs, and is not associated with glial and/or inflammatory response or detectable
neuronal degeneration.
How participants are grouped by cognitive status is particularly important for cross-
sectional studies which do not have the advantage of evaluating biomarker changes in the same
individuals over time. Thus, we next used an independent measure of cognitive status to determine
whether our CDR findings hold when cognitive dysfunction was now evaluated by
neuropsychological performance instead of CDR score. To do this, I analyzed CSF biomarkers
relative to the number of impaired cognitive domains that was determined using normalized scores
from 10 neuropsychological tests that evaluate memory, attention/executive function and
language, and global cognition functions, as described in Section 2.2.4 Neuropsychological
Evaluation and Domains of Impairment (see also Section 2.5 Acknowledgements). This
analysis indicated elevated CSF sPDGFRβ in participants with one cognitive domain impaired
relative to those with no domains impaired (Figure 2.6a). There was no difference, however, in
CSF Aβ1-42 between participants with one domain impaired and those with no domains impaired
(Figure 2.6b). Participants with one domain impaired showed, however, increased CSF pTau
relative to those with no domains impaired (Figure 2.6c).
I similarly evaluated whether increased CSF sPDGFRβ with cognitive domain impairment
was independent of Ab, tau, and VRFs status, and whether other NVU biomarkers were altered
with cognitive domain impairment. Participants were again stratified into those with and without
classic AD biomarker abnormalities which revealed that CSF sPDGFRβ increased in participants
with one or more domain impaired regardless of CSF Aβ1-42 (Figure 2.6d) or pTau (Figure 2.6e)
status. Higher CSF sPDGFRβ levels remained a significant predictor of cognitive impairment after
statistically controlling for CSF Aβ1-42 and pTau, as shown by estimated marginal means from
ANCOVA models (Figure 2.6f) indicating medium-to-large incremental effect sizes (η
2
partial range
= .07 -.14), which has been confirmed by logistic regression models at both sites (Table 2.4a-c).
Similar as for CDR analysis, in the subset of participants who underwent PiB-PET

74

Figure 2.6 Increased CSF sPDGFRβ with cognitive domain impairment is independent of
Aβ, tau and VRF status.
(a-c) CSF sPDGFRβ (a), Aβ1-42 (b) and pTau (c) levels in individuals with no cognitive domains
impaired 0 (n=83), and with 1 (n=39) or 2+ (n=39) cognitive domains impaired. (d) CSF sPDGFRβ
in individuals with no cognitive domains impaired that are CSF Aβ1-42 negative (Aβ-; n=35) or
positive (Aβ+; n=49) and with one or more cognitive domains impaired that are Aβ- (n=37) or
Aβ+ (n=47). (e) CSF sPDGFRβ in individuals with no cognitive domains impaired that are CSF
pTau negative (pTau-; n=63) or positive (pTau+; n=19) and with one or more cognitive domains
impaired that are pTau- (n=39) or pTau+ (n=38). (f) CSF sPDGFRβ controlled for CSF Aβ1-42 and
pTau levels in individuals with 0 domains (n=80) and 1+ domains (n=74) impaired. (g) CSF
sPDGFRβ is increased with cognitive domains impaired (0 domains, n=26; 1+ domains, n=9),
independent of amyloid positivity by (11)C-Pittsburgh compound B positron emission tomography
(PiB-PET). (h) No differences were observed in CSF Ab oligomer levels in individuals with 0
(n=20) or 1+ (n=15) cognitive domains impaired. (i) No differences were observed in CSF tau
oligomer levels in individuals with 0 (n=63) or 1+ (n=61) cognitive domains impaired. (j)
Increases in CSF sPDGFRb of individuals with 1+ versus 0 cognitive domain impairment remain
significant after statistically controlling for the impact of CSF tau oligomers. (k) CSF sPDGFRβ
is increased with increasing number of cognitive domains impaired, independent of vascular risk
factor (VRFs) burden (0 domains VRF 0-1, n=36; 0 domains VRF 2+, n=42; 1+ domains VRF 0-
1, n=28; 1+ domains VRF 2+, n=41). VRFs 0-1: no or 1 vascular risk factor; VRFs 2+: 2 or more
vascular risk factors; See Table 2.2 for the list of VRFs. Panels a-e, g-i, k: Box-and-whisker plot
lines indicate median values, boxes indicate interquartile range and whiskers indicate minimum
and maximum values. Panels a-e, g, k: Significance tests after FDR correction from ANCOVAs
with post-hoc Bonferroni comparisons. Panels h,i: Significance by two-tailed Student’s t-test at
a=0.05. Panels f and j: ANCOVA models representing estimated marginal means ± SEM.

75

Figure 2.7 Site-specific analysis confirming increased CSF sPDGFRβ with cognitive domain
impairment, independent of Aβ, tau and VRF status.
(a,b) Site-specific analysis of CSF sPDGFRβ and standard AD biomarkers, Aβ1-42 and pTau,
indicates an early increase in sPDGFRb with increasing domains impaired in both independent
clinical sites, USC (a) and Washington University (b); supports Figure 2.6a-c. (c,d) Site-specific
analysis of CSF sPDGFRβ indicates increases with the number of cognitive domains impaired,
independent of CSF Aβ1-42 and pTau status in two independent sites, USC (c) and Washington
University (d); supports Figure 2.6d-f. (e,f) Site-specific analysis CSF sPDGFRβ indicates
increases with the number of cognitive domains impaired, independent of VRFs burden in two
independent clinical sites, USC (e) and Washington University (f). VRFs 0-1: no or 1 vascular risk
factor. VRFs 2+: 2 or more vascular risk factors. See Table 2.2 for the list of VRFs. Supports
Figure 2.6k. Box-and-whisker plot lines indicate median values, boxes indicate interquartile range
and whiskers indicate minimum and maximum values. Significance tests from ANCOVAs.
Brackets denote sample size (n) in each analysis.

76
scans, participants with domain impairment exhibited increased CSF sPDGFRβ relative to those
without impairment, after statistically controlling for PiB-PET amyloid levels (Figure 2.6g)
corroborating CSF Aβ1-42 data (Figure 2.6d). There was no difference in CSF Ab and tau
oligomers between participants with impairment in one or more cognitive domains and those
without cognitive impairment (Figure 2.6h,i). CSF sPDGFRb remained significantly increased in
domain analysis after statistically controlling for CSF tau oligomers in ANCOVA models (Figure
2.6j). Similar as the CDR analysis, increased CSF sPDGFRβ with cognitive domain impairment
was independent of VRFs burden (Figure 2.6k). Analyses from the combined cohorts (Figure
2.6a-e, k) are also confirmed by site-specific analyses (Figure 2.7). Again, there were no
differences in CSF markers of glial and/or inflammatory response, or neuronal degeneration
182,409

between impaired and unimpaired participants on neuropsychological exams, as illustrated with a
few examples (Figure 2.8a; also confirmed by site-specific analysis in Figure 2.8b,c).
These data indicate that elevated CSF sPDGFRb is related to cognitive impairment
regardless of how cognitive impairment is determined (both by CDR and cognitive domains
impairment status) and is independent of Ab and tau status and VRFs burden. Next, I asked does
normal aging mediate the increase in CSF sPDGFRb? I found that CSF sPDGFRβ did not correlate
with age in either the cognitively normal (CDR 0) or early cognitive impairment (CDR 0.5) groups
(Figure 2.9a,b) and all CDR and domain impairment group differences in CSF sPDGFRβ values
were significant after age-corrections (Figure 2.2; Figure 2.6). This importantly indicates that
CSF sPDGFRβ reflects cognitive impairment independent of normal aging and therefore may be
good biomarkers of early cognitive dysfunction. Additionally, since brain capillary pericytes
critically maintain BBB integrity
11,397,449
, I wondered whether CSF sPDGFRb related to measures
of BBB breakdown in these individuals. I found a significant positive correlation between CSF
sPDGFRβ with classical biomarkers of BBB breakdown including CSF/plasma albumin ratio and
CSF fibrinogen (Figure 2.9c,d).

77

Figure 2.8 Other CSF biomarkers of the neurovascular unit are not altered with cognitive
domain impairment, confirmed by site-specific analysis.
(a-c) CSF markers of glial, inflammatory, or neuronal injury exhibited no significant differences
between unimpaired and impaired individuals on neuropsychological exams, including S100
calcium-binding protein B (S100B), interleukin-6* (IL-6), tumor necrosis factor-a
†
(TNFa), or
neuron-specific enolase
†
(NSE) in the combined site analysis (a) or in the site-specific analysis of
individuals from USC (b) or from Washington University (c). Panels a-c: Box-and-whisker plot
lines indicate median values, boxes indicate interquartile range and whiskers indicate minimum
and maximum values. Significance tests after FDR correction from ANCOVAs with post-hoc
Bonferroni comparisons. Brackets denote sample size (n) in each analysis. *Analysis did not
survive significance after FDR correction.
†
Individual group comparison p values reported because
omnibus test was p < 0.05 but post-hoc group comparisons were null.

Figure 2.9 CSF sPDGFRβ is not related to age but positively associates with CSF BBB
breakdown markers.
(a,b) Age does not correlate with CSF sPDGFRβ in individuals who are CDR 0 (n=84) (a) or CDR
0.5 (n=67) (b). (c,d) CSF sPDGFRβ is associated with BBB breakdown. CSF sPDGFRβ positively
correlates with conventional biochemical biomarkers of BBB breakdown including CSF/plasma
albumin ratio (Qalb; n=159) (c) and CSF fibrinogen (n=158) (d). Statistical significance by
Pearson correlation; r = Pearson correlation coefficient. Supports Figures 2.2 and 2.6.

78
A fundamental outstanding question is how does microvascular shedding of sPDGFRb
reach the CSF? A previous study found that ADAM10 (a disintegrin and metalloproteinase
domain-containing protein 10) sheds sPDGFRb in fibroblasts
435
, so I wondered whether ADAM10
was also involved in mediating sPDGFRb shedding from brain mural cells. I studied this using
cultured primary human brain pericytes and SMCs, and treated cells with ionomycin (IM, 2.5 µM),
a calcium ionophore that activates ADAM10. I found that, compared to pericytes, SMCs shed
extremely low levels of sPDGFRβ that was not significantly increased by IM. Pericytes shed high
basal levels of sPDGFRβ that was significantly increased by > 5-fold by treatment with IM, which
activated ADAM10. To further determine ADAM10’s involvement, IM treatment was conducted
in the presence of ADAM10 pharmacological inhibition with marimastat (MM, 4 µM) that inhibits
ADAM10 by binding to active site zinc, and genetic siRNA knockdown of ADAM10. Both
pharmacologic (MM) and genetic (siRNA) inhibition of ADAM10 significantly reduced
sPDGFRβ shedding activated by IM by > 90% and 75%, respectively. These data support found
that ADAM10 mediates sPDGFRβ shedding in human pericytes but not SMCs (Figure 2.10),

Figure 2.10 ADAM10 mediates sPDGFRβ shedding in human brain pericytes in vitro.
(a) Primary human brain vascular smooth muscle cells (SMCs) and pericytes were subjected to
treatment with ionomycin (IM, 2.5 µM) to activate ADAM10 or control treatment (media only),
and media was immunoprecipitated (IP) to measure sPDGFRβ by quantitative Western
immunoblot. IM treatment was conducted in the presence of ADAM10 pharmacological inhibition
with marimastat (MM, 4 µM) or genetic siRNA knockdown of ADAM10, which both significantly
reduced IM-induced sPDGFRβ shedding from pericytes. (b) The siRNA ADAM10 knockdown
efficiency in this study was 85% as shown by Western analysis. Data generated from n=3-6
independent culture experiments and plotted as means ± SEM. Statistical analyses: Panel a: SMC
data by two-tailed Student’s t-test; pericyte data by ANOVA with Tukey post-hoc test. Panel b:
Two-tailed Student’s t-test. Significance at a=0.05 for all analyses.

79
elucidating the mechanism of sPDGFRb shedding into the CSF and further supporting that CSF
sPDGFRb is primarily a biomarker of brain capillary pericytes
178,423
.
Beyond pericyte injury, we also wanted to investigate whether BBB breakdown was
similarly related to cognitive impairment in this cohort, and also determine regional changes. In
collaboration with colleagues in our group (see Section 2.5 Acknowledgements), DCE-MRI
analysis of BBB permeability to a gadolinium-based contrast agent was performed in a subset of
73 participants. Regional analysis indicated increased BBB permeability in individuals with
cognitive impairment (both by increasing CDR and cognitive domains impaired) in the
hippocampus (HC), parahippocampus (PHC) and CA1, CA3 and dentate gyrus (DG) HC subfields,
but not in other studied brain regions including frontal and temporal cortex, subcortical white
matter, corpus callosum, and internal capsule, and deep gray matter regions including thalamus,
and striatum (data not shown). These findings are consistent with a recent report demonstrating
that BBB breakdown during normal aging and MCI starts in the HC
178
.
In the present cohort, I observed a significant positive correlation between CSF sPDGFRβ
and BBB permeability in the HC and PHC (Figure 2.11a), consistent with previous findings
178
,
which was not the case for other studied brain regions without BBB permeability as illustrated
here for white matter regions (Figure 2.11b). This corroborates findings that CSF sPDGFRβ
positively correlates with BBB breakdown biomarkers in CSF (Figure 2.9c,d) and further supports
that CSF sPDGFRβ is related to BBB breakdown. Additionally, similar to sPDGFRβ findings, the
VRF burden did not influence BBB permeability changes in the HC, PHC or other studied brain
regions in individuals with early CDR or cognitive domain impairment (data not shown).

Figure 2.11 CSF sPDGFRβ relates to hippocampal gray matter regions.
(a,b) CSF sPDGFRβ is associated with BBB breakdown measured by neuroimaging in
hippocampal gray matter regions (a), but not in white matter (WM) regions including subcortical
WM and corpus callosum (CC) (b); n=69. Statistical significance determined by Pearson
correlation; r = Pearson correlation coefficient.

80
Surprisingly, we also found that individuals with early cognitive impairment (CDR 0.5 or
1+ domains impaired) compared to those who were cognitively normal (CDR 0 or 0 domains
impaired) exhibited BBB breakdown in the HC, PHC and HC subfields regardless of CSF Aβ1-42
and pTau status (Appendix Figure B.1a,b for CDR; Appendix Figure B.2a,b for cognitive
domains impaired). Increased regional BBB permeability in HC, PHC and HC subfields remained
a significant predictor of cognitive impairment after statistically controlling for CSF Aβ1-42 and
pTau, as shown by estimated marginal means from ANCOVA models (Appendix Figure B.1c for
CDR; Appendix Figure B.2c for cognitive domains impaired) indicating medium-to-large
incremental effect sizes (η
2
partial range = .09-.28 for CDR; η
2
partial range = .07-.18 for cognitive
domains impaired), also confirmed by logistic regression models (Table 2.3d-h for CDR; Table
2.4d-h for cognitive domains impaired). Moreover, the regional BBB analysis indicated that Ab
and tau status does not affect BBB integrity in other studied brain regions (data not shown).
Since we observe BBB permeability in the HC and PHC, I next addressed whether changes
in CSF sPDGFRβ depend on regional brain volume. To do this, we conducted ANCOVA analyses
and hierarchical logistic regression correcting for FreeSurfer-derived HC and PHC volumes (see
Section 2.5 Acknowledgements). Importantly, CSF sPDGFRβ increases remained significant
after controlling for HC and PHC volumes (estimated marginal means from ANCOVA models)
(Figure 2.12a for CDR; Figure 2.12d for cognitive domains impaired), and remained increased
when stratifying by CSF Aβ1-42 and pTau status (Figure 2.12b,c for CDR; Figure 2.12e,f for
cognitive domains impaired). Similarly, HC and PHC BBB permeability increases remained
significant after controlling for HC and PHC volumes, respectively, and when stratifying by Aβ1-
42 and pTau status (Appendix Figure B.1d,e for CDR; Appendix Figure B.2d,e for cognitive
domains impaired). All findings exhibited medium-to-large incremental effect sizes after
controlling for HC and PHC volume (η
2
partial range = .09-.31 for CDR; η
2
partial range = .19-.25 for
cognitive domains impaired) and were corroborated by logistic regression models (Table 2.5a-c
for CDR; Table 2.6a-c for cognitive domains impaired). Collectively, these data suggest that BBB
impairment that is represented by CSF sPDGFRβ and DCE-MRI measures is not only independent
of CSF AD biomarkers but is also not related to HC or PHC volume.

81

Figure 2.12 Increased CSF sPDGFRβ during cognitive impairment is independent of
(para)hippocampal volume.
(a-c) CDR analyses: CSF sPDGFRβ values controlled for hippocampal (HC) and parahippocampal
(PHC) volume in individuals with CDR 0 (n=39) and CDR 0.5 (n=18) (a), Aβ- (n=23) or Aβ+
(n=16) CDR 0 and Aβ- (n=10) or Aβ+ (n=8) CDR 0.5 (b), and pTau- (n=30) or pTau+ (n=12)
CDR 0 and pTau- (n=13) or pTau+ (n=7) CDR 0.5 (c). (d-f) Cognitive domain impairment
analyses: CSF sPDGFRβ controlled for HC and PHC volume in individuals with 0 domains (n=38)
and 1+ domains (n=21) impaired (d), and Aβ- (n=30) or Aβ+ (n=12) 0 domains impaired and Aβ-
(n=15) or Aβ+ (n=7) 1+ domains impaired (e), and pTau- (n=30) or pTau+ (n=12) 0 domains
impaired and pTau- (n=13) or pTau+ (n=7) 1+ domains impaired (f). Estimated marginal means ±
SEM from ANCOVA models. Box-and-whisker plot lines indicate median values, boxes indicate
interquartile range and whiskers indicate minimum and maximum values. Significance tests after
FDR correction from ANCOVAs with post-hoc Bonferroni comparisons.
Regional BBB permeability in the medial temporal lobe (HC and PHC) is a perfect
anatomical substrate for episodic memory impairment, and impairments in the HC and its
connecting pathways can also disrupt an array of cognitive functions/domains beyond just
memory. Numerous experimental studies in animals and observational human studies have found
that attention, working memory and executive function can become dysfunctional as a result of
HC-prefrontal pathway disruption
450–453
. Consistently, HC functional activation has been found to
underpin performance on semantic fluency tasks in both normal and pathological human
aging
454,455
. Moreover, since neither CSF sPDGFRβ (Figure 2.11) nor regional BBB permeability
in the HC and PHC (data not shown) were correlated with age and since all CDR and domain
impairment group differences in CSF sPDGFRβ and in HC and PHC BBB permeability were
significant after age-corrections (Figure 2.2; Figure 2.6; Appendix B), these data indicate that
CSF sPDGFRβ and HC and PHC BBB measures reflect cognitive impairment independent of

82
normal aging. Early brain capillary damage (increased CSF sPDGFRβ) and BBB breakdown
therefore may be good biomarkers of early cognitive dysfunction.
2.4 Discussion
In summary, I show that older adults with early cognitive dysfunction develop brain capillary
damage associated with mural cell pericyte injury and BBB breakdown in the HC irrespective of Aβ
and/or tau changes, suggesting that BBB breakdown is an independent, early biomarker of cognitive
impairment unrelated to Aβ and tau. The independence of the BBB breakdown pathway from Aβ/tau
pathway in predicting cognitive impairment is further supported by logistic regression models
indicating that BBB breakdown is not mediating the relationship between AD biomarkers and
cognitive impairment (Tables 2.7-10). Questions remain such as how do genetic factors influence
cerebrovascular dysfunction that contributes to cognitive and functional decline in humans? The
impact of APOE4, the major genetic risk factor for sporadic AD, and the impact of genetic mutations
contributing to early-onset autosomal dominant AD on cerebrovascular and cognitive function in
humans and will be studied next in Chapters 3-4, respectively. Importantly, why are pericytes injured
during cognitive and function decline? This is a key, open question that will be further discussed in
Chapter 7.
Biomarker-based diagnostic approaches, including the recent research recommendations for
AD
366
, mention vascular biomarkers, but suggest that CSF Aβ1-42 and pTau and amyloid PET and tau
PET are the key biomarkers defining AD pathology, although they may not be causal to the disease
process
7,366,456
. My present findings support that cerebrovascular dysfunction may represent a
previously underappreciated factor contributing to cognitive and functional decline, independent of
the classic pathophysiological hallmarks of AD. Moreover, these findings point to the brain vasculature
as an important new biomarker of cognitive dysfunction in both individuals without and with Aβ or
pTau positivity, the latter indicating individuals in the Alzheimer's continuum
366
.
2.5 Acknowledgements
I would like to acknowledge the contributions of several co-authors in this study who helped
with different aspects of data generation and analyses. Specifically: 1) Dr. Daniel Nation (Assistant
Professor in Psychology at USC) worked with our team to run the statistical models and he also

83
determined the participants’ cognitive domains impairment. 2) Dr. Axel Montagne (Assistant
Professor of Research in Physiology & Neuroscience and Assistant Director of the Functional
Biological Imaging Core at USC) analyzed the DCE-MRI scans to determine the participants’ regional
Ktrans BBB permeability, and he and Dr. Farshid Sepehrband (Assistant Professor at Stevens
Neuroimaging and Informatics Institute at USC) determined the hippocampal and parahippocampal
volumes. 3) Dr. Abhay Sagare (Assistant Professor of Research in Physiology & Neuroscience at USC)
with whom I worked to evaluate ADAM10’s involvement in the shedding of sPDGFRβ by human
brain pericytes, and to develop and optimize the CSF Aβ oligomer and CSF tau oligomer assays. 4)
Everyone involved with recruiting and enrolling participants at USC ADRC (including HMRI) and
Washington University Knight ADRC.
This study was supported by the National Institutes of Health (NIH) grant nos. 5P01AG052350
(Zlokovic/Toga) and 5P50AG005142 (Chui) in addition to the Alzheimer’s Association strategic grant
no. 509279 (Zlokovic/Toga), Cure Alzheimer’s Fund, and the Foundation Leducq Transatlantic
Network of Excellence for the Study of Perivascular Spaces in Small Vessel Disease reference no. 16
CVD 05 (Zlokovic/Wardlaw). Enrollment of participants into the Washington University Knight
ADRC is supported by NIH grant nos. P50AG05681 (Morris), P01AG03991 (Morris), and
P01AG026276 (Morris). Enrollment of participants into the USC ADRC is supported by NIH grant
no. 5P50AG005142 (Chui).

84
2.6 Supplementary Tables
Table 2.1 Total sample by level of Clinical Dementia Rating (CDR) score.
Participants with CDR 0 vs. 0.5 vs. 1 were compared on demographic and clinical characteristics.
P-values for age and education are from post-hoc Tukey tests where the one-way ANOVA
omnibus (F) was significant, and those for sex, education, APOE4 and all VRFs are from chi-
square (χ
2
) tests. Mean ± SD are displayed. Effect sizes are expressed by Cohen’s d and
proportional differences.

Abbreviations: VRFs, vascular risk factors; CDR, clinical dementia rating; APOE4, apolipoprotein
E4; TIA, transient ischemic attack; ANOVA, analysis of variance; USC, University of Southern
California.

85
Table 2.2 Total sample by number of cognitive domains impaired.
Participants with impairment in 0 vs. 1 vs. 2+ cognitive domains were compared on demographic
and clinical characteristics. P-values for age and education are from post-hoc Tukey tests where
the one-way ANOVA omnibus (F) was significant, and those for sex, education, APOE4 and all
VRFs are from chi-square (χ
2
) tests. Mean ± SD are displayed. Effect sizes are expressed by
Cohen’s d and proportional differences.

Abbreviations: VRFs, vascular risk factors; CDR, clinical dementia rating; APOE4, apolipoprotein
E4; TIA, transient ischemic attack; ANOVA, analysis of variance; USC, University of Southern
California.

86
Table 2.3 Logistic regression models of CSF sPDGFRβ or regional BBB Ktrans beyond
demographic factors and AD biomarkers in relation to CDR status.
Hierarchical logistic regression analyses examine the overall incremental predictive value of CSF
sPDGFRβ beyond demographic factors and CSF Aβ1-42 and pTau markers in relation to CDR 0.5
vs. 0 in the combined analysis (n=141) (a) and replicated at USC (n=68) (b) and Washington
University (n=73) (c), and the incremental predictive value of DCE-MRI beyond demographic
factors and CSF Aβ1-42 and pTau markers in relation to CDR 0.5 vs. 0 (n=65) based on BBB Ktrans
values from the hippocampus (d), parahippocampus (e), and hippocampal subfields including CA1
(f), CA3 (g) and DG (h). Effect sizes are indicated by logistic regression coefficients (B) expressed
in log-odds units.

2.3a. Total sample sPDGFRβ predicting CDR 0.5 vs. 0.
Chi-square df p-value
Step 2 6.786 1 0.009
Model 31.418 7 0.00005

Step Predictor B S.E. Wald p-value
1 Age (yrs) 0.036 0.022 2.608 0.106
1 Sex (freq male) -0.350 0.390 0.802 0.370
1 Education (yrs) 0.008 0.068 0.015 0.902
1 APOE4 status (freq ε4) 0.388 0.405 0.921 0.337
1 CSF Aβ 1-42 (pg/mL) -0.002 0.001 1.843 0.175
1 CSF pTau (pg/mL) 0.013 0.006 4.612 0.032
2 CSF sPDGFRβ (ng/mL) 0.001 0.001 6.212 0.013

2.3b. USC sample sPDGFRβ predicting CDR 0.5 vs. 0.
Chi-square df p-value
Step 2 7.508 1 0.006
Model 11.65 6 0.070

Step Predictor B S.E. Wald p-value
1 Age (yrs) -0.013 0.031 0.193 0.660
1 Sex (freq male) -0.077 0.571 0.018 0.893
1 Education (yrs) 0.240 0.135 3.150 0.076
1 CSF Aβ 1-42 (pg/mL) 0.000 0.002 0.055 0.815
1 CSF pTau (pg/mL) 0.002 0.010 0.051 0.822
2 CSF sPDGFRβ (ng/mL) 0.002 0.001 6.324 0.012

2.3c. Washington University sample sPDGFRβ predicting CDR 0.5 vs. 0.
Chi-square df p-value
Step 2 4.571 1 0.033
Model 28.774 6 0.00007

Step Predictor B S.E. Wald p-value
1 Age (yrs) 0.078 0.037 4.372 0.037
1 Sex (freq male) -0.297 0.638 0.217 0.642

87
1 Education (yrs) -0.088 0.126 0.484 0.487
1 CSF Aβ 1-42 (pg/mL) -0.003 0.002 2.390 0.122
1 CSF pTau (pg/mL) 0.008 0.009 0.728 0.394
2 CSF sPDGFRβ (ng/mL) 0.002 0.001 4.040 0.044

2.3d. Hippocampal DCE-MRI predicting CDR 0.5 vs. 0.
Chi-square df p-value
Step 2 18.193 1 0.00003
Model 23.362 6 0.0003

Step Predictor B S.E. Wald p-value
1 Age (yrs) -0.030 0.039 0.598 0.439
1 Sex (freq male) -0.491 0.645 0.580 0.446
1 Education (yrs) 0.200 0.136 2.173 0.140
1 CSF Aβ 1-42 (pg/mL) 0.001 0.002 0.077 0.781
1 CSF pTau (pg/mL) 0.003 0.012 0.087 0.769
2 BBB K trans (x 10
-3
min
-1
) 4.739 1.295 13.400 0.0003

2.3e. Parahippocampal DCE-MRI predicting CDR 0.5 vs. 0.
Chi-square df p-value
Step 2 21.974 1 0.000001
Model 27.143 6 0.00007

Step Predictor B S.E. Wald p-value
1 Age (yrs) -0.036 0.041 0.762 0.383
1 Sex (freq male) -0.620 0.679 0.835 0.361
1 Education (yrs) 0.257 0.148 3.024 0.082
1 CSF Aβ 1-42 (pg/mL) 0.000 0.002 0.038 0.846
1 CSF pTau (pg/mL) 0.004 0.012 0.128 0.721
2 BBB K trans (x 10
-3
min
-1
) 6.749 1.829 13.617 0.0002

2.3f. CA1 DCE-MRI predicting CDR 0.5 vs. 0.
Chi-square df p-value
Step 2 15.604 1 0.00008
Model 20.772 6 0.002

Step Predictor B S.E. Wald p-value
1 Age (yrs) -0.018 0.039 0.218 0.641
1 Sex (freq male) -0.503 0.625 0.648 0.421
1 Education (yrs) 0.217 0.138 2.476 0.116
1 CSF Aβ 1-42 (pg/mL) 0.000 0.002 0.038 0.845
1 CSF pTau (pg/mL) 0.003 0.011 0.050 0.824
2 BBB K trans (x 10
-3
min
-1
) 2.995 0.881 11.565 0.001

88

2.3g. CA3 DCE-MRI predicting CDR 0.5 vs. 0.
Chi-square df p-value
Step 2 5.753 1 0.016
Model 10.921 6 0.091

Step Predictor B S.E. Wald p-value
1 Age (yrs) -0.008 0.035 0.056 0.814
1 Sex (freq male) -0.481 0.568 0.717 0.397
1 Education (yrs) 0.166 0.125 1.781 0.182
1 CSF Aβ 1-42 (pg/mL) -0.001 0.002 0.186 0.667
1 CSF pTau (pg/mL) 0.010 0.010 1.077 0.299
2 BBB K trans (x 10
-3
min
-1
) 2.831 1.285 4.851 0.028

2.3h. DG DCE-MRI predicting CDR 0.5 vs. 0.
Chi-square df p-value
Step 2 12.501 1 0.0005
Model 17.669 6 0.007

Step Predictor B S.E. Wald p-value
1 Age (yrs) -0.025 0.036 0.501 0.479
1 Sex (freq male) -0.655 0.607 1.167 0.28
1 Education (yrs) 0.169 0.134 1.582 0.208
1 CSF Aβ 1-42 (pg/mL) -0.001 0.002 0.121 0.727
1 CSF pTau (pg/mL) 0.008 0.01 0.597 0.44
2 BBB K trans (x 10
-3
min
-1
) 6.865 2.306 8.865 0.003
Abbreviations: BBB, blood-brain barrier; DG, dentate gyrus; CA1, cornu amonis 1; CA3, cornu
amonis 3; DCE, dynamic contrast-enhanced; MRI, magnetic resonance imaging; CSF,
cerebrospinal fluid; Aβ1-42, amyloid-β; pTau, phosphorylated tau; APOE4, apolipoprotein E4;
CDR, clinical dementia rating; sPDGFRβ, soluble platelet-derived growth factor receptor-β.

89
Table 2.4 Logistic regression models of CSF sPDGFRβ or regional BBB Ktrans beyond AD
biomarkers in relation to cognitive domains impaired.
Hierarchical logistic regression analyses examine the overall incremental predictive value of CSF
sPDGFRβ beyond demographic factors and CSF Aβ1-42 and pTau markers in relation to 1+ vs. 0
impaired cognitive domains in the combined analysis (n=156) (a) and replicated at USC (n=71)
(b) and Washington University (n=85) (c), and the incremental predictive value of DCE-MRI
beyond demographic factors and CSF Aβ1-42 and pTau markers in relation to 1+ vs. 0 impaired
cognitive domains, n=70, based on BBB Ktrans values from the hippocampus (d), parahippocampus
(e), and hippocampal subfields including CA1 (f), CA3 (g), and DG (h). Effect sizes are indicated
by logistic regression coefficients (B) expressed in log-odds units.

2.4a. Total sample sPDGFRβ predicting Domain 1+ vs. 0.
Chi-square df p-value
Step 2 22.322 1 0.000002
Model 53.206 4 7.72 × 10
-11

Step Predictor B S.E. Wald p-value
1 APOE4 status (freq ε4) 0.446 0.415 1.156 0.282
1 CSF Aβ 1-42 (pg/mL) -0.003 0.001 6.165 0.013
1 CSF pTau (pg/mL) 0.023 0.007 12.388 0.0004
2 CSF sPDGFRβ (ng/mL) 0.003 0.001 17.471 0.00003

2.4b. USC sample sPDGFRβ predicting Domain 1+ vs. 0.
Chi-square df p-value
Step 2 14.359 1 0.0002
Model 20.638 3 0.0001

Step Predictor B S.E. Wald p-value
1 CSF Aβ 1-42 (pg/mL) -0.003 0.002 1.828 0.176
1 CSF pTau (pg/mL) 0.016 0.01 2.784 0.095
2 CSF sPDGFRβ (ng/mL) 0.003 0.001 10.721 0.001

2.4c. Washington University sample sPDGFRβ predicting Domain 1+ vs. 0.
Chi-square df p-value
Step 2 10.898 1 0.001
Model 35.021 3 1.21 × 10
-7

Step Predictor B S.E. Wald p-value
1 CSF Aβ 1-42 (pg/mL) -0.003 0.002 4.475 0.034
1 CSF pTau (pg/mL) 0.026 0.010 7.242 0.007
2 CSF sPDGFRβ (ng/mL) 0.003 0.001 8.315 0.004

90

2.4d. Hippocampal DCE-MRI predicting Domain 1+ vs. 0.
Chi-square df p-value
Step 2 12.975 1 0.0003
Model 19.550 3 0.0002

Step Predictor B S.E. Wald p-value
1 CSF Aβ 1-42 (pg/mL) -0.001 0.002 0.593 0.441
1 CSF pTau (pg/mL) 0.009 0.010 0.823 0.364
2 BBB K trans (x 10
-3
min
-1
) 3.735 1.142 10.697 0.001

2.4e. Parahippocampal DCE-MRI predicting Domain 1+ vs. 0.
Chi-square df p-value
Step 2 12.070 1 0.001
Model 18.645 3 0.0003

Step Predictor B S.E. Wald p-value
1 CSF Aβ 1-42 (pg/mL) -0.002 0.002 0.828 0.363
1 CSF pTau (pg/mL) 0.009 0.010 0.939 0.332
2 BBB K trans (x 10
-3
min
-1
) 4.268 1.360 9.849 0.002

2.4f. CA1 DCE-MRI predicting Domain 1+ vs. 0.
Chi-square df p-value
Step 2 8.871 1 0.003
Model 15.445 3 0.001

Step Predictor B S.E. Wald p-value
1 CSF Aβ 1-42 (pg/mL) -0.002 0.002 0.763 0.382
1 CSF pTau (pg/mL) 0.010 0.009 1.216 0.270
2 BBB K trans (x 10
-3
min
-1
) 2.072 0.752 7.584 0.006

2.4g. CA3 DCE-MRI predicting Domain 1+ vs. 0.
Chi-square df p-value
Step 2 3.235 1 0.072
Model 9.809 3 0.020

Step Predictor B S.E. Wald p-value
1 CSF Aβ 1-42 (pg/mL) -0.002 0.002 1.951 0.162
1 CSF pTau (pg/mL) 0.015 0.009 2.724 0.099
2 BBB K trans (x 10
-3
min
-1
) 1.955 1.123 3.033 0.082

91

2.4h. DG DCE-MRI predicting Domain 1+ vs. 0.
Chi-square df p-value
Step 2 5.329 1 0.021
Model 11.903 3 0.008

Step Predictor B S.E. Wald p-value
1 CSF Aβ 1-42 (pg/mL) -0.003 0.002 2.021 0.155
1 CSF pTau (pg/mL) 0.013 0.009 2.160 0.142
2 BBB K trans (x 10
-3
min
-1
) 3.952 1.811 4.761 0.029
Abbreviations: BBB, blood-brain barrier; DG, dentate gyrus; CA1, cornu amonis 1; CA3, cornu
amonis 3; DCE, dynamic contrast-enhanced; MRI, magnetic resonance imaging; CSF,
cerebrospinal fluid; Aβ1-42, amyloid-β; pTau, phosphorylated tau; APOE4, apolipoprotein E4;
sPDGFRβ, soluble platelet-derived growth factor receptor-β.

92
Table 2.5 Logistic regression models of CSF sPDGFRβ or regional BBB Ktrans beyond
demographic factors and regional brain volume in relation to CDR status.
Hierarchical logistic regression analyses examine the overall incremental predictive value of CSF
sPDGFRβ beyond demographic factors and hippocampal and parahippocampal volume in relation
to CDR 0.5 vs. 0 (n=57) (a), and the incremental predictive value of DCE-MRI (n=65) beyond
demographic factors and hippocampal volume for BBB Ktrans values from the hippocampus (b)
and parahippocampal volume for BBB Ktrans values the parahippocampus (c). Effect sizes are
indicated by logistic regression coefficients (B) expressed in log-odds units.

2.5a. CSF sPDGFRβ predicting CDR 0.5 vs. 0.
Chi-square df p-value
Step 2 7.231 1 0.007
Model 15.165 6 0.019

Step Predictor B S.E. Wald p-value
1 Age (years) -0.034 0.047 0.539 0.463
1 Sex (freq Male) 0.137 0.721 0.036 0.849
1 Education (years) 0.307 0.164 3.502 0.061
1 Hippocampal Volume (mm
3
/TICV) -3.115 6.837 0.208 0.649
1 Parahippocampal Volume (mm
3
/TICV) -15.672 10.064 2.425 0.119
2 CSF sPDGFRβ (ng/mL) 0.002 0.001 6.205 0.013

2.5b. Hippocampal DCE-MRI predicting CDR 0.5 vs. 0.
Chi-square df p-value
Step 2 17.033 1 0.000005
Model 24.452 5 0.00008

Step Predictor B S.E. Wald p-value
1 Age (yrs) -0.094 0.058 2.641 0.104
1 Sex (freq male) 0.195 0.811 0.058 0.810
1 Education (yrs) 0.285 0.172 2.730 0.098
1 Hippocampal Volume (mm
3
/TICV) -11.244 7.247 2.407 0.121
2 BBB K trans (x 10
-3
min
-1
) 5.579 1.689 10.917 0.001

2.5c. Parahippocampal DCE-MRI predicting CDR 0.5 vs. 0.
Chi-square df p-value
Step 2 20.47 1 0.000001
Model 29.942 5 0.000002

Step Predictor B S.E. Wald p-value
1 Age (yrs) -0.086 0.053 2.566 0.109
1 Sex (freq male) 0.721 0.876 0.678 0.410
1 Education (yrs) 0.481 0.194 6.145 0.013
1 Parahippocampal Volume (mm
3
/TICV) -19.429 12.335 2.481 0.115
2 BBB K trans (x 10
-3
min
-1
) 8.891 2.732 10.590 0.001

93
Abbreviations: TICV, total intracranial volume; BBB, blood-brain barrier; DCE, dynamic
contrast-enhanced; MRI, magnetic resonance imaging; CSF, cerebrospinal fluid; CDR, clinical
dementia rating; sPDGFRβ, soluble platelet-derived growth factor receptor-β.

94
Table 2.6 Logistic regression models of CSF sPDGFRβ or regional BBB Ktrans beyond
regional brain volume in relation to cognitive domains impaired.
Hierarchical logistic regression analyses examine the overall incremental predictive value of CSF
sPDGFRβ beyond demographic factors and hippocampal and parahippocampal volume in relation
to 1+ vs. 0 impaired cognitive domains (n=60) (a), and the incremental predictive value of DCE-
MRI beyond demographic factors and hippocampal volume for BBB Ktrans values from the
hippocampus (b) and parahippocampal volume for DCE-MRI (n=67) of the parahippocampus (c).
Effect sizes are indicated by logistic regression coefficients (B) expressed in log-odds units.

2.6a. CSF sPDGFRβ predicting Domain 1+ vs. 0.
Chi-square df p-value
Step 2 16.502 1 0.00005
Model 19.418 3 0.0002

Step Predictor B S.E. Wald p-value
1 Hippocampal Volume (mm
3
/TICV) -3.911 4.656 0.706 0.401
1 Parahippocampal Volume (mm
3
/TICV) -12.063 9.58 1.585 0.208
2 CSF sPDGFRβ (ng/mL) 0.003 0.001 10.675 0.001

2.6b. Hippocampal DCE-MRI predicting Domain 1+ vs. 0.
Chi-square df p-value
Step 2 14.607 1 0.0001
Model 16.38 2 0.0003

Step Predictor B S.E. Wald p-value
1 Hippocampal Volume (mm
3
/TICV) -4.948 3.955 1.565 0.211
2 BBB K trans (x 10
-3
min
-1
) 3.811 1.127 11.441 0.001

2.6c. Parahippocampal DCE-MRI predicting Domain 1+ vs. 0.
Chi-square df p-value
Step 2 13.157 1 0.0003
Model 17.536 2 0.0002

Step Predictor B S.E. Wald p-value
1 Parahippocampal Volume (mm
3
/TICV) -14.705 7.802 3.552 0.059
2 BBB K trans (x 10
-3
min
-1
) 4.489 1.414 10.075 0.002
Abbreviations: TICV, total intracranial volume; BBB, blood-brain barrier; DCE, dynamic
contrast-enhanced; MRI, magnetic resonance imaging; CSF, cerebrospinal fluid; sPDGFRβ,
soluble platelet-derived growth factor receptor-β.

95
Table 2.7 Logistic regression models of AD biomarkers with demographic factors and
beyond CSF sPDGFRβ in relation to CDR status.
Hierarchical logistic regression analyses (n=141) evaluate age, sex, education, APOE4 status, CSF
Aβ1-42 and pTau in relation to CDR 0.5 vs. 0 (a), and the incremental predictive value of CSF Aβ1-
42 and pTau markers beyond demographic factors and CSF sPDGFRβ in relation to CDR 0.5 vs. 0
(b). Effect sizes are indicated by logistic regression coefficients (B) expressed in log-odds units.

2.7a. Total sample Aβ 1-42 and pTau predicting CDR 0.5 vs. 0.
Chi-square df p-value
Model 24.633 6 0.0004

Predictor B S.E. Wald p-value
Age (yrs) 0.037 0.021 3.067 0.080
Sex (freq male) -0.481 0.377 1.623 0.203
Education (yrs) -0.010 0.067 0.023 0.879
APOE4 status (freq ε4) 0.586 0.388 2.285 0.131
CSF Aβ 1-42 (pg/mL) -0.002 0.001 2.222 0.136
CSF pTau (pg/mL) 0.014 0.006 6.025 0.014

2.7b. Total sample Aβ 1-42 and pTau predicting CDR 0.5 vs. 0 controlling for sPDGFRβ.
Chi-square df p-value
Step 2 7.234 2 0.027
Model 31.418 7 0.00005

Step Predictor B S.E. Wald p-value
1 Age (yrs) 0.036 0.022 2.608 0.106
1 Sex (freq male) -0.350 0.390 0.802 0.370
1 Education (yrs) 0.008 0.068 0.015 0.902
1 APOE4 status (freq ε4) 0.388 0.405 0.921 0.337
1 CSF sPDGFRβ (ng/mL) 0.001 0.001 6.212 0.013
2 CSF Aβ 1-42 (pg/mL) -0.002 0.001 1.843 0.175
2 CSF pTau (pg/mL) 0.013 0.006 4.612 0.032
Abbreviations: CSF, cerebrospinal fluid; Aβ1-42, amyloid-β; pTau, phosphorylated tau; APOE4,
apolipoprotein E4; CDR, clinical dementia rating; sPDGFRβ, soluble platelet-derived growth
factor receptor-β.

96
Table 2.8 Logistic regression models of AD biomarkers with demographic factors and
beyond regional BBB Ktrans in relation to CDR status.
Hierarchical logistic regression analyses within the DCE-MRI subsample (n=67) evaluate age, sex,
education, and CSF Aβ1-42 and pTau in relation to CDR 0.5 vs. 0 (a), and the incremental predictive
value of CSF Aβ1-42 and pTau markers beyond demographic factors and BBB Ktrans values from
the hippocampus (b), parahippocampus (c), and hippocampal subfields i.e. CA1 (d), CA3 (e), and
DG (f). Effect sizes indicated by logistic regression coefficients (B) expressed in log-odds units.

2.8a. DCE-MRI subsample Aβ 1-42 and pTau predicting CDR 0.5 vs. 0.
Chi-square df p-value
Model 5.168 5 0.396

Predictor B S.E. Wald p-value
Age (yrs) -0.017 0.033 0.268 0.605
Sex (freq male) -0.521 0.538 0.939 0.332
Education (yrs) 0.148 0.115 1.651 0.199
CSF Aβ 1-42 (pg/mL) -0.001 0.002 0.386 0.535
CSF pTau (pg/mL) 0.009 0.009 1.011 0.315

2.8b. Total sample Aβ 1-42 and pTau predicting CDR 0.5 vs. 0 controlling for hippocampal K trans.
Chi-square df p-value
Step 2 0.162 2 0.922
Model 23.362 6 0.001

Step Predictor B S.E. Wald p-value
1 Age (yrs) -0.030 0.039 0.598 0.439
1 Sex (freq male) -0.491 0.645 0.580 0.446
1 Education (yrs) 0.200 0.136 2.173 0.140
1 BBB K trans (x 10
-3
min
-1
) 4.739 1.295 13.400 <.001
2 CSF Aβ 1-42 (pg/mL) 0.001 0.002 0.077 0.781
2 CSF pTau (pg/mL) 0.003 0.012 0.087 0.769

2.8c. Total sample Aβ 1-42 and pTau predicting CDR 0.5 vs. 0 controlling for parahippocampal
K trans.
Chi-square df p-value
Step 2 0.167 2 0.92
Model 27.143 6 0.0001

Step Predictor B S.E. Wald p-value
1 Age (yrs) -0.036 0.041 0.762 0.383
1 Sex (freq male) -0.620 0.679 0.835 0.361
1 Education (yrs) 0.257 0.148 3.024 0.082
1 BBB K trans (x 10
-3
min
-1
) 6.749 1.829 13.617 0.0002
2 CSF Aβ 1-42 (pg/mL) 0.000 0.002 0.038 0.846
2 CSF pTau (pg/mL) 0.004 0.012 0.128 0.721

97

2.8d. Total sample Aβ 1-42 and pTau predicting CDR 0.5 vs. 0 controlling for CA1 K trans.
Chi-square df p-value
Step 2 0.088 2 0.957
Model 20.722 6 0.002

Step Predictor B S.E. Wald p-value
1 Age (yrs) -0.018 0.039 0.218 0.641
1 Sex (freq male) -0.503 0.625 0.648 0.421
1 Education (yrs) 0.217 0.138 2.476 0.116
1 BBB K trans (x 10
-3
min
-1
) 2.995 0.881 11.565 0.001
2 CSF Aβ 1-42 (pg/mL) 0.000 0.002 0.038 0.845
2 CSF pTau (pg/mL) 0.003 0.011 0.050 0.824

2.8e. Total sample Aβ 1-42 and pTau predicting CDR 0.5 vs. 0 controlling for CA3 K trans.
Chi-square df p-value
Step 2 1.212 2 0.546
Model 10.921 6 0.091

Step Predictor B S.E. Wald p-value
1 Age (yrs) -0.008 0.035 0.056 0.814
1 Sex (freq male) -0.481 0.568 0.717 0.397
1 Education (yrs) 0.166 0.125 1.781 0.182
1 BBB K trans (x 10
-3
min
-1
) 2.831 1.285 4.851 0.028
2 CSF Aβ 1-42 (pg/mL) -0.001 0.002 0.186 0.667
2 CSF pTau (pg/mL) 0.010 0.010 1.077 0.299

2.8f. Total sample Aβ 1-42 and pTau predicting CDR 0.5 vs. 0 controlling for DG K trans.
Chi-square df p-value
Step 2 0.684 2 0.71
Model 17.669 6 0.007

Step Predictor B S.E. Wald p-value
1 Age (yrs) -0.025 0.036 0.501 0.479
1 Sex (freq male) -0.655 0.607 1.167 0.280
1 Education (yrs) 0.169 0.134 1.582 0.208
1 BBB K trans (x 10
-3
min
-1
) 6.865 2.306 8.865 0.003
2 CSF Aβ 1-42 (pg/mL) -0.001 0.002 0.121 0.727
2 CSF pTau (pg/mL) 0.008 0.010 0.597 0.440
Abbreviations: BBB, blood-brain barrier; DG, dentate gyrus; CA1, cornu amonis 1; CA3, cornu
amonis 3; DCE, dynamic contrast-enhanced; MRI, magnetic resonance imaging; CSF,
cerebrospinal fluid; Aβ1-42, amyloid-β; pTau, phosphorylated tau; CDR, clinical dementia rating.

98
Table 2.9 Logistic regression models of AD biomarkers and predictive value beyond CSF
sPDGFRβ in relation to cognitive domains impaired.
Hierarchical logistic regression analyses (n=156) evaluate APOE4 status and CSF Aβ1-42 and pTau
in relation to impairment in 1+ vs. 0 cognitive domains (a), and the incremental predictive value
of CSF Aβ1-42 and pTau markers beyond APOE4 status and CSF sPDGFRβ in relation to
impairment in 1+ vs. 0 cognitive domains (b). Effect sizes are indicated by logistic regression
coefficients (B) expressed in log-odds units.

2.9a. Total sample Aβ 1-42 and pTau predicting Domain 1+ vs. 0.
Chi-square df p-value
Model 30.884 3 8.99 × 10
-7

Predictor B S.E. Wald p-value
APOE4 status (freq ε4) 0.085 0.367 0.053 0.817
CSF Aβ 1-42 (pg/mL) -0.003 0.001 7.288 0.007
CSF pTau (pg/mL) 0.023 0.006 14.636 0.0001

2.9b. Total sample Aβ 1-42 and pTau predicting Domain 1+ vs. 0 controlling for sPDGFRβ.
Chi-square df p-value
Step 2 25.671 2 0.000004
Model 53.206 4 7.72 × 10
-11

Step Predictor B S.E. Wald p-value
1 APOE4 status (freq ε4) 0.446 0.415 1.156 0.282
1 CSF sPDGFRβ (ng/mL) 0.003 0.001 17.471 0.00003
2 CSF Aβ 1-42 (pg/mL) -0.003 0.001 6.165 0.013
2 CSF pTau (pg/mL) 0.023 0.007 12.388 0.0004
Abbreviations: CSF, cerebrospinal fluid; Aβ1-42, amyloid-β; pTau, phosphorylated tau; APOE4,
apolipoprotein E4; sPDGFRβ, soluble platelet-derived growth factor receptor-β.

99
Table 2.10 Logistic regression models of AD biomarkers and predictive value beyond
regional BBB Ktrans in relation to cognitive domains impaired.
Hierarchical logistic regression analyses (n=70) evaluate CSF Aβ1-42 and pTau in relation to
impairment in 1+ vs. 0 cognitive domains (a), and the incremental predictive value of CSF Aβ1-42
and pTau markers in relation to impairment in 1+ vs. 0 cognitive domains beyond BBB Ktrans
values from the hippocampus (b), parahippocampus (c), and hippocampal subfields including CA1
(d), CA3 (e), and DG (f). Effect sizes are indicated by logistic regression coefficients (B) expressed
in log-odds units.

2.10a. DCE-MRI subsample Aβ 1-42 and pTau predicting Domain 1+ vs. 0.

Chi-square df p-value

Model 6.574 2 0.037

Predictor B S.E. Wald p-value

CSF Aβ 1-42 (pg/mL) -0.003 0.002 2.439 0.118

CSF pTau (pg/mL) 0.014 0.008 2.810 0.094

2.10b. Total sample Aβ 1-42 and pTau predicting Domain 1+ vs. 0 controlling for hippocampal
K trans.

Chi-square df p-value

Step 2 1.651 2 0.438

Model 19.55 3 0.0002

Step Predictor B S.E. Wald p-value
1 BBB K trans (x 10
-3
min
-1
) 3.735 1.142 10.697 0.001
2 CSF Aβ 1-42 (pg/mL) -0.001 0.002 0.593 0.441
2 CSF pTau (pg/mL) 0.009 0.010 0.823 0.364

2.10c. Total sample Aβ 1-42 and pTau predicting Domain 1+ vs. 0 controlling for parahippocampal
K trans.

Chi-square df p-value

Step 2 2.057 2 0.358

Model 18.645 3 0.0003

Step Predictor B S.E. Wald p-value
1 BBB K trans (x 10
-3
min
-1
) 4.268 1.360 9.849 0.002
2 CSF Aβ 1-42 (pg/mL) -0.002 0.002 0.828 0.363
2 CSF pTau (pg/mL) 0.009 0.010 0.939 0.332

2.10d. Total sample Aβ 1-42 and pTau predicting Domain 1+ vs. 0 controlling for CA1 K trans.

Chi-square df p-value

Step 2 2.35 2 0.309

Model 15.445 3 0.001

Step Predictor B S.E. Wald p-value

100
1 BBB K trans (x 10
-3
min
-1
) 2.072 0.752 7.584 0.006
2 CSF Aβ 1-42 (pg/mL) -0.002 0.002 0.763 0.382
2 CSF pTau (pg/mL) 0.010 0.009 1.216 0.270

2.10e. Total sample Aβ 1-42 and pTau predicting Domain 1+ vs. 0 controlling for CA3 K trans.

Chi-square df p-value

Step 2 5.875 2 0.053

Model 9.809 3 0.02

Step Predictor B S.E. Wald p-value
1 BBB K trans (x 10
-3
min
-1
) 1.955 1.123 3.033 0.082
2 CSF Aβ 1-42 (pg/mL) -0.002 0.002 1.951 0.162
2 CSF pTau (pg/mL) 0.015 0.009 2.724 0.099

2.10f. Total sample Aβ 1-42 and pTau predicting Domain 1+ vs. 0 controlling for DG K trans.

Chi-square df p-value

Step 2 5.077 2 0.079

Model 11.903 3 0.008

Step Predictor B S.E. Wald p-value
1 BBB K trans (x 10
-3
min
-1
) 3.952 1.811 4.761 0.029
2 CSF Aβ 1-42 (pg/mL) -0.003 0.002 2.021 0.155
2 CSF pTau (pg/mL) 0.013 0.009 2.160 0.142
Abbreviations: BBB, blood-brain barrier; DG, dentate gyrus; CA1, cornu amonis 1; CA3, cornu
amonis 3; DCE, dynamic contrast-enhanced; MRI, magnetic resonance imaging; CSF,
cerebrospinal fluid; Aβ1-42, amyloid-β; pTau, phosphorylated tau.

101
CHAPTER 3:
BBB BREAKDOWN PREDICTS EARLY COGNITIVE DYSFUNCTION
IN APOE4 CARRIERS

Adapted from:
Sweeney MD*, Montagne A*, Nation DA*…Zlokovic BV, In Preparation
3.1 Abstract
Vascular dysfunction is increasingly recognized in the pathophysiology of Alzheimer’s disease
(AD), and measures of cerebrovascular dysfunction can be evaluated using cerebrospinal fluid
(CSF) and imaging-based biomarker approaches. A clinical need exists to identify reliable
biomarkers for early AD diagnosis, early intervention, and evaluating the efficacy of clinical trials.
Human participants were recruited from the University of Southern California (USC) Alzheimer’s
Disease Research Center (ADRC) and the Washington University Knight ADRC. Here, I
quantified novel CSF biomarkers of responses and injury to the neurovascular unit (NVU) –
comprising vascular cells, glia, and neurons – using immunoassays. CSF biomarkers of the NVU
were analyzed in relation to subjects’ cognitive status (cognitively normal and early cognitive
impairment) and the major genetic risk factor for sporadic AD, apolipoprotein E-ε4 (APOE4). I
found that a CSF-based biomarker of microvascular pericyte injury, namely soluble platelet-
derived growth factor receptor-β (sPDGFRβ), is altered in APOE4 carriers versus noncarriers and
also increased with cognitive impairment. Biomarkers of blood-brain barrier (BBB) permeability
and dysfunction including albumin quotient and cyclophilin A, respectively, are similarly
increased in APOE4 carriers during cognitive impairment. Glial, inflammatory and neuronal injury
biomarkers were not significantly altered with APOE4 or cognitive impairment. The BBB and
cerebrovascular injury markers increased independent of CSF amyloid-β (Aβ)1-42 and
phosphorylated tau levels. Consistently, in a subgroup of participants, we also found a regional
increase in BBB permeability in the (para)hippocampus measured by dynamic contrast-enhanced
magnetic resonance imaging (DCE-MRI) that increased independent of Aβ and tau. These suggest
that CSF and MRI-based biomarkers of BBB and/or pericyte dysfunction are early detectable
changes that are altered early in cognitive impairment and accelerated in APOE4 carriers. Pilot

102
data also indicate that high baseline levels of CSF sPDGFRβ predicts longitudinal decline in
cognitive function in APOE4 carriers, but not in APOE4 noncarriers. Altogether, BBB breakdown
predicts early cognitive dysfunction in APOE4 carriers independent of Ab and tau, and
cerebrovascular biomarkers may be a useful predictor of subtle cognitive dysfunction.
3.2 Introduction
Cerebrovascular dysfunction in the pathophysiology of cognitive impairment and
Alzheimer’s disease (AD) has been an increasing focus in the field with notable support in the last
decade
5,7,43,373,374,412,421,457–460
, which has been gained from neuroimaging
2,43,178,181,344,345,461
,
biofluid
43,145,181,182,461
, and neuropathological
146,183,184
studies (presented in Chapter 1 above). In
fact, the vascular contributions to AD were identified as a research priority by the United States
government in the National Alzheimer’s Project Act (NAPA) in 2011
411
and the NAPA update in
2013 also addressed the need to identify novel biomarkers related to pre-symptomatic stages of
AD
411
to aid in diagnosis, treatment and prevention efforts for AD and/or cognitive impairment.
Unfortunately, vascular contributions to AD were underemphasized in the recent ‘Research
Framework’ update on the biological definition of AD put forth in 2018 by the National Institute
on Aging and Alzheimer’s Association
366
. This illustrates there is a critical need for more studies
supporting that cerebrovascular dysfunction is an important, measurable contributor to cognitive
impairment and AD. We recently reported that individuals with early cognitive dysfunction
develop brain capillary damage and regional blood-brain barrier (BBB) breakdown irrespective of
Alzheimer’s amyloid-β (Aβ) and/or tau biomarker changes, supporting that BBB breakdown is an
early biomarker of human cognitive dysfunction independent of Aβ and tau
461
(Chapter 2). An
open question that will be studied here is how do genetic risk factors impact BBB breakdown and
pericyte injury during cognitive impairment?
Genetic evidence including recent genome-wide association studies shows that
apolipoprotein E-ε4 (APOE4) is the strongest and most highly replicated genetic risk factor for
late-onset sporadic AD
201,204,205,277
. Individuals with one copy of APOE4 have a 3.7-fold increase
in AD risk and individuals with two copies of APOE4 have a 12-fold increase in AD risk relative
to APOE3/3 individuals
74,205,277,278
. APOE exerts allele-dependent (e4>e3>e2) toxic effects on the
cerebrovascular system
207
including accelerated BBB breakdown and pericyte degeneration
146,291–

103
293
, cerebral blood flow (CBF) dysregulation
294–296
, impaired cerebrovascular reactivity
297,298
, CBF
reductions
296,462
, and increased risk for cerebral amyloid angiopathy (CAA)
147,205,245,278,463,464
,
which is associated with AD, stroke and hemorrhage
7,278,301–303
. The precise mechanisms
underlying the BBB breakdown and cerebrovascular dysfunction in human APOE4 carriers remain
elusive.
Animal studies have shown that APOE4 increases BBB susceptibility to injury
304
and leads
to BBB breakdown and microvascular reductions in humanized APOE4 transgenic mice
306,420
.
Specifically, astrocyte-secreted APOE2 and APOE3 maintain BBB integrity by suppressing a
proinflammatory signaling pathway in pericytes involving cyclophilin A (CypA) and matrix
metalloproteinase-9 (MMP9)
420
, whereas astrocyte-secreted APOE4 fails to effectively suppress
the CypA-MMP9 pathway in APOE4 transgenic mice leading to CypA-MMP9-mediated
degradation of BBB tight junction and basement membrane proteins causing BBB breakdown
420
.
Consistently, evidence of CypA-MMP9 pathway activation in human APOE4 carriers has been
supported by neuropathological AD studies
146
and biofluid studies in small cohort study of
cognitively normal individuals
145
. Further determining how APOE4 impacts cerebrovascular
injury biomarkers during early cognitive impairment and how these changes relate to AD classical
biomarkers Ab and tau changes is timely and important. Here, I show that brain capillary damage
and BBB breakdown are increased during early cognitive dysfunction in APOE4 carriers
independent of Ab and tau, supporting that cerebrovascular biomarkers may be a useful predictor
of subtle cognitive dysfunction.
3.3 Methods
3.3.1 Study Participants
Participants were recruited from two sites, including the University of Southern California
(USC), Los Angeles, CA, and Washington University, St. Louis, MO. At the USC site, participants
were recruited through the USC Alzheimer’s Disease Research Center (ADRC): combined USC
and the Huntington Medical Research Institutes (HMRI), Pasadena, CA. At the Washington
University site, participants were recruited through the Washington University Knight ADRC.
The study and procedures were approved by the Institutional Review Board of USC ADRC
and Washington University Knight ADRC indicating compliance with all ethical regulations, and

104
informed consent was obtained from all participants prior to study enrollment. All participants
underwent neurological and neuropsychological evaluations performed using the Uniform Data
Set (UDS) and additional neuropsychological tests, as described below. All participants received
a venipuncture and lumbar puncture (LP) for collection of blood and cerebrospinal fluid (CSF),
respectively, and a subset of participants from both sites underwent the dynamic contrast-enhanced
(DCE)-MRI for assessment of BBB permeability if they had no contraindications for contrast
injection or MRI, as described below.
This study included 318 participants for CSF biomarker studies and a subgroup of 153
participants underwent DCE-MRI. 63% of participants were from USC and 37% were from
Washington University. All biomarker assays and quantitative MRI scans were conducted by
investigators blinded to the clinical status of the participant.
3.3.2 Participant Inclusion and Exclusion Criteria
Included participants (≥45 years of age) with neuropsychologically-confirmed no cognitive
dysfunction and/or early cognitive dysfunction had no current or prior history of any neurological
or psychiatric conditions that might better account for any observed cognitive impairment,
including organ failure, brain tumors, epilepsy, hydrocephalus, schizophrenia, major depression.
Participants were excluded if they were diagnosed with vascular cognitive impairment or vascular
dementia. These clinical diagnoses were conducted by neurologists and the criteria whether the
patient 1) had a known vascular brain injury and 2) the clinician judged that the vascular brain
injury played a role in their cognitive impairment, and/or pattern and course of symptoms. In
addition to clinical diagnosis, presence of vascular lesions was confirmed by moderate-to-severe
white matter changes and lacunar infarcts by fluid-attenuated inversion recovery (FLAIR) MRI
and/or subcortical microbleeds by T2*-weighted MRI
363
. Participants were also excluded if they
were diagnosed with Parkinson’s disease, Lewy body dementia or frontotemporal dementia.
History of a single stroke or transient ischemic attack was not an exclusion unless it was related to
symptomatic onset of cognitive impairment. Participants also did not have current
contraindications to MRI and were not currently using medications that might better account for
any observed cognitive impairment. Lastly, participants carrying an APOE2 allele were excluded
since APOE2 elicits protection and this study was interested in examining the impact of the APOE4
risk allele.

105
3.3.3 Neuropsychological Testing
Clinical Dementia Rating (CDR)
CDR assessments followed the standardized UDS procedures. Participants underwent
clinical interview, including health history, and a physical exam. Knowledgeable informants were
also interviewed. Given the lack of scientific consensus regarding the categorization of older adults
along the aging-to-MCI-to-AD dementia spectrum and the time course and sequence of biomarker
abnormalities, I did not use clinical diagnosis in our biomarker comparisons but rather stratified
participants along objective neuropsychological metrics of cognitive impairment and biological
metrics of AD biomarker status using established cutoffs
424,425
. Participant CDR score was
obtained through standardized interview and assessment with the participant and a knowledgeable
informant.
Cognitive Domain Impairment Evaluation
Neuropsychological performance was used to identify domain impairment, as we
previously reported
461
. All participants underwent neuropsychological testing using the UDS
battery (version 2.0 or 3.0) plus supplemental neuropsychological tests at each site. Test
impairment for UDS tests was determined using age-, sex- and education-corrected scores from
the National Alzheimer’s Coordinating Center (NACC) (www.alz.washington.edu). Normalized
scores from a total of 10 neuropsychological tests were used in determining domain impairment,
including three tests per cognitive domain (memory, attention/executive function and language)
and one test of global cognition. Domain impairment was determined using previously described
neuropsychological criteria
426
, and was defined as a score >1 standard deviation (SD) below norm-
referenced values on two or more tests within a domain
427
. Multiple domain impairment (2+) was
assigned when more than one domain fit the impairment criteria, or three or more tests were
impaired across domains
426,427
. Prior studies have established improved sensitivity and specificity
of these criteria relative to those employing a single test score, as well as adaptability of this
diagnostic approach to various neuropsychological batteries
426–428
. Cognition was presumed
normal unless multiple impaired tests were identified as specified by the criteria. Individuals with
low Mini Mental State Exam (MMSE) or Montreal Cognitive Assessment (MOCA) scores (<25)
who had multiple missing neuropsychological test scores due to difficulty completing testing were
considered to have domain impairment.

106
Test battery specifics for each UDS version and recruitment site are as follows. i) Global
cognition: Mini Mental State Exam (MMSE) for UDS version 2 and Montreal Cognitive
Assessment (MOCA) for UDS version 3. ii) Memory: The Logical Memory Story A Immediate
and Delayed free recall tests [modified from the original Wechsler Memory Scales – Third Edition
(WMS-III)] for UDS version 2 and the Craft Stories Immediate and Delayed free recall for UDS
version 3. For supplemental tests the USC participants underwent the California Verbal Learning
Test – Second Edition (CVLT-II) and the Selective Reminding Test (SRT) sum of free recall trials.
Norm-referenced scores for these supplemental test scores were derived from a nationally
representative sample published with the test manual (CVLT-II)
429
and in studies of normally
aging adults (SRT). iii) Attention and executive function: The Trails A, Trails B, and Wechsler
Adult Intelligence Scale - Revised (WAIS-R) Digit Span Backwards tests for UDS version 2 and
the Trails A, Trails B and Digit Span Backwards tests for UDS version 3. iv) Language: The
Animal Fluency, Vegetable Fluency, and Boston Naming Tests for UDS version 2 and Animal
Fluency, Vegetable Fluency, and Multilingual Naming Test (MINT) for UDS version 3.
3.3.4 Lumbar Puncture and Venipuncture
Participants underwent a lumbar puncture and venipuncture in the morning after an
overnight fast. The CSF was collected in polypropylene tubes, processed (centrifuged at 2000 g,
10 minutes, 4°C), aliquoted into polypropylene tubes and stored at -80°C until assay. Blood was
collected into ethylenediaminetetraacetic acid (EDTA) tubes and processed (centrifuged at 2000
g, 10 minutes, 4°C). Plasma and buffy coat were aliquoted in polypropylene tubes and stored at -
80°C; buffy coat was used for DNA extraction and APOE genotyping.
3.3.5 APOE Genotyping
DNA was extracted from buffy coat using the Quick-gDNA Blood MiniPrep (Cat. No.
D3024, Zymo Research, Irvine, CA). APOE genotyping was performed using polymerase chain
reaction restriction fragment length polymorphism approach (PCR-RFLP). The DNA was
amplified in a 50 µL reaction with Qiagen reagents (Cat. Nos. 201203 and 201900, Qiagen). Two
primers were used to amplify a 318 base pair fragment: upstream sequence (5’
ACTGACCCCGGTGGCGGAGGAGACGCGTGC) and downstream sequence (5’
TGTTCCACCAGGGGCCCCAGGCGCTCGCGG). The upstream primer introduces an AflIII
site in the amplified product, yielding a unique RFLP pattern for each APOE allele following

107
enzymatic digestion. The PCR reaction mixture was incubated at 94°C for 3 min, then 40 cycles
of amplification (94°C, 10 sec; 65°C, 30 sec; 72°C, 30 sec), and finally elongation at 72°C for 7
min. Restriction digests containing 10 μl amplicons and either 2.5 U AflIII or 1.5 U HaeII were
incubated at 37°C overnight. The digested products were analyzed on a 4% agarose gel. APOE
genotype was determined from the unique digestion pattern: APOE2/2 [A: 231; H: 267], APOE2/3
[A: 231; H: 231 and 267], APOE2/4 [A: 231 and 295; H: 231 and 267], APOE3/3 [A: 231; H:
231], APOE3/4 [A: 231 and 295; H: 231], and APOE4/4 [A: 295; H: 231]; the brackets denote
base pairs of amplicons following the AflIII (A) and HaeII (H) digestions.
3.3.6 Molecular Biofluid Assays
Quantitative Western Blotting of sPDGFRβ
The quantitative Western blot analysis was used to detect sPDGFRβ in human CSF
(ng/mL), as we previously reported
178,461
. Standard curves were generated using recombinant
human PDGFRβ (Cat. No. 385-PR-100/CF, R&D Systems, Minneapolis, MN). For additional
details, see Section 2.2.8 Molecular Biofluid Assays: Quantitative Western Blotting of
sPDGFRβ above.
BBB Breakdown Markers
Albumin quotient (Qalb, the ratio of CSF-to-plasma albumin levels) was determined using
enzyme-linked immunosorbent assay (ELISA) (Cat. No., E-80AL, Immunology Consultants
Laboratory, Inc., Portland, OR). CSF levels of fibrinogen levels (Cat. No. E-80FIB, Immunology
Consultants Laboratory, Inc., Portland, OR) and CSF plasminogen levels (Cat. No. E-80PMG,
Immunology Consultants Laboratory, Inc., Portland, OR) were determined by ELISA.
Cyclophilin A (CypA)
I developed a novel CypA assay on the MSD platform. Standard-bind 96-well plates
(Catalog no. L15XA-3 / L11XA-3, MSD, Rockville, MD) were spot-coated with 5 µL per well of
40 µg/mL mouse monoclonal cyclophilin A antibody (Catalog no. ab58144, Abcam, Cambridge,
MA) prepared in 0.03% Triton X-100 in 0.01 M PBS pH 7.4 solution. The plates were left
undisturbed overnight to dry. The next day, the plates were blocked with 150 µL per well of
Blocking One (Catalog no. 03953-95, Nacalai Tesque, Japan) and incubated for exactly one hour

108
with shaking. Meanwhile, samples and standards were prepared in 1:20 diluted blocking buffer.
Two recombinant human cyclophilin A proteins were used with comparable results (Catalog no.
10436-H08E, Sino Biological, Wayne, PA; Catalog no. 3589-CAB, R&D Systems, Minneapolis,
MN). Eight standard points ranging from 0.27 – 200 ng/mL were prepared using serial dilutions.
CSF samples were diluted 1:2. After blocking, the plates were manually washed 3 times with 200
µL per well of wash buffer (in 0.05% Tween-20 in 0.01 M PBS pH 7.4). The prepared samples or
standards were added at 25 µL per well, and the plates were incubated overnight at 4°C with
shaking.
The next day, the plates were washed 3 times, and 25 µL per well of 1 µg/mL detection
antibody, rabbit polyclonal cyclophilin A antibody (Catalog no. 10436-T52, Sino Biological,
Wayne, PA), prepared in 1:20 diluted block solution was added to each well and the plates
incubated for one hour at room temperature with shaking. Next, plates were washed 3 times, and
25 µL per well of 1 µg/mL of anti-rabbit sulfo-tag antibody (Catalog no. R32AB, MSD, Rockville,
MD) in 1:20 block buffer was prepared and added to each well and the plates incubated for one
hour at room temperature with shaking. The plates were washed 3 times, then 150 µL per well of
2x Read Buffer T with surfactant (Catalog no. R92TC-3, MSD, Rockville, MD) and the plates
were read immediately on the MSD SECTOR Imager 6000 (MSD, Rockville, MD) with
electrochemiluminescence detection.
The raw readings were analyzed by subtracting the average background value of the zero
standard from each recombinant standard and sample readings. A standard curve was constructed
by plotting the recombinant standard readings and their known concentrations and applying a
nonlinear four-parameter logistics curve fit. The CypA concentrations were calculated using the
samples’ reading and the standard curve equation; the result was corrected for the sample dilution
factor to arrive at the CypA concentration in the CSF samples.
Astrocyte Marker
CSF levels of the astrocytic cytokine, S100 calcium-binding protein B (S100B), were
determined using ELISA (Cat. No. EZHS100B-33K, EMD Millipore, Billerica, MA).

109
Inflammatory Markers
Meso Scale Discovery (MSD) multiplex assay was used to determine CSF levels of
interleukin-2 (IL-2), IL-4, IL-6, IL-8, IL-10, IL-12p70, IL-13, IL-1β, tumor necrosis factor α
(TNF-α), and interferon γ (IFN-γ) (Cat. No. K15049G, MSD, Rockville, MD).
Amyloid-β Peptides
MSD multiplex assay (Cat. No. K15200E, MSD, Rockville, MD) was used to determine
CSF levels of Aβ38, Aβ40 and Aβ42. Participants were stratified based on CSF analysis as either
Ab1-42-positive (Ab1-42+, <190 pg/mL) or Ab1-42-negative (Ab-, >190 pg/mL) using the accepted
cutoff values as previously reported for the MSD 6E10 Aβ peptide assay
424
.
Tau
MSD assay was used to determine CSF levels of total tau (Cat. No. K15121G, MSD,
Rockville, MD). Phosphorylated tau (pT181) was determined by ELISA (Cat. No. 81581, Innotest,
Belgium). Participants were stratified based on CSF analysis as either pTau181-positive (pTau+,
>78 pg/mL) or pTau181-negative (pTau-, <78 pg/mL), using the accepted cutoff value as previously
reported
425
.
Neuronal Marker
CSF levels of neuron specific enolase (NSE) were determined using ELISA (Cat. No. E-
80NEN, Immunology Consultant Laboratories, Portland, OR). The company no longer sells this
product; thus, this analyte has been measured in a majority of participants but not in individuals
that most recently enrolled in the study.
3.3.7 Magnetic Resonance Imaging and Analysis
The imaging protocol performed was developed to detect subtle BBB changes in patients
with cognitive impairment and is detailed in Montagne et al. 2015
178
. Prior to imaging, all
participants underwent a blood draw to ensure appropriate kidney function for gadolinium contrast
agent administration. To quantify subtle BBB permeability to the gadolinium contrast agent, post-
processing analysis was performed using Rocketship
347
running with Matlab. Each individuals’
arterial input function was fitted with a bi-exponential function prior to fitting with Patlak
model
349
. The Patlak linearized regression mathematical analysis was used to generate the BBB

110
permeability Ktrans maps
178,347,349
. Regional BBB Ktrans permeability was measured in 13 different
gray matter regions of interest including the hippocampus (HC), HC subfields (i.e., dentate gyrus
(DG), CA3, CA1), parahippocampus (PHC), caudate nucleus, superior frontal cortical gyri,
inferior temporal cortical gyri, thalamus, and striatum, and white matter regions of interest
including subcortical white matter fibers, corpus callosum, and internal capsule. For additional
details, see Section 2.2.10 Neuroimaging and Analysis above.
3.3.8 Statistical Analyses
Prior to performing statistical models, statistical outliers were screened and removed for
all continuous variables according to the criteria +/- 3 SDs from the mean. Continuous variables
were also evaluated for departures from normality through quantitative examination of skewness
and kurtosis, in addition to visual inspection of frequency distributions. Where departures of
normality were identified, log10-transformations were applied, and distribution normalization was
confirmed prior to parametric analyses.
All CSF biomarkers and DCE-MRI Ktrans data were compared in parallel analyses applied
across the entire sample stratified by the global CDR score and the number of impaired cognitive
domains using analysis of covariance (ANCOVA) with posthoc Bonferroni corrected
comparisons. For CDR analyses, model covariates included age, sex, and education. Cognitive of
domain impairment was determined using age, sex, and education-corrected values, so these
covariates were not additionally included in the analyses. For analysis of interactions by CSF Aβ1-
42 and pTau status, statistical interactions and main effects were examined in similar ANCOVA
models. Given the large number of analyses, false discovery rate (FDR)-correction was applied to
all ANCOVA omnibus p-values using the Benjamini-Hochberg method
445
. Where significant
primary findings (e.g., CSF sPDGFRβ and BBB Ktrans) were identified in the CDR or cognitive
domain impairment analyses, separate posthoc analyses of CSF sPDGFRβ and BBB Ktrans
differences controlling for CSF Aβ1-42 and pTau were performed utilizing ANCOVA models.
With regard to missing data, all participants had complete data for primary outcomes, and
the extent of missing data was capped at < 10% for all other CSF biomarkers and clinical measures
(i.e., >90% of participants had complete data).

111
3.4 Results
I studied individuals with normal cognition or early cognitive impairment from two clinical
sites (USC and Washington University, see Section 3.3.1 Study Participants) and related
biomarker changes to participants’ cognitive status. Table 3.1 shows participants’ demographic
data by level of Clinical Dementia Rating (CDR) score and also reports the number of cognitive
domains impaired in each group. I excluded individuals that were diagnosed with vascular
dementia, vascular cognitive impairment, or other disorders that might account for cognitive
impairment (see Section 3.3.2 Participant Inclusion and Exclusion Criteria). To investigate the
impact of APOE4 genetic risk on cerebrovascular biomarkers during early stages of cognitive
impairment and as related to the classic AD biomarkers, Ab and tau, I measured ~40 biomarkers
in CSF reflecting different NVU cell and system-specific injury.

Table 3.1 Participants’ demographic information for the biofluid cohort.

APOE4 accelerates pericyte degeneration and BBB breakdown in post-mortem
studies
146,291–293
, thus I wondered whether the same was true for pericyte and BBB biomarkers
measured in living humans carrying an APOE4 allele. I measured CSF sPDGFRb in this study to
assess pericyte injury, since pericytes predominantly express PDGFRb and shed sPDGFRb upon
injury (see Chapter 2 and Appendix A). I found that CSF sPDGFRb is increased in APOE4
carriers with cognitive impairment regardless of how cognitive status was determined, including
both individuals with CDR 0.5 (Figure 3.1a) and 1+ cognitive domains impaired (Figure 3.1b)
compared to cognitively normal controls (CDR 0 and 0 domains impaired, respectively).
Additionally, the most traditional biofluid-based marker of BBB breakdown is the

112

Figure 3.1 Pericyte injury and BBB breakdown increase in APOE4 carriers with cognitive
impairment.
(a) CSF soluble platelet-derived growth factor receptor-β (sPDGFRβ) levels in individuals with
clinical dementia rating (CDR) score 0 that are APOE4 noncarriers (black, n=114) and APOE4
carriers (red, n=62) and CDR 0.5 that are APOE4 noncarriers (black, n=42) and APOE4 carriers
(red, n=38). (b) CSF sPDGFRβ levels in individuals with 0 cognitive domains impaired that are
APOE4 noncarriers (black, n=116) and APOE4 carriers (red, n=57) and 1+ cognitive domains
impaired that are APOE4 noncarriers (black, n=58) and APOE4 carriers (red, n=48). (c) Albumin
quotient (Qalb, CSF to plasma albumin levels) in individuals with CDR 0 that are APOE4
noncarriers (black, n=100) and APOE4 carriers (red, n=56) and CDR 0.5 that are APOE4
noncarriers (black, n=44) and APOE4 carriers (red, n=36). (d) Qalb in individuals with 0 cognitive
domains impaired that are APOE4 noncarriers (black, n=104) and APOE4 carriers (red, n=53) and
1+ cognitive domains impaired that are APOE4 noncarriers (black, n=58) and APOE4 carriers
(red, n=47). (e-g) Positive correlation between CSF sPDGFRβ and albumin quotient (Qalb, ratio
of CSF to plasma albumin levels) (n=105, e), CSF fibrinogen (n=105, f), and CSF plasminogen
(n=62, g) levels in APOE4 carriers indicating that brain pericyte injury is related to blood-brain
barrier (BBB) breakdown. Panels a-d: Box-and-whisker plot lines indicate median values, boxes
indicate interquartile range and whiskers indicate minimum and maximum values. Significance
tests after FDR correction from ANCOVAs with Bonferroni posthoc comparisons. Panels e-g:
Statistical significance by Pearson correlation, a=0.05; r = Pearson correlation coefficient.

113
albumin quotient (Qalb) referring to the ratio of CSF to plasma albumin levels, which is reported
to be increased with mild cognitive impairment and AD, and AD genetic risk
178,145,405–408,189
. I
found that Qalb is increased in APOE4 carriers between CDR 0 and CDR 0.5 and is further
increased in CDR 0.5 APOE4 carriers compared to noncarriers (Figure 3.1c). Interestingly, when
participants were grouped by cognitive domain impairment status, I now detect an even earlier
increase in Qalb in cognitively normal (0 domains impaired) APOE4 carriers compared to
noncarriers (Figure 3.1d), consistent with a previous report in a small cohort study
145
. In APOE4
carriers, CSF sPDGFRb exhibits a strong positive correlation with Qalb as well as other blood-
derived molecules, fibrinogen and plasminogen, measured in CSF reflecting BBB breakdown
(Figure 3.1e-f), supporting that microvascular pericyte injury is related to BBB breakdown in
APOE4 carriers.
Next, I analyzed other NVU biomarkers reflecting glial, inflammatory, and neuronal injury
to determine whether they were differentially altered by APOE4 genotype or cognitive impairment
status. The astrocytic cytokine S100 calcium-binding protein B (S100B), the inflammatory
cytokine interleukin-6 (IL6), and a neuronal injury marker neuron specific enolase (NSE) were not
altered in individuals when analyzed by APOE genotype and CDR status (Figure 3.2a-c) or
cognitive domain impairment status (Figure 3.2d-f). I also determined levels of standard AD
pathological biomarkers, CSF Ab1-42 and pTau. CSF Ab1-42 was not differentially altered with
APOE genotype or cognitive status (Figure 3.3a,d) nor was CSF pTau with CDR impairment
although CSF pTau did increase in APOE3 carriers with cognitive domain impairment (Figure
3.3b,e). CSF total tau did increase in APOE3 carriers across CDR impairment and in APOE3 and
APOE4 carriers with domain impairment (Figure 3.3c,f).

114

Figure 3.2 Other CSF biomarkers of NVU cell and system injury are not differentially
altered during cognitive impairment.
(a) CSF S100 calcium-binding protein B (S100B) levels in individuals with clinical dementia
rating (CDR) score 0 that are APOE4 noncarriers (black, n=87) and APOE4 carriers (red, n=42)
and CDR 0.5 that are APOE4 noncarriers (black, n=43) and APOE4 carriers (red, n=35). (b) CSF
interleukin-6 (IL-6) levels in individuals with CDR 0 that are APOE4 noncarriers (black, n=79)
and APOE4 carriers (red, n=47) and CDR 0.5 that are APOE4 noncarriers (black, n=37) and
APOE4 carriers (red, n=31). (c) CSF neuron specific enolase (NSE) levels in individuals with CDR
0 that are APOE4 noncarriers (black, n=52) and APOE4 carriers (red, n=31) and CDR 0.5 that are
APOE4 noncarriers (black, n=32) and APOE4 carriers (red, n=30). (d) CSF S100B levels in
individuals with 0 cognitive domains impaired that are APOE4 noncarriers (black, n=98) and
APOE4 carriers (red, n=42) and 1+ cognitive domains impaired that are APOE4 noncarriers (black,
n=57) and APOE4 carriers (red, n=42). (e) CSF IL-6 levels in individuals with 0 cognitive domains
impaired that are APOE4 noncarriers (black, n=85) and APOE4 carriers (red, n=44) and 1+
cognitive domains impaired that are APOE4 noncarriers (black, n=46) and APOE4 carriers (red,
n=41). (f) CSF NSE levels in individuals with 0 cognitive domains impaired that are APOE4
noncarriers (black, n=52) and APOE4 carriers (red, n=34) and in 1+ cognitive domains impaired
that are APOE4 noncarriers (black, n=49) and APOE4 carriers (red, n=36). Box-and-whisker plot
lines indicate median values, boxes indicate interquartile range and whiskers indicate minimum
and maximum values. Significance tests after FDR correction from ANCOVAs with Bonferroni
posthoc comparisons. NS = non-significant.

115

Figure 3.3 CSF Aβ1-42 and pTau in APOE carriers during cognitive impairment.
(a) CSF Aβ1-42 levels in individuals with clinical dementia rating (CDR) score 0 that are APOE4
noncarriers (black, n=119) and APOE4 carriers (red, n=65) and CDR 0.5 that are APOE4
noncarriers (black, n=44) and APOE4 carriers (red, n=38). (b) CSF phosphorylated tau (pTau)
levels in individuals with CDR 0 that are APOE4 noncarriers (black, n=119) and APOE4 carriers
(red, n=65) and CDR 0.5 that are APOE4 noncarriers (black, n=44) and APOE4 carriers (red,
n=40). (c) CSF total tau levels in individuals with CDR 0 that are APOE4 noncarriers (black,
n=120) and APOE4 carriers (red, n=64) and CDR 0.5 that are APOE4 noncarriers (black, n=44)
and APOE4 carriers (red, n=39). (d) CSF Aβ1-42 levels in individuals with 0 cognitive domains
impaired that are APOE4 noncarriers (black, n=121) and APOE4 carriers (red, n=60) and 1+
cognitive domains impaired that are APOE4 noncarriers (black, n=57) and APOE4 carriers (red,
n=48). (e) CSF pTau levels in individuals with 0 cognitive domains impaired that are APOE4
noncarriers (black, n=120) and APOE4 carriers (red, n=60) and 1+ cognitive domains impaired
that are APOE4 noncarriers (black, n=58) and APOE4 carriers (red, n=49). (f) CSF total tau levels
in individuals with 0 cognitive domains impaired that are APOE4 noncarriers (black, n=121) and
APOE4 carriers (red, n=61) and in 1+ cognitive domains impaired that are APOE4 noncarriers
(black, n=57) and APOE4 carriers (red, n=47). Box-and-whisker plot lines indicate median values,
boxes indicate interquartile range and whiskers indicate minimum and maximum values.
Significance tests after FDR correction from ANCOVAs with Bonferroni posthoc comparisons.
NS = non-significant.

Since APOE4 is reported to modulate both vascular function
420
and Aβ accumulation
425
, I
then asked whether Ab and tau were mediating the increase observed in markers of pericyte injury
and BBB breakdown. Participants were stratified as either Ab-positive (Ab1-42+, <190 pg/mL) or
Ab-negative (Ab1-42-, >190 pg/mL), and pTau-positive (pTau+, >78 pg/mL) or pTau-negative
(pTau-, <78 pg/mL) determined by CSF analysis and applied using accepted cutoff values
424–426
.

116
CSF sPDGFRβ was increased in APOE4 carriers with CDR 0.5 relative to CDR 0 regardless of
CSF Aβ1-42 (Figure 3.4a) or pTau (Figure 3.4b) status. Moreover, upon controlling for CSF Aβ1-
42 and pTau, I still observed an increase in CSF sPDGFRβ in APOE4 carriers between CDR 0 and
0.5 (Figure 3.4c). Similarly, with cognitive domain impairment classification, I also found that
CSF sPDGFRβ was increased in APOE4 carriers with domain impairment

Figure 3.4 CSF sPDGFRβ increases in APOE4 carriers with cognitive impairment,
independent of Aβ and tau.
(a) CSF sPDGFRβ in APOE4 individuals with clinical dementia rating (CDR) score 0 that are CSF
Aβ1-42 negative (Aβ-; n=38) or positive (Aβ+; n=24), and with CDR 0.5 that are Aβ- (n=17) or
Aβ+ (n=19). (b) CSF sPDGFRβ in APOE4 individuals with CDR 0 that are CSF pTau negative
(pTau-; n=48) or positive (pTau+; n=14) and with CDR 0.5 that are pTau- (n=22) or pTau+ (n=16).
(c) CSF sPDGFRβ controlled for CSF Aβ1-42 and pTau levels in individuals with CDR 0 that are
APOE4 noncarriers (black, n=104) and APOE4 carriers (red, n=57) and CDR 0.5 that are APOE4
noncarriers (black, n=42) and APOE4 carriers (red, n=36). (d) CSF sPDGFRβ in APOE4
individuals with 0 domains impaired that are CSF Aβ- (n=32) or Aβ+ (n=24), and with 1+ domains
impaired that are Aβ- (n=21) or Aβ+ (n=26). (e) CSF sPDGFRβ in APOE4 individuals with 0
domains impaired that are CSF pTau- (n=41) or pTau+ (n=15) and with 1+ domains impaired that
are pTau- (n=28) or pTau+ (n=20). (f) CSF sPDGFRβ controlled for CSF Aβ1-42 and pTau levels
in individuals with 0 domains impaired that are APOE4 noncarriers (black, n=115) and APOE4
carriers (red, n=55) and 1+ cognitive domains impaired that are APOE4 noncarriers (black, n=57)
and APOE4 carriers (red, n=47). Panels a,b and d,e: Box-and-whisker plot lines indicate median
values, boxes indicate interquartile range and whiskers indicate minimum and maximum values.
Panels c and f: Estimated marginal means ± SEM from ANCOVA models. In all analyses,
significance tests after FDR correction from ANCOVAs with Bonferroni posthoc comparisons.

117
regardless of CSF Aβ1-42 (Figure 3.4d) or pTau (Figure 3.4e) status, and also when controlling
for CSF Aβ1-42 and pTau (Figure 3.4f). As for CSF sPDGFRβ, within each cognitive group there
was no difference in Qalb levels between CSF Aβ1-42 or pTau status, yet Qalb increased with CDR
impairment in APOE4 carriers (Figure 3.5a,b). Increased Qalb with APOE4 and CDR impairment
is also supported by analyses controlling for CSF Aβ1-42 and pTau (Figure 3.5c) confirming BBB
breakdown with CDR in APOE4 carriers (Figure 3.1c). Consistently, Qalb levels did not differ
between CSF Aβ1-42 or pTau status in APOE4 carriers in each cognitive domain impairment group
(Figure 3.5d,e), and controlling for CSF Aβ1-42 and pTau confirmed Qalb increases even in
cognitively normal APOE4 carriers that is further increased with domain impairment (Figure
3.5f).
Furthermore, to examine the regional effect of BBB breakdown, I analyzed regional Ktrans
BBB permeability values that were determined by a colleague in our group (see Section 3.7
Acknowledgements). DCE-MRI analysis was performed in a subset of participants from both
clinical sites (Appendix Table C.1). Since the hippocampus (HC) and parahippocampus (PHC)
are critical regions exhibiting increased BBB permeability in living humans with early cognitive
impairment
43,178,461
, we focused on these regions to determine the impact of APOE4. BBB Ktrans
permeability in the HC revealed a stepwise increase with APOE4 and cognitive impairment status
(CDR and domain impairment) (Appendix Figure C.1a). Upon controlling for CSF Aβ1-42 and
pTau, we similarly observed that BBB Ktrans permeability increased in the HC in individuals with
normal cognition (CDR 0; 0 domains impaired) that were APOE4 carriers vs. noncarriers, and also
increased in individuals with early cognitive impairment (CDR 0.5; 1+ domains impaired) that
were APOE4 carriers vs. noncarriers (Appendix Figure C.1b). In the PHC, we did not detect a
change in BBB permeability in cognitively normal individuals with different APOE genotypes,
but instead found an increase in APOE4 carriers between CDR 0 vs. 0.5 and between 0 vs. 1+
domains impaired (Appendix Figure C.1c) that remained significant when additionally
controlling for CSF Aβ1-42 and pTau (Appendix Figure C.1d). The HC BBB permeability data
corroborate the early increase observed in Qalb in cognitively normal APOE4 carriers with
increased genetic risk for AD.

118

Figure 3.5 Qalb increases in APOE4 carriers with cognitive impairment, independent of Aβ
and tau.
(a) Albumin quotient (Qalb, CSF to plasma albumin levels) in APOE4 individuals with clinical
dementia rating (CDR) score 0 that are CSF Aβ1-42 negative (Aβ-; n=32) or positive (Aβ+; n=24),
and with CDR 0.5 that are Aβ- (n=16) or Aβ+ (n=20). (b) Qalb in APOE4 individuals with CDR
0 that are CSF pTau negative (pTau-; n=41) or positive (pTau+; n=15) and with CDR 0.5 that are
pTau- (n=19) or pTau+ (n=17). (c) Qalb controlled for CSF Aβ1-42 and pTau levels in individuals
with CDR 0 that are APOE4 noncarriers (black, n=96) and APOE4 carriers (red, n=55) and CDR
0.5 that are APOE4 noncarriers (black, n=43) and APOE4 carriers (red, n=36). (d) Qalb in APOE4
individuals with 0 domains impaired that are CSF Aβ- (n=27) or Aβ+ (n=26), and with 1+ domains
impaired that are Aβ- (n=21) or Aβ+ (n=26). (e) Qalb in APOE4 individuals with 0 domains
impaired that are CSF pTau- (n=35) or pTau+ (n=17) and with 1+ domains impaired that are pTau-
(n=27) or pTau+ (n=20). (f) Qalb controlled for CSF Aβ1-42 and pTau levels in individuals with 0
domains impaired that are APOE4 noncarriers (black, n=103) and APOE4 carriers (red, n=52) and
1+ cognitive domains impaired that are APOE4 noncarriers (black, n=57) and APOE4 carriers
(red, n=47). Panels a,b and d,e: Box-and-whisker plot lines indicate median values, boxes indicate
interquartile range and whiskers indicate minimum and maximum values. Panels c and f:
Estimated marginal means ± SEM from ANCOVA models. In all analyses, significance tests after
FDR correction from ANCOVAs with Bonferroni posthoc comparisons.
Altogether these data support an early increase in pericyte injury and BBB breakdown in
APOE4 carriers with cognitive dysfunction, independent of Aβ and tau. Next, I set out to gain
mechanistic insight based on evidence in animal models
420
, human pathological tissue studies
146

and a small cohort of cognitively normal APOE4 carriers
145
reporting APOE4-mediated activation
of the proinflammatory CypA-MMP9 pathway. Currently, there is only one commercially

119
available CypA ELISA which I tested and found its performance to be subpar and unreliable.
Therefore, I developed and optimized a novel, ultrasensitive CypA assay on the MSD
electrochemiluminescence platform in order to address this mechanistic question. I validated the
new assay with dilution linearity tests and four different CypA recombinant proteins (all were
detectable and produced comparable results), but surprisingly my new CypA assay was unable to
detect the CypA recombinant standard provided in the commercially available ELISA which
further questions the legitimacy and usefulness of that product. Nevertheless, using my new assay,
I found that CSF cyclophilin A increases in APOE4 carriers with cognitive impairment (CDR 0
vs. CDR 0.5) (Figure 3.6a), and also in APOE4 carriers compared to noncarriers that have normal
cognition (0 domains impaired) (Figure 3.6b). These data support that the APOE4 and CypA
pathway reported in pericytes mediates BBB breakdown
420
in human APOE4 carriers.

Figure 3.6 CSF cyclophilin A increases in APOE4 carriers and with cognitive impairment.
(a) CSF cyclophilin A (CypA) levels in individuals with clinical dementia rating (CDR) score 0
that are APOE4 noncarriers (black, n=119) and APOE4 carriers (red, n=67) and CDR 0.5 that are
APOE4 noncarriers (black, n=30) and APOE4 carriers (red, n=22). (b) CSF CypA levels in
individuals with 0 cognitive domains impaired that are APOE4 noncarriers (black, n=119) and
APOE4 carriers (red, n=50) and 1+ cognitive domains impaired that are APOE4 noncarriers (black,
n=20) and APOE4 carriers (red, n=28). Box-and-whisker plot lines indicate median values, boxes
indicate interquartile range and whiskers indicate minimum and maximum values. Significance
tests after FDR correction from ANCOVAs with Bonferroni posthoc comparisons.

3.5 Discussion
In summary, I found that biomarkers of pericyte injury, BBB breakdown and CypA
pathway activation are altered in APOE4 carriers with normal cognition and/or early stages of
cognitive impairment. I also report that the observed pericyte and BBB dysfunction occurred
independent of Aβ and tau and precede detectable changes in other NVU biomarkers reflecting

120
astrocyte, inflammatory, neuronal, and AD-related injury.
Studies in transgenic murine models with targeted-replacement human APOE alleles have
shown that the CypA-MMP9 pathway in APOE4 transgenic mice leading to CypA-MMP9-
mediated degradation of BBB tight junction and basement membrane proteins causing BBB
breakdown
420
. Evidence of CypA-MMP9 pathway activation in human APOE4 carriers has been
supported by neuropathological AD studies
146
and biofluid studies in small cohort study of
cognitively normal individuals
145
. In this study I have confirmed in a larger cohort that CSF CypA,
measured with my novel ultrasensitive MSD assay, is increased in APOE4 carriers with cognitive
impairment. I also have pilot data showing that MMP9 activity is increased in APOE4 carriers that
is further increased during cognitive impairment (data not shown), and I am further exploring these
findings. CypA is currently being therapeutically targeted in humanized APOE4, AD transgenic
animal models with a Food and Drug Administration (FDA)-approved CypA inhibitor to
determine whether it can attenuate AD pathology and neuronal injury (Zlokovic laboratory,
unpublished). If translated to humans, pericyte injury and BBB breakdown biomarkers would be
useful determinants of the therapeutic efficacy.
Cognitively normal individuals with abnormal CSF Aβ1-42 and pTau levels develop
cognitive impairment faster than those with normal levels; however, many individuals with
abnormal AD-injury biomarkers remain cognitively normal for up to 7 years emphasizing that
these markers have poor sensitivity and discrimination to identify changes associated with
cognition
425
. Therefore, with the classical AD biomarkers alone, there is an inability to accurately
detect the onset of cognitive impairment and detect AD during preclinical stages. This supports a
pronounced need to identify molecular biomarkers associated with early cognitive impairment.
This study indicates that early pericyte injury and BBB breakdown are important contributors to
cognitive impairment in APOE4 carriers, independent of Aβ and tau. Routinely incorporating
biomarkers of cerebrovascular and BBB dysfunction will support efforts to generate an algorithm
of molecular and imaging changes during cognitive impairment and AD in individuals with
different genetic risk for AD, which will importantly aid in identifying novel treatment targets,
early diagnosis, early intervention, and evaluating the efficacy of clinical trials.

121
3.6 Current and Future Directions
We are currently following individuals longitudinally to assess how cerebrovascular
dysfunction changes over time as related to cognitive impairment, brain connectivity, and the
classic Aβ and tau pathological markers of AD. Participants in this study receive an annual
neuropsychological exam and bi- or triennial CSF collection and MRI. Thus far, a subgroup of the
APOE study cohort presented above in Chapter 3 has received several years of longitudinal
neuropsychological testing. In collaboration with investigators in our group (see Section 3.7
Acknowledgements), we have pilot data showing that CSF sPDGFRβ significantly predicts
subsequent cognitive and functional decline. Specifically, individuals with high CSF sPDGFRβ
levels at baseline experienced cognitive decline at later follow-up evaluations (the 1
st
time point
was modeled as 0.5-2.5 years after baseline; the 2
nd
time point was modeled as 2.5-4.5 years after
baseline), whereas individuals with low CSF sPDGFRβ at baseline did not experience cognitive
decline (Figure 3.7a). We next tested whether APOE genotype differentially impacted the
predictive potential of CSF sPDGFRβ on future cognitive function. Intriguingly, APOE4

Figure 3.7 High CSF sPDGFRβ at baseline predicts subsequent cognitive decline in APOE4
carriers.
(a) Linear mixed model analysis of all participants (CDR 0-0.5) age-, sex- and education-corrected
decline on mental status exams (MMSE or MOCA) indicated that higher CSF sPDGFRβ predicted
more rapid decline in global cognitive function over 2 follow-up intervals, p=0.01. (b,c) Posthoc
analysis by APOE4 status indicated no significant effect of CSF sPDGFRβ on cognitive decline
among APOE4- participants, p=0.50 (b) and a significant effect of CSF sPDGFRβ on cognitive
decline among APOE4+ participants was observed, p=0.02 (c). For graphical depiction, separate
lines indicate median split of CSF sPDGFRβ (black, sPDGFRβ- indicate below median; blue, CSF
sPDGFRβ+ indicates above median), but all models treated CSF sPDGFRβ as a continuous
predictor. Time was modeled as t0 = -1 to 0.5 years post-LP; t1 = > 0.5 to 2.5 years post-LP; t2 =
> 2.5 to 4.5 years post-LP. Numbers in brackets indicate participant n for each subgroup x time
point.

122
noncarriers did not experience cognitive decline (Figure 3.7b), whereas APOE4 carriers with high
CSF sPDGFRβ levels at baseline experienced significant cognitive decline while the APOE4
carriers with low CSF sPDGFRβ levels at baseline experienced improved cognitive function at
follow-up evaluations (Figure 3.7c). This raises the question whether resilience of pericytes and
the cerebrovascular system can protect individuals from APOE4-mediated cognitive and
functional decline associated with AD.
3.7 Acknowledgements
I would like to acknowledge the contributions of several co-authors in this study. Specifically:
1) Dr. Daniel Nation (Assistant Professor in Psychology at USC) ran the longitudinal models in Figure
3.7 and taught me how to determine participants’ number of cognitive domains impaired. 2) Dr. Axel
Montagne (Assistant Professor of Research in Physiology & Neuroscience and Assistant Director of
the Functional Biological Imaging Core at USC) determined participants’ regional Ktrans BBB
permeability from the DCE-MRI scans. 3) Dr. Abhay Sagare (Assistant Professor of Research in
Physiology & Neuroscience at USC) with whom I worked to develop and optimize the CSF cyclophilin
A assay. 4) Maricarmen Pachicano for help with biofluid sample processing and aliquoting, APOE
genotyping, and data entry. 5) Everyone involved with recruiting and enrolling participants at USC
ADRC (including HMRI) and Washington University Knight ADRC.
This study was supported by the National Institutes of Health (NIH) grant nos. 5P01AG052350
(Zlokovic/Toga) and 5P50AG005142 (Chui) in addition to the Alzheimer’s Association strategic grant
no. 509279 (Zlokovic/Toga), Cure Alzheimer’s Fund, and the Foundation Leducq Transatlantic
Network of Excellence for the Study of Perivascular Spaces in Small Vessel Disease reference no. 16
CVD 05 (Zlokovic/Wardlaw). Enrollment of participants into the Washington University Knight
ADRC is supported by NIH grant nos. P50AG05681 (Morris), P01AG03991 (Morris), and
P01AG026276 (Morris). Enrollment of participants into the USC ADRC is supported by NIH grant
no. 5P50AG005142 (Chui).

123
CHAPTER 4:
BBB DYSFUNCTION IN AUTOSOMAL DOMINANT ALZHEIMER’S DISEASE

Adapted from:
Sweeney MD…Zlokovic BV, Ringman JM, In Preparation
4.1 Abstract
Background: Cerebrovascular contributions to cognitive impairment including sporadic
Alzheimer’s disease (AD) are increasingly recognized. It is currently elusive whether
cerebrovascular disruption also contributes to the pathophysiology of autosomal dominant AD
(ADAD) that has a known genetic etiology. Studying cerebrospinal fluid (CSF) in ADAD provides
the opportunity to investigate cerebrovascular and blood-brain barrier (BBB) changes in a
population with a relatively well understood pathogenic cascade.
Methods: We performed lumbar punctures on 44 individuals from families harboring ADAD
mutations in PSEN1 or APP, including 32 mutation carriers (MCs) and 12 noncarrier controls.
Immunoassays were used to measure BBB permeability/dysfunction and standard AD biomarkers.
Specifically, BBB permeability was assessed by the albumin quotient and CSF fibrinogen. BBB
dysfunction was determined by CSF soluble intercellular adhesion molecule-1 (sICAM1), soluble
vascular cell adhesion molecule-1 (sVCAM1), and placental growth factor (PlGF). CSF ferritin
reflected brain iron load. Standard AD biomarkers included amyloid-β (Aβ) peptides and tau.
Results: I found that Qalb and CSF levels of fibrinogen, sICAM1, sVCAM1, and ferritin were
significantly increased in symptomatic MCs vs. noncarriers, and CSF fibrinogen, sICAM1,
sVCAM1, PlGF and ferritin were also significantly higher in symptomatic MCs vs. asymptomatic
MCs, indicating BBB dysfunction in ADAD that is further disrupted in symptomatic individuals.
As expected, CSF Aβ42, total tau and phosphorylated tau (pTau) were altered in MCs vs.
noncarriers. Moreover, there was a significant association between participants’ adjusted age with
Qalb and CSF fibrinogen, sICAM1, Aβ42 and pTau levels in MCs, but not in noncarriers,
suggesting that cerebrovascular and standard AD biomarkers both inform ADAD
pathophysiological progression.

124
Conclusions: These data indicate the presence of cerebrovascular dysfunction in ADAD that is
further disrupted in symptomatic vs. asymptomatic MCs. How the cerebrovasculature interacts
with Aβ and tau to promote disease progression is currently unknown. Nevertheless,
cerebrovascular biomarkers may prove useful to evaluate cerebrovascular function in treatment
trials and also to provide insight regarding new therapeutic targets.
4.2 Introduction
It is increasingly recognized that cerebrovascular dysfunction contributes to cognitive
impairment including sporadic Alzheimer’s disease (AD)
43,178,181,182,373,374,461
, as evidenced by
neuroimaging
178,181,461
, biofluid
43,178,182,461,465
, and neuropathological
146,183,184
studies. This raises
the question does cerebrovascular disruption also contribute to the pathophysiology of autosomal
dominant AD (ADAD) that has a known genetic etiology?
ADAD is relatively rare, accounting for ~1% of all AD cases, and exhibits an early age of
onset (<65 years of age)
202
. ADAD is caused by mutations in presenilin-1 (PSEN1), PSEN2, and
amyloid precursor protein (APP) genes
197–201
. PSEN1 is the catalytical component of g-
secretase
200
, and PSEN1 mutations alter g-secretase activity to increase the release of long amyloid-
β (Aβ) peptide species (i.e., Aβ42, Aβ43, and longer)
251–254
. PSEN1 mutations lead to faster
soluble-to-fibrillar conversion of Aβ42 promoting amyloid deposition in the brain
199,201,250
. APP
NH2- and COOH-terminal mutations affect β-secretase and g-secretase processing activities of
APP, respectively, leading to aberrant and increased Aβ production
199–201,218
.
Neuropathological studies in brains of individuals with different PSEN1 and APP
mutations report cerebrovascular dysfunction and blood-brain barrier (BBB) breakdown, in
addition to the hallmark pathology of Aβ plaques and tau tangles. Human PSEN1 mutation carriers
(MC) exhibit cerebellar and cerebral amyloid angiopathy (CAA)
249,255–265
, cerebral perivascular
amyloid deposits
255,257,266,267,258,261
, disrupted meningeal, subpial, and cortical arterioles
249,258,264
,
and degeneration of the perivascular mural cells – pericytes at capillaries and vascular smooth
muscle cells at arterioles/arteries
257
. CAA is caused by Aβ deposition in the vascular wall of small
brain arteries and capillaries and develops from an imbalance between Aβ production and
clearance, particularly faulty trans- and peri-vascular Aβ clearance from brain-to-blood
2,9,155,233
.
CAA is a major cause of mural cell vascular degeneration that is associated with BBB breakdown,

125
lobar microbleeds, infarcts, white matter changes, and cognitive impairment worsening AD
pathology
154,233
. These neuropathological findings in ADAD, in addition to evidence that BBB
breakdown occurs during early stages of sporadic AD
43,178,181,182,373,374,461
, provide rationale to
study cerebrovascular and BBB changes in living humans with ADAD.
The estimated temporal sequence of ADAD biomarker changes was reported in a study of
the DIAN (Dominantly Inherited Alzheimer Network) cohort that focused on changes in
cerebrospinal fluid (CSF) Aβ, CSF tau, Aβ deposition, hippocampal volume, and glucose uptake
relative to the progression of cognitive impairment
268
. Whether cerebrovascular and BBB
dysfunction is detectable and when it changes in the pathophysiological sequence is currently
elusive. Here, I studied CSF from ADAD MCs and noncarrier control participants to examine BBB
breakdown and dysfunction and standard AD biomarkers (i.e., Aβ and tau). This study indicates
the presence of cerebrovascular dysfunction in ADAD that is further disrupted in symptomatic
versus asymptomatic MCs. Finally, BBB biomarkers may prove useful to evaluate cerebrovascular
function in treatment trials and also to provide insight regarding novel therapeutic targets.
4.3 Methods
4.3.1 Study Participants
Participants known to carry or with a 50% risk of inheriting a pathogenic ADAD mutation
were enrolled in ADAD studies at the University of California, Los Angeles (UCLA), and a subset
of participants simultaneously participated in the Dominantly Inherited Alzheimer Network
(DIAN, U19 AG032438, Principal Investigator, John Morris, M.D., http://www.dian-info.org). A
total of 44 participants are included in this study, including 32 ADAD MCs and 12 control
participants without an ADAD mutation. The age of symptomatic onset is variable between
families but is fairly consistent within a family and mutation
466
, thus each participant’s age relative
to their estimated age of symptomatic onset was calculated and is hereby referred to as the
participant’s ‘adjusted age’
466
.
The study procedures were approved by the UCLA Institutional Review Board.
Participants underwent a lumbar puncture and venipuncture, and comprehensive clinical and
cognitive assessments including the Mini-Mental State Examination (MMSE)
467
, Cognitive
Abilities Screening Instrument (CASI), and the Clinical Dementia Rating (CDR) scale
468

126
conducted by trained physicians blind to their mutation status. The CDR evaluation is a structured
interview of the participant and knowledgeable informant in which participants are rated on the 5-
point scale: 0 (asymptomatic), 0.5 (equivocal or mild impairment), 1 (mild dementia), 2 (moderate
dementia), or 3 (severe dementia). Participants’ cognitive status (i.e., asymptomatic or
symptomatic) was determined by the following criteria. ADAD MCs were categorized as
asymptomatic if participants were CDR 0 or if participants were CDR 0.5 with an adjusted age of
£ -10 years and CASI > 90. ADAD MCs were categorized as symptomatic if participants were
CDR 0.5+. All noncarriers were cognitively normal (CDR 0; CASI > 90), which is comparable to
the asymptomatic ADAD MCs.
4.3.2 Lumbar Puncture and Venipuncture
Participants underwent lumbar puncture (LP) using a 22- or 24-gauge Sprotte needle and
venipuncture in the morning after an overnight fast. CSF was collected in polypropylene tubes and
blood was collected in ethylenediaminetetraacetic acid (EDTA) tubes. CSF and blood samples
were processed by centrifuging at 1200 g for 15 minutes at 4°C. CSF and plasma were aliquoted
into polypropylene tubes and stored at -80°C until analysis. Buffy coat was used for DNA
extraction and genetic testing.
4.3.3 Genetic Testing
DNA extraction and apolipoprotein E (APOE) genotyping were performed using standard
techniques. APOE genotyping was conducted by quantitative real-time polymerase chain reaction
(qRT-PCR) on an Applied Biosystems 7900HT Fast Real-Time PCR System (Applied
Biosystems, Foster City, CA) using Taqman SNP Genotyping Assays (#C_3084793_20 and
#C_904973_10 for rs429358 and rs7412, respectively). The SDS version 2.3 software was used to
analyze the raw data and determine the APOE genotype. Additionally, to determine the ADAD
MC or noncarrier status, each participant’s amplicon encompassing their family’s known gene
mutation was assessed by Sanger sequencing.
4.3.4 Molecular Biofluid Assays
Biomarkers of BBB breakdown and dysfunction were measured using the following
immunoassays. Albumin quotient (Qalb), namely the ratio of CSF-to-plasma albumin levels, was
determined for each participant by enzyme-linked immunosorbent assay (ELISA) (catalog no. E-

127
80AL, Immunology Consultants Laboratory, Inc., Portland, OR). CSF fibrinogen was measured
by ELISA (catalog no. E-80FIB, Immunology Consultants Laboratory, Inc., Portland, OR). CSF
endothelial cell injury markers including soluble intercellular adhesion molecule-1 (sICAM-1) and
soluble vascular cell adhesion molecule-1 (sVCAM-1) were measured by the Vascular Injury
Panel 2 Human multiplex assay on the Meso Scale Discovery (MSD) platform (catalog no.
K15198D, MSD, Rockville, MD). CSF placental growth factor (PlGF) was measured by the
Angiogenesis Panel 1 Human multiplex assay (catalog no. K15190D, MSD, Rockville, MD). CSF
ferritin levels were determined by ELISA (catalog no. ab108837, Abcam, Cambridge, MA).
Standard AD biomarkers, Aβ and tau, were measured using the following immunoassays.
CSF levels of amyloid-β peptides (i.e., Aβ38, Aβ40, Aβ42) were measured using the Aβ Peptide
Panel 1 (6E10) multiplex assay (catalog no. K15200E, MSD, Rockville, MD). CSF total tau levels
were measured using an MSD assay (catalog no. K151LAE, MSD, Rockville, MD), and CSF
phosphorylated tau (181P) was determined by ELISA (catalog no. 81581, Innotest, Belgium).
4.3.5 Statistical Analyses
CSF biomarkers were analyzed between ADAD noncarriers, asymptomatic MCs, and
symptomatic MCs using analysis of variance (ANOVA) with α=0.05 followed by Tukey posthoc
test. Regression analyses with α=0.05 were performed to relate CSF biomarker levels to
participants’ adjusted age (see above Section 4.3.1 Study Participants for a description of
adjusted age) in ADAD MCs and noncarriers. Pearson correlation coefficient, r, is reported. Single
data points are plotted in the figures.
4.4 Results
This study included 44 participants comprised of 12 noncarriers, 18 asymptomatic ADAD
MCs, and 14 symptomatic ADAD MCs. Table 4.1 presents demographic and clinical data of
participants grouped by cognitive and mutation status, with the following parameters reported:
PSEN1 or APP mutation carrier status, mean age at LP, adjusted age, mean MMSE score, APOE
genotype, percent females, and percent with high vascular risk factor burden.

Table 4.1 Participants’ demographic and clinical information.

128

Abbreviations: ADAD, autosomal dominant Alzheimer's disease; APOE, apolipoprotein E; APP,
amyloid precursor protein; DM, diabetes mellitus; HL, hyperlipidemia; HTN, hypertension; LP,
lumbar puncture; MCs, mutation carriers; MMSE, Mini-Mental State Examination; PSEN1,
presenilin-1; VRF, vascular risk factors.
I found that CSF biomarkers of cerebrovascular dysfunction are increased in ADAD MCs.
Traditional BBB breakdown biomarkers including Qalb (albumin quotient, the ratio of CSF-to-
plasma albumin levels) and CSF fibrinogen are elevated in MCs compared to noncarriers (Figure
4.1a,b). Qalb and CSF fibrinogen exhibit a strong positive correlation in MCs and a weaker
correlation in noncarriers (Figure 4.1c), indicating pronounced BBB breakdown in ADAD.
Additionally, I also investigated the endothelial cell injury markers soluble intercellular adhesion
molecule-1 (sICAM-1) and soluble vascular cell adhesion molecule-1 (sVCAM-1). Consistent
with the BBB breakdown markers, both sICAM1 and sVCAM1 are increased in CSF of MCs
compared to noncarriers (Figure 4.1d,e), reflecting vascular endothelial cell injury in ADAD. CSF
placental growth factor (PlGF) was also increased in symptomatic vs. asymptomatic MCs (Figure
4.1f), suggestive of aberrant angiogenesis. This is consistent with recent reports that increased CSF
PlGF reflects cerebrovascular dysfunction in early stages of sporadic AD
465
. Lastly, I determined
CSF ferritin levels as an indicator of brain iron load
469
and found that CSF ferritin was increased
in MCs compared to noncarriers (Figure 4.1g). Increased brain iron load may result from
microbleeds, which is often caused by CAA that is apparent in ADAD
154,233
.

129

Figure 4.1 CSF biomarkers of cerebrovascular dysfunction are increased in ADAD.
(a-c) Blood-brain barrier (BBB) breakdown biomarkers including albumin quotient (Qalb, ratio of
cerebrospinal fluid (CSF)-to-plasma albumin levels, a) and CSF fibrinogen (b) in autosomal
dominant Alzheimer’s disease (ADAD) noncarriers, asymptomatic mutation carriers (MC) and
symptomatic MCs. Qalb and CSF fibrinogen exhibit a strong positive correlation in MCs and a
weaker correlation in noncarriers (c). (d,e) Endothelial cell injury markers including soluble
intercellular adhesion molecule-1 (sICAM-1) (d) and soluble vascular cell adhesion molecule-1
(sVCAM-1) (e) in noncarriers, asymptomatic MCs, and symptomatic MCs. (f) Placental growth
factor (PlGF) in noncarriers, asymptomatic MCs, and symptomatic MCs. (g) CSF ferritin levels
of brain iron load in noncarriers, asymptomatic MCs, and symptomatic MCs. All panels plot single
data points. In panels a-b and d-g, the box and whisker plots indicate the median value (horizontal
line), the boxes indicate the interquartile range, and the whiskers indicate the minimum and
maximum values; significance by ANOVA with Tukey posthoc test, α=0.05. In panel c, Pearson
correlation coefficient, r; significance by linear regression analysis, α=0.05.

In line with the genetic etiology of ADAD that promotes increased Aβ42 generation
251–254

and/or faster soluble-to-fibrillar conversion
199,201,250
, I found that MCs (both asymptomatic and
symptomatic) had decreased CSF Aβ42 compared to noncarriers, but no differences in CSF Aβ38
or Aβ40 were observed (Figure 4.2a-c). Also, in confirmation with other studies
268
, I similarly
observed increased levels of both CSF total tau and pTau in MCs vs. noncarriers (Figure 4.2d,e).

130

Figure 4.2 Standard CSF Alzheimer’s biomarkers, Aβ42 and tau, are altered in ADAD.
(a-c) Cerebrospinal fluid (CSF) levels of amyloid-β (Aβ) peptides including Aβ38 (a), Aβ40 (b)
and Aβ42 (c) in autosomal dominant Alzheimer’s disease (ADAD) noncarriers, asymptomatic
mutation carriers (MCs), and symptomatic MCs. (d,e) CSF levels of total tau (d) and
phosphorylated tau (pTau, e) in noncarriers, asymptomatic MCs, and symptomatic MCs. All
panels plot single data points. The box and whisker plots indicate the median value (horizontal
line), the boxes indicate the interquartile range, and the whiskers indicate the minimum and
maximum values; significance by ANOVA with Tukey posthoc test, α=0.05.
Next, I analyzed CSF biomarkers by ‘adjusted age’, which is each participant’s age relative
to the median age of dementia diagnosis in the probands’ family. Adjusted age indicates the
individual’s estimated number of years from symptomatic onset and is used as a surrogate
longitudinal measure of biomarker changes relative to ADAD progression. Regression analysis
indicated that cerebrovascular biomarkers including Qalb and CSF levels of fibrinogen and
sICAM1 (Figure 4.3a-c) are significantly increased relative to the adjusted age of MCs, but no
association was observed in the noncarriers. Similarly, consistent with previous reports
268
, I also
found that standard AD biomarkers including CSF Aβ42 and CSF pTau are decreased and
increased, respectively, relative to the adjusted age of MCs, but no association was observed in the

131
noncarriers (Figure 4.3d,e). Altogether, these data indicate that cerebrovascular and standard AD
biomarkers both inform ADAD pathophysiological progression.

Figure 4.3 Cerebrovascular and standard AD biomarkers inform ADAD pathophysiological
progression.
Cerebrospinal fluid (CSF) biomarkers plotted by ‘adjusted age’, which is each participant’s age
relative to the median age of dementia diagnosis in the probands’ family, indicating the
individual’s estimated number of years (y) from symptomatic onset. (a-c) Cerebrovascular
biomarkers including albumin quotient (Qalb, ratio of CSF-to-plasma albumin levels, a) and CSF
levels of fibrinogen (b) and soluble intercellular adhesion molecule-1 (sICAM1, c) are increased
relative to the adjusted age of autosomal dominant Alzheimer’s disease (ADAD) mutation carriers
(MC), but no association is observed in the noncarriers. (d,e) Standard AD biomarkers including
CSF Aβ42 (d) and pTau (e) are decreased and increased, respectively, relative to the adjusted age
of MCs, but no association is observed in the noncarriers. All panels plot single data points;
Pearson correlation coefficient, r; significance by linear regression analysis, α=0.05.
4.5 Discussion
In this study, I found BBB dysfunction in ADAD that is further disrupted in symptomatic
individuals. Specifically, Qalb and CSF fibrinogen, sICAM1, sVCAM1, and ferritin levels are
increased in symptomatic MCs vs. noncarriers, and CSF fibrinogen, sICAM1, sVCAM1, PlGF
and ferritin are higher in symptomatic MCs vs. asymptomatic MCs. Moreover, I also found a
significant association between participants’ adjusted age with Qalb and CSF fibrinogen, sICAM1,

132
Aβ42 and pTau levels in MCs, but not in noncarriers, suggesting that cerebrovascular and standard
AD biomarkers both inform ADAD pathophysiological progression. Cerebrovascular dysfunction
in ADAD should be additionally included in the temporal sequence of clinical and pathologic Aβ
and tau biomarkers that was previously reported
268
.
To date, 228 PSEN1 mutations
197,249,250,200,199,201
and 40 APP mutations
216
are identified to
cause ADAD. This study included MCs with the following mutations: APP V717I* (n=5) and
PSEN1 M146L* (n=1), H163R* (n=1), E184D* (n=1), G206A (n=4), S212Y (n=2), S230N (n=1),
L235V (n=2), T245P (n=1), A260V* (n=1), R269H (n=4), E280A* (n=1), and A431E (n=8).
Cerebrovascular pathology with BBB breakdown has been shown for several of these mutations
(denoted with asterisks in the preceding list)
218,255,261,263,264
. The overall clinical and pathological
features associated with specific mutations has been reported in multiple small-scale and single
case studies, but how the pathophysiology and disrupted underlying mechanisms manifest from
specific mutations is still unclear, particularly for the hundreds of PSEN1 mutations.
CAA is prominent in ADAD with moderate-to-severe CAA reported in 63.3% of ADAD
cases relative to 39.2% in sporadic AD cases
265
. Interestingly, the genetic location of the PSEN1
mutation influences CAA severity, with the most severe CAA corresponding to mutations after
codon 200
470
, although the mechanism is currently unclear. CAA of small arteries and capillaries
in the brain causes vascular degeneration and lobar microbleeds, which contribute to BBB
breakdown, infarcts, white matter changes and cognitive impairment. Microbleeds are red blood
cell-derived iron-containing perivascular hemosiderin deposits that are detected in ADAD MCs
471
.
Iron is stored in the cytoplasmic protein, ferritin, and ferritin levels in CSF are an index of brain
iron load
469
. It is therefore unsurprising that I observed increased CSF ferritin in symptomatic MCs
compared to asymptomatic MCs and noncarriers. Future studies should similarly investigate CSF
ferritin and relate it to microbleeds and CAA measured by neuroimaging in individuals with
ADAD.
The CSF measures of BBB breakdown and dysfunction I found in ADAD is similar to what
is observed in sporadic AD, as recently reviewed
43,182
. In addition to biofluids, neuroimaging
approaches are a valuable tool to investigate regional changes in cerebrovascular function. For
example, the dynamic contrast-enhanced magnetic resonance imaging sequence (DCE-MRI)
detects regional Ktrans BBB permeability to a gadolinium contrast agent. Recent studies have
reported increased Ktrans BBB permeability in the hippocampus during early stages of cognitive

133
impairment related to sporadic AD
178,461
. It would be valuable to use the DCE-MRI sequence in
ADAD and relate the regional Ktrans BBB permeability to CSF markers of BBB breakdown and
dysfunction. Given the existing evidence of cerebrovascular changes in ADAD and sporadic AD,
biofluid- and imaging-based biomarkers of cerebrovascular dysfunction should be routinely
incorporated, as recently recommended
373
.
In summary, my data suggest that cerebrovascular dysfunction is an important contributor
to ADAD that has a known genetic etiology. There is a great need for longitudinal studies
following the same individuals over time to better understand the temporal pathophysiological
sequences in ADAD. Moreover, future studies should also investigate the mechanisms linking
cerebrovascular dysfunction with Aβ and tau, and mechanistic insight gained from ADAD studies
is valuable to better understand the relationship between vascular, Aβ and tau in sporadic AD.
Biomarkers of BBB integrity and function should be employed to better define the disease process
in both ADAD and sporadic AD, to evaluate the efficacy of clinical trials, and to identify novel
therapeutic targets.
4.6 Acknowledgements
I would like to acknowledge and thank Dr. John Ringman (Professor of Clinical Neurology
at USC) who initiated participant recruitment, lumbar punctures, and clinical testing of the ADAD
cohort, and with whom I collaborated on this study. Participants were recruited at UCLA and
genetic testing for this study was conducted at UCLA. I conducted all biofluid assays and analysis
at USC in the Zlokovic laboratory.
This study was supported by the National Institutes of Health (NIH) grant nos.
5P01AG052350 (Zlokovic/Toga), 5P50AG005142 (Chui), and 5U01AG051218 (Ringman), in
addition to the Alzheimer’s Association strategic grant no. 509279 (Zlokovic/Toga), Cure
Alzheimer’s Fund, and the Foundation Leducq Transatlantic Network of Excellence for the Study
of Perivascular Spaces in Small Vessel Disease reference no. 16 CVD 05 (Zlokovic/Wardlaw).

134
CHAPTER 5:
A NOVEL ASSAY TO VALIDATE A BIOMARKER OF PERICYTE INJURY

Adapted from:
Sweeney MD*, Sagare AP*…Zlokovic BV, Curr Pharm Des, Invited – In Preparation
5.1 Abstract
Vascular contributions to cognitive impairment are increasingly recognized. Brain vascular mural
cells, especially pericytes, abundantly express platelet-derived growth factor receptor-β
(PDGFRβ) which is a cell surface signaling receptor that is important for maintaining proper
pericyte and microvascular functioning. Using quantitative Western blot analysis, we found that
soluble PDGFRβ (sPDGFRβ) is detectable in human and murine cerebrospinal fluid (CSF), and is
an early new biomarker of human cognitive dysfunction. Now, I developed and validated a
reproducible, ultrasensitive quantitative self-sandwich immunoassay to measure sPDGFRβ in
human CSF from 147 individuals with normal cognition or early cognitive impairment that should
be easily used at different clinical sites. I identified a combination of antibodies and standards
yielding a highly sensitive and reproducible assay with inter- and intra-assay coefficient of
variation <5%. The assay’s performance was validated with measures of spike recovery,
parallelism and dilutional linearity. Using this assay, I found that CSF sPDGFRβ levels are
significantly elevated in individuals with mild cognitive impairment compared to cognitively
normal individuals, suggesting brain microvascular pericyte injury. CSF sPDGFRβ also positively
correlates with traditional fluid biomarkers of blood-brain barrier (BBB) breakdown including
Qalb and CSF fibrinogen levels. CSF sPDGFRβ classification may help identify individuals that
are at increased risk of developing cognitive impairment. This new assay offers a reproducible,
sensitive approach to quantify sPDGFRβ in human CSF and further validate CSF sPDGFRβ as a
biomarker of brain microvascular pericyte injury and BBB dysfunction. This can further determine
its diagnostic and prognostic value in predicting early cognitive impairment in individuals with
genetic and/or other risk factors for AD and/or different neurodegenerative disorders associated
with vascular dysfunction. The overall results support my previous work (Chapters 2-3) showing
that CSF sPDGFRβ is a promising and early biomarker of human cognitive dysfunction.

135
5.2 Introduction
Proper functioning of the central nervous system (CNS) requires highly coordinated actions
of the neurovascular unit which comprises vascular cells, glia, and neurons
5,7,8,74
. Increasing
evidence supports that cerebrovascular dysfunction contributes to complex neurodegenerative
disorders including Alzheimer’s disease (AD)
2,7,43,74,179,181,373,374
. This is been shown by human
neuroimaging and biofluid studies during different stages of AD pathophysiology, and as well as
in neuropathological analysis of AD brains. Multiple studies of AD brains reveal blood-brain
barrier (BBB) breakdown with accumulation of several blood-derived proteins in brain
tissue
146,147,292,293,387,389,472
and degeneration of brain capillary pericytes
146,389,391,392
that is
accelerated by apolipoprotein E ε4 (APOE4)
145–147,291–293
, the major genetic risk factor for sporadic
AD.
Pericytes and vascular smooth muscle cells (SMCs) are vascular mural cells that tightly
associate with the endothelium of brain capillaries and arteries/arterioles, respectively
8,374,473
.
Mural cell recruitment to the developing CNS vasculature is crucial for vascular angioarchitecture
formation and stability, and this process is mediated via signaling events between endothelia-
secreted platelet-derived growth factor (PDGF)-BB and PDGF receptor-β (PDGFRβ) expressed
by mural cells
8,473,474
. PDGFRβ is highly expressed by both pericytes and SMCs during
development
473,474
, but PDGFRβ is predominately expressed by pericytes in the adult brain as
reported in human tissue
390
, human primary cells
423
(Appendix A) and rodent studies
18
.
Pericytes are centrally positioned at the neurovascular unit (NVU) and are particularly
vulnerable to injury and dysfunction that can disrupt BBB integrity and cerebral blood
flow
2,8,11,152,374,397
. Pericyte injury results in cleavage of soluble PDGFRβ (sPDGFRβ)
423
that is
detectable in human and murine cerebrospinal fluid (CSF)
178,461
. Furthermore, CSF sPDGFRβ
levels are increased in humans during early stages of cognitive impairment and positively correlate
with hippocampal BBB Ktrans permeability measured via dynamic contrast-enhanced magnetic
resonance imaging
178,461
. These studies support that BBB breakdown and pericyte injury measured
by CSF sPDGFRβ are early biomarkers of human cognitive dysfunction
178,461
.
Here, I developed an ultrasensitive novel assay to quantify sPDGFRβ in human CSF using
electrochemiluminescence detection on the Meso Scale Discovery (MSD) platform. I used a self-
sandwich immunoassay and identified and validated a combination of reagents, antibodies and

136
standard that yields high sensitivity and reproducibility with inter- and intra-assay coefficient of
variation <5%. Using this new assay, I report increased CSF sPDGFRβ levels in individuals with
early cognitive impairment, supporting our earlier studies
178,461
(Chapters 2-3). CSF sPDGFRβ is
also increased in cognitively normal human APOE4 carriers compared to noncarriers, suggesting
the presence of BBB dysfunction as previously reported in APOE4 carriers
145–147,291–293
. This assay
is the first to offer a reproducible approach to quantify sPDGFRβ, and these results support that
sPDGFRβ is a promising and sensitive early biomarker of cognitive deficits in humans. This study
provides important data of a groundwork assay for the field to investigate sPDGFRβ as a biomarker
of brain microvascular injury and BBB dysfunction in human biofluids and determine the
diagnostic utility for different neurodegenerative disorders.
5.3 Methods
5.3.1 Novel sPDGFRβ Assay
I used the following reagents: Standard-bind 96-well plates (MSD catalog no. L15XA-3 /
L11XA-3); human PDGFRβ polyclonal goat IgG against amino acids 33-530 (R&D Systems
catalog no. AF385); human PDGFRβ biotinylated antibody against amino acids 33-530 (R&D
Systems catalog no. BAF385); PDGFRβ recombinant human protein without catalytic activity
domain (Invitrogen catalog no. 10514H08H50); Blocker B (MSD catalog no. R93BB-2); Sulfo-
tag labeled streptavidin (MSD catalog no. R32AD); Read Buffer T with surfactant (MSD catalog
no. R92TC-3); adhesive seal (Bio-Rad catalog no. MSB1001).
I developed a new self-sandwich immunoassay to quantify sPDGFRβ levels in human CSF
using the MSD platform. First, standard-bind 96-well plates were coated with a capture antibody
against the extracellular domain of human PDGFRβ. Each well was spot-coated with 5 μL of 40
μg/mL of human PDGFRβ polyclonal goat IgG prepared in 0.01 M phosphate-buffered saline
(PBS) pH 7.4 + 0.03% Triton X-100. The plate was placed uncovered on a flat surface to allow
the spot coating solution to air dry overnight at room temperature. The plates were blocked with
150 μL per well of 1% Blocker B or an equivalent milk-based solution prepared in 0.01 M PBS
pH 7.4 + 0.05% Tween-20. The plate was sealed with an adhesive seal and incubated at room
temperature for 1 hour on an orbital plate shaker (~500 rpm). The plate was washed 3 times with
200 μL/well of wash buffer (0.01 M PBS pH 7.4 + 0.05% Tween-20) and tapped on an absorbent

137
pad to remove residual wash buffer. 0.2% Blocker B diluent was prepared in wash buffer and used
to dilute standards and samples (prepared on ice immediately prior to use). For the standard, I used
PDGFRβ recombinant human protein without catalytic activity domain at a stock concentration of
0.5 μg/μL. The following standard concentrations were prepared and used in the assay: 6400, 3200,
1600, 800, 400, 200, 100 pg/mL. The diluent was used as the zero standard. Standards were mixed
well by vortexing between each step. For human CSF samples, I prepared 1:2 dilutions in 0.2%
Blocker B diluent in polypropylene protein low-bind tubes. 25 μL of prepared standards or samples
were pipetted into pre-designated wells in duplicate. The plate was sealed and incubated at 4°C
overnight on an orbital plate shaker (~500 rpm). The plate was washed 3 times with 200 μL/well
of wash buffer and tapped on an absorbent pad to remove residual wash buffer. The detection
antibody solution was prepared by combining 1 μg/mL of human PDGFRβ biotinylated antibody
and 1 μg/mL of Sulfo-tag labeled streptavidin in 0.2% Blocker B diluent; prepared on ice
immediately prior to use. 25 μL of the detection antibody solution was pipetted into each well, and
the sealed plate was incubated at room temperate for 1.5 hours on an orbital plate shaker (~500
rpm). The plate was washed 3 times with 200 μL/well of wash buffer and tapped on an absorbent
pad to remove residual wash buffer. 2x Read Buffer T with surfactant was prepared in ddH2O, and
150 μL was pipetted into each well (careful to avoid bubbles). The plate was read immediately on
the MSD SECTOR Imager 6000 with electrochemiluminescence detection. The raw readings were
analyzed by subtracting the average background value of the zero standard from each recombinant
standard and sample readings. A standard curve was constructed by plotting the recombinant
standard readings and their known concentrations, and applying a linear curve fit. The sPDGFRβ
concentrations were calculated using the samples’ reading and the linear standard curve equation;
the result was corrected for the sample dilution factor to arrive at the sPDGFRβ concentration in
the original CSF samples.
5.3.2 Study Participants
Participants were recruited through the University of Southern California (USC)
Alzheimer’s Disease Research Center (ADRC) in Los Angeles, CA and the Washington University
Knight ADRC in St. Louis, MO. A total of 147 individuals are included in this study. The study
procedures were approved by the Institutional Review Boards of USC and Washington University.
Participants received a lumbar puncture and venipuncture, and were evaluated using the Uniform

138
Data Set (UDS)
475
and additional neuropsychological tests. Participants’ Clinical Dementia Rating
(CDR) score was obtained through standardized interview and assessment with the participant
following UDS procedures, and interview with a knowledgeable informant.
I excluded volunteers with i) dementia (CDR >1), head injury with loss of consciousness
>15 minutes, stroke, or substance abuse, or ii) current: organ failure, psychiatric or neurological
disorders that might produce dementia symptoms, hydrocephalus, B12 deficiency,
hypothyroidism, and medication use likely to affect brain function.
5.3.3 Lumbar Puncture and Venipuncture
Participants underwent a lumbar puncture and venipuncture in the morning following an
overnight fast. The CSF was collected in polypropylene tubes, processed (centrifuged at 2000 g,
10 minutes, 4°C), aliquoted into polypropylene tubes, and immediately stored at -80°C until assay.
Blood was collected into ethylenediaminetetraacetic acid (EDTA) tubes and processed
(centrifuged at 2000 g, 10 minutes, 4°C). Plasma and buffy coat were aliquoted in polypropylene
tubes and stored at -80°C; buffy coat was used for DNA extraction and APOE genotyping.
5.3.4 APOE Genotyping
DNA was extracted from buffy coat using the Quick-gDNA Blood Miniprep Kit (Zymo
Research catalog no. D3024). APOE genotyping was performed via polymerase chain reaction
(PCR)-based retention fragment length polymorphism analysis, as previously reported
461
. See
above Section 2.2.7 APOE Genotyping for complete methodological details.
5.3.5 Molecular Biofluid Assays
Albumin quotient (Qalb, ratio of CSF-to-plasma albumin levels) was determined using
enzyme-linked immunosorbent assay (ELISA) (Cat. No., E-80AL, Immunology Consultants
Laboratory, Inc., Portland, OR). CSF fibrinogen levels (Cat. No. E-80FIB, Immunology
Consultants Laboratory, Inc., Portland, OR) and CSF plasminogen levels (Cat. No. E-80PMG,
Immunology Consultants Laboratory, Inc., Portland, OR) were determined by ELISA.
5.3.6 Statistical Analysis
For comparison between two groups, statistical significance was analyzed by unpaired two-
tailed Student’s t-test. For multiple comparisons, one-way analysis of variance (ANOVA)

139
followed by Tukey’s posthoc test was used. Linear regression analysis was used to assess the
significance of correlations, and Pearson correlation coefficient was determined. P < 0.05 was
considered significant. Statistical analyses were conducted using SPSS or GraphPad Prism 7.0
software, and figures were prepared using GraphPad Prism and Adobe Illustrator. Single data
points are plotted in the figures.
5.4 Results
I developed a new assay to detect the soluble extracellular domain of PDGFRβ using
electrochemiluminescence detection on the MSD platform. A combination of reagents and
conditions were tested, optimized and validated. Table 5.1 summarizes the reagents tested (i.e.,
plate types, block solutions, recombinant standards, capture antibodies, and detection antibodies)
and identifies the combination of conditions that yielded optimal results (denoted with asterisks).
Two different recombinant standard proteins exhibited a large, dynamic linear curve fit ranging
from 100-26,000 pg/mL with a coefficient of linearity (r
2
) of 0.9996 and 0.996 (Figure 5.1a). To
validate the assay, I tested dilutional linearity, spike recovery, and parallelism
476
. There was
excellent sample recovery (mean CV 2.55%) of CSF samples diluted from 1:2-1:16 (Figure 5.1b),
Table 5.1 Summary of reagents used to develop and optimize the sPDGFRβ assay on the
MSD platform.
The red asterisks denote the reagent combination that yielded optimal results.

140

Figure 5.1 Performance summary of the novel sPDGFRβ assay.
(a) Representative standard curves plotting concentration and electrochemiluminescence signal of
two recombinant standard proteins that both exhibit a linear curve fit over a large dynamic range
from 100-26,000 pg/mL; r
2
= 0.996-0.9996. (b) Dilution linearity test – CSF samples diluted 1:4,
1:8 and 1:16 have a low coefficient of variation across all sample dilutions. (c) Parallelism test –
the electrochemiluminescence signal of samples and the recombinant standard protein across a
range of dilutions from 1:2 to 1:128 is parallel. (d) Summary of assay performance detailing the
assay’s lower limit of sensitivity (100 pg/mL), sample linearity range (1:2-1:16 dilution of CSF
samples), and assay reproducibility (intra- and inter-assay variability <5%).
indicating that the dilutions yielded consistent results within the desirable assay range. Next,
parallelism measures revealed parallel response curves of samples and the standard across the
dilution range (Figure 5.1c), demonstrating that the test sample dilution does not result in a biased
measurement of the analyte concentration. In summary, the new sPDGFRβ assay yields
exceptional sensitivity with a lower detection limit of 100 pg/mL, and the assay produces
remarkable precision and reproducibility with an average intra-assay coefficient of variability
(CV) of 4.71% and an average inter-assay CV of 4.60% (Figure 5.1d).
The new assay was used to evaluate sPDGFRβ levels in human CSF to test its clinical
relevance. Individuals with normal cognition (CDR 0), mild cognitive impairment (CDR 0.5), and
mild dementia (CDR 1) were included in the study. Table 5.2 presents demographic and clinical
data of participants grouped by cognitive status, with the following parameters reported: CDR
score, number of participants, mean age at LP, percent female, and percent APOE4 carriers.

141
Table 5.2 Participants’ demographic information.

Figure 5.2 Validation of sPDGFRβ as a pericyte injury biomarker in human CSF.
(a) CSF sPDGFRβ levels measured by quantitative Western blot are significantly increased with
CDR impairment, as we reported
461
. (b) CSF sPDGFRβ levels measured by the novel MSD assay
are significantly increased in individuals with CDR 0.5 (n=35) and CDR 1 (n=36) compared to
cognitively normal CDR 0 individuals (n=73). (c) CSF sPDGFRβ levels measured by quantitative
Western blot (from panel a) and the new MSD assay (from panel b) exhibit a positive correlation
(n=93). (d-f) CSF sPDGFRβ relates to blood-brain barrier breakdown as shown by positive
correlations with albumin quotient (Qalb) of CSF-to-plasma albumin levels (n=143) (d), CSF
fibrinogen (n=144) (e), and CSF plasminogen (n=121) (f). (g) CSF sPDGFRβ levels are
significantly increased in cognitively normal APOE4 carriers (n=27) compared to APOE4
noncarriers (n=42). All panels plot single data points. In panels a,b and g, the box and whisker
plots indicate the median value (horizontal line), the boxes indicate the interquartile range, and the
whiskers indicate the minimum and maximum values. Significance by ANCOVA with Bonferroni
posthoc comparisons, α=0.05. In panels c-f, Pearson correlation coefficient, r; significance by
linear regression analysis.

142
My earlier work determined CSF sPDGFRβ levels by quantitative Western blot analysis,
reporting a 35% increase in CSF sPDGFRβ with early cognitive impairment (CDR 0 vs. 0.5)
(Figure 5.2a)
178,461
(Chapters 2-3). Now, using the new MSD assay I developed, we confirm my
earlier findings
178,461
that CSF sPDGFRβ levels are significantly elevated by 20% in individuals
with mild cognitive impairment (CDR 0.5) compared to cognitively normal (CDR 0) individuals
(Figure 5.2b), indicating brain microvascular pericyte injury during early stages of cognitive
impairment as previously reported
178,461
. I measured sPDGFRβ levels in the same CSF samples by
both quantitative Western blot and my new assay revealing a positive correlation (Figure 5.2c) as
further validation of the new assay’s performance and relevance to existing literature
178,461
.
Pericyte injury and BBB breakdown are related events, as shown by positive correlations of CSF
sPDGFRβ with traditional biofluid markers of BBB breakdown including Qalb and CSF
fibrinogen and plasminogen levels (Figure 5.2d-f). Lastly, I found that CSF sPDGFRβ levels are
significantly elevated in cognitively normal APOE4 carriers compared to noncarriers (Figure
5.2g), consistent with literature reporting BBB damage in APOE4 carriers
145–147,291–293
. The
increase in CSF sPDGFRβ in APOE4 carriers reported above in Chapter 3 (measured by
quantitative Western blot) occurred between cognitively normal and mildly impaired individuals,
whereas here with my new assay that has increased sensitivity I see an increase in CSF sPDGFRβ
in APOE4 carriers vs. noncarriers with normal cognition (Figure 5.2g), consistent with my
findings for Qalb, CSF cyclophilin A, and hippocampal Ktrans BBB permeability (Chapter 3).
CSF sPDGFRβ classification may help identify individuals that are at increased risk of
developing cognitive impairment. The overall results support my previous work
178,461
showing that
CSF sPDGFRβ is a promising and early biomarker of human cognitive dysfunction.
5.5 Discussion
The novel sPDGFRβ assay I developed on the MSD platform offers a reproducible
approach to quantify sPDGFRβ in human CSF, and these results provide important support that
CSF sPDGFRβ is a promising and sensitive biomarker for identifying individuals that are at
increased risk of developing early cognitive impairment, consistent with my earlier work
178,461

(Chapters 2-3). Compared with previous methods to detect CSF sPDGFRβ by quantitative
Western blot
178,461
or commercially available ELISAs, my new assay on the MSD platform offers

143
many advantages. Specifically, the assay is ultrasensitive due to a large dynamic range of detection
that allows use of small CSF sample volume (~7 µL). This is an invaluable advantage for limited,
precious samples, since traditional ELISAs typically require an enormous volume (~50-100 µL)
of CSF sample per well for detection. The new assay is extremely reproducible and more high-
throughput than quantitative Western blot. Importantly, use of the MSD platform enables
multiplexing of sPDGFRβ with other analytes of interest which will further save sample volume,
investigator’s time, and cost of materials. Altogether, this novel, ultrasensitive assay is easy to
incorporate at different sites to investigate pericyte injury in various cohorts.
PDGFRβ is predominantly expressed by pericytes in the adult brain of humans
390,423,461
and
mice
18
, and sPDGFRβ is primarily shed by pericytes
178,423,461
(Appendix A). Thus, increased CSF
sPDGFRβ levels reflect brain microvascular damage due to pericyte injury. My new assay detects
the soluble extracellular portion of PDGFRβ, which has 5 immunoglobulin (Ig)-like domains
477
.
Ligands predominantly bind to Ig-like domains 2 and 3 causing receptor dimerization, and the
receptor dimer is further stabilized by direct receptor-receptor interactions of Ig-like domain 4
477–
479
. To date the 3-dimensional structure of PDGFRβ has not been resolved, nor have the precise
mechanism(s) of PDGFRβ ectodomain shedding from pericytes been elucidated. Recent evidence
indicates that a disintegrin and metalloproteinase (ADAM) family member, ADAM10, can
mediate sPDGFRβ shedding from pericytes but not SMCs
461
, consistent with another study
showing ADAM10 sheds sPDGFRβ in fibroblasts
435
. While ADAM10 plays a role in PDGFRβ
shedding from pericytes
461
, it is currently elusive whether ADAM17 or other enzymes are also
involved. Further, it is presently unknown whether the extracellular domain of PDGFRβ is
internalized or cleaved into the soluble form prior to receptor internalization. Elucidating the exact
mechanism(s) underlying ectodomain shedding of PDGFRβ in response to pericyte injury would
inform the degree to which sPDGFRβ is detectable as a result of pericyte dysfunction versus
degeneration, and also has the potential to identify novel therapeutic targets.
The major genetic risk factor for sporadic AD, APOE4, is associated with a 1.9% and
17.5% increased lifetime risk for developing AD in individuals age 65 and 75 years,
respectively
277
. I show that cognitively normal APOE4 carriers vs. noncarriers have increased CSF
sPDGFRβ levels, and CSF sPDGFRβ positively correlates with hippocampal Ktrans BBB
permeability
178,461
and biofluid markers of BBB breakdown as reported here and previously
shown
461
. These data support that elevated CSF sPDGFRβ is related to BBB damage. APOE4 is

144
mechanistically linked to BBB breakdown in murine models via a proinflammatory signaling
pathway involving cyclophilin A (CypA) and matrix metalloproteinase-9 (MMP9)
420
, which was
also shown in human biofluid
145
and AD brain tissue
146
studies. The degree to which additional
pathways and mechanisms contribute to APOE4-mediated BBB breakdown during AD
pathophysiology is currently elusive. Nevertheless, CypA is currently being therapeutically
targeted in humanized APOE4, AD transgenic animal models with a Food and Drug
Administration (FDA)-approved CypA inhibitor to determine whether it can attenuate AD
pathology and neuronal injury (Zlokovic laboratory, unpublished). If translated to humans, CSF
sPDGFRβ and other related biofluid- and imaging-based biomarkers of BBB breakdown would be
useful determinants of the therapeutic efficacy.
In light of the growing evidence that cerebrovascular dysfunction contributes to cognitive
impairment and dementia including AD
2,7,43,74,179,181,373,374
, I urge different sites to adopt and
employ this assay to evaluate sPDGFRβ in their cohorts. Altogether this study provides important
supporting data of a groundwork assay for the field to investigate sPDGFRβ as a biomarker of
brain microvascular injury and BBB dysfunction in human biofluids and determine the diagnostic
utility for cognitive impairment in different neurodegenerative disorders.
5.6 Acknowledgements
I would like to acknowledge Dr. Abhay Sagare (Assistant Professor of Research in
Physiology & Neuroscience at USC) with whom I worked to develop and optimize the new
sPDGFRβ assay. I would also like to thank everyone involved with recruiting and enrolling
participants at USC ADRC (including HMRI) and Washington University Knight ADRC.
This study was supported by the National Institutes of Health (NIH) grant nos.
5P01AG052350 (Zlokovic/Toga) and 5P50AG005142 (Chui) in addition to the Alzheimer’s
Association strategic grant no. 509279 (Zlokovic/Toga), Cure Alzheimer’s Fund, and the
Foundation Leducq Transatlantic Network of Excellence for the Study of Perivascular Spaces in
Small Vessel Disease reference no. 16 CVD 05 (Zlokovic/Wardlaw). Enrollment of participants
into the Washington University Knight ADRC is supported by NIH grant nos. P50AG05681
(Morris), P01AG03991 (Morris), and P01AG026276 (Morris). Enrollment of participants into the
USC ADRC is supported by NIH grant no. 5P50AG005142 (Chui).

145
CHAPTER 6:
UNDERSTANDING THE MOLECULAR MECHANISMS / SIGNATURES OF THE
CEREBROVASCULATURE USING A MOUSE MODEL OF HYPOXIA

Adapted from:
Sweeney MD…Zlokovic BV, In Preparation
6.1 Abstract
Chronic mild hypoxia is a phenomenon observed in several central nervous system (CNS)
conditions including diabetic retinopathy, chronic hypoperfusion, and Alzheimer’s disease (AD).
Pericytes, mural cells that enwrap capillaries, become dysfunctional and degenerate in ischemic
stroke, diabetic retinopathy and AD. Pericytes are vital orchestrators of key neurovascular
functions including blood-brain barrier (BBB) integrity, cerebral blood flow (CBF) regulation, and
angiogenesis. CBF and angiogenesis are physiological responses in brain elicited by a chronic mild
hypoxic state. The transcriptional regulation of these cerebrovascular structural and functional
signatures has not been thoroughly investigated in the adult brain, nor has its corresponding impact
on brain microvascular health. Adult mice exposed to chronic mild hypoxia up to 21 days show an
acute increase in capillary diameter, transient BBB permeability, and a progressive increase in
angiogenesis. Bulk RNA-sequencing analysis of brain microvascular cells during chronic mild
hypoxia reveals significant differentially expressed genes of pathways regulating angiogenesis
(differentiation, proliferation, migration) and BBB permeability, determined via Ingenuity
Pathway Analysis. Altogether, elucidating the transcriptional, structural and functional responses
to chronic mild hypoxia in the adult brain is vital to inform therapeutic efforts to combat hypoxic
insults and/or microvascular dysfunction in numerous CNS disorders.

146
6.2 Introduction
The blood-brain barrier (BBB) physically separates the blood and brain, and functions as
a key regulator of central nervous system (CNS) physiology by precisely sanctioning entry of
oxygen and energy substrates and preventing entry of macromolecules, cells, and pathogens
74
.
This is accomplished by the expression of a myriad of transporters, receptors, active efflux pumps,
ion channels, regulatory molecules, junctional molecules, etc. and interactive signaling amongst
BBB cell types
74
. The BBB is comprised of a continuous endothelial monolayer ensheathed with
perivascular mural cells (specifically pericytes at the capillary level)
6,8,74
. Pericytes are centrally
positioned at the neurovascular interface, and are vital orchestrators of key neurovascular functions
including BBB integrity, angioarchitecture, and cerebral blood flow (CBF) regulation
8,57,152,397
. As
discussed in Chapters 1-5 above, BBB and cerebrovascular dysfunction are increasingly observed
in neurodegenerative diseases including Alzheimer’s disease (AD). Yet, transcriptional expression
and signaling mechanisms underlying cerebrovascular responses have not been fully elucidated
and therefore the cerebrovasculature has not become a primary target to help combat neurological
disorders and dementia. Here, I turn to a mouse model of chronic mild hypoxia as a tool to better
understand molecular mechanisms and signatures of the cerebrovasculature.
The cerebrovascular network has a remarkable ability to sense oxygen availability, control
oxygen delivery, and physiologically adapt to maintain adequate tissue oxygen tension levels
480–
482
. In response to chronic mild hypoxia, the mammalian CNS exhibits cerebrovascular adaptations
(i.e., acute CBF increase
483,484
and remodeling of the microvascular network
483,485–489
) to preserve
tissue oxygen and energy supply needed to support optimal neuronal function. Angiogenesis is the
formation of new blood vessels from existing vessels, in contrast to vasculogenesis that is de novo
blood vessel formation. Angiogenesis and vasculogenesis are dynamic processes that underlie
formation of the complex CNS microvascular network during embryonic and early postnatal
periods, and pericytes are particularly vital for generating proper, functional
angioarchitecture
446,490–495,447,473
. In the adult brain, angiogenesis is a physiologic response that
requires highly coordinated spatial and temporal disassembly of the neurovascular unit, and
guidance of endothelial cells and pericytes
473,495–498
.
Chronic mild hypoxia, microvascular dysfunction and/or aberrant angiogenesis are
phenomena observed in several CNS conditions including diabetic retinopathy
499,500
, chronic

147
hypoperfusion
501–503
, and AD
395
. During chronic mild hypoxia in the adult brain, how do brain
microvascular transcriptional changes and cellular crosstalk regulate physiological functional
responses? This will be studied here, in addition to elucidating the precise contributions of
pericytes in mediating the cerebrovascular responses.
6.3 Methods
6.3.1 Prolonged Hypoxia Paradigm
Mice were exposed to as low as 8% O2 for up to 21 days (Day 1: 10% O2; Day 2: 9% O2;
Days 3-21: 8% O2) in a normobaric hypoxia chamber with hypoxic conditions maintained via N2
gas infusion into the chamber regulated by an O2 controller (ProOx 110, BioSpherix). Normoxic
control mice were kept in the same location at 21% O2, as previously reported
395,221,504,481
. Mice
were examined at experimental days 0, 1, 2, 3, 7, 14 and 21.
6.3.2 Animals
Adult mice (2-4 months old; males and females) were used in this study. C57Bl/6 mice
were used for immunohistochemistry analyses. Tie2-eGFP mice were used for in vivo multiphoton
microscopy. Tie2-Cre; Ai14; Pdgfrb-eGFP mice with fluorescently labeled endothelial cells (red)
and pericyte-enriched mural cells (green) were used for in vivo MRI and RNA-sequencing. All
mice were monitored daily to ensure the animals’ well-being during the experimental paradigm.
Mice were removed from the hypoxia chamber for animal husbandry (i.e., cage change and food
and water replacement) and daily body weight measurements. The length of time mice were
exposed to normoxia was minimized to <5 min per day to ensure the hypoxic conditions were
maintained. Normoxic and hypoxic mice underwent in vivo experiments and/or were humanely
euthanized for terminal experiments as described in the following subsections.
6.3.3 Immunohistochemistry, Imaging and Quantification
Immunohistochemistry
Mice were anesthetized with an intraperitoneal injection of ketamine (100 mg/kg) and
xylazine (50 mg/kg), and transcardially perfused with 30 mL of ice-cold 0.01 M phosphate buffer
saline (PBS) pH 7.4 containing 1% 0.5 M EDTA. Brains were surgically removed and embedded

148
into optimal cutting temperature (OCT) compound (Tissue-Tek) on dry ice and stored at -80°C.
Fresh-frozen brains were serially cryosectioned coronally at 20 µm (Microm HM 500) and
mounted on glass Superfrost Plus slides (Cat. No. 48311-703, VWR). Sections were post-fixed
with 4% paraformaldehyde (PFA) for 10 min at 25°C, blocked (5% normal donkey serum (Vector
Laboratories) and 0.3% Triton X-100 in 0.01 M PBS) for 1 h at 25°C, then incubated with primary
antibody diluted in block solution overnight at 4°C. The next day, sections were incubated with
secondary antibody diluted in block solution for 1 h at 25°C. Sections were cover-slipped with
DAPI fluoromount mounting medium (Cat. No. 0100-20, SouthernBiotech).
To evaluate pericyte coverage, sections were immunostained with goat α-mouse CD13
primary antibody (1:100; Cat. No. AF2335, R&D Systems) followed by Alexa fluorophore-
conjugated donkey α-goat secondary antibody (1:500; Invitrogen) and DyLight fluorophore-
labeled L. esculentum lectin (1:200; Vector Laboratories).
To evaluate extravascular fibrinogen leakage, sections were immunostained with rabbit
polyclonal anti-human fibrinogen (1:400; Cat. No. A0080, Dako), which cross-reacts with mouse
fibrinogen
393,505
, followed by Alexa fluorophore-conjugated donkey anti-rabbit (1:500;
Invitrogen).
Confocal Microscopy
Slides were scanned on an A1R MP+ microscope (Nikon) coupled to a dual-beam laser
(InSight DS+, Spectra-Physics) with standardized gain, digital offset, and laser intensity, and
conditions were kept identical for each experimental group. I used a 488-nm argon laser to excite
Alexa Fluor and Dylight 488 and the emission was collected through a 500-550 nm band pass
filter; a 543 HeNe laser to excite Alexa Fluor 568 and Cy3 and the emission was collected through
a 560-615 nm band pass filter; a 633 HeNe laser to excite Alexa fluor 649 and the emission was
collected through a 650-700 nm band pass filter. Z-stack maximum intensity projections and
pseudo-coloring was performed with NIS-Elements Advanced Research software (Nikon).
Pericyte Coverage Analysis
The quantification analysis of pericyte coverage was restricted to CD13+ perivascular
mural cells that were associated with brain capillaries defined as vessels with <6 μm in diameter,
as previously described
393,506
. For pericyte coverage, 10 μm maximum projection Z-stacks (area
640 × 480 μm) were reconstructed, and the areas occupied by CD13+ (pericyte) and lectin+

149
(endothelium) fluorescent signals on vessels <6 μm were separately subjected to threshold
processing and analyzed using ImageJ. First, black and white 8-bit images for CD13 and lectin
signals were separately thresholded (Otsu) that minimize the intra-class variance of the thresholded
black and white pixels. After thresholding, the integrated signal density for each thresholded image
was calculated. In order to express the integrated signal density as the area of the image (in pixels)
occupied by the fluorescent signal, the integrated signal density was divided by 255 (the maximum
pixel intensity for an 8-bit image). The integrated pixel-based area ratios of CD13 and lectin
fluorescent signals were used to determine pericyte coverage as a percentage (%) of CD13+
surface area covering lectin+ endothelial capillary surface area per field, as previously
reported
393,506
. In each animal, five randomly selected fields from the cortex and hippocampus
were analyzed in six nonadjacent sections (∼100 μm apart) and averaged per mouse. Six animals
per group were analyzed.
Leakage Analysis
Quantification was performed from 10 μm maximum projection Z-stack (area 640 × 480
μm) reconstructed images subjected to threshold processing (Otsu) using ImageJ, as previously
described
393,505
. Briefly, the amount of perivascular fibrinogen deposits was determined as
integrated density of the deposits on the abluminal side of the lectin+ vessels. In each animal, five
randomly selected fields from the cortex and hippocampus were analyzed in six nonadjacent
sections (∼100 μm apart) and averaged per mouse. Six animals per group were analyzed.
Hemosiderin Deposits
Hemosiderin deposits in brain sections were detected by Prussian blue staining, as
previously described
505
. Percent area occupied by Prussian blue+ deposits was quantified using
ImageJ. In each animal, five randomly selected fields from the cortex and hippocampus were
analyzed in six nonadjacent sections (∼100 μm apart) and averaged per mouse. Six animals per
group were analyzed.
6.3.4 Analysis of Physiological and Biochemical Parameters
Arterial blood was collected in heparinized hematocrit capillary tubes (Cat. No. 505,
Kimble Chase), sealed with sigillum wax sealant (Cat. No. 51601, Globe Scientific Inc.), and spun
in a Micro-Capillary Centrifuge model MB (International Equipment Company, Boston, MA).

150
Hematocrit levels were determined by dividing the length of packed RBCs by the length of total
blood and expressed as a percentage.
Blood gases and electrolytes were determined from a small sample (∼90 μL) of arterial
blood, collected from the cannulated right femoral artery, using the i-STAT CG8+ moderately
complex panel (Cat. No. 03P88-25, Abbott). Measures of pH, PCO2, PO2, TCO2, HCO3, BEecf,
sO2, hemoglobin, sodium, potassium, ionized Ca
2+
, and hematocrit were determined.
For analysis of liver and kidney function, ∼200 μL serum was collected and sent to IDEXX
BioResearch for screening (Comprehensive Chemistry Panel, test code 6006). Measures of
alkaline phosphatase (ALP), aspartate aminotransferase (AST), alanine aminotransferase (ALT),
creatine kinase, albumin, total bilirubin, conjugated bilirubin, unconjugated bilirubin, total protein,
globulin, blood urea nitrogen, creatinine, cholesterol, glucose, calcium, phosphorus, bicarbonate,
chloride, potassium, albumin/globulin ratio sodium, blood urea nitrogen/creatinine ratio, and
sodium/potassium ratio were determined.
6.3.5 Longitudinal In Vivo Multiphoton Microscopy and Analysis
Chronic Cranial Window
An open-skull glass cover-slipped cranial window was prepared over the somatosensory
cortex hindlimb region (center at AP=-0.94 mm, L=1.5 mm) in anesthetized mice (1% isoflurane)
fixed in a stereotaxic frame. The window was filled with 2% agarose and covered with a 3 mm
round coverslip. Acute and chronic cranial windows offer good optical resolution for multiphoton
imaging, as commonly used by us
152,221,397,507
and others
508–511
. For a complete methodological
protocol of chronic cranial window preparation, see Kisler et al.
512
. Mice were returned to the
vivarium and given ≥3 weeks to recover from surgery prior to the longitudinal in vivo imaging
sessions.
In Vivo Multiphoton Microscopy
Longitudinal in vivo two-photon imaging was performed on an A1R MP+ multiphoton
microscope (Nikon) coupled to a dual-beam laser (InSight DS+, Spectra-Physics). Mice were
imaged while anesthetized with 1-1.1% isoflurane in air (normoxic conditions, day 0) or a custom
gas mixture of 10% O2 in balanced N2 (day 1) or 8% O2 in balanced N2 (days 3, 7, 14 and 21) to
maintain hypoxic conditions. Ambient temperature, humidity, and pressure and the mouse’s

151
physiological parameters (i.e., pulse, temperature) were monitored while imaging (MouseOx Plus,
Starr Life Sciences). The vasculature was labeled via retro-orbital injections (10 mg/mL) of 70
kDa Dextran TexasRed (Invitrogen). The cortical vasculature was imaged at a maximum depth of
~1 mm using dual-laser resonant scanning with wavelengths set at 800 nm to image eGFP and
1040 nm to image TexasRed (wavelengths were empirically determined). In vivo cortical
angiography stacks were acquired. Videos of individual cortical capillary segments (≤6 μm
diameter) located 4 branch points from the arteriole (~10 μm diameter) were acquired at ~100 μm
below the pial surface (cortical layers II and III). Capillary vessel diameter was evaluated as
described below. Moreover, we recently reported a complete methodological protocol for in vivo
multiphoton imaging and analysis of cerebrovascular hemodynamics
512
.
Capillary Diameter Analysis
Kymographs (denoting position versus time) were acquired perpendicular to the vessel, as
we
152,512
and others
513
have previously described. Images were post-processed with a 3x3 pixel
Gaussian filter to remove noise. Capillary (≤6 μm diameter) width was determined from the
kymograph, and a 1 s window box smoothing filter was applied to the resulting diameter data.
6.3.6 Longitudinal In Vivo Magnetic Resonance Imaging (MRI) and Analysis
In Vivo MRI
All MRI scans were performed using our MR Solutions 7T PET-MR system (bore size
~24-mm, up to 600 mT.m
-1
maximum gradient (MR Solutions Ltd., Guildford, UK) and a 20-mm
internal diameter quadrature bird cage mouse head coil. Normoxic (day 0) or hypoxic (days 1-21)
conditions were maintained for the duration of the MR imaging.
Mice were anesthetized by 1-1.1% isoflurane in air (normoxic conditions, day 0) or a
custom gas mixture of 10% O2 in balanced N2 (day 1) or 8% O2 in balanced N2 (days 3, 7, 14 and
21) to maintain hypoxic conditions. Respiration rate (72.0 ± 5.0 breaths per minute) and body
temperature (36.5 ± 0.5°C) were monitored during the experiments as we previously described
393
.
The sequences were collected in the following order: T2-weighted (2D-fast spin echo (FSE),
TR/TE 4,000/26 ms, 32 slices, slice thickness 300 μm, in-plane resolution 100x70 μm
2
) to obtain
structural images; diffusion weighted echo-planar imaging (EPI, TR/TE 5,000/32, 14 slices, 1 b-
value, 7 directions, slice thickness 300 μm, in-plane resolution 200x300 μm²) to assess brain

152
edema; dynamic contrast-enhanced (DCE) protocol for the capillary permeability assessment; and
finally, dynamic susceptibility-contrast (DSC) imaging for regional cerebral blood flow. Total
imaging time was approximately 40 min per mouse.
The DCE-MRI imaging protocol was performed within the dorsal hippocampus territory,
and included measurement of pre-contrast T1-values using a variable flip angle (VFA) fast low
angle shot (FLASH) sequence (FA = 5, 10, 15, 30 and 45°, TE = 3 ms, slice thickness 1 mm, in-
plane resolution 60x120 μm
2
), followed by a dynamic series of 180 T1-weighted images with
identical geometry and a temporal resolution of 5.1 s (FLASH, TR/TE = 20/3 ms, flip angle 15°,
slice thickness 1 mm, in-plane resolution 60x120 μm
2
). Using a power injector, a bolus dose (140
μL) of 0.5 mmol/kg Gd-DTPA (Gadolinium-diethylenetriamine pentaacetic acid, Magnevist®,
diluted in saline 1:4) was injected via the tail vein at a rate of 600 μL/min. DCE images were
collected over 10 min after the injection. The DSC-MRI imaging was performed on the exact same
geometry. A dynamic series of 80 T2*-weighted images is used, with a temporal resolution of 1.4
s (FLASH, TR/TE = 18/3 ms, slice thickness 1 mm, flip angle 15°, in-plane resolution 120x230
μm
2
). A second bolus dose of either (140 μL) of Gd-DTPA (1:1 dilution) or Ferimoxytol
(Ferahene®, 1:2 dilution) was injected via the tail vein at a rate of 1000 μL/min. DSC images were
collected over 85 s after the injection.
MRI Post-Processing Analysis
Capillary Permeability Assessment: Post-processing of the collected DCE-MRI data was
performed using in-house DCE processing software (Rocketship) implemented in Matlab
347
. The
unidirectional capillary transfer constant, Ktrans, to intravenously injected gadolinium-based
contrast agent was determined in both primary somatosensory cortex and hippocampus in mice
using a modified method as we reported in humans
178
with the post-processing Patlak
analysis
347,348,393
. We determined the arterial input function in each mouse from the common
carotid artery, as previously reported
178,393
.
6.3.7 Isolating Pericytes and Endothelial Cells
Brain Cell Dissociation
Tie2-Cre; Ai14; Pdgfrb-eGFP mice were used to isolate brain endothelial cells (Tie2-Cre;
Ai14) and pericyte-enriched mural cells (Pdgfrb-eGFP). Mice were anesthetized with an

153
intraperitoneal injection of ketamine (100 mg/kg) and xylazine (50 mg/kg), and quickly
transcardially perfused for 5 min with ice-cold 0.01 M phosphate buffer saline (PBS) pH 7.4.
Brains were surgically removed and placed in ice-cold Dulbecco’s Modified Eagle’s Medium
(DMEM, Thermo Fisher Scientific) supplemented with Penicillin and Streptomycin (Gibco). Prior
to generating a single cell suspension, the olfactory bulbs and brainstem were removed.
The Neural Tissue Dissociation Kit (P) (Cat. No. 130-092-628, Miltenyi Biotec) was used
with the following modifications to the manufacturer’s protocol, as previously reported
18
.
Specifically, 1) The first incubation was extended to 17 min instead of 15 min, with a rotation
speed of 20 RPM end-over-end. 2) The cell suspension was vigorously passed through a Pasteur
pipette for 15 times and the second incubation was extended to 12 minutes. 3) The cells were
passed through a 20G needle for a minimum of 10 times, and the final incubation was extended to
12 minutes. (4) All centrifugations steps were shortened to 5 minutes. All steps were performed at
room temperature, unless indicated otherwise.
Myelin was very detrimental to efficient sorting, so it was removed prior to sorting using
magnetic bead separation with Myelin Removal Beads II (Cat. No. 130-096-733, Miltenyi Biotec),
performed according to the manufacturer’s protocol. Briefly, for one mouse brain, 200 µL of beads
and three LS-columns (Cat. No. 130-042-401, Miltenyi Biotec) were used. The cells from three
columns were ultimately pooled and centrifuged at 300 g for 5 min at 4°C, after which the pellet
was resuspended in 500 µL FACS buffer (DMEM without phenol red (Thermo Fisher Scientific),
supplemented with 2% Fetal Bovine Serum (Thermo Fisher Scientific).
Fluorescence-Activated Cell Sorting (FACS)
Prior to FACS, all single cell suspensions were strained over a 35-μm mesh Cell-Strainer
capped 5 mL polystyrene round-bottom tubes (Cat. No. 352235, Corning). All cell suspensions
were sorted on a BD FACS Aria III (BD Biosciences) at the Flow Cytometry Facility at the Eli
and Edythe Broad CIRM Center for Regenerative Medicine and Stem Cell Research at USC. A
100-μm nozzle and 23-psi PBS sheet fluid pressure were used. First, cells gated for forward scatter
area/side scatter area (FSC-A/SSC-A, linear scale) to select single cells and discriminate against
doublets to avoid contamination of pericyte cells or fragments associated with single endothelial
cells, and vice versa. The single cell mask was used for all sorts. Subsequently, fluorescent events
based were selected based on the parent FSC-A/SSC-A gate. Specifically, GFP was excited with

154
a 488-nm laser, and emission was detected through a 510/20 filter, while tdTomato was excited
with a 561-nm laser and emission detected through a 582/15 filter. Cell suspensions generated
from mice lacking fluorescent reporters were used as a negative control. Single cells, positive for
GFP alone (Pdgfrb-eGFP) or positive for tdTomato alone (Tie2-Cre; Ai14) were bulk sorted into
5 mL polystyrene round-bottom collection tubes (Corning). All FACS data were analyzed using
FlowJo v10.1 (FlowJo, LLC) and FACS Diva version 8.0.2 (BD Biosciences). Bulk sorted
pericytes and endothelial cells were used for RNA isolation.
6.3.8 RNA-Sequencing and Analysis
RNA Isolation
Total RNA was isolated from sorted pericytes and endothelial cells using the Direct-zol
RNA MiniPrep Kit (Cat. No. R2050, Zymo Research). RNA concentration and RNA integrity
number (RIN) was determined with the Agilent 2100 Bioanalyzer (Agilent Technologies), and
samples with RIN >8 were used for library preparation and RNA-sequencing on Illumina Hi-Seq
2500 performed at the Technology Center for Genomics and Bioinformatics at the University of
California, Los Angeles.
Transcriptome Data Analysis
From the FASTQ raw data file, quality control of mapping rates was performed. Next, the
index was demultiplexed and adaptors and ribosomal reads were filtered out. The reads were
aligned to the Ensembl mouse gene assembly (NCBIM37) using Tophat2 software (version 2.0.4).
The number of read counts per gene was determined, quantile normalization was performed, and
read counts were log transformed. To identify the differential gene expression between cell types
(pericytes and endothelial cells) and experimental time points, statistical tests were performed
using ANOVA (Partek Genomics Suite Analysis Software). Functional pathways of differentially
expressed genes was determined using pathway analysis databases including Gene Ontology and
Ingenuity Pathway Analysis (Qiagen).

155
6.3.9 Statistical Analysis
Sample sizes were calculated using nQUERY, assuming a two-sided α level of 0.05, 80%
power, and homogenous variances for the two samples to be compared, with the means and
common SD for different parameters predicted from published work and our previous studies.
For all analyses, Shapiro-Wilk normality test was used to test normality of the data and
appropriate test to determine statistical significance was applied using GraphPad Prism 7.0
software. For comparison between two groups, an F test was conducted to determine similarity in
the variances between the groups that are statistically compared, and statistical significance was
analyzed by Student’s t-test (parametric) or Mann-Whitney U test (nonparametric). For multiple
comparisons, one-way ANOVA followed by Tukey’s posthoc (parametric) or Kruskal-Wallis test
(nonparametric) were used. For longitudinal in vivo multiphoton and MRI quantification, repeated
measures ANOVA followed by Tukey posthoc was used. In all analyses, α=0.05.
6.4 Results
I used a chronic mild hypoxia paradigm (Figure 6.1a) for up to 21-days in young adult
mice to investigate cerebrovascular signatures and their underlying molecular mechanisms. I found
a progressive increase in capillary microvascular length as a result of hypoxia-induced
angiogenesis (Fig 6.1b,c). Increased capillary density is reported to be completed by week 3 of
hypoxia exposure
483,485–489
, in agreement with my findings. Additionally, using longitudinal in vivo
multiphoton imaging of the same capillary segments during normoxic (day 0) or hypoxic (days 1-
21) conditions, I observed an acute robust increase in capillary diameter with a maximal 55%
increase at day 3 that then begins to stabilize by day 7 back to pre-hypoxic conditions (Figure
6.1d-f). This is consistent with evidence that chronic mild hypoxia elicits an acute CBF
increase
483,484
. As a compensatory response to low tissue oxygen availability, hematocrit (the
proportion of packed red blood cell (RBC) volume relative to total blood volume) levels increased
progressively with chronic mild hypoxia (Figure 6.1g). Erythropoiesis is a well-reported response
to hypoxic insult
483,514,515
, and it is also a possible explanation for the CBF renormalization that
coincides with increased hematocrit. By increasing the oxygen carrier, the oxygen content of
circulating blood is restored at the pre-hypoxic CBF rate.

156

Figure 6.1 Cerebral microvascular remodeling during hypoxia.
(a) Schematic of the 21-day chronic mild hypoxia paradigm and analysis time points. (b,c)
Representative images showing lectin+ endothelia profiles (red) (b) and quantification of
microvascular length (c) in the hippocampus at days 0, 1, 3, 7, 14, and 21 of chronic mild hypoxia
exposure. (d-f) Longitudinal in vivo multiphoton imaging during the chronic mild hypoxia
paradigm. Representative angiography image of the same cortical depth 250 µm stack (d) and
representative image of the same capillary longitudinally at days 0, 3, 7, and 14 of chronic mild
hypoxia exposure. Quantification of capillary diameter of the same capillary segment
longitudinally at days 0, 3, 7, and 14 of chronic mild hypoxia exposure (f). (g) Percent hematocrit
(ratio of red blood cell volume to total blood volume) at days 0, 1, 3, 7, 14, and 21 of chronic mild
hypoxia exposure. In panels c and g, significance by one-way ANOVA with Tukey posthoc text,
α=0.05; n=6 mice/group; mean ± SEM. In panel f, significance by repeated measures ANOVA
with Tukey posthoc test, α=0.05; n=20 vessels/time point; mean ± SEM.

At the cellular level, angiogenesis involves pericyte-endothelial cell detachment,
migration, reattachment, and stabilization of the newly remodeled microvasculature
473
. Since
pericytes detach from the vasculature in order to sprout new cerebral blood vessels, I hypothesized
there would be disrupted pericyte coverage and related BBB breakdown during hypoxia. As
expected, I observed a 38% decrease in pericyte coverage with the maximal decrease occurring at
day 7 (Figure 6.2a,b). Extravascular leakage of blood-derived fibrinogen was also transiently
increased with the largest leakage occurring at day 7 (Figure 6.2c,d), indicating BBB breakdown.
We next used a dynamic contrast-enhanced magnetic resonance imaging sequence to assess BBB
permeability to a gadolinium contrast agent as another independent measure of BBB breakdown.
Employing this MRI sequence longitudinally while maintaining hypoxic conditions, we similarly
observed a 250% increase in BBB permeability with the maximal increase at day 7 (Figure 6.2e)

157

Figure 6.2 Transient blood-brain barrier (BBB) breakdown during hypoxia.
(a,b) Representative images showing CD13+ pericytes (red) and lectin+ endothelia profiles
(green) (a) and quantification of pericyte coverage measuring the percentage of CD13+/lectin+
signal (b) in the hippocampus at days 0, 1, 3, 7, 14, and 21 of chronic mild hypoxia exposure. (c,d)
Representative images showing blood-derived fibrinogen (green) and lectin+ endothelia profiles
(purple) (e) and quantification of fibrinogen leakage (f) in the hippocampus at days 0, 1, 3, 7, 14,
and 21 of chronic mild hypoxia exposure. (e) Longitudinal in vivo dynamic contrast-enhanced
magnetic resonance imaging measuring Ktrans BBB permeability in the cortex (left) and
hippocampus (right) at days 0, 1, 3, 7, 14, and 21 of chronic mild hypoxia exposure. (f)
Representative images of Prussian blue staining of hemosiderin deposits (blue) at days 0 and 7 of
chronic mild hypoxia exposure. In panels b and d, significance by one-way ANOVA with Tukey
posthoc test, α=0.05. In panel e, significance by repeated measures ANOVA with Tukey posthoc
test, α=0.05. Analyses include n=5-6 mice/group; mean ± SEM (panels b and d) and mean ± SD
(panel e).

which mirrors the time course of altered pericyte coverage. Prussian blue staining revealed
pronounced hemosiderin deposits at day 7 (Figure 6.2f). The T2*-MRI sequence would be useful
to further assess the regional distribution of microbleeds; however, it cannot be accurately
quantified in vivo since microbleeds and the hypoxic blood both appear as hypointense signals.
Although the entire body is exposed to chronic mild hypoxic conditions, there is not a
major impact on peripheral physiological functions. Mice exposed to hypoxia experience an acute
body weight loss that then stabilizes and begins to recover (Figure 6.3a), as reported
504
, and all
mice survived the hypoxia paradigm. Chronic mild hypoxia does not affect liver and kidney
analyses (Figure 6.3b,c) or systemic physiological and biochemical parameters including
electrolytes and blood gases (Figure 6.3d). This supports that chronic mild hypoxia is a valid tool
to investigate molecular regulation and signatures of the cerebrovasculature in the adult brain.

158

Figure 6.3 Chronic mild hypoxia does not affect peripheral physiological functions.
(a) Daily body weight of mice during the 21-day duration of normoxia (21% O2) and hypoxia (8-
10% O2) exposure; for details see Section 6.3.1 Prolonged Hypoxia Paradigm. n=6-36
mice/group; mean ± SEM; significance by two-way ANOVA, α=0.05. (b) Peripheral measures of
liver function include aspartate aminotransferase (AST), alanine aminotransferase (ALT),
globulin, and cholesterol. (c) A peripheral measure of kidney function includes blood urea
nitrogen. (d) Peripheral measures of electrolytes and blood gases include sodium, potassium,
ionized calcium (Ca
2+
), pH, and bicarbonate (HCO3). In panels b-d, n=3 mice/group; mean ± SEM;
significance by two-tailed Student’s t-test, α=0.05.
The major cerebrovascular responses to chronic mild hypoxia presented in Figures 6.1-6.3
include cerebrovascular remodeling (increased capillary density), acute increase in capillary
diameter, and transient BBB breakdown. But, what are the molecular signaling events that underlie
these cerebrovascular responses? Using Tie2-Cre; Ai14; Pdgfrb-eGFP mice with fluorescently
labeled endothelial cells and pericyte-enriched mural cells (Figure 6.4a), I conducted bulk RNA-
sequencing to investigate transcriptional profiles during hypoxia. There are 950 differentially
expressed genes in pericytes during hypoxia at day 3 compared to normoxic controls (Figure
6.4b). Pathway analysis informs that these differentially expressed genes group into functional and
disease categories underlying angiogenesis, cell stress, BBB permeability and immune responses
(Figure 6.4c). I also found that downstream gene expression changes were mediated by upstream
molecules and pathways, specifically the endothelial cell-secreted growth factors including
vascular endothelial growth factor A (VEGFA), transforming growth factor-b1 (TGFB1),
angiopoietin-2 (ANGPT2), and platelet-derived growth factor-B (PDGFB), and transcription
factors including forkhead box F2 (FOXF2), kruppel like factor 2 (KLF2), and KLF4 (Figure
6.4d). This provides new insights into mechanisms that regulate cerebrovascular responses in the

159
adult brain which will aid in identifying novel therapeutic targets for neurological conditions with
hypoxia and/or BBB and cerebrovascular dysfunction.

Figure 6.4 Molecular regulation of cerebrovascular responses during hypoxia.
(a) Representative images showing GFP+ pericyte-enriched mural cells (green) and tdTomato+
endothelial cells (red) in the cortex of Pdgfrb-eGFP; Tie2-Cre; Ai14 mice. (b) Volcano plot of the
differentially expressed genes in pericytes at 3 days versus 0 days of hypoxia exposure. (c)
Ingenuity Pathway Analysis (IPA) reporting the categories of diseases and functions of the
differentially expressed genes. (d) Upstream analysis using IPA that reveals the key molecules that
are predicted to induce the downstream differential changes in gene expression. See 6.3.8 RNA-
Sequencing and Analysis for more details.

6.5 Discussion
I report the temporal course of cerebrovascular signatures in response to chronic mild
hypoxia including cerebrovascular remodeling (increased capillary density), acute increase in
capillary diameter, and transient BBB breakdown. I further explored the transcriptional changes
during hypoxia to gain insight into the molecular signaling events that underlie these
cerebrovascular responses. Associating the functional/pathophysiological data with molecular data
raises the question – what occurs first? Cell type-specific molecular changes can drive the
functional responses which, in turn, drive additional molecular changes.

160
I observed that VEGFA, PDGFB, TGFB, EGR, and KLF are key molecules expressed
during chronic mild hypoxia that mediate downstream transcriptional changes in pericytes to
regulate cerebrovascular responses. The two major pathways reported during hypoxia including
hypoxia-inducible factor-1 (HIF-1)
480,504
dependent upregulation of VEGF
516,517
and PDGFB
518–
520
, and HIF-1 independent, cyclooxygenase-2 (COX-2) dependent upregulation of
ANGPT2
489,504,521,522
. Furthermore, my data also reveals the importance of FOXF2 and KLF2/4
expression during chronic mild hypoxia. FOXF2 was recently reported to be required for brain
pericyte differentiation and BBB maintenance during development
523
, and KLF2/4 control
endothelial identity and vascular integrity
524
and KLF4 was recently reported to protect the brain
microvasculature from ischemia-induced apoptosis
525
. These data support that multiple, redundant
pathways are responsible for regulating cerebrovascular signatures, and brain microvascular
remodeling depends on the balance between pro- and anti-angiogenic processes
526–528
.
Continuous hypoxic exposure can be produced both by decreasing atmospheric pressure
(hypobaric) or decreasing the fractional oxygen component in the inspired gas mixture, and there
do not appear to be important differences in the physiological adaptive response between the two
types of hypoxia
482
. The atmospheric altitude equivalent of 10% O2 is ~5,000 m which is
considered mild hypoxia
482
. Mild hypoxia is not associated with permanent neuronal damage, but
moderate and severe degrees of hypoxia lead to neuronal loss in a time of exposure/degree of
hypoxia-dependent manner
482
.
Elucidating microvascular transcriptional and cellular regulation of cerebrovascular
signatures in the adult brain is of upmost importance to target and counteract the many CNS
conditions plagued by microvascular dysfunction, BBB breakdown, aberrant angiogenesis, and
hypoxic insult that promotes pathophysiology. Whether the molecular makeup and signaling
pathways gained from animal models is similar in the human brain vasculature remains unclear,
illustrating a major gap between animal models and human studies. Nevertheless, knowledge
informed from animal studies offers a roadmap to examine the link between neurovascular
responses/adaptations and neurological conditions with cerebrovascular dysfunction in humans.

161
6.6 Future Directions
My data so far provide a solid platform and the framework for future studies aimed at
understanding the complex relationship between molecular and functional changes, including
molecular initiators and negative feedback in response to the functional changes, as well as better
understanding cell non-autonomous interactions between endothelial cells and pericytes.
Additionally, my data on transcriptional changes between normoxia and hypoxia gained
from bulk RNA-sequencing informs transcriptional regulation of vascular cell types (i.e.,
endothelial cells and pericytes) and their interactions during hypoxia, but it is unable to inform
regional transcriptional regulation along the vascular arteriovenous axis. Moreover, upon
comparing my specific gene changes observed with a recent molecular atlas of cell types and
zonation in the brain vasculature performed by Dr. Christer Betsholtz’s group and recently
published in Nature
18
, I noticed potential contamination from fragments of fibroblasts and/or
endothelial cells in my dataset. Thus, to validate my bulk RNA-sequencing data and to further
explore the transcriptional changes along the arteriovenous axis in the brain, I am currently
performing single cell RNA-sequencing. To do this, I am FACS-ing single cells from Tie2-Cre;
Ai14; Pdgfrb-eGFP mice into 384-well plates for single cell RNA-sequencing using Smart-seq2
529

in collaboration with Dr. Christer Betsholtz at Karolinska Institute who pioneered this approach to
generate a molecular atlas of cell types and zonation in the brain vasculature in his recent Nature
paper
18
. Using this approach, we are able to evaluate vascular zonation by backspin clustering
analysis that uses known arterial, capillary, and venous genes, as reported
18
. This will inform
expression of contractile proteins, microvascular remodeling (i.e., TJ, AJ, basement membrane
molecules), transcytosis, and metabolomic changes along the vascular tree during hypoxia.
I am confirming select candidate molecules from the RNA-sequencing studies by in situ
hybridization, immunohistochemistry, and Western blot analysis. Additionally, while RNA-
sequencing is a powerful tool to reveal differential transcriptional responses, transcripts are
translated to proteins at different rates. Thus, proteomic studies are important to confirm the
functional impact of molecules and signaling mechanisms informed from the RNA-sequencing
analysis, and I am conducting proteome analysis during hypoxia to further confirm my results.
Furthermore, future studies will attempt to establish and differentiate molecular changes
that drive functional responses, and molecular changes that result from functional responses

162
(negative feedback). This will require studies at very early timepoints of hypoxia (e.g., 1 hour, 3
hours, 6 hours, 12 hours, etc.) during which molecular changes have been initiated but functional
responses have not yet occurred or are just beginning to take place. Moreover, there is redundancy
in some signaling pathways, for example VEGF-A is secreted by both pericytes (paracrine
signaling) and endothelial cells (autocrine signaling) to activate the VEGFR2 pathway in
endothelial cells
8
. Thus, it would be useful to characterize the degree to which endothelial cell
responses result from autonomous versus cell non-autonomous interactions between pericytes and
endothelial cells. Future directions also include investigating molecular regulation of
cerebrovascular adaptions to hypoxia during aging, in pericyte-deficient mouse models, and in
mouse models of neurological conditions (i.e., AD). These studies will further inform therapeutic
efforts to combat hypoxic insults and/or microvascular dysfunction in numerous CNS disorders.
6.7 Acknowledgements
I would like to acknowledge Dr. Mikko Huuskonen (Postdoctoral Scholar at USC) and Dr.
Axel Montagne (Assistant Professor of Research in Physiology & Neuroscience and Assistant
Director of the Functional Biological Imaging Core at USC) for their contributions in this study
conducting and analyzing MRI sequences, particularly DCE-MRI to assess BBB permeability. I
would also like to thank Dr. Zhen Zhao (Assistant Professor in Physiology & Neuroscience at
USC) for initially breeding the Pdgfrb-eGFP; Tie2-Cre; Ai14 mice. Finally, thanks to Dr. Christer
Betsholtz (Professor at the Karolinska Institute in Sweden) and colleagues for our collaboration
conducting single cell RNA-sequencing which is adding valuable insight into the vascular
transcriptional changes along the arteriovenous axis during hypoxia – see above Section 6.6
Future Directions.
This research was supported by National Institutes of Health (NIH) grant nos. NS100459,
AG039452, NS034467 and AG023084 to B.V.Z. and the Foundation Leducq Transatlantic
Network of Excellence for the Study of Perivascular Spaces in Small Vessel Disease reference no.
16 CVD 05.

163
CHAPTER 7:
CONCLUSIONS, SYNTHESIS AND FUTURE IMPLICATIONS

Adapted from:
Sweeney MD…Zlokovic BV, Physiological Reviews, 2019
Sweeney MD…Zlokovic BV, Alzheimer’s & Dementia, 2019
Sweeney MD…Zlokovic BV, Nature Neuroscience, 2018
Sweeney MD…Zlokovic BV, Nature Reviews Neurology, 2018
Sweeney MD…Zlokovic BV, Nature Neuroscience, 2016
Sweeney MD…Zlokovic BV, Journal of Cerebral Blood Flow and Metabolism, 2015
7.1 Summary of Findings
The preceding chapters present the complexity of BBB maintenance and regulation during
normal physiology and also discuss existing evidence of BBB breakdown and dysfunction in AD
(Chapter 1). This was followed by original evidence that BBB breakdown is an early biomarker
of cognitive impairment in humans (Chapter 2) that is further accelerated by APOE4 genetic risk
(Chapter 3). Future directions extending from this work include evaluating the predictive potential
of CSF sPDGFRb and BBB breakdown biomarkers to determine their ability to predict
longitudinal cognitive decline in individuals with and without genetic risk factors, e.g., APOE4.
Also, further confirming mechanisms of BBB breakdown (e.g., the APOE4-mediated CypA-MMP
pathway), for example by validating activation of these molecules in CSF is important and opens
the door for new therapeutic targets. Additionally, evidence of cerebrovascular dysfunction in CSF
in ADAD individuals carrying PSEN1 or APP mutations is reported (Chapter 4). Understanding
how the pathophysiology and disrupted underlying mechanisms manifest from specific mutations
in PSEN1 and APP has yet to be elucidated. Also, whether or not cerebrovascular dysfunction
similarly occurs in individuals with PSEN2 mutations is currently unknown since these mutations
are extremely rare and account for only ~5% of all ADAD cases
272
. Whether targeting
cerebrovascular dysfunction in ADAD has the potential to delay disease onset and/or progress is
an open question that should be studied.

164
A novel assay to detect pericyte injury (sPDGFRb) was developed and validated as a
sensitive, reliable and clinically-relevant biomarker of pericyte injury during cognitive impairment
(Chapter 5). This new assay offers many advantages. It is ultrasensitive which enables use of
small CSF sample volume, which is an invaluable advantage for limited, precious samples. Also,
since it was developed on the MSD platform, sPDGFRb can be multiplexed with other analytes of
interest to create a single panel of vascular biomarkers for research and clinical evaluation. This
novel, ultrasensitive assay is easy to incorporate at different sites to investigate pericyte injury in
various cohorts and should be employed across neurodegenerative diseases.
Lastly, a mouse model of hypoxia that exhibits cerebrovascular responses was used to
better understand molecular mechanisms and signatures of the cerebrovasculature (Chapter 6).
This study provides a solid platform and the framework for future studies aimed at understanding
the complex relationship between molecular and functional changes, including molecular initiators
and negative feedback in response to the functional changes, as well as better understanding cell
non-autonomous interactions between endothelial cells and pericytes. Additionally, my future
work includes using single cell RNA-sequencing to explore the transcriptional changes occurring
along the vascular tree on the arterio-capillary-venous axis. The final chapter below will synthesize
the original findings within the current perspectives in the field of neurodegenerative diseases,
BBB-directed therapeutic efforts, and discuss gaps in the field and future directions.
Overall, this dissertation has provided compelling evidence that the BBB and brain
capillary pericytes are injured during cognitive and functional decline in humans (Chapters 2-5),
but it is still unknown why pericytes are injured. Does chronic mild hypoxia contribute to pericyte
injury? Animal studies suggest hypoxia may be an underlying mechanism responsible for
cerebrovascular dysfunction in AD
395
, but there is not direct evidence of this in humans. Currently,
we can measure CBF changes with MRI sequences (e.g., arterial spin labelling, ASL), and recent
studies support that hypoperfusion is among the earliest detectable change in AD
pathophysiology
181
. Moreover, in our studied cohort, CSF sPDGFRβ correlated with CSF pTau
(unpublished observation) although pTau did not mediate the relationship between CSF sPDGFRβ
and cognitive impairment (Chapter 2). Tau tangles that have been linked to vasculotoxicity so it
is possible that pericyte injury could result from hyperphosphorylated tau. Since tau tangles cannot
be detected by CSF, future studies should use second-generation tau tracers (e.g., Flortaucipir)
with improved sensitivity by PET to relate to CSF sPDGFRβ and regional BBB Ktrans changes

165
The recent 2018 NIA-AA Research Framework attempts to unify language of biomarker-
based definition of AD
366
, however it underrecognizes AD as a heterogeneous disease and does
not clearly define AD in the context of multifactorial and functional systems contributing to
disease pathophysiology. In living humans, it is difficult to determine the degree to which AD
exists alone or as a mixed dementia (e.g., concomitant with vascular dementia, etc.). Nevertheless,
the data presented in this dissertation support that cerebrovascular dysfunction is an important
biomarker of early cognitive impairment that reflects a core pathophysiological contributor to
cognitive impairment that occurs independent of systemic vascular risk factors. Rather than
focusing only on amyloid and tau, broadening the perspective and study of contributing factors to
AD including cerebrovascular dysfunction will aid in patient-directed therapeutic efforts to apply
the right drug(s) - at the right dose - at the right time - in the right study design - and with the right
outcome measures for successful intervention to delay, prevent and/or reverse dementia including
AD. Individualized, targeted therapies for AD will be successful when the complexity of AD
pathophysiology is fully appreciated so that multidisciplinary team efforts can be mounted to
successfully address one of the most challenging diseases in the 21
st
century.
7.1.1 Update on Dynamic Biomarkers in Alzheimer’s Disease
The recent 2018 NIA-AA Research Framework acknowledges vascular biomarkers could
be added when they are defined
366
, but unfortunately does not fully appreciate that several vascular
biomarkers “ready-to-be-used” already exist and are well defined. Since amyloid and tau deposits
may not be causal in AD pathogenesis, as recognized by the Research Framework
366
, it is the right
time to encourage inclusion of biomarkers of vascular dysfunction in observational and
interventional research studies as we recently recommended
373
. I urge different research and
clinical sites to incorporate biomarkers of vascular dysfunction to assess vascular contributions to
cognitive impairment and AD using molecular biomarkers of vascular damage and imaging
biomarkers such as fluid-attenuated inversion recovery (FLAIR), diffusion tensor imaging (DTI),
T2*-weighted sequences, DCE, ASL, time of flight (TOF) angiography, and blood oxygen level-
dependent (BOLD) MRI sequences in individuals with AD or dementia risk or with suspected
dementia. Whenever and whichever possible, vascular biomarkers should be adopted in AD
research studies, large epidemiological studies and interventional trials
530
. Integration of vascular

166
dysfunction biomarkers into the diagnostic process may allow for earlier diagnosis of AD in some
patient subsets. Recognizing and including the wealth of knowledge on how to prevent and treat
vascular disease and on interventions to modify vascular dysfunction could significantly advance
research in AD and dementia, thus ultimately helping patients.
The first hypothetical model of in vivo AD dynamic biomarkers, often called the Jack
model, was proposed in 2010 and updated in 2013 to include interim evidence and present inter-
subject variability in cognitive impairments
531
. This model was intended to summarize literature
showing temporal evolution of in vivo AD biomarkers relative to each other and to the onset and
progression of clinical symptoms. This initial model, however, did not include currently available
evidence that vascular dysfunction contributes to AD pathophysiology in a significant way
21
.
Therefore, it is time now to update the hypothetical biomarkers model of AD and include the
impact of cerebral blood vessels on AD pathophysiology, as previously suggested
21,181
. Figure 7.1
presents an updated revised model of AD biomarkers to show the role of brain vasculature and
early changes in CBF and BBB biomarkers in AD that, according to some studies (as we

Figure 7.1 Hypothetical updated Jack model of AD biomarkers to include the role of brain
vasculature.
Hypothetical model of AD biomarker changes illustrating that early CBF and BBB biomarkers
and vascular dysfunction may contribute to initial stages of AD pathophysiological progression
from no cognitive impairment (NCI) to MCI to AD, which is followed by cerebrospinal fluid and
brain changes in Aβ, amyloid, and tau biomarkers. All biomarker curves converge at the point of
maximum abnormality. The horizontal axis of disease progression is expressed as time. Cognitive
response is illustrated as a zone (blue filled area) with low and high-risk borders. Subjects with
high risk of AD-related cognitive impairment are shown with a cognitive response curve that is
shifted to the left. In contrast, the cognitive response curve is shifted to the right in subjects with a
protective genetic profile, high cognitive reserve, and the absence of comorbid brain pathologies.

167
reviewed
374
), are altered prior to cognitive decline, brain atrophy, neurodegeneration, and/or
amyloid and tau biomarker abnormalities. The sigmoid shape of the vascular curve reflects
growing evidence obtained from neuroimaging, biofluid and post-mortem tissue studies indicating
an initial acceleration followed by deceleration of brain vascular changes, which do not plateau,
similar as shown for the other biomarkers. However, future longitudinal studies may warrant
amendment of the proposed trajectories of the vascular curve as well as of the other biomarker
curves and/or the order of their respective appearance.
7.2 Commonalities of Cerebrovascular Dysfunction in Neurodegenerative Disorders
To date, AD is the primary neurodegenerative disease linked to neurovascular dysfunction,
but emerging evidence is also implicating the brain vasculature in the pathogenesis of other
neurological diseases with dementia and/or neurodegeneration, including frontotemporal
dementia, Parkinson’s disease (PD), Huntington’s disease (HD), amyotrophic lateral sclerosis
(ALS), multiple sclerosis (MS), HIV-associated dementia (HAD), and chronic traumatic
encephalopathy (CTE)
5,8,43
. Existing evidence of BBB breakdown by neuroimaging, biofluid, and
neuropathological findings in PD, HD, ALS, MS, HAD, and CTE is presented in Appendix D
Tables D.1-D.3. Also, see my recent review for discussion of BBB breakdown and CBF
dysregulation in these diseases
43
.
Since pericyte injury and BBB breakdown are associated with early cognitive impairment
as shown above in Chapters 2-5, it would be extremely valuable to assess pericyte injury by CSF
sPDGFRb in these additional neurodegenerative diseases, which has never been done before.
Consistently assessing BBB breakdown by DCE-MRI and biofluid analyses should also be studied
to confirm the limited reports of BBB breakdown using these techniques in living humans with
some of these neurodegenerative diseases (summarized in Appendix Tables D.1 and D.2). Many
more studies are needed to better understand how BBB and cerebrovascular dysfunction contribute
to the pathophysiology in each disease. To date, the majority of BBB/cerebrovascular dysfunction
evidence across neurodegenerative diseases has come from neuropathological brain tissue analysis
(and spinal cord in ALS) (Appendix Table D.3), which cannot provide temporal or
pathophysiological insight. Nevertheless, existing evidence indicates numerous neurodegenerative
diseases share pathological alterations of the vessel wall resulting in BBB disruption. Endothelial

168
degeneration with loss of tight junction proteins and/or increased transendothelial bulk flow
transcytosis
6,7
, and the associated pericyte degeneration causing BBB breakdown
8,11,57,397,449
,
initiates multiple pathways to neuronal injury and neurodegeneration; these pathways are
summarized in Appendix Figure D.1 and discussed briefly below.
Entry of several neurotoxic blood-derived proteins including plasminogen, thrombin and
fibrinogen into different CNS regions is evident in different neurodegenerative disorders. Plasmin,
which is generated from circulating plasminogen, degrades the neuronal matrix protein laminin,
thereby promoting neuronal injury
532
. High concentrations of thrombin mediate neurotoxicity and
memory impairment
533
and accelerates BBB disruption
534
. Fibrinogen leads to axonal retraction
535

and BBB damage, which promotes neuroinflammation
228
. Additionally, fibrin promotes
neuroinflammation and demyelination in a mouse model of MS
536
and stimulates activation and
induction of antigen presenting genes in primary microglia and bone marrow-derived
macrophages
537
. The role of coagulation and fibrinolysis proteins on development of brain
pathology in MS has been recently reviewed
538
.
Influx of albumin leads to perivascular edema obstructing brain microcirculation and blood
flow, which generates hypoxic conditions leading to neuronal injury and impaired hemodynamic
responses contributing to neurodegeneration
2,152
. RBCs extravasation causing microbleeds, as seen
in almost all neurodegenerative disorders
359,362
, leads to accumulation of toxic, iron-containing
proteins (e.g. hemoglobin) and release of free iron (Fe
2+
)
539–541
generating reactive oxygen species
(ROS) and oxidant stress to neurons
542
.
In neurodegenerative diseases such as AD, PD, and HAD
384,385,543,544
, dysfunction in P-gp
active efflux transport at the BBB leads to accumulation of toxic xenobiotics (e.g., environmental
pollutants, food additives, pesticides, drugs) in brain. Reduced P-gp and LRP1 expression at the
BBB
113,119,127
, and increased RAGE expression in brain microvessels
6,121,122,124,125
, lead to faulty
clearance of Alzheimer’s Aβ toxin and its accumulation in brain. Reduced blood flow and Aβ can
both promote tau pathology, another key pathological hallmark of AD
2
. Whether faulty clearance
of tau in AD and CTE, α-synuclein in PD and/or huntingtin in HD can also contribute to their
respective accumulations in CNS is not clear at present.
Interestingly, recent experimental studies suggest that α-synuclein is transported in and out
of the brain across the BBB as a free peptide
545
. Moreover, systemically administered α-synuclein
oligomers, ribbons and fibrils cause distinct synucleinopathies implying they can cross the BBB
546
.

169
In humans, extracellular vesicles containing α-synuclein were found in the CSF and blood
suggesting bidirectional transport of α-synuclein between blood and CSF
547,548
. As RBCs are a
source of α-synuclein-containing extracellular vesicles
547
, and their extravasation is reported in the
striatum in PD
549,550
, it is possible that RBCs may contribute to development of synucleinopathy
in humans. Since circulating α-synuclein levels are two orders of magnitude higher than in the
CNS
548
, α-synuclein transport across the BBB might be implicated in PD pathogenesis and could
be a novel therapeutic target.
Neurotoxic accumulates and reduced blood flow can activate microglia and astrocytes
inflammatory response including secretion of neurotoxic cytokines and chemokines
180
.
Additionally, in some diseases such as AD, infiltration of peripheral macrophages
147,398
and
neutrophils
399
suggests activation of innate immune response. Besides peripheral macrophage
infiltration, influx of T and B lymphocytes across the BBB was found in MS indicating adaptive
immune response
160
. These studies suggest breakdown of the BBB to circulating leukocytes.
Moreover, BBB breakdown leads to generation of several anti-CNS autoantibodies in humans
193
,
but their role in pathogenesis of neurodegenerative disorders has not been fully explored. Also,
BBB breakdown lowers the microbial barrier, which can allow circulating pathogens to enter the
brain and injure neurons, and/or provoke amyloid response aggravating β-amyloidosis, as shown
in AD models
551,552
. The ‘gut-brain axis’ is a recently emerging concept with important
implications for brain health and age-related neurodegenerative diseases.
Altogether there are commonalities of vascular dysfunction across neurodegenerative
diseases, yet CNS regions are differentially affected depending on the disease. This concept is
illustrated in Figure 7.2 that focuses on specific regions where BBB dysfunction and CBF
shortfalls are apparent during early stages of each disease
374
. Given the importance of the
vasculature for CNS health, aging, and neurodegenerative diseases, efforts to evaluate the clinical
relevance of BBB/pericyte dysfunction (Chapters 2-5), to understand mechanisms of
cerebrovascular regulation (Chapter 6), and to target the BBB for treatments (discussed below)
are all critical avenues for future research.

170

Figure 7.2 Commonality of an early involvement of the CNS vasculature in different
neurodegenerative disorders.
Region-specific brain vascular dysfunction including CBF shortfalls (reductions and
dysregulation) and/or BBB breakdown (increased vascular permeability and transporter
dysfunction) is a common pathway seen early in multiple neurodegenerative disorders, including
AD, PD, HD, ALS, and MS. See Appendix D for details. Specifically, some studies suggest that
vascular dysfunction (CBF and/or BBB) in the hippocampus, gray matter, and entorhinal cortex in
AD may precede dementia, brain atrophy, and/or detectable Aβ and tau biomarker changes.
Similarly, vascular dysfunction in the white matter and corpus callosum in MS, basal ganglia (the
caudate nucleus, thalamus, putamen, globus pallidus, and substantia nigra) in PD and HD, spinal
cord white matter pyramidal tract in MS, and motor cortex and spinal cord in ALS are found by
some studies in early stages of these disorders before progression of neurological symptoms
including motor deficits.
7.3 Targeting the BBB for Treatments
Pericyte injury and BBB breakdown are related to early stages of cognitive impairment and
AD (Chapters 2-5), and cerebrovascular dysfunction is apparent in numerous neurodegenerative
diseases (discussed in the preceding Section 7.2 Commonalities of Cerebrovascular
Dysfunction in Neurodegenerative Disorders and Appendix D) and in acute neurological
injuries, as reviewed elsewhere
74
. Thus, it is timely and important to target the BBB for treatments.
Improving cerebrovascular function has the potential to delay and/or prevent neurodegenerative
diseases. Moreover, even targeting the classic AD pathology (e.g., Ab and tau) still requires the
BBB since drug delivery to the CNS is problematic in both physiological and pathological states.

171
The BBB poses two major challenges for the development of therapeutics for CNS
disorders. First, the BBB’s specialized gatekeeping nature hinders therapeutic drug delivery into
the CNS under physiological conditions. Due to its highly effective active efflux systems, the BBB
rejects most of the small molecule drugs and large therapeutic molecules including growth factors
and antibodies
6,553,554
. Second, pathological BBB breakdown does not provide easy, open access
into the brain – this is a common misconception. Focal BBB breakdown in the disease leads to
perivascular accumulation of blood-derived toxic products and macromolecules, immune and/or
inflammatory responses, vascular regression and local CBF reductions
2,7
. These focal vascular
changes limit CNS distribution of neurotherapeutics in disease-affected regions by disrupting
diffusional transport across brain extracellular spaces and/or by blocking normal ISF flow
dynamics
43,555
.
It is incorrectly assumed that disease-initiated BBB breakdown presents an opportunity to
deliver therapeutic antibodies, proteins, peptides, small molecules, and/or gene medicines to cell
types (e.g. neurons) without a further need to manipulate the BBB. However, brain regions affected
by neurodegeneration develop a pathological BBB breakdown with functional and structural
changes in blood vessels often before neurodegeneration, which persist as the disease progresses.
These vascular changes include endothelial degeneration, reduced tight and adherens junction
proteins at the BBB, increased endothelial bulk flow transcytosis, disrupted BBB transporter
expression, pericyte degeneration, perivascular accumulation of toxic products, inflammation, and
immune response (Appendix Figure D.1), all of which hinder therapeutic delivery. Under
pathological conditions, blood-derived products, water and electrolytes accumulate into enlarged
perivascular spaces, which blocks normal diffusion of solutes across brain extracellular spaces,
ISF formation and ISF flow resulting in impaired distribution of solutes throughout the CNS
(Figure 7.3). As a consequence of disease-driven BBB disruption, the therapeutic antibodies,
proteins and peptides, and small drug molecules will likely get trapped in pathologically altered
brain tissue within the enlarged perivascular spaces along with other blood-derived debris, and
will not effectively reach their targets (e.g., neurons) because of impaired transport across
parenchymal extracellular spaces and diminished ISF regional flow. Low expression of CMT and
RMT systems additionally complicates their use for therapeutic drug delivery. Therefore, relying
on regions with healthy blood vessels and/or stabilizing damaged vasculature in disease-affected

172

Figure 7.3 BBB dysfunction – implications for drug delivery.
In a healthy BBB (left), strategies to breach the BBB and deliver neuropharmaceuticals to the brain
rely on carrier-mediated transport (CMT), receptor-mediated transcytosis (RMT), nanoparticles
and/or transient opening of the BBB (for example, using focused ultrasound). Under pathological
conditions (right), the disrupted BBB enables blood-derived debris and cells to accumulate in
enlarged perivascular spaces. These accumulations prevent the normal distribution of molecules
throughout the CNS by concentration-gradient-driven diffusion across brain extracellular spaces
(ECSs) and interrupt the regional formation of interstitial fluid (ISF) and ISF flow, which prevent
therapeutic antibodies, proteins, peptides, gene medicine and other drugs from efficiently reaching
their neuronal targets. Abbreviations: L‑DOPA, L-3,4‑dihydroxyphenylalanine.

regions to improve cerebrovascular integrity and re-establish diffusion across extracellular spaces
and ISF circulation, is important for successful delivery of neurotherapeutics to a disease-affected
brain.
Successful therapeutic delivery across the BBB requires functionally and structurally
healthy blood vessels, normal vascularization, blood flow, and expression of CMT and RMT
systems that can be utilized to facilitate drug delivery to the CNS. Strategies utilizing existing
CMT and RMT BBB systems have been explored to increase brain penetration and potency of
neurotherapeutics (Figure 1.4). Successful examples include the LAT1 large neutral amino acid
CMT transporter for delivery of L-3,4-dihydroxyphenylalanine (L-DOPA, as known as levodopa)
in PD
66
and targeting transferrin BBB RMT for delivery of therapeutic antibodies for various
neurological conditions
4,104,556,557
. On the other hand, approaches to circumvent the BBB have been
utilized clinically for treatment of some CNS disorders. For example, these include
intracerebroventricular administration of Cerliponase for Batten’s disease
558
, intrathecal
administration of antisense oligo Spinraza targeting the survival motor neuron 1 gene (SMN1) for

173
infantile spinal muscular atrophy
559
, intrathecal Ziconited peptide for chronic pain
560
, and
intranasal administration of insulin, leptin, and oxytocin
561
. Other approaches using
nanoparticles
562
and/or opening the BBB by focused ultrasound
563–565
have been recently attempted
to deliver therapeutics to the CNS. Recent approaches developed to protect the BBB and eliminate
secondary vascular-mediated CNS changes, and to effectively traverse the BBB to improve CNS
drug delivery are summarized in Appendix Table E.1 and select approaches are briefly discussed
below and reviewed in detail elsewhere
74
.
A few pharmacological agents have been reported to restore BBB function in animal
models of acute neurological disorders
566,567
and neurodegeneration
9
. Activated protein C (APC),
for example, exerts pleiotropic beneficial activities including protection of the BBB integrity, anti-
inflammatory effects, direct neuroprotection, and pro-neurogenic and pro-angiogenic effects, as
shown in rodent models of stroke, traumatic brain injury, ALS and AD
566
. APC cleaves protease-
activated receptor-1 (PAR1) in brain endothelium and subsequently activates β-arrestin-2
dependent biased signaling pathway, which targets phosphatidylinositol 3-kinase (PI3K) for
cytoprotection and Rac1 GTPase for sealing the BBB
566
. 3K3A-APC, a recombinant variant of
APC with reduced anticoagulant activity, has advanced from bench to bedside, and has completed
phase 2 clinical trial for stroke (NCT02222714)
568
. Recently, 3K3A-APC was effective at
improving cerebrovascular integrity, preventing Ab deposits, and improving functional outcomes
in a mouse model of AD
505
. 3K3A-APC also holds promise for traumatic brain injury, ALS, AD
and possibly other neurodegenerative disorders
566
.
Inhibition of CypA-MMP9 BBB degrading pathway by cyclosporine, a CypA inhibitor,
restored BBB integrity and reversed secondary neurodegenerative changes in transgenic
humanized APOE4 mice
144
. Activation of CypA-MMP9 pathway associated with BBB breakdown
has also been shown in human APOE4 carriers compared to noncarriers, as indicated by CSF
145

and postmortem BBB tissue
146
analyses. Whether a non-immunosuppressive cyclosporine analog
Debio 025, which is used in humans in phase III trial for hepatitis C (NCT01318694), can also
protect cerebrovascular integrity and improve cognitive impairment in human APOE4 carriers at
risk for AD remains worth exploring.
The BBB is a major clearance site for many brain-produced potentially toxic substances,
which is particularly important in maintaining brain Aβ homeostasis
6,7
. Targeting the BBB
clearance machinery in AD is an emerging therapeutic approach to shift the balance between Aβ

174
production and clearance. For example, LRP1 minigene delivery to the BBB by viral vectors
facilitates Aβ clearance and attenuates Aβ pathology
73
. On the other hand, blocking RAGE at the
BBB
124
effectively reduces Aβ re-entry into the brain, inhibits neuroinflammation and improves
CBF in a mouse model AD
124,125
. Based on better understanding of RAGE biology, a small
molecule RAGE inhibitor has advanced to phase 3 clinical trials in AD patients (NCT02080364).
The PICALM-dependent transcytotic machinery at the BBB can also be therapeutically targeted
by gene therapy to enhance Aβ clearance across the BBB
6
. Additionally, allopregnanolone
promotes Aβ and cholesterol clearance in preclinical studies in animal models
569
and is currently
in phase 1 clinical trials for MCI and early AD (NCT02221622).
RMT is part of the highly specialized transport system at the BBB, allowing the exchange
of certain macromolecules between circulating blood and brain ISF
6
. A properly functioning RMT
is highly selective due to specific interaction between ligands and their preferred receptors, as well
as the spatial distribution of the receptors (luminal vs. abluminal), which ensures exclusive entry
of essential peptides and proteins into the brain and effective clearance of toxic waste products
from brain to blood. Targeting the RMT systems offers a tremendous opportunity for CNS drug
delivery, especially at the receptor level by selecting or even engineering therapeutic ligands. In
fact, approaches such as molecular Trojan horses, bispecific or brain shuttle monoclonal antibodies
targeting the BBB RMT systems have shown great promises for BBB penetrance in preclinical
animal models
66
, including anti-TfR-BACE1 antibody in Alzheimer’s Tg2576 mice
570
and
nonhuman primates
557
or anti-TfR-Aβ antibody in PS2APP mouse model
556
.
Numerous clinical trials for BBB protection have been used in various acute and chronic
neurological conditions, as discussed in section above. Briefly, relating to BBB integrity, trials
aim to visualize BBB permeability using DCE-MRI in ischemic stroke (NCT00715533;
NCT02077582), MS (NCT01836055), and epileptic regions (NCT02531880; NCT00419874).
Additional trials relating to BBB function aim to prevent anti-epileptic drug resistance with
ABCB1 (P-gp) inhibitors in epilepsy (NCT02144792; NCT01126307; NCT01663545;
NCT00605254), and block leukocyte infiltration across the BBB with anti-VLA-4 antibodies
(NCT00859482) or anti-CD52 antibodies (NCT03193086) in relapsing remitting MS.
Lastly, temporal opening of the BBB can be achieved with transient up-rise of osmotic
pressure
571
or physically with focused ultrasound
563
, which is supposed to grant a short therapeutic
window for CNS drug delivery. However, the impact of BBB opening on the diseased brain has

175
yet to be carefully examined, especially the long-term effects. Particularly, clinical safety issues
associated with osmotic challenge limit its current use to rare and specific situations such as certain
brain tumors. In addition, due to the preclinical success of focused ultrasound with microbubble
for noninvasive, transient and targeted delivery of therapeutics through the BBB
563
and use in brain
tumors (NCT02343991; NCT01473485), this technology is rapidly moving towards clinical
testing for early AD (NCT02986932) and PD (NCT02347254; NCT02252380)
564
. It is
noteworthy, however, that focused ultrasound with microbubbles may induce brain inflammatory
responses that are comparable with ischemic injury or mild traumatic brain injury
572
. The exact
mechanism of action in neurodegenerative disorders and long-term side effects remain unclear at
present, and this approach should proceed with extreme caution or not at all.
7.4 Lessons Learned and Future Directions
A healthy brain needs a healthy vascular system to function properly. Recent research has
raised awareness about the role of BBB and cerebrovascular dysfunction in the pathophysiology
of different CNS diseases, particularly AD. Strides have also been made towards better
understanding the molecular and cellular measures of BBB (dys)function. At the same time, these
advances have uncovered gaps in our knowledge of vascular health and have provided us with a
roadmap to ask new questions that should be addressed by future studies.
BBB breakdown is a key pathogenic feature of neurodegenerative diseases (Appendix D).
Development of new molecular biomarkers of vascular injury and/or repair in CSF and blood, and
how they relate to other systemic and cell-specific biomarkers of the neurovascular unit including
astrocytes, neurons, oligodendrocytes, microglia and inflammatory biomarkers, and/or standard
disease biomarkers such as Ab and tau in AD
182
, will further advance our understanding of vascular
contributions to neurodegeneration and cognitive decline. Additionally, the development of
advanced neuroimaging techniques to interrogate regional changes in BBB integrity
178,179,347,348
,
CBF and hemodynamic responses
2
, enlarged perivascular spaces
363
, incidence and distribution of
microbleeds
361,362
, and relating these to the molecular biofluid measures, hold significant promise
for future neurovascular research in humans. Moreover, developing novel molecular ligands, for
example, ligands to visualize MMP activity at the BBB in vivo
573
and/or activity of other BBB
transporters, receptors and/or junctional proteins affected by disease process, will provide

176
important mechanistic insights into the role of the vascular system in neurodegeneration. How
genetics, vascular risk factors, environment and lifestyle influence cerebrovascular/BBB functions
during normal aging and disease, and how this relates to neurological disorders, are another
important focus for future studies.
Besides the key question what is the role of the vascular system in the pathogenesis of
neurodegenerative disorders, dementia and/or motor CNS changes, an emerging question is what
is the prognostic and diagnostic value of neurovascular imaging and molecular biofluid markers
in predicting neurodegenerative processes and cognitive decline? If CNS vascular changes drive
the initial pathophysiologic events contributing to onset and/or progression of neurodegeneration,
loss of brain connectivity, and neuronal injury and loss in complex neurodegenerative disorders,
the question persists will therapeutic targeting of the BBB arrest and/or reverse the course of
neurological disorder in humans as shown in some animal models
124,125,420,505,541,574
? We still have
limited knowledge about molecular mechanisms underlying cerebrovascular/BBB dysfunction in
complex human neurodegenerative disorders, and much of the mechanistic insights have been
gained from animal models of these disorders
6,9
. Future advances in molecular and imaging
biomarkers in humans will hopefully reveal untapped novel targets of disease-modifying
therapeutics for multiple neurodegenerative disorders.
BBB breakdown is also found in rare, inherited monogenic neurological disorders with a
primary genetic deficit in brain vascular cell types
95,96,575–580
, discussed in detail elsewhere
6,74
.
Albeit rare, the known genetic etiology of these diseases offers valuable insights into the causal
mechanisms between BBB genetic defects and neurodegeneration and neurological deficits.
Examples include mutations in an endothelial tight junction protein (OCLN) and endothelial-
specific transporters for glucose (GLUT1) and omega-3 fatty acid docosahexaenoic acid
(MFSD2A), as mentioned briefly above in Section 1.2.3 BBB Transport Systems. These genetic
diseases provide a direct link between BBB dysfunction and neurodegeneration, and raise the
question: what is the role of altered expression of BBB transporters, receptors, tight junction
proteins, active efflux systems and ion channels in common, sporadic neurodegenerative diseases
such as AD and others? To name one of the many examples: sporadic AD is characterized with an
early loss of BBB transporters including GLUT1 transporter for glucose
375,376,381,377,378
, which we
know leads to severe neurological phenotype in the GLUT1-deficiency syndrome
73,575
. This begs
the question whether changes in the BBB molecular make-up like what is seen with the GLUT1

177
transporter in AD are innocent bystanders in the disease process or may have a major influence on
the course of the disease, as in monogenic disorders, by synergistically promoting AD
pathophysiology? These important questions remain to be answered.
The questions posed above require an accurate and in-depth understanding of
transcriptomic and proteomic definitions of vessel-associated cell types, and the differences
between endothelial and mural cells in different brain regions and along the arterio-capillary-
venous axis. The transcriptional atlas of cerebral blood vessels and the BBB is being elucidated in
animal models
18,15,20,581
, however we lack the molecular definition of the human brain vasculature,
and perhaps NVU, to generate an atlas of blood vessels in the human brain during health and
disease. Comparative knowledge of molecular makeup between humans and animals is essential
to bridge the translational gap, and for future studies to take advantage of discoveries in animal
models. This has the potential to reveal why regional changes in the brain vasculature may lead to
disease-specific neurological phenotypes in different neurodegenerative diseases and inform gene
networks and upstream regulators driving the link between cerebrovascular dysfunction and
neurodegeneration. Moreover, using stem cell technology to develop in vitro human BBB models
derived from induced pluripotent stem cells (iPSCs) from patients with different
neurodegenerative disorders carrying genetic risk and from those with sporadic forms of disease
will advance drug discovery to stabilize vascular function in neurodegenerative disorders and/or
develop new drug delivery approaches targeting the BBB. Going forward, the BBB should be
regarded as an important therapeutic opportunity, in combination with other approaches, to
prevent, arrest and/or ultimately reverse the neurodegenerative process and clinical deficits.

178
APPENDIX A:
HUMAN BRAIN PERICYTES SHED SOLUBLE PDGFRb

Adapted from:
Sagare AP, Sweeney MD…Zlokovic BV, Neuroscience Letters, 2015

Specific Contributions to Data Presented in Appendix A:
Dr. Abhay P. Sagare (Assistant Professor of Research in Physiology & Neuroscience at
USC) performed the cell lysate blots and I performed the quantitative reverse transcription
polymerase chain reaction (qRT-PCR) in Figure A.1. For data in Figure A.2, I assisted Dr. Abhay
P. Sagare with the in vitro cell culture experiments with hypoxia and Ab treatment and running
the quantitative Western blots.

Figure A.1 Abundant PDGFRb expression in human brain pericytes compared to other
vascular cell types.
(a,b) Immunoblotting for PDGFRb and b-actin (a) and relative abundance of PDGFRb protein
levels compared to b-actin (b) in cultured primary human brain endothelial cells, brain arterial
vascular smooth muscle cells (SMCs), and brain pericytes by Western blot analysis. (c)
Quantitative real-time polymerase chain reaction (qRT-PCR) analysis of PDGFRb mRNA levels
in SMCs and pericytes. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA was used
as an internal control. Means ± SEM from 6 independent cultures from 6 donors in triplicate.
Significance by one-way ANOVA with Tukey posthoc test (panel b) and two-tailed Student’s t-
test (panel c) both at a=0.05.

179

Figure A.2 Hypoxia and Ab peptide lead to shedding of sPDGFRb from human brain
pericytes, but not from SMCs.
(a,b) Immunoblotting for sPDGFRb (a) and quantification of sPDGFRb levels (ng/mL) by
quantitative Western blot analysis (b) in the culture medium from primary human vascular smooth
muscle cells (SMCs) and pericytes cultured under normoxic (21% O2) or hypoxic (1% O2)
conditions, or incubated with human synthetic Ab40 (25 µM) for 48 h. (c,d) Immunoblotting for
cell-associated PDGFRb (c) and relative abundance of cellular PDGFRb levels, (d) in primary
human SMCs and pericytes cultured under normoxic (21% O2) or hypoxic (1% O2) conditions, or
incubated with human synthetic Ab40 (25 µM) for 48 h. b-actin was used as an internal loading
control. Means ± SEM from 3 independent cultures from 3 donors in triplicates. In panels b and
d, significance in each cell type by one-way ANOVA with Tukey posthoc test, a=0.05.

180
APPENDIX B:
REGIONAL BBB PERMEABILITY BY MRI

Adapted from:
Nation DA*, Sweeney MD*, Montagne A*…Zlokovic BV, Nature Medicine, 2019

Specific Contributions to Data Presented in Appendix B:
Participants’ regional Ktrans BBB permeability was determined from DCE-MRI scans by
Dr. Axel Montagne (Assistant Professor of Research in Physiology & Neuroscience and Assistant
Director of the Functional Biological Imaging Core at USC). Hippocampal and parahippocampal
volumes were determined from MRI scans by Dr. Axel Montagne and Dr. Farshid Sepehrband
(Assistant Professor at Stevens Neuroimaging and Informatics Institute at USC). I measured CSF
Ab1-42 and pTau that are used in all analyses below.

Figure B.1 BBB breakdown in the (para)hippocampus with increased CDR is independent
of Aβ and tau status and regional volume.
(a) Quantification of BBB Ktrans values in the hippocampus (HC), parahippocampus (PHC), and
CA1, CA3 and dentate gyrus (DG) hippocampus subfields in CDR 0 individuals that are Aβ-
(n=24) or Aβ+ (n=20) and CDR 0.5 participants that are Aβ- (n=11) or Aβ+ (n=12). (b)
Quantification of BBB Ktrans values in the HC, PHC, and CA1, CA3 and DG hippocampus
subfields in individuals with CDR 0 that are pTau- (n=32) or pTau+ (n=12), and with CDR 0.5
that are pTau- (n=14) or pTau+ (n=8). (c) Regional Ktrans values controlled for CSF Aβ and pTau
levels in CDR 0 (n=44) and CDR 0.5 (n=23) individuals. (d,e) BBB Ktrans values in the HC and
PHC controlled for the respective HC or PHC volume in individuals [continued on the next page]

181
that are Aβ- (n=23) or Aβ+ (n=19) CDR 0 and Aβ- (n=11) or Aβ+ (n=9) CDR 0.5 (d), and pTau-
(n=30) or pTau+ (n=12) CDR 0 and pTau- (n=13) or pTau+ (n=7) CDR 0.5 (e). Panels a,b: Box-
and-whisker plot lines indicate median values, boxes indicate interquartile range and whiskers
indicate minimum and maximum values. Panels c-e: Estimated marginal means ± SEM from
ANCOVA models. For all analyses, significance after FDR correction from ANCOVAs with
Bonferroni posthoc comparisons.

Figure B.2 BBB breakdown in the (para)hippocampus in individuals with increased
cognitive domain impairment is independent of Aβ and tau status and regional volume.
(a) Quantification of BBB Ktrans values in the hippocampus (HC), parahippocampus (PHC), and
CA1, CA3 and dentate gyrus (DG) hippocampus subfields in individuals with no cognitive
domains impaired that are Aβ- (n=25) or Aβ+ (n=20) and with one or more cognitive domains
impaired that are Aβ- (n=12) or Aβ+ (n=13). (b) Quantification of BBB Ktrans values in the HC,
PHC, and CA1, CA3 and DG hippocampus subfields in individuals with no cognitive domains
impaired that are pTau- (n=33) or pTau+ (n=12) and with one or more cognitive domains impaired
that are pTau- (n=15) or pTau+ (n=9). (c) Regional Ktrans values controlled for CSF Aβ and pTau
levels in individuals with 0 domains (n=45) and 1+ domains (n=22) impaired. (d,e) BBB Ktrans
values in the HC and PHC controlled for the respective HC or PHC volume in participants that are
Aβ- (n=24) or Aβ+ (n=18) 0 domains impaired and Aβ- (n=12) or Aβ+ (n=10) 1+ domains
impaired (d), and pTau- (n=30) or pTau+ (n=12) 0 domains impaired and pTau- (n=13) or pTau+
(n=7) 1+ domains impaired (e). Panels a,b: Box-and-whisker plot lines indicate median values,
boxes indicate interquartile range and whiskers indicate minimum and maximum values. Panels c-
e: Estimated marginal means ± SEM from ANCOVA. For all analyses, significance after FDR
correction from ANCOVAs with Bonferroni posthoc comparisons.

182
APPENDIX C:
REGIONAL BBB PERMEABILITY BY MRI IN APOE4 CARRIERS

Adapted from:
Sweeney MD*, Montagne A*, Nation DA*…Zlokovic BV, In Preparation

Specific Contributions to Data Presented in Appendix C:
Participants’ regional Ktrans BBB permeability was determined from DCE-MRI scans by
Dr. Axel Montagne (Assistant Professor of Research in Physiology & Neuroscience and Assistant
Director of the Functional Biological Imaging Core at USC). I measured CSF Ab1-42 and pTau and
performed all statistical analyses.

Table C.1 Participants’ demographic information for the MRI cohort.

183

Figure C.1 BBB Ktrans in the (para)hippocampus increases in APOE4 carriers, independent
of Aβ and tau.
(a) Blood-brain barrier (BBB) Ktrans permeability in the hippocampus (HC) in individuals with
normal cognition that are APOE4 noncarriers (n=79, CDR 0; n=61, 0 domains impaired) and
APOE4 carriers (n=42, CDR 0; n=33, 0 domains impaired) and in individuals with early cognitive
impairment that are APOE4 noncarriers (n=13, CDR 0.5; n=16, 1+ domains impaired) and APOE4
carriers (n=19, CDR 0.5; n=20, 1+ domains impaired). (b) BBB Ktrans permeability in the HC
controlled for CSF Aβ1-42 and pTau levels in individuals with normal cognition that are APOE4
noncarriers (n=56, CDR 0; n=46, 0 domains impaired) and APOE4 carriers (n=33, CDR 0; n=24,
0 domains impaired) and individuals with early cognitive impairment that are APOE4 noncarriers
(n=12, CDR 0.5; n=15, 1+ domains impaired) and APOE4 carriers (n=14, CDR 0.5; n=19, 1+
domains impaired). (c) BBB Ktrans permeability in the parahippocampus (PHC) in individuals with
normal cognition that are APOE4 noncarriers (n=79, CDR 0; n=61, 0 domains impaired) and
APOE4 carriers (n=42, CDR 0; n=33, 0 domains impaired) and in individuals with early cognitive
impairment that are APOE4 noncarriers (n=13, CDR 0.5; n=16, 1+ domains impaired) and APOE4
carriers (n=19, CDR 0.5; n=20, 1+ domains impaired). (d) BBB Ktrans permeability in the PHC
controlled for CSF Aβ1-42 and pTau levels in individuals with normal cognition that are APOE4
noncarriers (n=56, CDR 0; n=46, 0 domains impaired) and APOE4 carriers (n=33, CDR 0; n=24,
0 domains impaired) and individuals with early cognitive impairment that are APOE4 noncarriers
(n=12, CDR 0.5; n=15, 1+ domains impaired) and APOE4 carriers (n=14, CDR 0.5; n=19, 1+
domains impaired). Panels a and c: Box-and-whisker plot lines indicate median values, boxes
indicate interquartile range and whiskers indicate minimum and maximum values. Panels b and d:
Estimated marginal means ± SEM from ANCOVA models. In all analyses, significance tests after
FDR correction from ANCOVAs with Bonferroni posthoc comparisons.

184
APPENDIX D:
BBB BREAKDOWN IN NEURODEGENERATIVE DISORDERS

Adapted from:
Sweeney MD…Zlokovic BV, Nature Reviews Neurology, 2018

Table D.1 BBB disruption by neuroimaging in neurodegenerative disorders.
Abbreviations: AD, Alzheimer’s disease; ALS, amyotrophic lateral sclerosis; BBB, blood-brain
barrier; DCE, dynamic contrast-enhanced; FDG, fluorodeoxyglucose; PET; positron emission
tomography; HD, Huntington disease; Ktrans, the regional CNS blood-brain barrier permeability
constant; MCI, mild cognitive impairment; MMP, matrix metalloproteinase; MS, multiple
sclerosis; PD, Parkinson disease; SWI, susceptibility-weighted imaging.
Disease Region and Details References
Increased leakage of gadolinium (DCE-MRI; K trans)
MCI Hippocampus
178,179,344–346,351

Early AD Several grey and white matter regions
179,344–346

PD Basal ganglia
582

HD Caudate nucleus
42

MS Perivascular growth of lesions in white matter regions
178,583–587

Microbleeds (T2*-weighted and SWI-MRI)
MCI 25% of patients
354,362

AD 45-78% of patients
291,354–362

PD Several deep and cortical grey matter regions and white matter
588,589

ALS Deep cortical layers
539

Diminished glucose transport (FDG-PET)
Normal
cognition with
AD genetic or
parental risk
Entorhinal cortex, hippocampus, posterior cingulate cortex and
whole brain
402,403,590

MCI
Precuneus, cingulate cortex and temporal cortex (prior to
conversion to AD)
375,400,591

Early AD
Hippocampus, parietal, temporal and cingulate cortex (prior to
development of atrophy and neurodegeneration)
370,375,400,401,591,592

Diminished P-glycoprotein function (verapamil-PET)
Mild AD
Parietotemporal, frontal and posterior cingulate cortices and
hippocampus
384,385

AD
Frontal, parietal, temporal and occipital cortices and posterior and
anterior cingulate
384

PD Mid-brain
543

CNS leukocyte infiltration (MMP inhibitor-PET)
MS Leukocyte infiltration (MMP activation) in lesions
573

185
Table D.2 BBB disruption by CSF analysis in neurodegenerative disorders.
Abbreviations: AD, Alzheimer’s disease; ALS, amyotrophic lateral sclerosis; CSF, cerebrospinal
fluid; IgG; immunoglobulin G; MCI, mild cognitive impairment; MS, multiple sclerosis; PD,
Parkinson disease; Qalb, ratio of CSF albumin to serum albumin. Note: Analytes in the table were
measured by enzyme-linked immunosorbent assay (ELISA).

Analyte Disease Details References
Albumin
Preclinical AD or
MCI
Increased Qalb
145,178,182

AD Increased or no change in Qalb
405–407,409

AD with vascular
risk factors
Increased Qalb
189,408,410

PD Increased Qalb
406,589,593,594

ALS Increased Qalb in 40% of patients
540,595

MS Increased Qalb
416,583

HIV-associated
dementia
Increased Qalb
596

Plasminogen
Preclinical AD or
MCI
Increased CSF levels of blood-derived proteins
(plasminogen)
418

Fibrinogen
Preclinical AD or
MCI
Increased CSF levels of blood-derived proteins
(fibrinogen)
417

IgG PD Increased CSF to serum levels of IgG
593

ALS
Increased CSF levels of blood-derived proteins
(IgG)
540

186

Table D.3 BBB disruption by post-mortem tissue analysis in neurodegenerative disorders.
Disease Region and Details References
Brain capillary leakages*
‡

AD
Accumulation of blood-derived fibrinogen, thrombin,
albumin, IgG and hemosiderin in the cortex and hippocampus
146,147,293,386–390

PD
Accumulation of blood-derived proteins in striatum:
fibrinogen, IgG and hemosiderin in the globus pallidus
549,550

HD Leakage of blood-derived proteins: fibrin in the putamen
42

ALS
Leakage of blood-derived proteins: fibrinogen, thrombin, IgG,
collagen type IV and iron-containing proteins
540,539,597

MS Leakage of blood-derived proteins: fibrinogen
598

Chronic
traumatic
encephalopathy
Perivascular hemosiderin-laden macrophages and histiocytes
599,600

Pericyte degeneration
AD*
‡§

Ultrastructural changes in the cortex: accumulation of
osmiophilic material; mitochondrial changes; pinocytosis and
loss of pericyte capillary coverage and numbers in the cortex
and hippocampus
146,389–392

ALS*
‡§

Pericyte loss in the medulla and reduced pericyte capillary
coverage and numbers in the cervical spinal cord
540,597

HIV-associated
dementia*
Loss of pericyte coverage in the frontal cortex
601

Chronic
traumatic
encephalopathy*
Mural cell mineralization in deep penetrating vessels
600

Endothelial degeneration
AD*
§

Microvascular reductions, reduced tight junction proteins and
capillary basement membrane changes in the cortex and
hippocampus
146,292,389,392,395,400

PD
§

Microvascular degeneration, reduced and disrupted tight
junctions and capillary basement membrane changes in the
subthalamic nucleus
550

HD
‡

Decreased and disrupted tight junction protein expression in
the putamen
42

ALS*
¥

Microvascular degeneration and intracellular vacuolization;
reduced or disrupted tight junctions, capillary basement
membrane changes and enlarged perivascular spaces in the
medulla, cervical spinal cord and lumbar spinal cord
597,602,603

MS* Decreased and disrupted tight junctions
598

HIV-associated
dementia*
Decreased numbers of or disrupted tight junctions
604,605

187
Abbreviations: Ab, amyloid-b; AD, Alzheimer’s disease; ALS, amyotrophic lateral sclerosis;
APOE4, apolipoprotein E e4 allele; GLUT1, solute carrier family 2, facilitated glucose transporter
member 1; HD, Huntington disease; IgG; immunoglobulin G; LRP1, low-density lipoprotein
receptor-related protein 1; MMP9, matrix metalloproteinase-9; MS, multiple sclerosis; PD,
Parkinson disease; RAGE, receptor for advanced glycation endproducts. Note: *Detected by
immunohistochemistry.
‡
Detected by immunoblotting.
§
Detected by electron microscopy.
¥
Detected by quantitative reverse transcription polymerase chain reaction.

Cellular infiltration*
AD
Extravasation of red blood cells, infiltration of peripheral
macrophages and neutrophils
147,386,398,399

PD Red blood cell extravasation in striatum
549

ALS Red blood cell extravasation
540

MS Leukocyte infiltration
573

HIV-associated
dementia
Peripheral macrophages infiltration
604

Chronic
traumatic
encephalopathy
Lymphocyte infiltration in Virchow-Robin spaces
599

Aberrant angiogenesis*
PD
Increased endothelial cell number in the substantia nigra pars
compacta; increased angiogenic endothelial integrin avb3
expression in the substantia nigra pars compacta, locus
coeruleus and putamen
606,607

HD
Increased vessel density, particularly of capillaries in the
cortex, caudate or putamen and substantia nigra
42,608

Molecular changes
AD*
‡

Reduced GLUT1 levels (diminished brain glucose uptake);
reduced LRP1 levels (diminished Ab clearance); upregulation
of RAGE (increased Ab re-entry and neurovascular
inflammation); activation of the proinflammatory cyclophilin
A-MMP9 pathway in APOE4 carriers (blood-brain barrier
breakdown owing to degradation of tight junction and
basement membrane proteins); increased levels of angiogenic
proteins
113,115,121,122,124,146,367–
369,372,396

HIV-associated
dementia*
‡

Reduced P-glycoprotein expression
544

HD*
§

Huntingtin protein aggregation in endothelial cells,
perivascular macrophages, vascular smooth muscle cells and
vascular basal lamina in the cortex and putamen
42,609

188

Figure D.1 BBB breakdown promotes neurodegeneration.
BBB breakdown is characterized by pericyte and endothelial degeneration with loss of tight and
adherens junctions and increased bulk flow transcytosis. BBB breakdown leads to brain entry of
microbial pathogens, accumulation of neurotoxic material, faulty BBB transport, red blood cell
(RBC) extravasation and the release of neurotoxic Fe
2+
, which generates reactive oxygen species
(ROS) and oxidative stress. Inflammatory and immune responses lead to the generation of
autoantibodies. Dashed line: P-glycoprotein has an indirect role in Aβ accumulation mediated by
Aβ efflux at the luminal endothelium. Abbreviations: Aβ, amyloid-β; LRP1, low-density
lipoprotein receptor-related protein-1; RAGE, receptor for advanced glycosylation end products.

189
APPENDIX E:
BBB-BASED THERAPEUTIC APPROACHES

Adapted from:
Sweeney MD…Zlokovic BV, Physiological Reviews, 2019

Table E.1 Circumventing, protecting and traversing the BBB for treatments.
Therapeutics Mechanism Disease Animal model
Clinical
trials
Refs
Circumventing the BBB for CNS Drug Delivery
Alternative routes of administration
Cerliponase
Intracerebro-
ventricular
Batten’s disease
Multiple
species/models
FDA
approved
558

Spinraza Intrathecal Infantile SMA
Multiple
species/models
FDA
approved
559

Ziconited
peptide
Intrathecal Chronic pain
Multiple
species/models
FDA
approved
560

Insulin Intranasal Cognitive impairment
Multiple
species/models
Phase II/III
561

Leptin Intranasal Obesity
Multiple
species/models
Phase I
561

Oxytocin Intranasal Autism
Multiple
species/models
Phase II
561

Protecting a Damaged BBB
BBB sealing
APC and its
analogs
β-arrestin-
mediated PAR1-
biased
signaling
Stroke
Rodent stroke
models
(arterial
occlusion,
embolic
stroke)
Phase II
566

APC and its
analogs
β-arrestin-
mediated PAR1-
biased
signaling
ALS
SOD1 mutant
models
NA
566

190
Glucocorticoids
Upregulation of
intercellular
junctional
proteins,
suppression of
MMPs and
inflammation
Niemann-Pick disease,
Type C
NPC1 NA
610

Eliminating consequences of BBB breakdown
Ancrod
Depleting
fibrinogen
AD TgCRND8 NA
228

Ancrod
Depleting
fibrinogen
MS EAE NA
611

Deferoxamine Iron chelation ALS SOD1 (G93A) NA
541

Glutathione
monoethyl ester
Antioxidant ALS SOD1 (G93A) NA
541

APC and its
analogs
PI3K/Akt
mediated
neuroprotection,
endothelial
protection
Stroke
MCAO,
dMCAO,
embolic stroke
Phase II
566

APC and its
analogs
PI3K/Akt
mediated
neuroprotection,
endothelial
protection
ALS SOD1 mutants NA
566

Enhancing clearance functions
LRP1 minigene Improve efflux AD Tg2576 NA
541

RAGE inhibitor
(Azeliragon)
Reduce influx AD Tg2576 Phase III
124

Allopreg-
nanolone
Promoting Ab
and cholesterol
clearance
AD 3xTgAD Phase I
569

Cell therapy
Mesenchymal
stem cells
transplantation
Improve BBB
functions
CNS injuries
Rodent
experimental
models
NA
612–
614

Pericytes
transplantation
Improve BBB
functions
ALS SOD1
615

191

Other BBB-targeted clinical trials
DCE-MRI
Identifying and
tracking sites of
BBB
permeability
Ischemic stroke
Rodent
experimental
models
Observational
trials
616

DCE-MRI
Identifying and
tracking sites of
BBB
permeability
MS
Rodent
experimental
models
Observational
trials
617

DCE-MRI
Identifying and
tracking sites of
BBB
permeability
Epilepsy
Rodent
experimental
models
Phase I
618,619

P-gp inhibitor
Prevent anti-
epileptic drug
resistance
Epilepsy
Rodent
experimental
models
Phase II
620

Anti-VLA-4
humanized
monoclonal
antibody
(Natalizumab)
Block CNS
leukocyte
infiltration
Relapsing remitting
MS
Rodent
experimental
models
Phase IV
621

Anti-CD52
humanized
monoclonal
antibody
(Alemtuzumab)
Block CNS
leukocyte
infiltration
Relapsing remitting
MS
Rodent
experimental
models
FDA
approved
622

Traversing the BBB for CNS Drug Delivery
Direct opening of the BBB
Focused
ultrasound
Doxorubicin
delivery
Brain tumor
Multiple
species and
models
Phase I
623

Focused
ultrasound
To promote
therapeutic
delivery
AD Phase I
624

Focused
ultrasound
To promote
therapeutic
delivery
PD Phase I
624

Colloidal carriers
Nanoparticles
Entrap within or
covalently bind
to drugs
A broad spectrum of
CNS diseases
Multiple
species and
models
Phase 1
555

192
Exosome
Entrap within or
covalently bind
to drugs
A broad spectrum of
CNS diseases
Multiple
species and
models
NA
555

Carrier-medicated transport
L-DOPA
LAT-1 Large
amino acid
transporter
PD MPTP
FDA
approved
6

Receptor-mediated transport
Bispecific
antibodies
Anti-TfR-
BACE1
AD Tg2576
570

Bispecific
antibodies
Anti-TfR-Aβ AD PS2APP
556

Molecular
Trojan horses
L-iduronidase
fused with anti-
TfR
Mycopolysacharoidosis
I
Rhesus
Monkey
Phase II
625

Molecular
Trojan horses
Iduronate 2
sulfatase fused
with anti-IR
Mycopolysacharoidosis
II
Rhesus
Monkey
Phase I
625

Viral vectors and variants
Gene delivery
Brain tropic
AAV9 variants
PD
TgSNCA-
A53T mouse
NA
626

Abbreviations: AAV9, adeno-associated virus 9; AD, Alzheimer’s disease; ALS, amyotrophic
lateral sclerosis; APC, activated protein C; Aβ, amyloid-β; BBB, blood-brain barrier; CNS, central
nervous system; DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; EAE,
experimental autoimmune encephalomyelitis; FDA, Food and Drug Administration; L-DOPA,
levodopa; LRP1, low-density lipoprotein receptor-related protein-1; MMP, matrix
metalloproteinase; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; MS, multiple sclerosis;
NA, not applicable; PAR1, proteinase-activated receptor 1; PD, Parkinson’s disease; RAGE,
receptor for advanced glycation endproducts; SMA, spinal muscular atrophy; SOD1, superoxide
dismutase 1; TfR, transferrin receptor; VLA-4, very late antigen-4.

193
REFERENCES

1. Zlokovic, B. V. The blood-brain barrier in health and chronic neurodegenerative disorders.
Neuron 57, 178–201 (2008).
2. Kisler, K., Nelson, A. R., Montagne, A. & Zlokovic, B. V. Cerebral blood flow regulation
and neurovascular dysfunction in Alzheimer disease. Nat. Rev. Neurosci. 18, 419–434
(2017).
3. Iadecola, C. The pathobiology of vascular dementia. Neuron 80, 844–866 (2013).
4. Pardridge, W. M. Targeted delivery of protein and gene medicines through the blood-brain
barrier. Clin. Pharmacol. Ther. 97, 347–361 (2015).
5. Iadecola, C. The Neurovascular Unit Coming of Age: A Journey through Neurovascular
Coupling in Health and Disease. Neuron 96, 17–42 (2017).
6. Zhao, Z., Nelson, A. R., Betsholtz, C. & Zlokovic, B. V. Establishment and Dysfunction of
the Blood-Brain Barrier. Cell 163, 1064–1078 (2015).
7. Zlokovic, B. V. Neurovascular pathways to neurodegeneration in Alzheimer’s disease and
other disorders. Nat. Rev. Neurosci. 12, 723–738 (2011).
8. Sweeney, M. D., Ayyadurai, S. & Zlokovic, B. V. Pericytes of the neurovascular unit: key
functions and signaling pathways. Nat. Neurosci. 19, 771–783 (2016).
9. Montagne, A., Zhao, Z. & Zlokovic, B. Alzheimer’s disease: a matter of blood-brain barrier
dysfunction? J Exp Med 214, 3151–3169 (2017).
10. Mann, G. E., Zlokovic, B. V. & Yudilevich, D. L. Evidence for a lactate transport system in
the sarcolemmal membrane of the perfused rabbit heart: kinetics of unidirectional influx,
carrier specificity and effects of glucagon. Biochim. Biophys. Acta 819, 241–248 (1985).
11. Armulik, A. et al. Pericytes regulate the blood-brain barrier. Nature 468, 557–561 (2010).
12. Bondjers, C. et al. Microarray analysis of blood microvessels from PDGF-B and PDGF-
Rbeta mutant mice identifies novel markers for brain pericytes. FASEB J. Off. Publ. Fed.
Am. Soc. Exp. Biol. 20, 1703–1705 (2006).
13. Calabria, A. R. & Shusta, E. V. A genomic comparison of in vivo and in vitro brain
microvascular endothelial cells. J. Cereb. Blood Flow Metab. Off. J. Int. Soc. Cereb. Blood
Flow Metab. 28, 135–148 (2008).

194
14. Daneman, R. et al. The mouse blood-brain barrier transcriptome: a new resource for
understanding the development and function of brain endothelial cells. PloS One 5, e13741
(2010).
15. He, L. et al. Analysis of the brain mural cell transcriptome. Sci. Rep. 6, 35108 (2016).
16. Li, J. Y., Boado, R. J. & Pardridge, W. M. Blood-brain barrier genomics. J. Cereb. Blood
Flow Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab. 21, 61–68 (2001).
17. Li, J. Y., Boado, R. J. & Pardridge, W. M. Rat blood-brain barrier genomics. II. J. Cereb.
Blood Flow Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab. 22, 1319–1326 (2002).
18. Vanlandewijck, M. et al. A molecular atlas of cell types and zonation in the brain vasculature.
Nature 554, 475–480 (2018).
19. Zeisel, A. et al. Brain structure. Cell types in the mouse cortex and hippocampus revealed by
single-cell RNA-seq. Science 347, 1138–1142 (2015).
20. Zhang, Y. et al. An RNA-sequencing transcriptome and splicing database of glia, neurons,
and vascular cells of the cerebral cortex. J. Neurosci. Off. J. Soc. Neurosci. 34, 11929–11947
(2014).
21. Iadecola, C. Neurovascular regulation in the normal brain and in Alzheimer’s disease. Nat.
Rev. Neurosci. 5, 347–360 (2004).
22. Reese, T. S. & Karnovsky, M. J. Fine structural localization of a blood-brain barrier to
exogenous peroxidase. J. Cell Biol. 34, 207–217 (1967).
23. Tietz, S. & Engelhardt, B. Brain barriers: Crosstalk between complex tight junctions and
adherens junctions. J. Cell Biol. 209, 493–506 (2015).
24. Siegenthaler, J. A., Sohet, F. & Daneman, R. ‘Sealing off the CNS’: cellular and molecular
regulation of blood-brain barriergenesis. Curr. Opin. Neurobiol. 23, 1057–1064 (2013).
25. Davson, H. Review lecture. The blood-brain barrier. J. Physiol. 255, 1–28 (1976).
26. Hartmann, D. A. et al. Pericyte structure and distribution in the cerebral cortex revealed by
high-resolution imaging of transgenic mice. Neurophotonics 2, 041402 (2015).
27. Nguyen, L. N. et al. Mfsd2a is a transporter for the essential omega-3 fatty acid
docosahexaenoic acid. Nature 509, 503–506 (2014).
28. Ben-Zvi, A. et al. Mfsd2a is critical for the formation and function of the blood-brain barrier.
Nature 509, 507–511 (2014).

195
29. Alarcon-Martinez, L. et al. Capillary pericytes express α-smooth muscle actin, which
requires prevention of filamentous-actin depolymerization for detection. eLife 7, (2018).
30. Dejana, E. & Vestweber, D. The role of VE-cadherin in vascular morphogenesis and
permeability control. Prog. Mol. Biol. Transl. Sci. 116, 119–144 (2013).
31. Vorbrodt, A. W. & Dobrogowska, D. H. Molecular anatomy of intercellular junctions in
brain endothelial and epithelial barriers: electron microscopist’s view. Brain Res. Brain Res.
Rev. 42, 221–242 (2003).
32. Giri, R. et al. beta-amyloid-induced migration of monocytes across human brain endothelial
cells involves RAGE and PECAM-1. Am. J. Physiol. Cell Physiol. 279, C1772-1781 (2000).
33. Turowski, P. et al. Phosphorylation of vascular endothelial cadherin controls lymphocyte
emigration. J. Cell Sci. 121, 29–37 (2008).
34. Allport, J. R., Muller, W. A. & Luscinskas, F. W. Monocytes induce reversible focal changes
in vascular endothelial cadherin complex during transendothelial migration under flow. J.
Cell Biol. 148, 203–216 (2000).
35. Wessel, F. et al. Leukocyte extravasation and vascular permeability are each controlled in
vivo by different tyrosine residues of VE-cadherin. Nat. Immunol. 15, 223–230 (2014).
36. Little, T. L., Beyer, E. C. & Duling, B. R. Connexin 43 and connexin 40 gap junctional
proteins are present in arteriolar smooth muscle and endothelium in vivo. Am. J. Physiol.
268, H729-739 (1995).
37. Nagasawa, K. et al. Possible involvement of gap junctions in the barrier function of tight
junctions of brain and lung endothelial cells. J. Cell. Physiol. 208, 123–132 (2006).
38. Traub, O. et al. Characterization of the gap junction protein connexin37 in murine
endothelium, respiratory epithelium, and after transfection in human HeLa cells. Eur. J. Cell
Biol. 77, 313–322 (1998).
39. Garrido-Urbani, S., Bradfield, P. F. & Imhof, B. A. Tight junction dynamics: the role of
junctional adhesion molecules (JAMs). Cell Tissue Res. 355, 701–715 (2014).
40. Lamagna, C. et al. Dual interaction of JAM-C with JAM-B and alpha(M)beta2 integrin:
function in junctional complexes and leukocyte adhesion. Mol. Biol. Cell 16, 4992–5003
(2005).
41. Nitta, T. et al. Size-selective loosening of the blood-brain barrier in claudin-5-deficient mice.
J. Cell Biol. 161, 653–660 (2003).

196
42. Drouin-Ouellet, J. et al. Cerebrovascular and blood-brain barrier impairments in
Huntington’s disease: Potential implications for its pathophysiology. Ann. Neurol. 78, 160–
177 (2015).
43. Sweeney, M. D., Sagare, A. P. & Zlokovic, B. V. Blood-brain barrier breakdown in
Alzheimer disease and other neurodegenerative disorders. Nat. Rev. Neurol. 14, 133–150
(2018).
44. Pfeiffer, F. et al. Claudin-1 induced sealing of blood-brain barrier tight junctions ameliorates
chronic experimental autoimmune encephalomyelitis. Acta Neuropathol. (Berl.) 122, 601–
614 (2011).
45. Saitou, M. et al. Complex phenotype of mice lacking occludin, a component of tight junction
strands. Mol. Biol. Cell 11, 4131–4142 (2000).
46. O’Driscoll, M. C. et al. Recessive mutations in the gene encoding the tight junction protein
occludin cause band-like calcification with simplified gyration and polymicrogyria. Am. J.
Hum. Genet. 87, 354–364 (2010).
47. Van Itallie, C. M. & Anderson, J. M. Architecture of tight junctions and principles of
molecular composition. Semin. Cell Dev. Biol. 36, 157–165 (2014).
48. Tornavaca, O. et al. ZO-1 controls endothelial adherens junctions, cell-cell tension,
angiogenesis, and barrier formation. J. Cell Biol. 208, 821–838 (2015).
49. Sohet, F. et al. LSR/angulin-1 is a tricellular tight junction protein involved in blood-brain
barrier formation. J. Cell Biol. 208, 703–711 (2015).
50. Ueda, H. et al. Immunolocalization of dystrobrevin in the astrocytic endfeet and endothelial
cells in the rat cerebellum. Neurosci. Lett. 283, 121–124 (2000).
51. Nico, B. et al. Severe alterations of endothelial and glial cells in the blood-brain barrier of
dystrophic mdx mice. Glia 42, 235–251 (2003).
52. Gerhardt, H., Wolburg, H. & Redies, C. N-cadherin mediates pericytic-endothelial
interaction during brain angiogenesis in the chicken. Dev. Dyn. Off. Publ. Am. Assoc. Anat.
218, 472–479 (2000).
53. Zheng, Y. et al. Angiomotin like-1 is a novel component of the N-cadherin complex affecting
endothelial/pericyte interaction in normal and tumor angiogenesis. Sci. Rep. 6, 30622 (2016).
54. Choudhary, M. et al. Tumor-induced loss of mural Connexin 43 gap junction activity
promotes endothelial proliferation. BMC Cancer 15, 427 (2015).

197
55. Hirschi, K. K., Burt, J. M., Hirschi, K. D. & Dai, C. Gap junction communication mediates
transforming growth factor-beta activation and endothelial-induced mural cell
differentiation. Circ. Res. 93, 429–437 (2003).
56. Larson, D. M., Wrobleski, M. J., Sagar, G. D., Westphale, E. M. & Beyer, E. C. Differential
regulation of connexin43 and connexin37 in endothelial cells by cell density, growth, and
TGF-beta1. Am. J. Physiol. 272, C405-415 (1997).
57. Winkler, E. A., Bell, R. D. & Zlokovic, B. V. Central nervous system pericytes in health and
disease. Nat. Neurosci. 14, 1398–1405 (2011).
58. Ezan, P. et al. Deletion of astroglial connexins weakens the blood-brain barrier. J. Cereb.
Blood Flow Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab. 32, 1457–1467 (2012).
59. Lutz, S. E., Raine, C. S. & Brosnan, C. F. Loss of astrocyte connexins 43 and 30 does not
significantly alter susceptibility or severity of acute experimental autoimmune
encephalomyelitis in mice. J. Neuroimmunol. 245, 8–14 (2012).
60. Yamamoto, T., Ochalski, A., Hertzberg, E. L. & Nagy, J. I. On the organization of astrocytic
gap junctions in rat brain as suggested by LM and EM immunohistochemistry of connexin43
expression. J. Comp. Neurol. 302, 853–883 (1990).
61. Boulay, A.-C. et al. Immune quiescence of the brain is set by astroglial connexin 43. J.
Neurosci. Off. J. Soc. Neurosci. 35, 4427–4439 (2015).
62. Baeten, K. M. & Akassoglou, K. Extracellular matrix and matrix receptors in blood-brain
barrier formation and stroke. Dev. Neurobiol. 71, 1018–1039 (2011).
63. Yamamoto, H. et al. Integrin β1 controls VE-cadherin localization and blood vessel stability.
Nat. Commun. 6, 6429 (2015).
64. Yao, Y., Chen, Z.-L., Norris, E. H. & Strickland, S. Astrocytic laminin regulates pericyte
differentiation and maintains blood brain barrier integrity. Nat. Commun. 5, (2014).
65. Menezes, M. J. et al. The extracellular matrix protein laminin α2 regulates the maturation
and function of the blood-brain barrier. J. Neurosci. Off. J. Soc. Neurosci. 34, 15260–15280
(2014).
66. Pardridge, W. M. Drug transport across the blood-brain barrier. J. Cereb. Blood Flow Metab.
Off. J. Int. Soc. Cereb. Blood Flow Metab. 32, 1959–1972 (2012).
67. Abbott, N. J., Patabendige, A. A. K., Dolman, D. E. M., Yusof, S. R. & Begley, D. J.
Structure and function of the blood-brain barrier. Neurobiol. Dis. 37, 13–25 (2010).
68. Daneman, R. & Prat, A. The blood-brain barrier. Cold Spring Harb. Perspect. Biol. 7,
a020412 (2015).

198
69. Pardridge, W. M. Blood-brain barrier endogenous transporters as therapeutic targets: a new
model for small molecule CNS drug discovery. Expert Opin. Ther. Targets 19, 1059–1072
(2015).
70. Alvarez, J. I. The Hedgehog pathway promotes blood-brain barrier integrity and CNS
immune quiescence. Science 334, 1727–1731 (2011).
71. Deng, D. et al. Crystal structure of the human glucose transporter GLUT1. Nature 510, 121–
125 (2014).
72. Mann, G. E., Yudilevich, D. L. & Sobrevia, L. Regulation of amino acid and glucose
transporters in endothelial and smooth muscle cells. Physiol. Rev. 83, 183–252 (2003).
73. Winkler, E. A. et al. GLUT1 reductions exacerbate Alzheimer’s disease vasculo-neuronal
dysfunction and degeneration. Nat. Neurosci. 18, 521–530 (2015).
74. Sweeney, M. D., Zhao, Z., Montagne, A., Nelson, A. R. & Zlokovic, B. V. Blood-Brain
Barrier: From Physiology to Disease and Back. Physiol. Rev. 99, 21–78 (2019).
75. Farrell, C. L. & Pardridge, W. M. Blood-brain barrier glucose transporter is asymmetrically
distributed on brain capillary endothelial lumenal and ablumenal membranes: an electron
microscopic immunogold study. Proc. Natl. Acad. Sci. U. S. A. 88, 5779–5783 (1991).
76. Simpson, I. A., Vannucci, S. J., DeJoseph, M. R. & Hawkins, R. A. Glucose transporter
asymmetries in the bovine blood-brain barrier. J. Biol. Chem. 276, 12725–12729 (2001).
77. Quistgaard, E. M., Löw, C., Moberg, P., Trésaugues, L. & Nordlund, P. Structural basis for
substrate transport in the GLUT-homology family of monosaccharide transporters. Nat.
Struct. Mol. Biol. 20, 766–768 (2013).
78. Sun, L. et al. Crystal structure of a bacterial homologue of glucose transporters GLUT1-4.
Nature 490, 361–366 (2012).
79. Elfeber, K. et al. Localization of the Na+-D-glucose cotransporter SGLT1 in the blood-brain
barrier. Histochem. Cell Biol. 121, 201–207 (2004).
80. Yu, A. S. et al. Regional distribution of SGLT activity in rat brain in vivo. Am. J. Physiol.
Cell Physiol. 304, C240-247 (2013).
81. O’Kane, R. L. & Hawkins, R. A. Na+-dependent transport of large neutral amino acids
occurs at the abluminal membrane of the blood-brain barrier. Am. J. Physiol. Endocrinol.
Metab. 285, E1167-1173 (2003).
82. Stoll, J., Wadhwani, K. C. & Smith, Q. R. Identification of the cationic amino acid transporter
(System y+) of the rat blood-brain barrier. J. Neurochem. 60, 1956–1959 (1993).

199
83. Hawkins, R. A., O’Kane, R. L., Simpson, I. A. & Viña, J. R. Structure of the blood-brain
barrier and its role in the transport of amino acids. J. Nutr. 136, 218S–26S (2006).
84. Lee, W. J., Hawkins, R. A., Viña, J. R. & Peterson, D. R. Glutamine transport by the blood-
brain barrier: a possible mechanism for nitrogen removal. Am. J. Physiol. 274, C1101-1107
(1998).
85. O’Kane, R. L., Martínez-López, I., DeJoseph, M. R., Viña, J. R. & Hawkins, R. A. Na(+)-
dependent glutamate transporters (EAAT1, EAAT2, and EAAT3) of the blood-brain barrier.
A mechanism for glutamate removal. J. Biol. Chem. 274, 31891–31895 (1999).
86. Oldendorf, W. H. & Szabo, J. Amino acid assignment to one of three blood-brain barrier
amino acid carriers. Am. J. Physiol. 230, 94–98 (1976).
87. Simpson, I. A., Carruthers, A. & Vannucci, S. J. Supply and demand in cerebral energy
metabolism: the role of nutrient transporters. J. Cereb. Blood Flow Metab. Off. J. Int. Soc.
Cereb. Blood Flow Metab. 27, 1766–1791 (2007).
88. Puchowicz, M. A. et al. Diet-induced ketosis increases capillary density without altered
blood flow in rat brain. Am. J. Physiol. Endocrinol. Metab. 292, E1607-1615 (2007).
89. Mayerl, S. et al. Transporters MCT8 and OATP1C1 maintain murine brain thyroid hormone
homeostasis. J. Clin. Invest. 124, 1987–1999 (2014).
90. Roberts, L. M. et al. Expression of the thyroid hormone transporters monocarboxylate
transporter-8 (SLC16A2) and organic ion transporter-14 (SLCO1C1) at the blood-brain
barrier. Endocrinology 149, 6251–6261 (2008).
91. Dumitrescu, A. M., Liao, X.-H., Best, T. B., Brockmann, K. & Refetoff, S. A novel syndrome
combining thyroid and neurological abnormalities is associated with mutations in a
monocarboxylate transporter gene. Am. J. Hum. Genet. 74, 168–175 (2004).
92. Friesema, E. C. H. et al. Association between mutations in a thyroid hormone transporter and
severe X-linked psychomotor retardation. Lancet Lond. Engl. 364, 1435–1437 (2004).
93. Capri, Y. et al. Relevance of different cellular models in determining the effects of mutations
on SLC16A2/MCT8 thyroid hormone transporter function and genotype-phenotype
correlation. Hum. Mutat. 34, 1018–1025 (2013).
94. Mitchell, R. W., On, N. H., Del Bigio, M. R., Miller, D. W. & Hatch, G. M. Fatty acid
transport protein expression in human brain and potential role in fatty acid transport across
human brain microvessel endothelial cells. J. Neurochem. 117, 735–746 (2011).
95. Alakbarzade, V. et al. A partially inactivating mutation in the sodium-dependent
lysophosphatidylcholine transporter MFSD2A causes a non-lethal microcephaly syndrome.
Nat. Genet. 47, 814–817 (2015).

200
96. Guemez-Gamboa, A. et al. Inactivating mutations in MFSD2A, required for omega-3 fatty
acid transport in brain, cause a lethal microcephaly syndrome. Nat. Genet. 47, 809–813
(2015).
97. Cass, C. E., Young, J. D. & Baldwin, S. A. Recent advances in the molecular biology of
nucleoside transporters of mammalian cells. Biochem. Cell Biol. Biochim. Biol. Cell. 76,
761–770 (1998).
98. Thompson, B. J. et al. Hypoxia/reoxygenation stress signals an increase in organic anion
transporting polypeptide 1a4 (Oatp1a4) at the blood-brain barrier: relevance to CNS drug
delivery. J. Cereb. Blood Flow Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab. 34, 699–
707 (2014).
99. Gao, B., Vavricka, S. R., Meier, P. J. & Stieger, B. Differential cellular expression of organic
anion transporting peptides OATP1A2 and OATP2B1 in the human retina and brain:
implications for carrier-mediated transport of neuropeptides and neurosteriods in the CNS.
Pflugers Arch. 467, 1481–1493 (2015).
100. Ganapathy, M. E. et al. beta-lactam antibiotics as substrates for OCTN2, an organic
cation/carnitine transporter. J. Biol. Chem. 275, 1699–1707 (2000).
101. Lin, C.-J. et al. Cellular localization of the organic cation transporters, OCT1 and OCT2, in
brain microvessel endothelial cells and its implication for MPTP transport across the blood-
brain barrier and MPTP-induced dopaminergic toxicity in rodents. J. Neurochem. 114, 717–
727 (2010).
102. Zlokovic, B. V., Begley, D. J. & Chain-Eliash, D. G. Blood-brain barrier permeability to
leucine-enkephalin, D-alanine2-D-leucine5-enkephalin and their N-terminal amino acid
(tyrosine). Brain Res. 336, 125–132 (1985).
103. Jefferies, W. A. et al. Transferrin receptor on endothelium of brain capillaries. Nature 312,
162–163 (1984).
104. Bray, N. Biologics: Transferrin’ bispecific antibodies across the blood-brain barrier. Nat.
Rev. Drug Discov. 14, 14–15 (2015).
105. Pardridge, W. M., Buciak, J. L. & Friden, P. M. Selective transport of an anti-transferrin
receptor antibody through the blood-brain barrier in vivo. J. Pharmacol. Exp. Ther. 259, 66–
70 (1991).
106. Pardridge, W. M. Molecular biology of the blood-brain barrier. Mol. Biotechnol. 30, 57–70
(2005).
107. Ito, S., Yanai, M., Yamaguchi, S., Couraud, P.-O. & Ohtsuki, S. Regulation of Tight-Junction
Integrity by Insulin in an In Vitro Model of Human Blood-Brain Barrier. J. Pharm. Sci. 106,
2599–2605 (2017).

201
108. Banks, W. A. Leptin transport across the blood-brain barrier: implications for the cause and
treatment of obesity. Curr. Pharm. Des. 7, 125–133 (2001).
109. Golden, P. L., Maccagnan, T. J. & Pardridge, W. M. Human blood-brain barrier leptin
receptor. Binding and endocytosis in isolated human brain microvessels. J. Clin. Invest. 99,
14–18 (1997).
110. Zlokovic, B. V. et al. Differential regulation of leptin transport by the choroid plexus and
blood-brain barrier and high affinity transport systems for entry into hypothalamus and
across the blood-cerebrospinal fluid barrier. Endocrinology 141, 1434–1441 (2000).
111. Zlokovic, B. V. et al. Kinetics of circulating vasopressin uptake by choroid plexus. Am. J.
Physiol. 260, F216-224 (1991).
112. Zlokovic, B. V. et al. Kinetics of arginine-vasopressin uptake at the blood-brain barrier.
Biochim. Biophys. Acta 1025, 191–198 (1990).
113. Deane, R. et al. LRP/amyloid beta-peptide interaction mediates differential brain efflux of
Abeta isoforms. Neuron 43, 333–344 (2004).
114. Méresse, S., Delbart, C., Fruchart, J. C. & Cecchelli, R. Low-density lipoprotein receptor on
endothelium of brain capillaries. J. Neurochem. 53, 340–345 (1989).
115. Shibata, M. et al. Clearance of Alzheimer’s amyloid-ss(1-40) peptide from brain by LDL
receptor-related protein-1 at the blood-brain barrier. J. Clin. Invest. 106, 1489–1499 (2000).
116. Urich, E., Lazic, S. E., Molnos, J., Wells, I. & Freskgård, P.-O. Transcriptional profiling of
human brain endothelial cells reveals key properties crucial for predictive in vitro blood-
brain barrier models. PloS One 7, e38149 (2012).
117. Bell, R. D. et al. Transport pathways for clearance of human Alzheimer’s amyloid beta-
peptide and apolipoproteins E and J in the mouse central nervous system. J. Cereb. Blood
Flow Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab. 27, 909–918 (2007).
118. Storck, S. E. et al. Endothelial LRP1 transports amyloid-β(1-42) across the blood-brain
barrier. J. Clin. Invest. 126, 123–136 (2016).
119. Zhao, Z. et al. Central role for PICALM in amyloid-β blood-brain barrier transcytosis and
clearance. Nat. Neurosci. 18, 978–987 (2015).
120. Deane, R. et al. apoE isoform-specific disruption of amyloid beta peptide clearance from
mouse brain. J. Clin. Invest. 118, 4002–4013 (2008).
121. Donahue, J. E. et al. RAGE, LRP-1, and amyloid-beta protein in Alzheimer’s disease. Acta
Neuropathol. (Berl.) 112, 405–415 (2006).

202
122. Miller, M. C. et al. Hippocampal RAGE immunoreactivity in early and advanced
Alzheimer’s disease. Brain Res. 1230, 273–280 (2008).
123. Zlokovic, B. V. et al. Glycoprotein 330/megalin: probable role in receptor-mediated transport
of apolipoprotein J alone and in a complex with Alzheimer disease amyloid beta at the blood-
brain and blood-cerebrospinal fluid barriers. Proc. Natl. Acad. Sci. U. S. A. 93, 4229–4234
(1996).
124. Deane, R. et al. RAGE mediates amyloid-beta peptide transport across the blood-brain
barrier and accumulation in brain. Nat. Med. 9, 907–913 (2003).
125. Deane, R. et al. A multimodal RAGE-specific inhibitor reduces amyloid β-mediated brain
disorder in a mouse model of Alzheimer disease. J. Clin. Invest. 122, 1377–1392 (2012).
126. Qosa, H., Miller, D. S., Pasinelli, P. & Trotti, D. Regulation of ABC efflux transporters at
blood-brain barrier in health and neurological disorders. Brain Res. 1628, 298–316 (2015).
127. Cirrito, J. R. et al. P-glycoprotein deficiency at the blood-brain barrier increases amyloid-
beta deposition in an Alzheimer disease mouse model. J. Clin. Invest. 115, 3285–3290
(2005).
128. Wang, W., Bodles-Brakhop, A. M. & Barger, S. W. A Role for P-Glycoprotein in Clearance
of Alzheimer Amyloid β -Peptide from the Brain. Curr. Alzheimer Res. 13, 615–620 (2016).
129. Taylor, C. J., Nicola, P. A., Wang, S., Barrand, M. A. & Hladky, S. B. Transporters involved
in regulation of intracellular pH in primary cultured rat brain endothelial cells. J. Physiol.
576, 769–785 (2006).
130. Funck, V. R. et al. Contrasting effects of Na+, K+-ATPase activation on seizure activity in
acute versus chronic models. Neuroscience 298, 171–179 (2015).
131. Vorbrodt, A. W. Ultrastructural cytochemistry of blood-brain barrier endothelia. Prog.
Histochem. Cytochem. 18, 1–99 (1988).
132. Foroutan, S., Brillault, J., Forbush, B. & O’Donnell, M. E. Moderate-to-severe ischemic
conditions increase activity and phosphorylation of the cerebral microvascular endothelial
cell Na+-K+-Cl- cotransporter. Am. J. Physiol. Cell Physiol. 289, C1492-1501 (2005).
133. O’Donnell, M. E., Lam, T. I., Tran, L. Q., Foroutan, S. & Anderson, S. E. Estradiol reduces
activity of the blood-brain barrier Na-K-Cl cotransporter and decreases edema formation in
permanent middle cerebral artery occlusion. J. Cereb. Blood Flow Metab. Off. J. Int. Soc.
Cereb. Blood Flow Metab. 26, 1234–1249 (2006).
134. Yuen, N. et al. Ischemic factor-induced increases in cerebral microvascular endothelial cell
Na/H exchange activity and abundance: evidence for involvement of ERK1/2 MAP kinase.
Am. J. Physiol. Cell Physiol. 306, C931-942 (2014).

203
135. Sonkusare, S. K. et al. Elementary Ca2+ signals through endothelial TRPV4 channels
regulate vascular function. Science 336, 597–601 (2012).
136. Balbuena, P., Li, W., Rzigalinski, B. A. & Ehrich, M. Malathion/oxon and lead acetate
increase gene expression and protein levels of transient receptor potential canonical channel
subunits TRPC1 and TRPC4 in rat endothelial cells of the blood-brain barrier. Int. J. Toxicol.
31, 238–249 (2012).
137. Hicks, K., O’Neil, R. G., Dubinsky, W. S. & Brown, R. C. TRPC-mediated actin-myosin
contraction is critical for BBB disruption following hypoxic stress. Am. J. Physiol. Cell
Physiol. 298, C1583-1593 (2010).
138. Longden, T. A. et al. Capillary K(+)-sensing initiates retrograde hyperpolarization to
increase local cerebral blood flow. Nat. Neurosci. 20, 717–726 (2017).
139. Millar, I. D., Wang, S., Brown, P. D., Barrand, M. A. & Hladky, S. B. Kv1 and Kir2
potassium channels are expressed in rat brain endothelial cells. Pflugers Arch. 456, 379–391
(2008).
140. Yamamura, H., Suzuki, Y., Yamamura, H., Asai, K. & Imaizumi, Y. Hypoxic stress up-
regulates Kir2.1 expression and facilitates cell proliferation in brain capillary endothelial
cells. Biochem. Biophys. Res. Commun. 476, 386–392 (2016).
141. Longden, T. A., Hill-Eubanks, D. C. & Nelson, M. T. Ion channel networks in the control of
cerebral blood flow. J. Cereb. Blood Flow Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab.
36, 492–512 (2016).
142. Sagare, A. P. et al. Pericyte loss influences Alzheimer-like neurodegeneration in mice. Nat.
Commun. 4, 2932 (2013).
143. Wilhelmus, M. M. M. et al. Lipoprotein receptor-related protein-1 mediates amyloid-beta-
mediated cell death of cerebrovascular cells. Am. J. Pathol. 171, 1989–1999 (2007).
144. Bell, R. D. et al. Apolipoprotein E controls cerebrovascular integrity via cyclophilin A.
Nature 485, 512–516 (2012).
145. Halliday, M. R. et al. Relationship between cyclophilin a levels and matrix metalloproteinase
9 activity in cerebrospinal fluid of cognitively normal apolipoprotein e4 carriers and blood-
brain barrier breakdown. JAMA Neurol. 70, 1198–1200 (2013).
146. Halliday, M. R. et al. Accelerated pericyte degeneration and blood-brain barrier breakdown
in apolipoprotein E4 carriers with Alzheimer’s disease. J. Cereb. Blood Flow Metab. Off. J.
Int. Soc. Cereb. Blood Flow Metab. 36, 216–227 (2016).
147. Hultman, K., Strickland, S. & Norris, E. H. The APOE ɛ4/ɛ4 genotype potentiates vascular
fibrin(ogen) deposition in amyloid-laden vessels in the brains of Alzheimer’s disease

204
patients. J. Cereb. Blood Flow Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab. 33, 1251–
1258 (2013).
148. Hamilton, N. B., Attwell, D. & Hall, C. N. Pericyte-mediated regulation of capillary
diameter: a component of neurovascular coupling in health and disease. Front.
Neuroenergetics 2, (2010).
149. Biesecker, K. R. et al. Glial Cell Calcium Signaling Mediates Capillary Regulation of Blood
Flow in the Retina. J. Neurosci. Off. J. Soc. Neurosci. 36, 9435–9445 (2016).
150. Fernández-Klett, F., Offenhauser, N., Dirnagl, U., Priller, J. & Lindauer, U. Pericytes in
capillaries are contractile in vivo, but arterioles mediate functional hyperemia in the mouse
brain. Proc. Natl. Acad. Sci. U. S. A. 107, 22290–22295 (2010).
151. Hall, C. N. et al. Capillary pericytes regulate cerebral blood flow in health and disease.
Nature 508, 55–60 (2014).
152. Kisler, K. et al. Pericyte degeneration leads to neurovascular uncoupling and limits oxygen
supply to brain. Nat. Neurosci. 20, 406–416 (2017).
153. Peppiatt, C. M., Howarth, C., Mobbs, P. & Attwell, D. Bidirectional control of CNS capillary
diameter by pericytes. Nature 443, 700–704 (2006).
154. Saito, S. & Ihara, M. Interaction between cerebrovascular disease and Alzheimer pathology.
Curr. Opin. Psychiatry 29, 168–173 (2016).
155. Tarasoff-Conway, J. M. et al. Clearance systems in the brain--implications for Alzheimer
diseaser. Nat. Rev. Neurol. 12, 248 (2016).
156. Bakker, E. N. T. P. et al. Lymphatic Clearance of the Brain: Perivascular, Paravascular and
Significance for Neurodegenerative Diseases. Cell. Mol. Neurobiol. 36, 181–194 (2016).
157. Bradbury, M. W., Cserr, H. F. & Westrop, R. J. Drainage of cerebral interstitial fluid into
deep cervical lymph of the rabbit. Am. J. Physiol. 240, F329-336 (1981).
158. Ichimura, T., Fraser, P. A. & Cserr, H. F. Distribution of extracellular tracers in perivascular
spaces of the rat brain. Brain Res. 545, 103–113 (1991).
159. Aspelund, A. et al. A dural lymphatic vascular system that drains brain interstitial fluid and
macromolecules. J. Exp. Med. 212, 991–999 (2015).
160. Engelhardt, B. et al. Vascular, glial, and lymphatic immune gateways of the central nervous
system. Acta Neuropathol. (Berl.) 132, 317–338 (2016).
161. Engelhardt, B., Vajkoczy, P. & Weller, R. O. The movers and shapers in immune privilege
of the CNS. Nat. Immunol. 18, 123–131 (2017).

205
162. Louveau, A. et al. Structural and functional features of central nervous system lymphatic
vessels. Nature 523, 337–341 (2015).
163. Absinta, M. et al. Human and nonhuman primate meninges harbor lymphatic vessels that can
be visualized noninvasively by MRI. eLife 6, (2017).
164. Da Mesquita, S. et al. Functional aspects of meningeal lymphatics in ageing and Alzheimer’s
disease. Nature 560, 185–191 (2018).
165. Sweeney, M. D. & Zlokovic, B. V. A lymphatic waste-disposal system implicated in
Alzheimer’s disease. Nature 560, 172–174 (2018).
166. Xie, L. et al. Sleep drives metabolite clearance from the adult brain. Science 342, 373–377
(2013).
167. Rennels, M. L., Gregory, T. F., Blaumanis, O. R., Fujimoto, K. & Grady, P. A. Evidence for
a ‘paravascular’ fluid circulation in the mammalian central nervous system, provided by the
rapid distribution of tracer protein throughout the brain from the subarachnoid space. Brain
Res. 326, 47–63 (1985).
168. Iliff, J. J. et al. A paravascular pathway facilitates CSF flow through the brain parenchyma
and the clearance of interstitial solutes, including amyloid β. Sci. Transl. Med. 4, 147ra111
(2012).
169. Jessen, N. A., Munk, A. S. F., Lundgaard, I. & Nedergaard, M. The Glymphatic System: A
Beginner’s Guide. Neurochem. Res. 40, 2583–2599 (2015).
170. Smith, A. J., Yao, X., Dix, J. A., Jin, B.-J. & Verkman, A. S. Test of the ‘glymphatic’
hypothesis demonstrates diffusive and aquaporin-4-independent solute transport in rodent
brain parenchyma. eLife 6, (2017).
171. Holter, K. E. et al. Interstitial solute transport in 3D reconstructed neuropil occurs by
diffusion rather than bulk flow. Proc. Natl. Acad. Sci. U. S. A. 114, 9894–9899 (2017).
172. Hladky, S. B. & Barrand, M. A. Mechanisms of fluid movement into, through and out of the
brain: evaluation of the evidence. Fluids Barriers CNS 11, 26 (2014).
173. Spector, R., Robert Snodgrass, S. & Johanson, C. E. A balanced view of the cerebrospinal
fluid composition and functions: Focus on adult humans. Exp. Neurol. 273, 57–68 (2015).
174. Asgari, N., Berg, C. T., Mørch, M. T., Khorooshi, R. & Owens, T. Cerebrospinal fluid
aquaporin-4-immunoglobulin G disrupts blood brain barrier. Ann. Clin. Transl. Neurol. 2,
857–863 (2015).
175. Asgari, M., de Zélicourt, D. & Kurtcuoglu, V. Glymphatic solute transport does not require
bulk flow. Sci. Rep. 6, 38635 (2016).

206
176. Jin, B.-J., Smith, A. J. & Verkman, A. S. Spatial model of convective solute transport in brain
extracellular space does not support a ‘glymphatic’ mechanism. J. Gen. Physiol. 148, 489–
501 (2016).
177. Hochwald, G. M., Wald, A. & Malhan, C. The sink action of cerebrospinal fluid volume
flow. Effect on brain water content. Arch. Neurol. 33, 339–344 (1976).
178. Montagne, A. et al. Blood-brain barrier breakdown in the aging human hippocampus. Neuron
85, 296–302 (2015).
179. Montagne, A. et al. Brain imaging of neurovascular dysfunction in Alzheimer’s disease. Acta
Neuropathol. (Berl.) 131, 687–707 (2016).
180. Nelson, A. R., Sweeney, M. D., Sagare, A. P. & Zlokovic, B. V. Neurovascular dysfunction
and neurodegeneration in dementia and Alzheimer’s disease. Biochim. Biophys. Acta 1862,
887–900 (2016).
181. Iturria-Medina, Y. et al. Early role of vascular dysregulation on late-onset Alzheimer’s
disease based on multifactorial data-driven analysis. Nat. Commun. 7, 11934 (2016).
182. Sweeney, M. D., Sagare, A. P. & Zlokovic, B. V. Cerebrospinal fluid biomarkers of
neurovascular dysfunction in mild dementia and Alzheimer’s disease. J. Cereb. Blood Flow
Metab. 35, 1055–1068 (2015).
183. Arvanitakis, Z., Capuano, A. W., Leurgans, S. E., Bennett, D. A. & Schneider, J. A. Relation
of cerebral vessel disease to Alzheimer’s disease dementia and cognitive function in elderly
people: a cross-sectional study. Lancet Neurol. 15, 934–943 (2016).
184. Toledo, J. B. et al. Contribution of cerebrovascular disease in autopsy confirmed
neurodegenerative disease cases in the National Alzheimer’s Coordinating Centre. Brain J.
Neurol. 136, 2697–2706 (2013).
185. Faraco, G. et al. Hypertension enhances Aβ-induced neurovascular dysfunction, promotes β-
secretase activity, and leads to amyloidogenic processing of APP. J. Cereb. Blood Flow
Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab. 36, 241–252 (2016).
186. Iadecola, C. et al. Impact of Hypertension on Cognitive Function: A Scientific Statement
From the American Heart Association. Hypertens. Dallas Tex 1979 68, e67–e94 (2016).
187. Barbagallo, M. & Dominguez, L. J. Type 2 diabetes mellitus and Alzheimer’s disease. World
J. Diabetes 5, 889–893 (2014).
188. Cheng, G., Huang, C., Deng, H. & Wang, H. Diabetes as a risk factor for dementia and mild
cognitive impairment: a meta-analysis of longitudinal studies. Intern. Med. J. 42, 484–491
(2012).

207
189. Bowman, G. L., Kaye, J. A. & Quinn, J. F. Dyslipidemia and blood-brain barrier integrity in
Alzheimer’s disease. Curr. Gerontol. Geriatr. Res. 2012, 184042 (2012).
190. Toledo, J. B. et al. Cardiovascular risk factors, cortisol, and amyloid-β deposition in
Alzheimer’s Disease Neuroimaging Initiative. Alzheimers Dement. J. Alzheimers Assoc. 8,
483–489 (2012).
191. Calderón-Garcidueñas, L. et al. Decreases in Short Term Memory, IQ, and Altered Brain
Metabolic Ratios in Urban Apolipoprotein ε4 Children Exposed to Air Pollution. J.
Alzheimers Dis. JAD 45, 757–770 (2015).
192. Calderón-Garcidueñas, L. & Villarreal-Ríos, R. Living close to heavy traffic roads, air
pollution, and dementia. Lancet Lond. Engl. 389, 675–677 (2017).
193. Calderón-Garcidueñas, L. et al. Air pollution and children: neural and tight junction
antibodies and combustion metals, the role of barrier breakdown and brain immunity in
neurodegeneration. J. Alzheimers Dis. JAD 43, 1039–1058 (2015).
194. González-Maciel, A., Reynoso-Robles, R., Torres-Jardón, R., Mukherjee, P. S. & Calderón-
Garcidueñas, L. Combustion-Derived Nanoparticles in Key Brain Target Cells and
Organelles in Young Urbanites: Culprit Hidden in Plain Sight in Alzheimer’s Disease
Development. J. Alzheimers Dis. JAD 59, 189–208 (2017).
195. Ishii, M. & Iadecola, C. Adipocyte-derived factors in age-related dementia and their
contribution to vascular and Alzheimer pathology. Biochim. Biophys. Acta 1862, 966–974
(2016).
196. Wheeler, M. J. et al. Sedentary behavior as a risk factor for cognitive decline? A focus on
the influence of glycemic control in brain health. Alzheimers Dement. N. Y. N 3, 291–300
(2017).
197. Basun, H. et al. Clinical and neuropathological features of the arctic APP gene mutation
causing early-onset Alzheimer disease. Arch. Neurol. 65, 499–505 (2008).
198. Filley, C. M. et al. The genetics of very early onset Alzheimer disease. Cogn. Behav. Neurol.
Off. J. Soc. Behav. Cogn. Neurol. 20, 149–156 (2007).
199. Karch, C. M., Cruchaga, C. & Goate, A. M. Alzheimer’s disease genetics: from the bench to
the clinic. Neuron 83, 11–26 (2014).
200. Karch, C. M. & Goate, A. M. Alzheimer’s disease risk genes and mechanisms of disease
pathogenesis. Biol. Psychiatry 77, 43–51 (2015).
201. Tanzi, R. E. The genetics of Alzheimer disease. Cold Spring Harb. Perspect. Med. 2, (2012).

208
202. Alzheimer’s Association. 2017 Alzheimer’s disease facts and figures. Alzheimers Dement.
J. Alzheimers Assoc. 13, (2017).
203. Holtzman, D. M., Herz, J. & Bu, G. Apolipoprotein E and apolipoprotein E receptors: normal
biology and roles in Alzheimer disease. Cold Spring Harb. Perspect. Med. 2, a006312 (2012).
204. Liu, C.-C., Liu, C.-C., Kanekiyo, T., Xu, H. & Bu, G. Apolipoprotein E and Alzheimer
disease: risk, mechanisms and therapy. Nat. Rev. Neurol. 9, 106–118 (2013).
205. Verghese, P. B., Castellano, J. M. & Holtzman, D. M. Apolipoprotein E in Alzheimer’s
disease and other neurological disorders. Lancet Neurol. 10, 241–252 (2011).
206. Zhao, N., Liu, C.-C., Qiao, W. & Bu, G. Apolipoprotein E, Receptors, and Modulation of
Alzheimer’s Disease. Biol. Psychiatry 83, 347–357 (2017).
207. Zlokovic, B. V. Cerebrovascular effects of apolipoprotein E: implications for Alzheimer
disease. JAMA Neurol. 70, 440–444 (2013).
208. Guerreiro, R. et al. TREM2 variants in Alzheimer’s disease. N. Engl. J. Med. 368, 117–127
(2013).
209. Harold, D. et al. Genome-wide association study identifies variants at CLU and PICALM
associated with Alzheimer’s disease. Nat. Genet. 41, 1088–1093 (2009).
210. Hollingworth, P. et al. Common variants at ABCA7, MS4A6A/MS4A4E, EPHA1, CD33
and CD2AP are associated with Alzheimer’s disease. Nat. Genet. 43, 429–435 (2011).
211. Jonsson, T. et al. Variant of TREM2 associated with the risk of Alzheimer’s disease. N. Engl.
J. Med. 368, 107–116 (2013).
212. Lambert, J. C. et al. Meta-analysis of 74,046 individuals identifies 11 new susceptibility loci
for Alzheimer’s disease. Nat. Genet. 45, 1452–1458 (2013).
213. Lambert, J.-C. et al. Genome-wide association study identifies variants at CLU and CR1
associated with Alzheimer’s disease. Nat. Genet. 41, 1094–1099 (2009).
214. Seshadri, S. et al. Genome-wide analysis of genetic loci associated with Alzheimer disease.
JAMA 303, 1832–1840 (2010).
215. Sims, R. et al. Rare coding variants in PLCG2, ABI3, and TREM2 implicate microglial-
mediated innate immunity in Alzheimer’s disease. Nat. Genet. 49, 1373–1384 (2017).
216. Bateman, R. J. et al. Autosomal-dominant Alzheimer’s disease: a review and proposal for
the prevention of Alzheimer’s disease. Alzheimers Res. Ther. 3, 1 (2011).

209
217. Grabowski, T. J., Cho, H. S., Vonsattel, J. P., Rebeck, G. W. & Greenberg, S. M. Novel
amyloid precursor protein mutation in an Iowa family with dementia and severe cerebral
amyloid angiopathy. Ann. Neurol. 49, 697–705 (2001).
218. Levy, E., Prelli, F. & Frangione, B. Studies on the first described Alzheimer’s disease
amyloid beta mutant, the Dutch variant. J. Alzheimers Dis. JAD 9, 329–339 (2006).
219. Zarranz, J. J. et al. Iowa APP mutation-related hereditary cerebral amyloid angiopathy
(CAA): A new family from Spain. J. Neurol. Sci. 363, 55–56 (2016).
220. Beckmann, N., Gérard, C., Abramowski, D., Cannet, C. & Staufenbiel, M. Noninvasive
magnetic resonance imaging detection of cerebral amyloid angiopathy-related microvascular
alterations using superparamagnetic iron oxide particles in APP transgenic mouse models of
Alzheimer’s disease: application to passive Abeta immunotherapy. J. Neurosci. Off. J. Soc.
Neurosci. 31, 1023–1031 (2011).
221. Bell, R. D. et al. SRF and myocardin regulate LRP-mediated amyloid-beta clearance in brain
vascular cells. Nat. Cell Biol. 11, 143–153 (2009).
222. Klohs, J. et al. Detection of cerebral microbleeds with quantitative susceptibility mapping in
the ArcAbeta mouse model of cerebral amyloidosis. J. Cereb. Blood Flow Metab. Off. J. Int.
Soc. Cereb. Blood Flow Metab. 31, 2282–2292 (2011).
223. Klohs, J. et al. Longitudinal Assessment of Amyloid Pathology in Transgenic ArcAβ Mice
Using Multi-Parametric Magnetic Resonance Imaging. PloS One 8, e66097 (2013).
224. Kruyer, A., Soplop, N., Strickland, S. & Norris, E. H. Chronic Hypertension Leads to
Neurodegeneration in the TgSwDI Mouse Model of Alzheimer’s Disease. Hypertens. Dallas
Tex 1979 66, 175–182 (2015).
225. Kumar-Singh, S. et al. Dense-core plaques in Tg2576 and PSAPP mouse models of
Alzheimer’s disease are centered on vessel walls. Am. J. Pathol. 167, 527–543 (2005).
226. Merlini, M., Meyer, E. P., Ulmann-Schuler, A. & Nitsch, R. M. Vascular β-amyloid and
early astrocyte alterations impair cerebrovascular function and cerebral metabolism in
transgenic arcAβ mice. Acta Neuropathol. (Berl.) 122, 293–311 (2011).
227. Park, L. et al. Brain and circulating levels of Aβ1-40 differentially contribute to vasomotor
dysfunction in the mouse brain. Stroke 44, 198–204 (2013).
228. Paul, J., Strickland, S. & Melchor, J. P. Fibrin deposition accelerates neurovascular damage
and neuroinflammation in mouse models of Alzheimer’s disease. J. Exp. Med. 204, 1999–
2008 (2007).
229. Sagare, A. et al. Clearance of amyloid-beta by circulating lipoprotein receptors. Nat. Med.
13, 1029–1031 (2007).

210
230. Tanifum, E. A., Starosolski, Z. A., Fowler, S. W., Jankowsky, J. L. & Annapragada, A. V.
Cerebral vascular leak in a mouse model of amyloid neuropathology. J. Cereb. Blood Flow
Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab. 34, 1646–1654 (2014).
231. Ujiie, M., Dickstein, D. L., Carlow, D. A. & Jefferies, W. A. Blood-brain barrier permeability
precedes senile plaque formation in an Alzheimer disease model. Microcirc. N. Y. N 1994
10, 463–470 (2003).
232. Hecht, M., Krämer, L. M., von Arnim, C. A. F., Otto, M. & Thal, D. R. Capillary cerebral
amyloid angiopathy in Alzheimer’s disease: association with allocortical/hippocampal
microinfarcts and cognitive decline. Acta Neuropathol. (Berl.) 135, 681–694 (2018).
233. Weller, R. O., Boche, D. & Nicoll, J. A. R. Microvasculature changes and cerebral amyloid
angiopathy in Alzheimer’s disease and their potential impact on therapy. Acta Neuropathol.
(Berl.) 118, 87–102 (2009).
234. Kumar-Singh, S. et al. Dense-core senile plaques in the Flemish variant of Alzheimer’s
disease are vasocentric. Am. J. Pathol. 161, 507–520 (2002).
235. Sellal, F. et al. APP Mutations in Cerebral Amyloid Angiopathy with or without Cortical
Calcifications: Report of Three Families and a Literature Review. J. Alzheimers Dis. JAD
56, 37–46 (2017).
236. Hawkes, C. A. et al. MK886 reduces cerebral amyloid angiopathy severity in TgCRND8
mice. Neurodegener. Dis. 13, 17–23 (2014).
237. Kamp, J. A. et al. Amyloid β in hereditary cerebral hemorrhage with amyloidosis-Dutch type.
Rev. Neurosci. 25, 641–651 (2014).
238. Natté, R. et al. Dementia in hereditary cerebral hemorrhage with amyloidosis-Dutch type is
associated with cerebral amyloid angiopathy but is independent of plaques and
neurofibrillary tangles. Ann. Neurol. 50, 765–772 (2001).
239. Monro, O. R. et al. Substitution at codon 22 reduces clearance of Alzheimer’s amyloid-beta
peptide from the cerebrospinal fluid and prevents its transport from the central nervous
system into blood. Neurobiol. Aging 23, 405–412 (2002).
240. Jonsson, T. et al. A mutation in APP protects against Alzheimer’s disease and age-related
cognitive decline. Nature 488, 96–99 (2012).
241. Kero, M. et al. Amyloid precursor protein (APP) A673T mutation in the elderly Finnish
population. Neurobiol. Aging 34, 1518.e1–3 (2013).
242. Peacock, M. L., Warren, J. T., Roses, A. D. & Fink, J. K. Novel polymorphism in the A4
region of the amyloid precursor protein gene in a patient without Alzheimer’s disease.
Neurology 43, 1254–1256 (1993).

211
243. Chen, Z.-L. et al. Depletion of coagulation factor XII ameliorates brain pathology and
cognitive impairment in Alzheimer disease mice. Blood 129, 2547–2556 (2017).
244. Domnitz, S. B. et al. Progression of cerebral amyloid angiopathy in transgenic mouse models
of Alzheimer disease. J. Neuropathol. Exp. Neurol. 64, 588–594 (2005).
245. Fryer, J. D. et al. Human apolipoprotein E4 alters the amyloid-beta 40:42 ratio and promotes
the formation of cerebral amyloid angiopathy in an amyloid precursor protein transgenic
model. J. Neurosci. Off. J. Soc. Neurosci. 25, 2803–2810 (2005).
246. Holtzman, D. M. et al. Apolipoprotein E facilitates neuritic and cerebrovascular plaque
formation in an Alzheimer’s disease model. Ann. Neurol. 47, 739–747 (2000).
247. Thakker, D. R. et al. Intracerebroventricular amyloid-beta antibodies reduce cerebral
amyloid angiopathy and associated micro-hemorrhages in aged Tg2576 mice. Proc. Natl.
Acad. Sci. U. S. A. 106, 4501–4506 (2009).
248. Webster, S. J., Bachstetter, A. D., Nelson, P. T., Schmitt, F. A. & Van Eldik, L. J. Using
mice to model Alzheimer’s dementia: an overview of the clinical disease and the preclinical
behavioral changes in 10 mouse models. Front. Genet. 5, 88 (2014).
249. Niwa, A. et al. Clinical and neuropathological findings in a patient with familial Alzheimer
disease showing a mutation in the PSEN1 gene. Neuropathol. Off. J. Jpn. Soc. Neuropathol.
33, 199–203 (2013).
250. Sun, L., Zhou, R., Yang, G. & Shi, Y. Analysis of 138 pathogenic mutations in presenilin-1
on the in vitro production of Aβ42 and Aβ40 peptides by γ-secretase. Proc. Natl. Acad. Sci.
U. S. A. 114, E476–E485 (2017).
251. Chávez-Gutiérrez, L. et al. The mechanism of γ-Secretase dysfunction in familial Alzheimer
disease. EMBO J. 31, 2261–2274 (2012).
252. Fernandez, M. A., Klutkowski, J. A., Freret, T. & Wolfe, M. S. Alzheimer presenilin-1
mutations dramatically reduce trimming of long amyloid β-peptides (Aβ) by γ-secretase to
increase 42-to-40-residue Aβ. J. Biol. Chem. 289, 31043–31052 (2014).
253. Quintero-Monzon, O. et al. Dissociation between the processivity and total activity of γ-
secretase: implications for the mechanism of Alzheimer’s disease-causing presenilin
mutations. Biochemistry 50, 9023–9035 (2011).
254. Szaruga, M. et al. Qualitative changes in human γ-secretase underlie familial Alzheimer’s
disease. J. Exp. Med. 212, 2003–2013 (2015).
255. Lemere, C. A. et al. The E280A presenilin 1 Alzheimer mutation produces increased A beta
42 deposition and severe cerebellar pathology. Nat. Med. 2, 1146–1150 (1996).

212
256. Dermaut, B. et al. Cerebral amyloid angiopathy is a pathogenic lesion in Alzheimer’s disease
due to a novel presenilin 1 mutation. Brain J. Neurol. 124, 2383–2392 (2001).
257. Szpak, G. M. et al. Small cerebral vessel disease in familial amyloid and non-amyloid
angiopathies: FAD-PS-1 (P117L) mutation and CADASIL. Immunohistochemical and
ultrastructural studies. Folia Neuropathol. 45, 192–204 (2007).
258. Ikeda, M. et al. The clinical phenotype of two missense mutations in the presenilin I gene in
Japanese patients. Ann. Neurol. 40, 912–917 (1996).
259. Singleton, A. B. et al. Pathology of early-onset Alzheimer’s disease cases bearing the
Thr113-114ins presenilin-1 mutation. Brain J. Neurol. 123 Pt 12, 2467–2474 (2000).
260. Wegiel, J. et al. Cell-type-specific enhancement of amyloid-beta deposition in a novel
presenilin-1 mutation (P117L). J. Neuropathol. Exp. Neurol. 57, 831–838 (1998).
261. Yasuda, M. et al. A novel missense mutation in the presenilin-1 gene in a familial
Alzheimer’s disease pedigree with abundant amyloid angiopathy. Neurosci. Lett. 232, 29–
32 (1997).
262. Fox, N. C. et al. Clinicopathological features of familial Alzheimer’s disease associated with
the M139V mutation in the presenilin 1 gene. Pedigree but not mutation specific age at onset
provides evidence for a further genetic factor. Brain J. Neurol. 120 ( Pt 3), 491–501 (1997).
263. Ishii, K. et al. Increased A beta 42(43)-plaque deposition in early-onset familial Alzheimer’s
disease brains with the deletion of exon 9 and the missense point mutation (H163R) in the
PS-1 gene. Neurosci. Lett. 228, 17–20 (1997).
264. Mann, D. M. et al. Amyloid beta protein (Abeta) deposition in chromosome 14-linked
Alzheimer’s disease: predominance of Abeta42(43). Ann. Neurol. 40, 149–156 (1996).
265. Ringman, J. M. et al. Neuropathology of Autosomal Dominant Alzheimer Disease in the
National Alzheimer Coordinating Center Database. J. Neuropathol. Exp. Neurol. 75, 284–
290 (2016).
266. Janssen, J. C. et al. Autopsy-confirmed familial early-onset Alzheimer disease caused by the
l153V presenilin 1 mutation. Arch. Neurol. 58, 953–958 (2001).
267. Armstrong, R. A. Spatial correlations between beta-amyloid (Abeta) deposits and blood
vessels in familial Alzheimer’s disease. Folia Neuropathol. 46, 241–248 (2008).
268. Bateman, R. J. et al. Clinical and biomarker changes in dominantly inherited Alzheimer’s
disease. N. Engl. J. Med. 367, 795–804 (2012).
269. Benzinger, T. L. S. et al. Regional variability of imaging biomarkers in autosomal dominant
Alzheimer’s disease. Proc. Natl. Acad. Sci. U. S. A. 110, E4502-4509 (2013).

213
270. Gama Sosa, M. A. et al. Age-related vascular pathology in transgenic mice expressing
presenilin 1-associated familial Alzheimer’s disease mutations. Am. J. Pathol. 176, 353–368
(2010).
271. Wen, P. H. et al. Selective expression of presenilin 1 in neural progenitor cells rescues the
cerebral hemorrhages and cortical lamination defects in presenilin 1-null mutant mice. Dev.
Camb. Engl. 132, 3873–3883 (2005).
272. Marín-Muñoz, J. et al. Novel Mutation (Gly212Val) in the PS2 Gene Associated with Early-
Onset Familial Alzheimer’s Disease. J. Alzheimers Dis. JAD 53, 73–78 (2016).
273. Canevelli, M. et al. Familial Alzheimer’s disease sustained by presenilin 2 mutations:
systematic review of literature and genotype-phenotype correlation. Neurosci. Biobehav.
Rev. 42, 170–179 (2014).
274. Potter, R. et al. Increased in vivo amyloid-β42 production, exchange, and loss in presenilin
mutation carriers. Sci. Transl. Med. 5, 189ra77 (2013).
275. Scheuner, D. et al. Secreted amyloid beta-protein similar to that in the senile plaques of
Alzheimer’s disease is increased in vivo by the presenilin 1 and 2 and APP mutations linked
to familial Alzheimer’s disease. Nat. Med. 2, 864–870 (1996).
276. Nochlin, D., Bird, T. D., Nemens, E. J., Ball, M. J. & Sumi, S. M. Amyloid angiopathy in a
Volga German family with Alzheimer’s disease and a presenilin-2 mutation (N141I). Ann.
Neurol. 43, 131–135 (1998).
277. Genin, E. et al. APOE and Alzheimer disease: a major gene with semi-dominant inheritance.
Mol. Psychiatry 16, 903–907 (2011).
278. Kim, J., Basak, J. M. & Holtzman, D. M. The role of apolipoprotein E in Alzheimer’s disease.
Neuron 63, 287–303 (2009).
279. Mahley, R. W. & Huang, Y. Apolipoprotein e sets the stage: response to injury triggers
neuropathology. Neuron 76, 871–885 (2012).
280. Mahley, R. W., Weisgraber, K. H. & Huang, Y. Apolipoprotein E: structure determines
function, from atherosclerosis to Alzheimer’s disease to AIDS. J. Lipid Res. 50 Suppl, S183-
188 (2009).
281. Bien-Ly, N. et al. C-terminal-truncated apolipoprotein (apo) E4 inefficiently clears amyloid-
beta (Abeta) and acts in concert with Abeta to elicit neuronal and behavioral deficits in mice.
Proc. Natl. Acad. Sci. U. S. A. 108, 4236–4241 (2011).
282. Castellano, J. M. et al. Human apoE isoforms differentially regulate brain amyloid-β peptide
clearance. Sci. Transl. Med. 3, 89ra57 (2011).

214
283. Holtzman, D. M. et al. Apolipoprotein E isoform-dependent amyloid deposition and neuritic
degeneration in a mouse model of Alzheimer’s disease. Proc. Natl. Acad. Sci. U. S. A. 97,
2892–2897 (2000).
284. Hudry, E. et al. Gene transfer of human Apoe isoforms results in differential modulation of
amyloid deposition and neurotoxicity in mouse brain. Sci. Transl. Med. 5, 212ra161 (2013).
285. Irizarry, M. C. et al. Apolipoprotein E affects the amount, form, and anatomical distribution
of amyloid beta-peptide deposition in homozygous APP(V717F) transgenic mice. Acta
Neuropathol. (Berl.) 100, 451–458 (2000).
286. Ji, Y., Permanne, B., Sigurdsson, E. M., Holtzman, D. M. & Wisniewski, T. Amyloid
beta40/42 clearance across the blood-brain barrier following intra-ventricular injections in
wild-type, apoE knock-out and human apoE3 or E4 expressing transgenic mice. J.
Alzheimers Dis. JAD 3, 23–30 (2001).
287. Kim, J. et al. Haploinsufficiency of human APOE reduces amyloid deposition in a mouse
model of amyloid-β amyloidosis. J. Neurosci. Off. J. Soc. Neurosci. 31, 18007–18012
(2011).
288. Bales, K. R. et al. Apolipoprotein E is essential for amyloid deposition in the APP(V717F)
transgenic mouse model of Alzheimer’s disease. Proc. Natl. Acad. Sci. U. S. A. 96, 15233–
15238 (1999).
289. Sanan, D. A. et al. Apolipoprotein E associates with beta amyloid peptide of Alzheimer’s
disease to form novel monofibrils. Isoform apoE4 associates more efficiently than apoE3. J.
Clin. Invest. 94, 860–869 (1994).
290. Shi, Y. et al. ApoE4 markedly exacerbates tau-mediated neurodegeneration in a mouse
model of tauopathy. Nature 549, 523–527 (2017).
291. Zonneveld, H. I. et al. Prevalence of cortical superficial siderosis in a memory clinic
population. Neurology 82, 698–704 (2014).
292. Salloway, S. et al. Effect of APOE genotype on microvascular basement membrane in
Alzheimer’s disease. J. Neurol. Sci. 203–204, 183–187 (2002).
293. Zipser, B. D. et al. Microvascular injury and blood-brain barrier leakage in Alzheimer’s
disease. Neurobiol. Aging 28, 977–986 (2007).
294. Reiman, E. M. et al. Functional brain abnormalities in young adults at genetic risk for late-
onset Alzheimer’s dementia. Proc. Natl. Acad. Sci. U. S. A. 101, 284–289 (2004).
295. Sheline, Y. I. et al. APOE4 allele disrupts resting state fMRI connectivity in the absence of
amyloid plaques or decreased CSF Aβ42. J. Neurosci. Off. J. Soc. Neurosci. 30, 17035–
17040 (2010).

215
296. Thambisetty, M., Beason-Held, L., An, Y., Kraut, M. A. & Resnick, S. M. APOE epsilon4
genotype and longitudinal changes in cerebral blood flow in normal aging. Arch. Neurol. 67,
93–98 (2010).
297. Hajjar, I., Sorond, F. & Lipsitz, L. A. Apolipoprotein E, carbon dioxide vasoreactivity, and
cognition in older adults: effect of hypertension. J. Am. Geriatr. Soc. 63, 276–281 (2015).
298. Suri, S. et al. Reduced cerebrovascular reactivity in young adults carrying the APOE ε4
allele. Alzheimers Dement. J. Alzheimers Assoc. 11, 648-657.e1 (2015).
299. Ossenkoppele, R. et al. Differential effect of APOE genotype on amyloid load and glucose
metabolism in AD dementia. Neurology 80, 359–365 (2013).
300. Protas, H. D. et al. Posterior cingulate glucose metabolism, hippocampal glucose
metabolism, and hippocampal volume in cognitively normal, late-middle-aged persons at 3
levels of genetic risk for Alzheimer disease. JAMA Neurol. 70, 320–325 (2013).
301. Ruitenberg, A. et al. Cerebral hypoperfusion and clinical onset of dementia: the Rotterdam
Study. Ann. Neurol. 57, 789–794 (2005).
302. Snowdon, D. A. et al. Brain infarction and the clinical expression of Alzheimer disease. The
Nun Study. JAMA 277, 813–817 (1997).
303. Vermeer, S. E. et al. Silent brain infarcts and the risk of dementia and cognitive decline. N.
Engl. J. Med. 348, 1215–1222 (2003).
304. Nishitsuji, K., Hosono, T., Nakamura, T., Bu, G. & Michikawa, M. Apolipoprotein E
regulates the integrity of tight junctions in an isoform-dependent manner in an in vitro blood-
brain barrier model. J. Biol. Chem. 286, 17536–17542 (2011).
305. Cacciottolo, M. et al. The APOE4 allele shows opposite sex bias in microbleeds and
Alzheimer’s disease of humans and mice. Neurobiol. Aging 37, 47–57 (2016).
306. Alata, W., Ye, Y., St-Amour, I., Vandal, M. & Calon, F. Human apolipoprotein E ɛ4
expression impairs cerebral vascularization and blood-brain barrier function in mice. J.
Cereb. Blood Flow Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab. 35, 86–94 (2015).
307. Castillo-Gomez, E. et al. The brain as immunoprecipitator of serum autoantibodies against
N-Methyl-D-aspartate receptor subunit NR1. Ann. Neurol. 79, 144–151 (2016).
308. Di Cataldo, V. et al. Exercise Does Not Protect against Peripheral and Central Effects of a
High Cholesterol Diet Given Ad libitum in Old ApoE(-/-) Mice. Front. Physiol. 7, 453
(2016).

216
309. Fullerton, S. M., Shirman, G. A., Strittmatter, W. J. & Matthew, W. D. Impairment of the
blood-nerve and blood-brain barriers in apolipoprotein e knockout mice. Exp. Neurol. 169,
13–22 (2001).
310. Hafezi-Moghadam, A., Thomas, K. L. & Wagner, D. D. ApoE deficiency leads to a
progressive age-dependent blood-brain barrier leakage. Am. J. Physiol. Cell Physiol. 292,
C1256-1262 (2007).
311. Hammer, C. et al. Neuropsychiatric disease relevance of circulating anti-NMDA receptor
autoantibodies depends on blood-brain barrier integrity. Mol. Psychiatry 19, 1143–1149
(2014).
312. Methia, N. et al. ApoE deficiency compromises the blood brain barrier especially after injury.
Mol. Med. Camb. Mass 7, 810–815 (2001).
313. Mulder, M. et al. Apolipoprotein E protects against neuropathology induced by a high-fat
diet and maintains the integrity of the blood-brain barrier during aging. Lab. Investig. J. Tech.
Methods Pathol. 81, 953–960 (2001).
314. Soto, I. et al. APOE Stabilization by Exercise Prevents Aging Neurovascular Dysfunction
and Complement Induction. PLoS Biol. 13, e1002279 (2015).
315. Jiang, Q. et al. ApoE promotes the proteolytic degradation of Abeta. Neuron 58, 681–693
(2008).
316. Bruinsma, I. B. et al. Apolipoprotein E protects cultured pericytes and astrocytes from D-
Abeta(1-40)-mediated cell death. Brain Res. 1315, 169–180 (2010).
317. Casey, C. S. et al. Apolipoprotein E Inhibits Cerebrovascular Pericyte Mobility through a
RhoA Protein-mediated Pathway. J. Biol. Chem. 290, 14208–14217 (2015).
318. Wilhelmus, M. M. M. et al. Apolipoprotein E genotype regulates amyloid-beta cytotoxicity.
J. Neurosci. Off. J. Soc. Neurosci. 25, 3621–3627 (2005).
319. Xu, Q. et al. Profile and regulation of apolipoprotein E (ApoE) expression in the CNS in
mice with targeting of green fluorescent protein gene to the ApoE locus. J. Neurosci. Off. J.
Soc. Neurosci. 26, 4985–4994 (2006).
320. Carrasquillo, M. M. et al. Replication of CLU, CR1, and PICALM associations with
alzheimer disease. Arch. Neurol. 67, 961–964 (2010).
321. Corneveaux, J. J. et al. Association of CR1, CLU and PICALM with Alzheimer’s disease in
a cohort of clinically characterized and neuropathologically verified individuals. Hum. Mol.
Genet. 19, 3295–3301 (2010).

217
322. Huang, K.-L. et al. A common haplotype lowers PU.1 expression in myeloid cells and delays
onset of Alzheimer’s disease. Nat. Neurosci. 20, 1052–1061 (2017).
323. Jones, E. L. et al. Evidence that PICALM affects age at onset of Alzheimer’s dementia in
Down syndrome. Neurobiol. Aging 34, 2441.e1–5 (2013).
324. Thomas, R. S. et al. Decreasing the expression of PICALM reduces endocytosis and the
activity of β-secretase: implications for Alzheimer’s disease. BMC Neurosci. 17, 50 (2016).
325. Itoh, T. et al. Role of the ENTH domain in phosphatidylinositol-4,5-bisphosphate binding
and endocytosis. Science 291, 1047–1051 (2001).
326. Miller, M. A. Patterning with Diffusion Barriers. Dev. Cell 35, 395–396 (2015).
327. Miller, S. E. et al. The molecular basis for the endocytosis of small R-SNAREs by the clathrin
adaptor CALM. Cell 147, 1118–1131 (2011).
328. Moreau, K. et al. PICALM modulates autophagy activity and tau accumulation. Nat.
Commun. 5, 4998 (2014).
329. Treusch, S. et al. Functional links between Aβ toxicity, endocytic trafficking, and
Alzheimer’s disease risk factors in yeast. Science 334, 1241–1245 (2011).
330. Xiao, Q. et al. Role of phosphatidylinositol clathrin assembly lymphoid-myeloid leukemia
(PICALM) in intracellular amyloid precursor protein (APP) processing and amyloid plaque
pathogenesis. J. Biol. Chem. 287, 21279–21289 (2012).
331. Tian, Y., Chang, J. C., Fan, E. Y., Flajolet, M. & Greengard, P. Adaptor complex
AP2/PICALM, through interaction with LC3, targets Alzheimer’s APP-CTF for terminal
degradation via autophagy. Proc. Natl. Acad. Sci. U. S. A. 110, 17071–17076 (2013).
332. Calero, M. et al. Functional and structural properties of lipid-associated apolipoprotein J
(clusterin). Biochem. J. 344 Pt 2, 375–383 (1999).
333. Nuutinen, T., Suuronen, T., Kauppinen, A. & Salminen, A. Clusterin: a forgotten player in
Alzheimer’s disease. Brain Res. Rev. 61, 89–104 (2009).
334. Trougakos, I. P. et al. Intracellular clusterin inhibits mitochondrial apoptosis by suppressing
p53-activating stress signals and stabilizing the cytosolic Ku70-Bax protein complex. Clin.
Cancer Res. Off. J. Am. Assoc. Cancer Res. 15, 48–59 (2009).
335. Zhang, H. et al. Clusterin inhibits apoptosis by interacting with activated Bax. Nat. Cell Biol.
7, 909–915 (2005).
336. Rogaeva, E. et al. The neuronal sortilin-related receptor SORL1 is genetically associated
with Alzheimer disease. Nat. Genet. 39, 168–177 (2007).

218
337. Reitz, C. et al. Meta-analysis of the association between variants in SORL1 and Alzheimer
disease. Arch. Neurol. 68, 99–106 (2011).
338. Hermey, G., Sjøgaard, S. S., Petersen, C. M., Nykjaer, A. & Gliemann, J. Tumour necrosis
factor alpha-converting enzyme mediates ectodomain shedding of Vps10p-domain receptor
family members. Biochem. J. 395, 285–293 (2006).
339. Gliemann, J. et al. The mosaic receptor sorLA/LR11 binds components of the plasminogen-
activating system and platelet-derived growth factor-BB similarly to LRP1 (low-density
lipoprotein receptor-related protein), but mediates slow internalization of bound ligand.
Biochem. J. 381, 203–212 (2004).
340. Zlokovic, B. V., Deane, R., Sagare, A. P., Bell, R. D. & Winkler, E. A. Low-density
lipoprotein receptor-related protein-1: a serial clearance homeostatic mechanism controlling
Alzheimer’s amyloid β-peptide elimination from the brain. J. Neurochem. 115, 1077–1089
(2010).
341. Fonseca, M. I. et al. Analysis of the Putative Role of CR1 in Alzheimer’s Disease: Genetic
Association, Expression and Function. PloS One 11, e0149792 (2016).
342. Kuo, Y. M. et al. Amyloid-beta peptides interact with plasma proteins and erythrocytes:
implications for their quantitation in plasma. Biochem. Biophys. Res. Commun. 268, 750–
756 (2000).
343. Wang, Y. et al. TREM2 lipid sensing sustains the microglial response in an Alzheimer’s
disease model. Cell 160, 1061–1071 (2015).
344. van de Haar, H. J. et al. Blood-Brain Barrier Leakage in Patients with Early Alzheimer
Disease. Radiology 281, 527–535 (2016).
345. van de Haar, H. J. et al. Neurovascular unit impairment in early Alzheimer’s disease
measured with magnetic resonance imaging. Neurobiol. Aging 45, 190–196 (2016).
346. van de Haar, H. J. et al. Subtle blood-brain barrier leakage rate and spatial extent:
Considerations for dynamic contrast-enhanced MRI. Med. Phys. 44, 4112–4125 (2017).
347. Barnes, S. R. et al. ROCKETSHIP: a flexible and modular software tool for the planning,
processing and analysis of dynamic MRI studies. BMC Med. Imaging 15, 19 (2015).
348. Barnes, S. R. et al. Optimal acquisition and modeling parameters for accurate assessment of
low Ktrans blood-brain barrier permeability using dynamic contrast-enhanced MRI. Magn.
Reson. Med. 75, 1967–1977 (2016).
349. Patlak, C. S. & Blasberg, R. G. Graphical evaluation of blood-to-brain transfer constants
from multiple-time uptake data. Generalizations. J. Cereb. Blood Flow Metab. Off. J. Int.
Soc. Cereb. Blood Flow Metab. 5, 584–590 (1985).

219
350. Heye, A. K. et al. Tracer kinetic modelling for DCE-MRI quantification of subtle blood-
brain barrier permeability. NeuroImage 125, 446–455 (2016).
351. Wang, H., Golob, E. J. & Su, M.-Y. Vascular volume and blood-brain barrier permeability
measured by dynamic contrast enhanced MRI in hippocampus and cerebellum of patients
with MCI and normal controls. J. Magn. Reson. Imaging JMRI 24, 695–700 (2006).

352. Starr, J. M., Farrall, A. J., Armitage, P., McGurn, B. & Wardlaw, J. Blood-brain barrier
permeability in Alzheimer’s disease: a case-control MRI study. Psychiatry Res. 171, 232–
241 (2009).
353. Viswanathan, A. & Greenberg, S. M. Cerebral amyloid angiopathy in the elderly. Ann.
Neurol. 70, 871–880 (2011).
354. Yates, P. A. et al. Incidence of cerebral microbleeds in preclinical Alzheimer disease.
Neurology 82, 1266–1273 (2014).
355. Goos, J. D. C. et al. Patients with Alzheimer disease with multiple microbleeds: relation with
cerebrospinal fluid biomarkers and cognition. Stroke J. Cereb. Circ. 40, 3455–3460 (2009).
356. Uetani, H. et al. Prevalence and topography of small hypointense foci suggesting
microbleeds on 3T susceptibility-weighted imaging in various types of dementia. AJNR Am.
J. Neuroradiol. 34, 984–989 (2013).
357. Olazarán, J. et al. Pattern of and risk factors for brain microbleeds in neurodegenerative
dementia. Am. J. Alzheimers Dis. Other Demen. 29, 263–269 (2014).
358. Heringa, S. M. et al. Multiple microbleeds are related to cerebral network disruptions in
patients with early Alzheimer’s disease. J. Alzheimers Dis. JAD 38, 211–221 (2014).
359. Shams, S. et al. Cerebral microbleeds: different prevalence, topography, and risk factors
depending on dementia diagnosis—the Karolinska Imaging Dementia Study. AJNR Am. J.
Neuroradiol. 36, 661–666 (2015).
360. Poliakova, T., Levin, O., Arablinskiy, A., Vasenina, E. & Zerr, I. Cerebral microbleeds in
early Alzheimer’s disease. J. Neurol. 263, 1961–1968 (2016).
361. Brundel, M. et al. High prevalence of cerebral microbleeds at 7Tesla MRI in patients with
early Alzheimer’s disease. J. Alzheimers Dis. JAD 31, 259–263 (2012).
362. Shams, S. & Wahlund, L.-O. Cerebral microbleeds as a biomarker in Alzheimer’s disease?
A review in the field. Biomark. Med. 10, 9–18 (2016).

220
363. Wardlaw, J. M. et al. Neuroimaging standards for research into small vessel disease and its
contribution to ageing and neurodegeneration. Lancet Neurol. 12, 822–838 (2013).
364. Vernooij, M. W. et al. Prevalence and risk factors of cerebral microbleeds: the Rotterdam
Scan Study. Neurology 70, 1208–1214 (2008).
365. Kantarci, K. et al. Focal hemosiderin deposits and β-amyloid load in the ADNI cohort.
Alzheimers Dement. J. Alzheimers Assoc. 9, S116-123 (2013).
366. Jack, C. R. et al. NIA-AA Research Framework: Toward a biological definition of
Alzheimer’s disease. Alzheimers Dement. J. Alzheimers Assoc. 14, 535–562 (2018).
367. Cunnane, S. et al. Brain fuel metabolism, aging, and Alzheimer’s disease. Nutr. Burbank Los
Angel. Cty. Calif 27, 3–20 (2011).
368. Sokoloff, L. et al. The [14C]deoxyglucose method for the measurement of local cerebral
glucose utilization: theory, procedure, and normal values in the conscious and anesthetized
albino rat. J. Neurochem. 28, 897–916 (1977).
369. McDougal, D. B. et al. Use of nonradioactive 2-deoxyglucose to study compartmentation of
brain glucose metabolism and rapid regional changes in rate. Proc. Natl. Acad. Sci. U. S. A.
87, 1357–1361 (1990).
370. Rokka, J., Grönroos, T. J., Viljanen, T., Solin, O. & Haaparanta-Solin, M. HPLC and TLC
methods for analysis of [(18)F]FDG and its metabolites from biological samples. J.
Chromatogr. B Analyt. Technol. Biomed. Life. Sci. 1048, 140–149 (2017).
371. Sols, A. & Crane, R. K. Substrate specificity of brain hexokinase. J. Biol. Chem. 210, 581–
595 (1954).
372. Crane, R. K. & Sols, A. The non-competitive inhibition of brain hexokinase by glucose-6-
phosphate and related compounds. J. Biol. Chem. 210, 597–606 (1954).
373. Sweeney, M. D. et al. Vascular dysfunction-The disregarded partner of Alzheimer’s disease.
Alzheimers Dement. J. Alzheimers Assoc. 15, 158–167 (2019).
374. Sweeney, M. D., Kisler, K., Montagne, A., Toga, A. W. & Zlokovic, B. V. The role of brain
vasculature in neurodegenerative disorders. Nat. Neurosci. 21, 1318–1331 (2018).
375. Hunt, A. et al. Reduced cerebral glucose metabolism in patients at risk for Alzheimer’s
disease. Psychiatry Res. 155, 147–154 (2007).
376. Samuraki, M. et al. Partial volume effect-corrected FDG PET and grey matter volume loss
in patients with mild Alzheimer’s disease. Eur. J. Nucl. Med. Mol. Imaging 34, 1658–1669
(2007).

221
377. Mosconi, L. et al. Multicenter standardized 18F-FDG PET diagnosis of mild cognitive
impairment, Alzheimer’s disease, and other dementias. J. Nucl. Med. Off. Publ. Soc. Nucl.
Med. 49, 390–398 (2008).
378. Landau, S. M. et al. Associations between cognitive, functional, and FDG-PET measures of
decline in AD and MCI. Neurobiol. Aging 32, 1207–1218 (2011).
379. Mosconi, L. et al. Amyloid and metabolic positron emission tomography imaging of
cognitively normal adults with Alzheimer’s parents. Neurobiol. Aging 34, 22–34 (2013).
380. Landau, S. M. et al. Comparing predictors of conversion and decline in mild cognitive
impairment. Neurology 75, 230–238 (2010).
381. Mosconi, L. et al. Hypometabolism exceeds atrophy in presymptomatic early-onset familial
Alzheimer’s disease. J. Nucl. Med. Off. Publ. Soc. Nucl. Med. 47, 1778–1786 (2006).
382. Niwa, K., Kazama, K., Younkin, S. G., Carlson, G. A. & Iadecola, C. Alterations in cerebral
blood flow and glucose utilization in mice overexpressing the amyloid precursor protein.
Neurobiol. Dis. 9, 61–68 (2002).
383. McInerney, M. P., Short, J. L. & Nicolazzo, J. A. Neurovascular Alterations in Alzheimer’s
Disease: Transporter Expression Profiles and CNS Drug Access. AAPS J. 19, 940–956
(2017).
384. van Assema, D. M. E. et al. Blood-brain barrier P-glycoprotein function in Alzheimer’s
disease. Brain J. Neurol. 135, 181–189 (2012).
385. Deo, A. K. et al. Activity of P-Glycoprotein, a β-Amyloid Transporter at the Blood-Brain
Barrier, Is Compromised in Patients with Mild Alzheimer Disease. J. Nucl. Med. Off. Publ.
Soc. Nucl. Med. 55, 1106–1111 (2014).
386. Cullen, K. M., Kócsi, Z. & Stone, J. Pericapillary haem-rich deposits: evidence for
microhaemorrhages in aging human cerebral cortex. J. Cereb. Blood Flow Metab. Off. J. Int.
Soc. Cereb. Blood Flow Metab. 25, 1656–1667 (2005).
387. Ryu, J. K. & McLarnon, J. G. A leaky blood-brain barrier, fibrinogen infiltration and
microglial reactivity in inflamed Alzheimer’s disease brain. J. Cell. Mol. Med. 13, 2911–
2925 (2009).
388. Cortes-Canteli, M. et al. Fibrinogen and beta-amyloid association alters thrombosis and
fibrinolysis: a possible contributing factor to Alzheimer’s disease. Neuron 66, 695–709
(2010).
389. Sengillo, J. D. et al. Deficiency in mural vascular cells coincides with blood-brain barrier
disruption in Alzheimer’s disease. Brain Pathol. Zurich Switz. 23, 303–310 (2013).

222
390. Miners, J. S., Schulz, I. & Love, S. 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.
Off. J. Int. Soc. Cereb. Blood Flow Metab. 271678X17690761 (2017).
doi:10.1177/0271678X17690761
391. Farkas, E. & Luiten, P. G. Cerebral microvascular pathology in aging and Alzheimer’s
disease. Prog. Neurobiol. 64, 575–611 (2001).
392. Baloyannis, S. J. & Baloyannis, I. S. The vascular factor in Alzheimer’s disease: a study in
Golgi technique and electron microscopy. J. Neurol. Sci. 322, 117–121 (2012).
393. Montagne, A. et al. Pericyte degeneration causes white matter dysfunction in the mouse
central nervous system. Nat. Med. 24, 326–337 (2018).
394. Bailey, T. L., Rivara, C. B., Rocher, A. B. & Hof, P. R. The nature and effects of cortical
microvascular pathology in aging and Alzheimer’s disease. Neurol. Res. 26, 573–578 (2004).
395. Wu, Z. et al. Role of the MEOX2 homeobox gene in neurovascular dysfunction in Alzheimer
disease. Nat. Med. 11, 959–965 (2005).
396. Grammas, P., Tripathy, D., Sanchez, A., Yin, X. & Luo, J. Brain microvasculature and
hypoxia-related proteins in Alzheimer’s disease. Int. J. Clin. Exp. Pathol. 4, 616–627 (2011).
397. Bell, R. D. et al. Pericytes control key neurovascular functions and neuronal phenotype in
the adult brain and during brain aging. Neuron 68, 409–427 (2010).
398. Fiala, M. et al. Cyclooxygenase-2-positive macrophages infiltrate the Alzheimer’s disease
brain and damage the blood-brain barrier. Eur. J. Clin. Invest. 32, 360–371 (2002).
399. Zenaro, E. et al. Neutrophils promote Alzheimer’s disease-like pathology and cognitive
decline via LFA-1 integrin. Nat. Med. 21, 880–886 (2015).
400. Simpson, I. A., Chundu, K. R., Davies-Hill, T., Honer, W. G. & Davies, P. Decreased
concentrations of GLUT1 and GLUT3 glucose transporters in the brains of patients with
Alzheimer’s disease. Ann. Neurol. 35, 546–551 (1994).
401. Mooradian, A. D., Chung, H. C. & Shah, G. N. GLUT-1 expression in the cerebra of patients
with Alzheimer’s disease. Neurobiol. Aging 18, 469–474 (1997).
402. Kalaria, R. N. & Harik, S. I. Reduced glucose transporter at the blood-brain barrier and in
cerebral cortex in Alzheimer disease. J. Neurochem. 53, 1083–1088 (1989).
403. Horwood, N. & Davies, D. C. Immunolabelling of hippocampal microvessel glucose
transporter protein is reduced in Alzheimer’s disease. Virchows Arch. Int. J. Pathol. 425, 69–
72 (1994).

223
404. Conejero-Goldberg, C. et al. APOE2 enhances neuroprotection against Alzheimer’s disease
through multiple molecular mechanisms. Mol. Psychiatry 19, 1243–1250 (2014).
405. Skoog, I. et al. A population study on blood-brain barrier function in 85-year-olds: relation
to Alzheimer’s disease and vascular dementia. Neurology 50, 966–971 (1998).
406. Janelidze, S. et al. Increased blood-brain barrier permeability is associated with dementia and
diabetes but not amyloid pathology or APOE genotype. Neurobiol. Aging 51, 104–112
(2017).
407. Skillbäck, T. et al. CSF/serum albumin ratio in dementias: a cross-sectional study on 1861
patients. Neurobiol. Aging 59, 1–9 (2017).
408. Blennow, K. et al. Blood-brain barrier disturbance in patients with Alzheimer’s disease is
related to vascular factors. Acta Neurol. Scand. 81, 323–326 (1990).
409. Wallin, A., Blennow, K. & Rosengren, L. Cerebrospinal fluid markers of pathogenetic
processes in vascular dementia, with special reference to the subcortical subtype. Alzheimer
Dis. Assoc. Disord. 13 Suppl 3, S102-105 (1999).
410. Blennow, K., Wallin, A., Uhlemann, C. & Gottfries, C. G. White-matter lesions on CT in
Alzheimer patients: relation to clinical symptomatology and vascular factors. Acta Neurol.
Scand. 83, 187–193 (1991).
411. Montine, T. J. et al. Recommendations of the Alzheimer’s Disease-Related Dementias
Conference. Neurology 83, 851–860 (2014).
412. Snyder, H. M. et al. Vascular contributions to cognitive impairment and dementia including
Alzheimer’s disease. Alzheimers Dement. J. Alzheimers Assoc. 11, 710–717 (2015).
413. Faraco, G. & Iadecola, C. Hypertension: a harbinger of stroke and dementia. Hypertens.
Dallas Tex 1979 62, 810–817 (2013).
414. Ivens, S. et al. TGF-beta receptor-mediated albumin uptake into astrocytes is involved in
neocortical epileptogenesis. Brain J. Neurol. 130, 535–547 (2007).
415. Braganza, O. et al. Albumin is taken up by hippocampal NG2 cells and astrocytes and
decreases gap junction coupling. Epilepsia 53, 1898–1906 (2012).
416. LeVine, S. M. Albumin and multiple sclerosis. BMC Neurol. 16, 47 (2016).
417. Craig-Schapiro, R. et al. Multiplexed immunoassay panel identifies novel CSF biomarkers
for Alzheimer’s disease diagnosis and prognosis. PloS One 6, e18850 (2011).

224
418. Hanzel, C. E. et al. Analysis of matrix metallo-proteases and the plasminogen system in mild
cognitive impairment and Alzheimer’s disease cerebrospinal fluid. J. Alzheimers Dis. JAD
40, 667–678 (2014).
419. Blair, L. J. et al. Tau depletion prevents progressive blood-brain barrier damage in a mouse
model of tauopathy. Acta Neuropathol. Commun. 3, 8 (2015).
420. Bell, R. D. et al. Apolipoprotein E controls cerebrovascular integrity via cyclophilin A.
Nature 485, 512–516 (2012).
421. Gottesman, R. F. et al. Association Between Midlife Vascular Risk Factors and Estimated
Brain Amyloid Deposition. JAMA 317, 1443–1450 (2017).
422. Bennett, R. E. et al. Tau induces blood vessel abnormalities and angiogenesis-related gene
expression in P301L transgenic mice and human Alzheimer’s disease. Proc. Natl. Acad. Sci.
U. S. A. 115, E1289–E1298 (2018).
423. Sagare, A. P., Sweeney, M. D., Makshanoff, J. & Zlokovic, B. V. Shedding of soluble
platelet-derived growth factor receptor-β from human brain pericytes. Neurosci. Lett. 607,
97–101 (2015).
424. Pan, C. et al. Diagnostic Values of Cerebrospinal Fluid T-Tau and Aβ₄₂ using Meso Scale
Discovery Assays for Alzheimer’s Disease. J. Alzheimers Dis. JAD 45, 709–719 (2015).
425. Roe, C. M. et al. Amyloid imaging and CSF biomarkers in predicting cognitive impairment
up to 7.5 years later. Neurology 80, 1784–1791 (2013).
426. Bondi, M. W. et al. Neuropsychological criteria for mild cognitive impairment improves
diagnostic precision, biomarker associations, and progression rates. J. Alzheimers Dis. JAD
42, 275–289 (2014).
427. Jak, A. J. et al. Quantification of five neuropsychological approaches to defining mild
cognitive impairment. Am. J. Geriatr. Psychiatry Off. J. Am. Assoc. Geriatr. Psychiatry 17,
368–375 (2009).
428. Jak, A. J. et al. Neuropsychological Criteria for Mild Cognitive Impairment and Dementia
Risk in the Framingham Heart Study. J. Int. Neuropsychol. Soc. JINS 22, 937–943 (2016).
429. Delis, D., Kramer, J., Kaplan, E. & Ober, B. California Verbal Learning Test. (PsychCorp,
2000).
430. Nation, D. A. et al. Antemortem pulse pressure elevation predicts cerebrovascular disease in
autopsy-confirmed Alzheimer’s disease. J. Alzheimers Dis. JAD 30, 595–603 (2012).

225
431. Bangen, K. J. et al. Aggregate effects of vascular risk factors on cerebrovascular changes in
autopsy-confirmed Alzheimer’s disease. Alzheimers Dement. J. Alzheimers Assoc. 11, 394-
403.e1 (2015).
432. Hölttä, M. et al. Evaluating amyloid-β oligomers in cerebrospinal fluid as a biomarker for
Alzheimer’s disease. PloS One 8, e66381 (2013).
433. Sengupta, U. et al. Tau oligomers in cerebrospinal fluid in Alzheimer’s disease. Ann. Clin.
Transl. Neurol. 4, 226–235 (2017).
434. Verbeek, M. M., Otte-Höller, I., Wesseling, P., Ruiter, D. J. & de Waal, R. M. Induction of
alpha-smooth muscle actin expression in cultured human brain pericytes by transforming
growth factor-beta 1. Am. J. Pathol. 144, 372–382 (1994).
435. Mendelson, K., Swendeman, S., Saftig, P. & Blobel, C. P. Stimulation of platelet-derived
growth factor receptor beta (PDGFRbeta) activates ADAM17 and promotes
metalloproteinase-dependent cross-talk between the PDGFRbeta and epidermal growth
factor receptor (EGFR) signaling pathways. J. Biol. Chem. 285, 25024–25032 (2010).
436. Fischl, B. FreeSurfer. NeuroImage 62, 774–781 (2012).
437. Fischl, B. et al. Whole brain segmentation: automated labeling of neuroanatomical structures
in the human brain. Neuron 33, 341–355 (2002).
438. Desikan, R. S. et al. An automated labeling system for subdividing the human cerebral cortex
on MRI scans into gyral based regions of interest. NeuroImage 31, 968–980 (2006).
439. Fischl, B. & Dale, A. M. Measuring the thickness of the human cerebral cortex from
magnetic resonance images. Proc. Natl. Acad. Sci. U. S. A. 97, 11050–11055 (2000).
440. Fischl, B., Sereno, M. I., Tootell, R. B. & Dale, A. M. High-resolution intersubject averaging
and a coordinate system for the cortical surface. Hum. Brain Mapp. 8, 272–284 (1999).
441. Dinov, I. et al. Neuroimaging study designs, computational analyses and data provenance
using the LONI pipeline. PloS One 5, (2010).
442. Sepehrband, F. et al. Neuroanatomical morphometric characterization of sex differences in
youth using statistical learning. NeuroImage 172, 217–227 (2018).
443. Su, Y. et al. Quantitative analysis of PiB-PET with FreeSurfer ROIs. PloS One 8, e73377
(2013).
444. Gordon, B. A. et al. Longitudinal β-Amyloid Deposition and Hippocampal Volume in
Preclinical Alzheimer Disease and Suspected Non-Alzheimer Disease Pathophysiology.
JAMA Neurol. 73, 1192–1200 (2016).

226
445. Glickman, M. E., Rao, S. R. & Schultz, M. R. False discovery rate control is a recommended
alternative to Bonferroni-type adjustments in health studies. J. Clin. Epidemiol. 67, 850–857
(2014).
446. Lindahl, P., Johansson, B. R., Levéen, P. & Betsholtz, C. Pericyte loss and microaneurysm
formation in PDGF-B-deficient mice. Science 277, 242–245 (1997).
447. Winkler, E. A., Bell, R. D. & Zlokovic, B. V. Pericyte-specific expression of PDGF beta
receptor in mouse models with normal and deficient PDGF beta receptor signaling. Mol.
Neurodegener. 5, 32 (2010).
448. Trost, A. et al. Brain and Retinal Pericytes: Origin, Function and Role. Front. Cell. Neurosci.
10, 20 (2016).
449. Daneman, R., Zhou, L., Kebede, A. A. & Barres, B. A. Pericytes are required for blood-brain
barrier integrity during embryogenesis. Nature 468, 562–566 (2010).
450. Spellman, T. et al. Hippocampal-prefrontal input supports spatial encoding in working
memory. Nature 522, 309–314 (2015).
451. Tan, Z. et al. Dynamic ErbB4 Activity in Hippocampal-Prefrontal Synchrony and Top-Down
Attention in Rodents. Neuron 98, 380-393.e4 (2018).
452. Croxson, P. L. et al. Quantitative investigation of connections of the prefrontal cortex in the
human and macaque using probabilistic diffusion tractography. J. Neurosci. Off. J. Soc.
Neurosci. 25, 8854–8866 (2005).
453. O’Shea, A., Cohen, R. A., Porges, E. C., Nissim, N. R. & Woods, A. J. Cognitive Aging and
the Hippocampus in Older Adults. Front. Aging Neurosci. 8, 298 (2016).
454. Glikmann-Johnston, Y., Oren, N., Hendler, T. & Shapira-Lichter, I. Distinct functional
connectivity of the hippocampus during semantic and phonemic fluency. Neuropsychologia
69, 39–49 (2015).
455. Rodríguez-Aranda, C. et al. Neuroanatomical correlates of verbal fluency in early
Alzheimer’s disease and normal aging. Brain Lang. 155–156, 24–35 (2016).
456. Drachman, D. A. The amyloid hypothesis, time to move on: Amyloid is the downstream
result, not cause, of Alzheimer’s disease. Alzheimers Dement. J. Alzheimers Assoc. 10, 372–
380 (2014).
457. Zlokovic, B. V. Neurovascular mechanisms of Alzheimer’s neurodegeneration. Trends
Neurosci. 28, 202–208 (2005).
458. Zlokovic, B. V. Neurodegeneration and the neurovascular unit. Nat. Med. 16, 1370–1371
(2010).

227
459. Grammas, P., Sanchez, A., Tripathy, D., Luo, E. & Martinez, J. Vascular signaling
abnormalities in Alzheimer disease. Cleve. Clin. J. Med. 78 Suppl 1, S50-53 (2011).
460. Kalaria, R. N., Akinyemi, R. & Ihara, M. Does vascular pathology contribute to Alzheimer
changes? J. Neurol. Sci. 322, 141–147 (2012).
461. Nation, D. A. et al. Blood-brain barrier breakdown is an early biomarker of human cognitive
dysfunction. Nat. Med. 25, 270–276 (2019).
462. Michels, L. et al. Arterial spin labeling imaging reveals widespread and Aβ-independent
reductions in cerebral blood flow in elderly apolipoprotein epsilon-4 carriers. J. Cereb. Blood
Flow Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab. 36, 581–595 (2016).
463. Liao, F., Yoon, H. & Kim, J. Apolipoprotein E metabolism and functions in brain and its role
in Alzheimer’s disease. Curr. Opin. Lipidol. 28, 60–67 (2017).
464. Huynh, T.-P. V., Davis, A. A., Ulrich, J. D. & Holtzman, D. M. Apolipoprotein E and
Alzheimer’s disease: the influence of apolipoprotein E on amyloid-β and other
amyloidogenic proteins. J. Lipid Res. 58, 824–836 (2017).
465. Janelidze, S. et al. CSF biomarkers of neuroinflammation and cerebrovascular dysfunction
in early Alzheimer disease. Neurology 91, e867–e877 (2018).
466. Ryman, D. C. et al. Symptom onset in autosomal dominant Alzheimer disease: a systematic
review and meta-analysis. Neurology 83, 253–260 (2014).
467. Folstein, M. F., Folstein, S. E. & McHugh, P. R. ‘Mini-mental state’. A practical method for
grading the cognitive state of patients for the clinician. J. Psychiatr. Res. 12, 189–198 (1975).
468. Morris, J. C. Clinical dementia rating: a reliable and valid diagnostic and staging measure
for dementia of the Alzheimer type. Int. Psychogeriatr. 9 Suppl 1, 173–176; discussion 177-
178 (1997).
469. Ayton, S., Faux, N. G., Bush, A. I. & Alzheimer’s Disease Neuroimaging Initiative. Ferritin
levels in the cerebrospinal fluid predict Alzheimer’s disease outcomes and are regulated by
APOE. Nat. Commun. 6, 6760 (2015).
470. Mann, D. M., Pickering-Brown, S. M., Takeuchi, A., Iwatsubo, T. & Members of the
Familial Alzheimer’s Disease Pathology Study Group. Amyloid angiopathy and variability
in amyloid beta deposition is determined by mutation position in presenilin-1-linked
Alzheimer’s disease. Am. J. Pathol. 158, 2165–2175 (2001).
471. Lee, S. et al. White matter hyperintensities and the mediating role of cerebral amyloid
angiopathy in dominantly-inherited Alzheimer’s disease. PloS One 13, e0195838 (2018).

228
472. Farrall, A. J. & Wardlaw, J. M. Blood-brain barrier: ageing and microvascular disease--
systematic review and meta-analysis. Neurobiol. Aging 30, 337–352 (2009).
473. Armulik, A., Genové, G. & Betsholtz, C. Pericytes: developmental, physiological, and
pathological perspectives, problems, and promises. Dev. Cell 21, 193–215 (2011).
474. Hellström, M., Kalén, M., Lindahl, P., Abramsson, A. & Betsholtz, C. Role of PDGF-B and
PDGFR-beta in recruitment of vascular smooth muscle cells and pericytes during embryonic
blood vessel formation in the mouse. Dev. Camb. Engl. 126, 3047–3055 (1999).
475. Weintraub, S. et al. The Alzheimer’s Disease Centers’ Uniform Data Set (UDS): the
neuropsychologic test battery. Alzheimer Dis. Assoc. Disord. 23, 91–101 (2009).
476. Andreasson, U. et al. A Practical Guide to Immunoassay Method Validation. Front. Neurol.
6, 179 (2015).
477. Heldin, C.-H. Targeting the PDGF signaling pathway in tumor treatment. Cell Commun.
Signal. CCS 11, 97 (2013).
478. Yang, Y., Yuzawa, S. & Schlessinger, J. Contacts between membrane proximal regions of
the PDGF receptor ectodomain are required for receptor activation but not for receptor
dimerization. Proc. Natl. Acad. Sci. U. S. A. 105, 7681–7686 (2008).
479. Omura, T., Heldin, C. H. & Ostman, A. Immunoglobulin-like domain 4-mediated receptor-
receptor interactions contribute to platelet-derived growth factor-induced receptor
dimerization. J. Biol. Chem. 272, 12676–12682 (1997).
480. Pugh, C. W. & Ratcliffe, P. J. Regulation of angiogenesis by hypoxia: role of the HIF system.
Nat. Med. 9, 677–684 (2003).
481. LaManna, J. C., Chavez, J. C. & Pichiule, P. Structural and functional adaptation to hypoxia
in the rat brain. J. Exp. Biol. 207, 3163–3169 (2004).
482. Xu, K. & Lamanna, J. C. Chronic hypoxia and the cerebral circulation. J. Appl. Physiol.
Bethesda Md 1985 100, 725–730 (2006).
483. LaManna, J. C., Vendel, L. M. & Farrell, R. M. Brain adaptation to chronic hypobaric
hypoxia in rats. J. Appl. Physiol. Bethesda Md 1985 72, 2238–2243 (1992).
484. Xu, K., Puchowicz, M. A. & LaManna, J. C. Renormalization of regional brain blood flow
during prolonged mild hypoxic exposure in rats. Brain Res. 1027, 188–191 (2004).
485. Boero, J. A., Ascher, J., Arregui, A., Rovainen, C. & Woolsey, T. A. Increased brain
capillaries in chronic hypoxia. J. Appl. Physiol. Bethesda Md 1985 86, 1211–1219 (1999).

229
486. Harik, N. et al. Time-course and reversibility of the hypoxia-induced alterations in cerebral
vascularity and cerebral capillary glucose transporter density. Brain Res. 737, 335–338
(1996).
487. Opitz, E. Increased vascularization of the tissue due to acclimatization to high altitude and
its significance for the oxygen transport. Exp. Med. Surg. 9, 389–403 (1951).
488. Patt, S., Sampaolo, S., Théallier-Jankó, A., Tschairkin, I. & Cervós-Navarro, J. Cerebral
angiogenesis triggered by severe chronic hypoxia displays regional differences. J. Cereb.
Blood Flow Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab. 17, 801–806 (1997).
489. Pichiule, P. & LaManna, J. C. Angiopoietin-2 and rat brain capillary remodeling during
adaptation and deadaptation to prolonged mild hypoxia. J. Appl. Physiol. Bethesda Md 1985
93, 1131–1139 (2002).
490. Balabanov, R. & Dore-Duffy, P. Role of the CNS microvascular pericyte in the blood-brain
barrier. J. Neurosci. Res. 53, 637–644 (1998).
491. Lindblom, P. et al. Endothelial PDGF-B retention is required for proper investment of
pericytes in the microvessel wall. Genes Dev. 17, 1835–1840 (2003).
492. von Tell, D., Armulik, A. & Betsholtz, C. Pericytes and vascular stability. Exp. Cell Res.
312, 623–629 (2006).
493. Hellström, M. et al. Dll4 signalling through Notch1 regulates formation of tip cells during
angiogenesis. Nature 445, 776–780 (2007).
494. Feng, Y. et al. Impaired pericyte recruitment and abnormal retinal angiogenesis as a result of
angiopoietin-2 overexpression. Thromb. Haemost. 97, 99–108 (2007).
495. Gaengel, K., Genové, G., Armulik, A. & Betsholtz, C. Endothelial-mural cell signaling in
vascular development and angiogenesis. Arterioscler. Thromb. Vasc. Biol. 29, 630–638
(2009).
496. Carmeliet, P. Angiogenesis in health and disease. Nat. Med. 9, 653–660 (2003).
497. Dore-Duffy, P., Wang, X., Mehedi, A., Kreipke, C. W. & Rafols, J. A. Differential
expression of capillary VEGF isoforms following traumatic brain injury. Neurol. Res. 29,
395–403 (2007).
498. Carmeliet, P. & Jain, R. K. Principles and mechanisms of vessel normalization for cancer
and other angiogenic diseases. Nat. Rev. Drug Discov. 10, 417–427 (2011).
499. Liu, N., Zhao, N., Chen, L. & Cai, N. Survivin contributes to the progression of diabetic
retinopathy through HIF-1α pathway. Int. J. Clin. Exp. Pathol. 8, 9161–9167 (2015).

230
500. Hall, A. P. Review of the pericyte during angiogenesis and its role in cancer and diabetic
retinopathy. Toxicol. Pathol. 34, 763–775 (2006).
501. Zhang, Z. G. et al. Correlation of VEGF and angiopoietin expression with disruption of
blood-brain barrier and angiogenesis after focal cerebral ischemia. J. Cereb. Blood Flow
Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab. 22, 379–392 (2002).
502. Lou, Y.-L. et al. miR-210 activates notch signaling pathway in angiogenesis induced by
cerebral ischemia. Mol. Cell. Biochem. 370, 45–51 (2012).
503. Zechariah, A. et al. Vascular endothelial growth factor promotes pericyte coverage of brain
capillaries, improves cerebral blood flow during subsequent focal cerebral ischemia, and
preserves the metabolic penumbra. Stroke J. Cereb. Circ. 44, 1690–1697 (2013).
504. Benderro, G. F. & Lamanna, J. C. Hypoxia-induced angiogenesis is delayed in aging mouse
brain. Brain Res. 1389, 50–60 (2011).
505. Lazic, D. et al. 3K3A-activated protein C blocks amyloidogenic BACE1 pathway and
improves functional outcome in mice. J. Exp. Med. 216, 279–293 (2019).
506. Nikolakopoulou, A. M., Zhao, Z., Montagne, A. & Zlokovic, B. V. Regional early and
progressive loss of brain pericytes but not vascular smooth muscle cells in adult mice with
disrupted platelet-derived growth factor receptor-β signaling. PloS One 12, e0176225 (2017).
507. Zhu, D. et al. Protein S controls hypoxic/ischemic blood-brain barrier disruption through the
TAM receptor Tyro3 and sphingosine 1-phosphate receptor. Blood 115, 4963–4972 (2010).
508. Takano, T. et al. Cortical spreading depression causes and coincides with tissue hypoxia.
Nat. Neurosci. 10, 754–762 (2007).
509. Brown, C. E., Li, P., Boyd, J. D., Delaney, K. R. & Murphy, T. H. Extensive turnover of
dendritic spines and vascular remodeling in cortical tissues recovering from stroke. J.
Neurosci. Off. J. Soc. Neurosci. 27, 4101–4109 (2007).
510. Meyer-Luehmann, M. et al. Rapid appearance and local toxicity of amyloid-beta plaques in
a mouse model of Alzheimer’s disease. Nature 451, 720–724 (2008).
511. Garcia-Alloza, M. et al. Matrix metalloproteinase inhibition reduces oxidative stress
associated with cerebral amyloid angiopathy in vivo in transgenic mice. J. Neurochem. 109,
1636–1647 (2009).
512. Kisler, K. et al. In vivo imaging and analysis of cerebrovascular hemodynamic responses and
tissue oxygenation in the mouse brain. Nat. Protoc. 13, 1377–1402 (2018).

231
513. Gebala, V., Collins, R., Geudens, I., Phng, L.-K. & Gerhardt, H. Blood flow drives lumen
formation by inverse membrane blebbing during angiogenesis in vivo. Nat. Cell Biol. 18,
443–450 (2016).
514. Haase, V. H. Hypoxic regulation of erythropoiesis and iron metabolism. Am. J. Physiol.
Renal Physiol. 299, F1-13 (2010).
515. Lenfant, C. & Sullivan, K. Adaptation to high altitude. N. Engl. J. Med. 284, 1298–1309
(1971).
516. Ferrara, N., Gerber, H.-P. & LeCouter, J. The biology of VEGF and its receptors. Nat. Med.
9, 669–676 (2003).
517. Semenza, G. L. Targeting HIF-1 for cancer therapy. Nat. Rev. Cancer 3, 721–732 (2003).
518. Hirota, K. & Semenza, G. L. Regulation of angiogenesis by hypoxia-inducible factor 1. Crit.
Rev. Oncol. Hematol. 59, 15–26 (2006).
519. Kelly, B. D. et al. Cell type-specific regulation of angiogenic growth factor gene expression
and induction of angiogenesis in nonischemic tissue by a constitutively active form of
hypoxia-inducible factor 1. Circ. Res. 93, 1074–1081 (2003).
520. Kuwabara, K. et al. Hypoxia-mediated induction of acidic/basic fibroblast growth factor and
platelet-derived growth factor in mononuclear phagocytes stimulates growth of hypoxic
endothelial cells. Proc. Natl. Acad. Sci. U. S. A. 92, 4606–4610 (1995).

521. Pichiule, P., Chavez, J. C. & LaManna, J. C. Hypoxic regulation of angiopoietin-2 expression
in endothelial cells. J. Biol. Chem. 279, 12171–12180 (2004).
522. Feng, Y. et al. Angiopoietin-2 deficiency decelerates age-dependent vascular changes in the
mouse retina. Cell. Physiol. Biochem. Int. J. Exp. Cell. Physiol. Biochem. Pharmacol. 21,
129–136 (2008).
523. Reyahi, A. et al. Foxf2 Is Required for Brain Pericyte Differentiation and Development and
Maintenance of the Blood-Brain Barrier. Dev. Cell 34, 19–32 (2015).
524. Sangwung, P. et al. KLF2 and KLF4 control endothelial identity and vascular integrity. JCI
Insight 2, e91700 (2017).
525. Yang, H., Xi, X., Zhao, B., Su, Z. & Wang, Z. KLF4 protects brain microvascular endothelial
cells from ischemic stroke induced apoptosis by transcriptionally activating MALAT1.
Biochem. Biophys. Res. Commun. 495, 2376–2382 (2018).

232
526. Dore-Duffy, P. & LaManna, J. C. Physiologic angiodynamics in the brain. Antioxid. Redox
Signal. 9, 1363–1371 (2007).
527. Eklund, L. & Olsen, B. R. Tie receptors and their angiopoietin ligands are context-dependent
regulators of vascular remodeling. Exp. Cell Res. 312, 630–641 (2006).
528. LaManna, J. C., Sun, X., Ivy, A. D. & Ward, N. L. Is cycloxygenase-2 (COX-2) a major
component of the mechanism responsible for microvascular remodeling in the brain? Adv.
Exp. Med. Biol. 578, 297–303 (2006).
529. Picelli, S. et al. Full-length RNA-seq from single cells using Smart-seq2. Nat. Protoc. 9, 171–
181 (2014).
530. Wardlaw, J. M., Smith, C. & Dichgans, M. Mechanisms of sporadic cerebral small vessel
disease: insights from neuroimaging. Lancet Neurol. 12, 483–497 (2013).
531. Jack, C. R. et al. Tracking pathophysiological processes in Alzheimer’s disease: an updated
hypothetical model of dynamic biomarkers. Lancet Neurol. 12, 207–216 (2013).
532. Chen, Z. L. & Strickland, S. Neuronal death in the hippocampus is promoted by plasmin-
catalyzed degradation of laminin. Cell 91, 917–925 (1997).
533. Mhatre, M. et al. Thrombin, a mediator of neurotoxicity and memory impairment. Neurobiol.
Aging 25, 783–793 (2004).
534. Chen, B., Cheng, Q., Yang, K. & Lyden, P. D. Thrombin mediates severe neurovascular
injury during ischemia. Stroke 41, 2348–2352 (2010).
535. Schachtrup, C. et al. Fibrinogen inhibits neurite outgrowth via beta 3 integrin-mediated
phosphorylation of the EGF receptor. Proc. Natl. Acad. Sci. U. S. A. 104, 11814–11819
(2007).
536. Akassoglou, K. et al. Fibrin depletion decreases inflammation and delays the onset of
demyelination in a tumor necrosis factor transgenic mouse model for multiple sclerosis. Proc.
Natl. Acad. Sci. U. S. A. 101, 6698–6703 (2004).
537. Ryu, J. K. et al. Blood coagulation protein fibrinogen promotes autoimmunity and
demyelination via chemokine release and antigen presentation. Nat. Commun. 6, 8164
(2015).
538. Bardehle, S., Rafalski, V. A. & Akassoglou, K. Breaking boundaries-coagulation and
fibrinolysis at the neurovascular interface. Front. Cell. Neurosci. 9, 354 (2015).
539. Kwan, J. Y. et al. Iron accumulation in deep cortical layers accounts for MRI signal
abnormalities in ALS: correlating 7 tesla MRI and pathology. PloS One 7, e35241 (2012).

233
540. Winkler, E. A. et al. Blood-spinal cord barrier breakdown and pericyte reductions in
amyotrophic lateral sclerosis. Acta Neuropathol. (Berl.) 125, 111–120 (2013).
541. Winkler, E. A. et al. Blood-spinal cord barrier disruption contributes to early motor-neuron
degeneration in ALS-model mice. Proc. Natl. Acad. Sci. U. S. A. 111, E1035-1042 (2014).
542. Zhong, Z. et al. Activated protein C therapy slows ALS-like disease in mice by
transcriptionally inhibiting SOD1 in motor neurons and microglia cells. J. Clin. Invest. 119,
3437–3449 (2009).
543. Kortekaas, R. et al. Blood-brain barrier dysfunction in parkinsonian midbrain in vivo. Ann.
Neurol. 57, 176–179 (2005).
544. Langford, D. et al. Altered P-glycoprotein expression in AIDS patients with HIV
encephalitis. J. Neuropathol. Exp. Neurol. 63, 1038–1047 (2004).
545. Sui, Y.-T., Bullock, K. M., Erickson, M. A., Zhang, J. & Banks, W. A. Alpha synuclein is
transported into and out of the brain by the blood-brain barrier. Peptides 62, 197–202 (2014).
546. Peelaerts, W. et al. α-Synuclein strains cause distinct synucleinopathies after local and
systemic administration. Nature 522, 340–344 (2015).
547. Matsumoto, J. et al. Transmission of α-synuclein-containing erythrocyte-derived
extracellular vesicles across the blood-brain barrier via adsorptive mediated transcytosis:
another mechanism for initiation and progression of Parkinson’s disease? Acta Neuropathol.
Commun. 5, 71 (2017).
548. Shi, M. et al. Plasma exosomal α-synuclein is likely CNS-derived and increased in
Parkinson’s disease. Acta Neuropathol. (Berl.) 128, 639–650 (2014).
549. Gray, M. T. & Woulfe, J. M. Striatal blood-brain barrier permeability in Parkinson’s disease.
J. Cereb. Blood Flow Metab. Off. J. Int. Soc. Cereb. Blood Flow Metab. 35, 747–750 (2015).
550. Pienaar, I. S. et al. Deep-brain stimulation associates with improved microvascular integrity
in the subthalamic nucleus in Parkinson’s disease. Neurobiol. Dis. 74, 392–405 (2015).
551. Kumar, D. K. V. et al. Amyloid-β peptide protects against microbial infection in mouse and
worm models of Alzheimer’s disease. Sci. Transl. Med. 8, 340ra72 (2016).
552. Soscia, S. J. et al. The Alzheimer’s disease-associated amyloid beta-protein is an
antimicrobial peptide. PloS One 5, e9505 (2010).
553. Banks, W. A. From blood–brain barrier to blood–brain interface: new opportunities for CNS
drug delivery. Nat. Rev. Drug Discov. 15, 275–292 (2016).

234
554. Sun, W., Hu, Q., Ji, W., Wright, G. & Gu, Z. Leveraging Physiology for Precision Drug
Delivery. Physiol. Rev. 97, 189–225 (2017).
555. Saraiva, C. et al. Nanoparticle-mediated brain drug delivery: Overcoming blood–brain
barrier to treat neurodegenerative diseases. J. Controlled Release 235, 34–47 (2016).
556. Niewoehner, J. et al. Increased brain penetration and potency of a therapeutic antibody using
a monovalent molecular shuttle. Neuron 81, 49–60 (2014).
557. Yu, Y. J. et al. Therapeutic bispecific antibodies cross the blood-brain barrier in nonhuman
primates. Sci. Transl. Med. 6, 261ra154 (2014).
558. Markham, A. Cerliponase Alfa: First Global Approval. Drugs 77, 1247–1249 (2017).
559. Dolgin, E. Spinal muscular atrophy approval boosts antisense drugs. Nat. Biotechnol. 35,
99–100 (2017).
560. McDowell, G. C. & Pope, J. E. Intrathecal Ziconotide: Dosing and Administration Strategies
in Patients With Refractory Chronic Pain: INTRATHECAL ZICONOTIDE DOSING
STRATEGIES. Neuromodulation Technol. Neural Interface 19, 522–532 (2016).
561. Chapman, C. D. et al. Intranasal Treatment of Central Nervous System Dysfunction in
Humans. Pharm. Res. 30, 2475–2484 (2013).
562. Yemisci, M. et al. Systemically administered brain-targeted nanoparticles transport peptides
across the blood-brain barrier and provide neuroprotection. J. Cereb. Blood Flow Metab. Off.
J. Int. Soc. Cereb. Blood Flow Metab. 35, 469–475 (2015).
563. Burgess, A. & Hynynen, K. Microbubble-Assisted Ultrasound for Drug Delivery in the Brain
and Central Nervous System. Adv. Exp. Med. Biol. 880, 293–308 (2016).
564. Poon, C., McMahon, D. & Hynynen, K. Noninvasive and targeted delivery of therapeutics
to the brain using focused ultrasound. Neuropharmacology 120, 20–37 (2017).
565. Chen, H. & Konofagou, E. E. The size of blood-brain barrier opening induced by focused
ultrasound is dictated by the acoustic pressure. J. Cereb. Blood Flow Metab. Off. J. Int. Soc.
Cereb. Blood Flow Metab. 34, 1197–1204 (2014).
566. Griffin, J. H., Zlokovic, B. V. & Mosnier, L. O. Activated protein C: biased for translation.
Blood 125, 2898–2907 (2015).
567. Zlokovic, B. V. & Griffin, J. H. Cytoprotective protein C pathways and implications for
stroke and neurological disorders. Trends Neurosci. 34, 198–209 (2011).
568. Lyden, P. et al. Final Results of the RHAPSODY Trial: A Multi-Center, Phase 2 Trial Using
a Continual Reassessment Method to Determine the Safety and Tolerability of 3K3A-APC,

235
A Recombinant Variant of Human Activated Protein C, in Combination with Tissue
Plasminogen Activator, Mechanical Thrombectomy or both in Moderate to Severe Acute
Ischemic Stroke. Ann. Neurol. 85, 125–136 (2019).
569. Brinton, R., Swanson, H. & Irwin, R. Allopregnanolone promotes cholesterol and amyloid-
beta clearance mechanisms: Assessment of a regenerative therapeutic for Alzheimer’s
disease. Alzheimers Dement 12, 1024–5 (2016).
570. Atwal, J. K. et al. A Therapeutic Antibody Targeting BACE1 Inhibits Amyloid- Production
in Vivo. Sci. Transl. Med. 3, 84ra43-84ra43 (2011).
571. Rapoport, S. I. Osmotic opening of the blood-brain barrier: principles, mechanism, and
therapeutic applications. Cell. Mol. Neurobiol. 20, 217–230 (2000).
572. Kovacs, Z. I. et al. Disrupting the blood–brain barrier by focused ultrasound induces sterile
inflammation. Proc. Natl. Acad. Sci. 114, E75–E84 (2017).
573. Gerwien, H. et al. Imaging matrix metalloproteinase activity in multiple sclerosis as a
specific marker of leukocyte penetration of the blood-brain barrier. Sci. Transl. Med. 8,
364ra152 (2016).
574. Sagare, A. P., Deane, R. & Zlokovic, B. V. Low-density lipoprotein receptor-related protein
1: a physiological Aβ homeostatic mechanism with multiple therapeutic opportunities.
Pharmacol. Ther. 136, 94–105 (2012).
575. Wang, D., Kranz-Eble, P. & De Vivo, D. C. Mutational analysis of GLUT1 (SLC2A1) in
Glut-1 deficiency syndrome. Hum. Mutat. 16, 224–231 (2000).
576. Novara, F. et al. Clinical and Molecular Characteristics of SLC16A2 (MCT8) Mutations in
Three Families with the Allan-Herndon-Dudley Syndrome. Hum. Mutat. 38, 260–264
(2017).
577. Abdel-Hamid, M. S., Abdel-Salam, G. M. H., Issa, M. Y., Emam, B. A. & Zaki, M. S. Band-
like calcification with simplified gyration and polymicrogyria: report of 10 new families and
identification of five novel OCLN mutations. J. Hum. Genet. 62, 553–559 (2017).
578. Akawi, N. A. et al. Delineation of the clinical, molecular and cellular aspects of novel JAM3
mutations underlying the autosomal recessive hemorrhagic destruction of the brain,
subependymal calcification, and congenital cataracts. Hum. Mutat. 34, 498–505 (2013).
579. Keller, A. et al. Mutations in the gene encoding PDGF-B cause brain calcifications in humans
and mice. Nat. Genet. 45, 1077–1082 (2013).
580. Nicolas, G. et al. Mutation of the PDGFRB gene as a cause of idiopathic basal ganglia
calcification. Neurology 80, 181–187 (2013).

236
581. Sabbagh, M. F. et al. Transcriptional and epigenomic landscapes of CNS and non-CNS
vascular endothelial cells. eLife 7, (2018).
582. Al-Bachari, S. MRI Assessment of Neurovascular Changes in Idiopathic Parkinson’s
disease. (University of Manchester, 2016).
583. Taheri, S., Gasparovic, C., Shah, N. J. & Rosenberg, G. A. Quantitative measurement of
blood-brain barrier permeability in human using dynamic contrast-enhanced MRI with fast
T1 mapping. Magn. Reson. Med. Off. J. Soc. Magn. Reson. Med. Soc. Magn. Reson. Med.
65, 1036–1042 (2011).
584. Cramer, S. P., Simonsen, H., Frederiksen, J. L., Rostrup, E. & Larsson, H. B. W. Abnormal
blood-brain barrier permeability in normal appearing white matter in multiple sclerosis
investigated by MRI. NeuroImage Clin. 4, 182–189 (2014).
585. Cramer, S. P., Modvig, S., Simonsen, H. J., Frederiksen, J. L. & Larsson, H. B. W.
Permeability of the blood-brain barrier predicts conversion from optic neuritis to multiple
sclerosis. Brain J. Neurol. 138, 2571–2583 (2015).
586. Gaitán, M. I. et al. Evolution of the blood-brain barrier in newly forming multiple sclerosis
lesions. Ann. Neurol. 70, 22–29 (2011).
587. Ingrisch, M. et al. Quantification of perfusion and permeability in multiple sclerosis:
dynamic contrast-enhanced MRI in 3D at 3T. Invest. Radiol. 47, 252–258 (2012).
588. Ham, J. H. et al. Cerebral microbleeds in patients with Parkinson’s disease. J. Neurol. 261,
1628–1635 (2014).
589. Janelidze, S. et al. Increased CSF biomarkers of angiogenesis in Parkinson disease.
Neurology 85, 1834–1842 (2015).
590. Sokoloff, L. Measurement of local cerebral glucose utilization and its relation to local
functional activity in the brain. Adv. Exp. Med. Biol. 291, 21–42 (1991).
591. Sperling, R. A. et al. Toward defining the preclinical stages of Alzheimer’s disease:
recommendations from the National Institute on Aging-Alzheimer’s Association
workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement. J.
Alzheimers Assoc. 7, 280–292 (2011).
592. Southworth, R., Parry, C. R., Parkes, H. G., Medina, R. A. & Garlick, P. B. Tissue-specific
differences in 2-fluoro-2-deoxyglucose metabolism beyond FDG-6-P: a 19F NMR
spectroscopy study in the rat. NMR Biomed. 16, 494–502 (2003).
593. Pisani, V. et al. Increased blood-cerebrospinal fluid transfer of albumin in advanced
Parkinson’s disease. J. Neuroinflammation 9, 188 (2012).

237
594. Liguori, C. et al. Cerebrospinal-fluid Alzheimer’s Disease Biomarkers and Blood-Brain
Barrier Integrity in a Natural Population of Cognitive Intact Parkinson’s Disease Patients.
CNS Neurol. Disord. Drug Targets 16, 339–345 (2017).
595. Brettschneider, J., Petzold, A., Süssmuth, S. D., Ludolph, A. C. & Tumani, H. Axonal
damage markers in cerebrospinal fluid are increased in ALS. Neurology 66, 852–856 (2006).
596. Jessen Krut, J. et al. Biomarker evidence of axonal injury in neuroasymptomatic HIV-1
patients. PloS One 9, e88591 (2014).
597. Garbuzova-Davis, S. et al. Impaired blood-brain/spinal cord barrier in ALS patients. Brain
Res. 1469, 114–128 (2012).
598. Kirk, J., Plumb, J., Mirakhur, M. & McQuaid, S. Tight junctional abnormality in multiple
sclerosis white matter affects all calibres of vessel and is associated with blood-brain barrier
leakage and active demyelination. J. Pathol. 201, 319–327 (2003).
599. Omalu, B. I., Fitzsimmons, R. P., Hammers, J. & Bailes, J. Chronic traumatic
encephalopathy in a professional American wrestler. J. Forensic. Nurs. 6, 130–136 (2010).
600. Omalu, B. I. et al. Chronic traumatic encephalopathy in a National Football League player.
Neurosurgery 57, 128–134; discussion 128-134 (2005).
601. Niu, F., Yao, H., Zhang, W., Sutliff, R. L. & Buch, S. Tat 101-mediated enhancement of
brain pericyte migration involves platelet-derived growth factor subunit B homodimer:
implications for human immunodeficiency virus-associated neurocognitive disorders. J.
Neurosci. Off. J. Soc. Neurosci. 34, 11812–11825 (2014).
602. Henkel, J. S., Beers, D. R., Wen, S., Bowser, R. & Appel, S. H. Decreased mRNA expression
of tight junction proteins in lumbar spinal cords of patients with ALS. Neurology 72, 1614–
1616 (2009).
603. Miyazaki, K. et al. Disruption of neurovascular unit prior to motor neuron degeneration in
amyotrophic lateral sclerosis. J. Neurosci. Res. 89, 718–728 (2011).
604. Persidsky, Y. et al. Rho-mediated regulation of tight junctions during monocyte migration
across the blood-brain barrier in HIV-1 encephalitis (HIVE). Blood 107, 4770–4780 (2006).
605. Yamamoto, M. et al. Phosphorylation of claudin-5 and occludin by rho kinase in brain
endothelial cells. Am. J. Pathol. 172, 521–533 (2008).
606. Wada, K. et al. Expression levels of vascular endothelial growth factor and its receptors in
Parkinson’s disease. Neuroreport 17, 705–709 (2006).

238
607. Desai Bradaric, B., Patel, A., Schneider, J. A., Carvey, P. M. & Hendey, B. Evidence for
angiogenesis in Parkinson’s disease, incidental Lewy body disease, and progressive
supranuclear palsy. J. Neural Transm. Vienna Austria 1996 119, 59–71 (2012).
608. Lin, C.-Y. et al. Neurovascular abnormalities in humans and mice with Huntington’s disease.
Exp. Neurol. 250, 20–30 (2013).
609. Cicchetti, F. et al. Mutant huntingtin is present in neuronal grafts in Huntington disease
patients. Ann. Neurol. 76, 31–42 (2014).
610. Salvador, E., Shityakov, S. & Förster, C. Glucocorticoids and endothelial cell barrier
function. Cell Tissue Res. 355, 597–605 (2014).
611. Davalos, D. et al. Fibrinogen-induced perivascular microglial clustering is required for the
development of axonal damage in neuroinflammation. Nat. Commun. 3, 1227 (2012).
612. Horie, N. et al. Transplanted stem cell-secreted vascular endothelial growth factor effects
poststroke recovery, inflammation, and vascular repair. Stem Cells Dayt. Ohio 29, 274–285
(2011).
613. Park, H. J. et al. Mesenchymal stem cells stabilize the blood–brain barrier through regulation
of astrocytes. Stem Cell Res. Ther. 6, (2015).
614. Tang, G. et al. Mesenchymal stem cells maintain blood-brain barrier integrity by inhibiting
aquaporin-4 upregulation after cerebral ischemia. Stem Cells Dayt. Ohio 32, 3150–3162
(2014).
615. Coatti, G. C. et al. Pericytes Extend Survival of ALS SOD1 Mice and Induce the Expression
of Antioxidant Enzymes in the Murine Model and in IPSCs Derived Neuronal Cells from an
ALS Patient. Stem Cell Rev. (2017). doi:10.1007/s12015-017-9752-2
616. Khalil, A. A. et al. Relationship Between Changes in the Temporal Dynamics of the Blood-
Oxygen-Level-Dependent Signal and Hypoperfusion in Acute Ischemic Stroke. Stroke 48,
925–931 (2017).
617. Alvarez, J. I., Cayrol, R. & Prat, A. Disruption of central nervous system barriers in multiple
sclerosis. Biochim. Biophys. Acta 1812, 252–264 (2011).
618. Deshpande, T. et al. Subcellular reorganization and altered phosphorylation of the astrocytic
gap junction protein connexin43 in human and experimental temporal lobe epilepsy. Glia 65,
1809–1820 (2017).
619. Rigau, V. et al. Angiogenesis is associated with blood-brain barrier permeability in temporal
lobe epilepsy. Brain J. Neurol. 130, 1942–1956 (2007).

239
620. Zhang, C., Kwan, P., Zuo, Z. & Baum, L. The transport of antiepileptic drugs by P-
glycoprotein. Adv. Drug Deliv. Rev. 64, 930–942 (2012).
621. Schwab, N., Schneider-Hohendorf, T. & Wiendl, H. Therapeutic uses of anti-α4-integrin
(anti-VLA-4) antibodies in multiple sclerosis. Int. Immunol. 27, 47–53 (2015).
622. Ziemssen, T. & Thomas, K. Alemtuzumab in the long-term treatment of relapsing-remitting
multiple sclerosis: an update on the clinical trial evidence and data from the real world. Ther.
Adv. Neurol. Disord. 10, 343–359 (2017).
623. Burgess, A. & Hynynen, K. Microbubble-Assisted Ultrasound for Drug Delivery in the Brain
and Central Nervous System. in Therapeutic Ultrasound (eds. Escoffre, J.-M. & Bouakaz,
A.) 880, 293–308 (Springer International Publishing, 2016).
624. Poon, C., McMahon, D. & Hynynen, K. Noninvasive and targeted delivery of therapeutics
to the brain using focused ultrasound. Neuropharmacology 120, 20–37 (2017).
625. Boado, R. J., Hui, E. K.-W., Lu, J. Z. & Pardridge, W. M. Re-engineering iduronate 2-
sulfatase for penetration of the blood-brain barrier via transport on the insulin receptor. Mol.
Genet. Metab. 111, S27 (2014).
626. Deverman, B. E. et al. Cre-dependent selection yields AAV variants for widespread gene
transfer to the adult brain. Nat. Biotechnol. 34, 204–209 (2016).

Abstract (if available)

Abstract The blood-brain barrier (BBB) is a complex, dynamic structure that maintains cerebrovascular integrity and brain health by functioning as a gatekeeper. The BBB sanctions entry of oxygen and energy substrates while denying entry of macromolecules, cells and pathogens under proper physiological conditions. Existing evidence of BBB dysfunction in humans with Alzheimer’s disease (AD) has been shown by neuroimaging and biofluid findings and post-mortem brain tissue analysis. Consistently, neuropathological studies have shown that cerebrovascular pathology is a major risk factor for cognitive impairment in clinically diagnosed AD-type dementia. While evidence of cerebrovascular dysfunction is increasingly reported in AD and associated with early stages of AD, vascular/BBB dysfunction is not yet widely accepted as an important contributor to AD pathophysiology. The goal of this dissertation is to evaluate and validate biomarkers of BBB dysfunction that are clinically relevant to cognitive impairment and dementia including AD, and also to explore basic mechanisms underlying cerebrovascular dysfunction as related to pathophysiology that will ultimately aid in vascular-directed therapeutic efforts. First, this dissertation presents original evidence that BBB breakdown is an early biomarker of cognitive impairment in humans. BBB breakdown is determined by a capillary pericyte injury marker, soluble platelet-derived growth factor receptor-β (sPDGFRβ), in cerebrospinal fluid (CSF) and regional BBB permeability to a contrast agent by dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), which are independent yet related measures of BBB breakdown. These markers of BBB breakdown are further shown to be impacted by genetic factors including APOE4, the major genetic risk factor for sporadic AD. Additionally, CSF evidence of cerebrovascular dysfunction is also shown in autosomal dominant AD (ADAD) individuals carrying PSEN1 or APP mutations. Next, a novel assay to detect CSF sPDGFRβ was developed and validated as a reliable and clinically-relevant biomarker of pericyte injury during cognitive impairment. Then, in order to elucidate and better understand molecular mechanisms and signatures of cerebrovascular dysfunction, I turn to a mouse model of hypoxia that exhibits cerebrovascular dysfunction including BBB breakdown. The final chapter will synthesize my original findings with the current perspective in the AD field and also relate the observed cerebrovascular dysfunction to what is similarly seen in other dementias and/or neurodegenerative disorders including Parkinson’s disease, Huntington’s disease, amyotrophic lateral sclerosis, multiple sclerosis, human immunodeficiency virus-1-associated dementia, and chronic traumatic encephalopathy. In conclusion, opportunities to target the BBB for treatments will be presented, and finally gaps in the field and future directions will be discussed.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Blood-brain barrier breakdown and vessel integrity in Alzheimer’s disease

PDF

Neural crest-derived cranial pericytes and their dysfunction in disease

PDF

The role of vascular dysfunction in cognitive impairment

PDF

Understanding the role of APP and DYRK1A in human brain pericytes

PDF

Long-term blood pressure variability across the clinical and biomarker spectrum of Alzheimer’s disease

PDF

Vascular contributions to brain aging along the Alzheimer's disease continuum

PDF

Renin-angiotensin system antihypertensive medicines with and without blood-brain-barrier crossing potential: effects on cognitive outcomes in the elderly

PDF

Mapping water exchange rate across the blood-brain barrier

PDF

Neuroimaging markers of risk & resilience to brain aging and dementia

PDF

Alzheimer’s disease: dysregulated genes, ethno-racial disparities, and environmental pollution

PDF

Affective neuropsychiatric symptoms and neural connectivity in the early stages of Alzheimer’s disease

PDF

A virtual reality exergaming system to enhance brain health in older adults at risk for Alzheimer’s disease

PDF

The role of the locus coeruleus in Alzheimer’s disease and cerebrovascular function: insights from neuroimaging, neuropsychology, and biofluid markers

PDF

Perimenopausal transition increases blood brain permeability: implications for neurodegenerative diseases

PDF

Identification of therapeutic targets in human cerebral brain organoid models of neurodegeneration

PDF

Role of oxidative stress in age-associated mild cognitive impairment and Alzheimer's disease

PDF

The opening of the blood brain barrier by homogenized perillyl alcohol

PDF

A transgenic mouse model with overexpression of human RAGE in endothelial cells presents enhanced amyloid-beta transport into the brain

PDF

Insulin sensitivity in cognition, Alzheimer's disease and brain aging

PDF

Indexing cerebrovascular resistance in cognitive decline and Alzheimer's disease

Asset Metadata

Creator Sweeney, Melanie Danielle (author)

Core Title Blood-brain barrier pathophysiology in cognitive impairment and injury

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Neuroscience

Publication Date 04/27/2021

Defense Date 03/19/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Alzheimer's disease,biomarkers,blood-brain barrier,cognitive impairment,CSF,hypoxia,OAI-PMH Harvest,pericytes,vascular

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Pike, Christian (committee chair), Ichida, Justin (committee member), Jung, Jae (committee member), Mack, William (committee member), Zlokovic, Berislav (committee member)

Creator Email mdsweene@usc.edu,melaniedsweeney@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-154177

Unique identifier UC11660981

Identifier etd-SweeneyMel-7331.pdf (filename),usctheses-c89-154177 (legacy record id)

Legacy Identifier etd-SweeneyMel-7331.pdf

Dmrecord 154177

Document Type Dissertation

Format application/pdf (imt)

Rights Sweeney, Melanie Danielle

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA