BUILDING AND BREAKING DOWN THE BLOOD-BRAIN BARRIER by
STEPHANIE BONNEY
B.S., University of Colorado Denver, 2010
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment
of the requirements for the degree of Doctor of Philosophy
Cell Biology, Stem Cells and Development Program 2019
ii This thesis for the Doctor of Philosophy degree by
Stephanie Bonney has been approved for the
Cell Biology, Stem Cells, and Development program by
Joseph Brzezinski, Chair Kenneth Tyler
Bruce Appel Tobias Eckle Peter Dempsey Julie Siegenthaler, Advisor
Date: May 17th 2019
iii Bonney, Stephanie (Ph.D., Cell Biology, Stem Cells, and Development Program)
Building and Breaking Down the Blood-Brain Barrier Thesis directed by Julie A. Siegenthaler
ABSTRACT
The fine tuning of cellular events and regulatory signals during vascular development in the central nervous system (CNS) is emerging as an important facet of brain vascular health and function. Through these complex events and crosstalk with the maturing neural environment, the brain endothelial cells that make up the blood vessels create a vast and supportive vascular network. Central to CNS vascular development is development and maturation of a blood-brain barrier (BBB) that protects the CNS from potentially damaging blood contents. Impairment in the developmental acquisition of BBB properties or loss of BBB integrity in disease undermines the ability of the vasculature to support CNS function. However we are just beginning to uncover the different cellular and signaling events that regulate BBB development, maturation, and its breakdown in disease. In this dissertation, I have: 1) Identified a complex signaling convergence of the retinoic acid (RA) and WNT pathways important for proper brain vascular growth and maturation, two features required for vascular stability. 2) Provided substantial evidence that RA is not an essential factor in promoting BBB properties during development. 3) Demonstrated that BBB leakage is driven by IFN-mediated deterioration of BBB properties in a mouse model of viral encephalitis, a disease where the events and signals resulting in BBB leakage is poorly understood.
I began my thesis work aiming to understand how RA regulates vascular growth in the embryonic brain. In using mice incapable of generating RA and therefore deficient in RA and RA signaling (Rdh10 mutants), I showed that RA functions to drive angiogenesis in the CNS by
iv ensuring WNT signaling occurs in the growing brain vasculature, a pathway crucial for CNS angiogenesis. This involves the non-cell autonomous function of RA in suppressing the
expression of WNT antagonists, and therefore is a likely contributing event to the avascular brain regions observed in Rdh10 mutants. In studying Rdh10 mutants and embryos with endothelial- specific disruption of RA signaling (PdgfbiCre/+; dnRAR403fl/fl) I found that RA also plays an important cell autonomous function in limiting endothelial-WNT signaling. The prevention of ectopic endothelial-WNT signaling by RA appeared to be an important step for regulating vascular stability during brain vascular growth. In the second part of my thesis, I extended on this work to identify the underlying mechanism of how RA regulates endothelial-WNT
signaling. I found that RA plays an important role in modulating endothelial -catenin expression, a major WNT signaling effector, through both transcriptional and unique, post- translational mechanisms that regulate -catenin degradation. Interestingly, ectopic -catenin expression, and WNT signaling, within the vasculature of Rdh10 and PdgfbiCre/+;
dnRAR403fl/fl mutants led to increases in pericytes in the brain vasculature, a cell type important for vascular stability. I then demonstrated that endothelial-WNT signaling and the WNT target gene, Sox17, are crucial in establishing the high pericyte numbers that are normally found in the brain vasculature. However in the absence of endothelial-RA signaling, uncontrolled pericyte coverage, likely due to ectopic WNT--catenin-Sox17 signaling, appears to be a consequence of and correlates with vascular instability. In concurrence with these studies and the third part of my dissertation, I demonstrated that RA is not required for the establishment of the barrier properties in the brain vasculature during development. This is in contrast to work by other groups that showed high, non-physiological concentrations of RA induce barrier properties in cultured brain endothelial cells. Instead, I showed that these high concentrations of RA used in
v BBB culture models has off target effects and the BBB-inducing effect in this system may
involve activation of LXR/RXR signaling.
In the fourth and final part of my dissertation, I aimed to elucidate cellular events and signaling pathways that contribute to BBB leakage in a mouse model of viral encephalitis. I found that BBB leakage is a late event during CNS viral infection and correlates with pericyte loss and disorganization in cell-cell junctions, a property that is crucial to maintain BBB
integrity. By performing RNA-sequencing on brain endothelial cells I identified Interferon (IFN) signaling as a major pathway activated in the endothelium during CNS viral infection. In vitro studies showed that IFN reduces barrier properties in brain endothelial cells by inducing cytoskeletal contractions, altering junctional protein localization, and therefore disrupting cell- cell connections in a Rho-kinase (ROCK) dependent manner. I found that neutralization of IFN
during CNS viral infection improves vascular junctional organization, attenuates endothelial- ROCK activity and BBB leakage.
In summary, my thesis revealed previously unknown signals and events in establishing vascular stability during brain vascular growth, BBB acquisition during development, and breakdown of the BBB in disease. The knowledge gained from my work will lend direction in future studies to understand crucial signals, events and interactions between endothelial cells and pericytes during brain vascular development and maturation. Furthermore, identification of major pathways, such as IFN signaling, and cellular events that result in BBB leakage may assist in understanding the consequences of poor vascular function in CNS disease with hopes of improving patient treatment in the future.
The form and content of this abstract are approved. I recommend its’ publication.
Approved by: Julie Siegenthaler
vi DEDICATION
For my Mom, Dad, wonderful siblings- Rian, Carissa, Devon, Korey, Cassie, and my twin sister, Megan, who has been by my side since day one. My step parents, Joe and Misty, and my
grandparents, Raymond, Lorena, Ralph, Linda, and Rick, have all provided immense encouragement throughout my life. Without all of your unconditional love, support, and
inspiration none of this would have been possible.
vii ACKNOWLEDGMENTS
I would like to thank my Ph.D. advisor, Dr. Julie Siegenthaler, for being an incredible mentor for the last 5 years. Without her patience, support, and knowledge I fully believe I would not be the
scientist I am today. I am very grateful for the amazing journey I have had in her lab. I would like to thank my committee members – Dr. Joseph Brzezinski, Dr. Bruce Appel, Dr. Kenneth Tyler, Dr. Tobias Eckle and Dr. Peter Dempsey for their guidance in my research projects and
career advice. I would especially like to thank Dr. Kenneth Tyler for supporting me in my projects investigating BBB leakage in viral encephalitis. His guidance was invaluable in obtaining my F31 and for the success of this particular project. A special thanks to the lab members of the Siegenthaler lab that were present for the majority of my time, Dr. Swati Mishra,
Dr. Kathleen Kelly, and Amber MacPherson, for their assistance and enjoyable demeanor that made for a delightful experience in the lab. Rotation students and interns in the Siegenthaler lab,
Brenna Dennison, Caitlin Ryan and Dr. Megan Wendlandt, provided immense assistance that was crucial for the success of my projects. I would also like to thank Dr. Kenneth Jones for data analysis in my RNA-sequencing experiments. I also appreciate all the help from the UC Denver EM core, specifically Dr. Jennifer Bourne. Thank you to the National Institute of Neurological
Disorders & Stroke for generously funding my research on vascular instability in viral encephalitis (award F31-NS100565). Finally I would like to thank the faculty and students in the
Cell Biology, Stem Cells, and Development graduate program for not only providing a scientifically stimulating environment but also for their personal and professional support
throughout the years.
viii TABLE OF CONTENTS
CHAPTER
I. Building and breaking down the blood-brain barrier……….………..…………18 Part I: Blood vessel growth and maturation in the CNS..………...………...18
Developmental origin, timing and cellular events of blood vessel
growth in the CNS……….………...18 Regulatory signals that direct blood vessel growth in the CNS……..……...20 Maturation of the CNS vasculature – Properties important for
vascular stability……….……....24 Exploring the role of RA in WNT-mediated blood vessel growth and
maturation………...……….……..26 Part II: Acquisition of blood-brain barrier properties by the growing
CNS vasculature……...….………...………..………....27 Blood-brain barrier properties – a developmental perspective….…..………...….27
Enrichment of barrier properties in brain endothelial cells
during development………....28 Suppression of peripheral endothelial properties in the
developing brain vasculature....………..…29 Regulatory signals that drive blood-brain barrier properties
during development………....31
Investigating the role of RA in the acquisition of blood-brain barrier
properties………....33
ix
Part III: Breakdown of the blood-brain barrier in viral encephalitis……….………...34
Impact of blood-brain barrier leakage in disease – emphasis on viral infection in the CNS……….……….….34
Disease pathogenesis and blood-brain barrier leakage in viral encephalitis ………..……..35
Alterations in blood-brain barrier properties in viral encephalitis………36
Major pathways involved in blood-brain barrier breakdown in viral encephalitis……….….………..37
Investigating the breakdown of the blood-brain barrier in viral encephalitis……….…………..….…38
Figures…...……….………..……..39
II. Diverse functions of retinoic acid in brain vascular development………...…...44
Introduction………..………..44
Contributions……….………46
Materials and Methods……….…..47
Results………....55
Discussion………...….69
Figures………...75
III. Retinoic acid regulates endothelial -catenin expression and pericyte numbers in the developing brain vasculature……….85
Introduction………85
Materials and Methods……….…..87
x
Results………95
Discussion………104
Figures………..…108
IV. Differential effects of retinoic acid concentrations in regulating blood-brain barrier properties………...…………...120
Introduction………..120
Materials and Methods……….122
Results………..128
Discussion………...….135
Figures………..141
V. Interferon-gamma contributes to blood-brain barrier leakage during experimental viral encephalitis………...……….……….151
Introduction………..151
Materials and Methods………...………..153
Results………..165
Discussion………...…….180
Figures………....……..186
VI. Conclusions and future directions in understanding blood-brain barrier development, maturation, and breakdown in disease………...…226
Identification of retinoic acid as a major regulator of endothelial-WNT signaling during brain vascular development………...226
Importance of retinoic acid in modulating WNT--catenin-Sox17 signaling during brain vascular development………..………...….229
xi Can modulation of retinoic acid help identify significant
endothelial-pericyte interactions required for vascular stability?...232 Retinoic acid does not regulate blood-brain barrier properties during brain
vascular development – new role for LXR/RXR signaling?...233 Pericyte loss as a major event during CNS viral infection – Does this
impact vascular stability and transcytosis?...234 Disorganization of junctional complexes in the vasculature during viral
encephalitis – highlighting the importance of the endothelial cytoskeleton……..236 REFERENCES……….………..…………239
xii LIST OF FIGURES
Figure 1.1: Blood vessel development in the CNS………39
Figure 1.2: Pathways involved in blood vessel growth and maturation………40
Figure 1.3: Blood brain barrier properties and regulatory signals……….41
Figure 1.4: Blood-brain barrier breakdown in viral encephalitis…..……….42
Figure 1.5: Gaps in knowledge – Regulation of blood vessel growth, BBB properties and breakdown………..…43
Figure 2.1: Neocortical vascular development in E14.5 Rdh10 mutant embryos………...……..75
Figure 2.2: Hypoxia inducible targets VEGFA and GLUT-1 are elevated in Rdh10 mutant neocortices……….76
Figure 2.3: Diminished WNT signaling in Rdh10 mutant cerebrovasculature………..77
Figure 2.4: Elevated WNT signaling in non-cortical Rdh10 mutant vasculature and neurovascular development in PdgfbiCre; dnRAR403-fl/fl animals……….….78
Figure 2.5: Endothelial WNT signaling in increased in fetal brain vasculature of PdgfbiCre; dnRAR403-fl/fl mutants……….….79
Figure 2.6: RA inhibits endothelial WNT signaling in vivo and in vitro……….….80
Figure 2.7: Elevated expression of Sox17 in PdgfbiCre; dnRAR403-fl/fl neurovasculature……82
Figure 2.8: Elevated Sox17 expression in PdgfbiCre; dnRAR403-fl/fl venous and arterial vessels………83
Figure 2.9: Model of RA functions during neurovascular development………...84
Figure 3.1. Retinoic acid regulates -catenin expression in brain endothelial cells………108
Figure 3.2. Retinoic acid regulates -catenin expression via transcription and proteasomal degradation in brain endothelial cells………..110
Figure 3.3. Retinoic acid induces the phosphorylation of -catenin through RAR and PKC activity………..111 Figure 3.4. Retinoic acid induces interactions between -catenin with RAR and
PKC
xiii Figure 3.5. Ectopic WNT signaling in endothelial RA mutants does not alter vascular
expression of the WNT target gene, Claudin5………...…..114
Figure 3.6. Ectopic vascular WNT signaling in RA mutants result in increased pericytes along the developing brain vasculature………115
Figure 3.7. Vascular WNT signaling and Sox17 regulate pericyte numbers along the developing brain vasculature………...……117
Figure 3.8. Retinoic acid regulates endothelial WNT signaling through -catenin to appropriately modulate WNT-driven pericyte numbers along the developing brain vasculature……….……….119
Figure 4.1: BBB function and protein expression is not affected in the non-neocortical vasculature of Rdh10 mutants………..141
Figure 4.2: BBB protein expression is altered following in utero atRA exposure………..142
Figure 4.3: Extended data of 4.2 – BBB protein expression is altered following in utero atRA exposure………..143
Figure 4.4: In utero atRA exposure does not overtly affect BBB integrity……….144
Figure 4.5: Differential effects of RA concentrations on BBB gene expression……….145
Figure 4.6: Extended data of 4.5 and 4.7 – RA in BBB integrity and gene expression…...147
Figure 4.7: Induction of BBB genes by pharmacological RA correlates with activation of LXR/RXR signaling………....…148
Figure 4.8: Differential concentrations of RA in regulating BBB properties………..150
Figure 5.1: Blood-brain barrier breakdown occurs at late-stages of CNS reovirus infection………...186
Figure 5.2: Intracranial injection does not induce BBB leakage in mock animals………….….187
Figure 5.3: Blood-brain barrier breakdown correlates with substantial inflammation and cell death at late stages of reovirus-induced encephalitis………...188
Figure 5.4: Presence of CD3 and CD8 positive T cells during late stages of reovirus infection………...189
Figure 5.5: The brain vasculature is not infected during CNS reovirus infection………...190
xiv Figure 5.6: Morphological changes in the vasculature occur at late-stages of CNS reovirus
infection………...191 Figure 5.7: Diameter increase in Glut1-expressing blood vessels at late-stages of
CNS reovirus infection………192
Figure 5.8: CNS reovirus infection may suppress postnatal blood vessel growth……….…..…193 Figure 5.9: Pericyte numbers are reduced at late-stages of CNS reovirus infection………...…194 Figure 5.10: Electron microscopy reveal intact BBB properties and continuous endothelial
junctions 6 days post CNS reovirus infection………..…196 Figure 5.11: Blood vessel junctions become disorganized during late-stages of CNS
reovirus infection……….…198
Figure 5.12: Blood-brain barrier dysfunction occurs through capillaries and arterioles
during reovirus infection………..……200
Figure 5.13: Endothelial specific expression of Tdtomato in the brain vasculature in
Cdh5creERT2/+;Ai14fl/+ mice………..….202 Figure 5.14: Transcript expression in Tdtomato-sorted endothelial cells during CNS
reovirus infection……….…203
Figure 5.15: IFN reduces barrier properties in brain endothelial cells through
ROCK-mediated cytoskeletal contractions………..204 Figure 5.16: IFN ligands are highly expressed in various brain regions during
reovirus-induced encephalitis……….…….206 Figure 5.17: IFN exposure induces acellular formations between brain endothelial
cells through ROCK activity………207
Figure 5.18: IFN requires ROCK activity to reorganize junctional protein localization
in brain endothelial cells……….……...…..208 Figure 5.19: Fluorescent intensity of junctional proteins at the junction and cytoplasm
in brain endothelial cells following IFN exposure………...210 Figure 5.20: Neutralization of IFN dampens vascular IFN signaling during reovirus
infection………...212
xv Figure 5.21: Neutralization of IFN does not alter viral replication or immune responses
during reovirus-induced encephalitis……….………..213 Figure 5.22: Neutralization of IFN during CNS reovirus infection attenuates BBB
leakage……….214
Figure 5.23: Neutralization of IFN attenuates Fibrinogen leakage during CNS reovirus
infection………...216
Figure 5.24: Neutralization of IFN attenuates blood vessel dilation during CNS reovirus
infection………...217 Figure 5.25: Neutralization of IFN partially attenuates disease progression during CNS
reovirus infection……….218
Figure 5.26: Hypoxia is not a major event that occurs during CNS reovirus infection………..220 Figure 5.27: Neutralization of IFN attenuates pericyte loss during CNS reovirus
infection………...221
Figure 5.28: IFN-pSTAT signaling likely does not directly induce pericyte death during
CNS reovirus infection………222
Figure 5.29: Neutralization of IFN attenuates junctional disorganization and ROCK
activation in the brain vasculature during CNS reovirus infection………..223 Figure 5.30: IFN contributes to BBB leakage through ROCK-mediated junctional
reorganization during experimental viral encephalitis………225
xvi LIST OF ABBREVIATIONS
9cRA 9-cis retinoic acid
AJ Adherens Junction
AMT Absorptive-mediated transcytosis
SMA -smooth muscle actin atRA All-trans retinoic acid BBB Blood-brain barrier
-galactosidase -gal
BIM Bisindoylmaleimide I
BV Blood vessel
CC3 Cleaved-caspase 3
CCM Cerebral cavernous malformation ChIP Chromatin Immunoprecipitation
Cldn Claudin
CNS Central nervous system
ColIV Collagen-IV
Dll4 Delta-like canonical ligand-4
dnRAR Dominant-negative retinoic acid receptor ECs Endothelial cells
ELISA Enzyme-linked immunosorbent assay
EM Electron microscopy
GFP Green fluorescent protein Glut-1 Glucose transporter 1 HSV Herpes simplex virus
HUVECs Human umbilical vein endothelial cells
IB4 Isolectin-b4
ICC Immunocytochemistry
IFN Interferon
IHC Immunohistochemistry
iPSC Induced-pluripotent stem cells JEV Japanese encephalitis virus
LACV La Crosse virus
LXR Liver X receptor
NaF Sodium Fluorescein
NPCs Neural progenitor cells
PAMPs Pathogen-associated molecular patterns p--catenin Phospho--catenin
PBS Phosphate buffered saline
PCs Pericytes
PDGFB Platelet-derived growth factor B
PDGFr Platelet-derived growth factor receptor-
PFA Paraformaldehyde
PFU Plaque forming units
PKC Protein kinase C
xvii PLA Proximity ligation assay
PLVAP Plasmalemma vesicle associated protein pMLC Phospho-myosin light chain
PNVP Perineural vascular plexus
PPAR Peroxisome proliferator-activated receptor qPCR Quantitative polymerase chain reaction
RA Retinoic acid
RAR Retinoic acid receptor
RMT Receptor-mediated transcytosis ROCK Rho-associated protein kinase
ROCKi ROCK inhibitor
RXR Retinoid X receptor
SFRP Secreted frizzled related protein SMCs Smooth muscle cells
STAT Signal transducer and activator of transcription
T3A Type 3 Abney
TGF Transforming growth factor-
TGFR Transforming growth factor- receptor
TJ Tight Junction
TRAP-RNA seq Translating ribosome affinity purification – RNA sequencing TSA Tyramide System Amplification
VEGF Vascular endothelial growth factor
VZ Ventricular zone
Wingless/Integrated WNT
WNV West Nile Virus
ZO-1 Zona occludin-1
18 CHAPTER I
INTRODUCTION1
The vasculature in the brain grows in parallel with the development of the central nervous system (CNS). Soon after entry into the neural tissue, CNS vessels become specialized to form the blood-brain barrier (BBB) that supports neuronal function, meets high energy demands and provides protection from potentially damaging components in the blood. Successful growth and maturation of the vasculature in the CNS, along with specialization of the BBB, relies on neural- derived growth and maturation cues and important cellular interactions with vascular supporting cells. Maintaining a healthy vascular network with functioning BBB properties in the CNS is essential for proper neural function and health. Disruption to this vascular network and BBB properties is a hallmark of many CNS diseases where BBB damaging signals and their effects on BBB properties are only beginning to be uncovered.
Part I: Blood vessel growth and maturation in the CNS
Developmental origin, timing and cellular events of blood vessel growth in the CNS Vascularization of the CNS begins after a primitive vascular network is established within the entire embryo through processes known as vasculogenesis and angiogenesis. In mice, vasculogenesis or the de novo formation of vessels begins around embryonic day 6.5 (E6.5) where endothelial precursors called angioblasts fuse together, differentiate into endothelial cells and form an early vascular net. In regards to CNS vascularization, angioblasts derived from the somatic and cephalic mesoderm form the perineural vascular plexus (PNVP) that surrounds the
1 The Introduction is adapted from sections that I have written for Chapter 43 in the 2nd edition of Comprehensive Developmental Neuroscience and is currently under review. Bonney S., Mishra S., Pleasure S.J., Siegenthaler J.A.
“Meninges and Vasculature”.
19 spinal cord and developing brain respectively (Bagnall et al., 1989, Couly et al., 1995). The PNVP is established around E9.5 in which blood vessels ingress into the CNS tissue through angiogenesis where new vessels sprout and grow off of existing vessels. Blood vessel ingression begins in a caudal-to-rostral direction, starting in the spinal cord and hindbrain and ultimately reaching the forebrain. In the forebrain, angiogenesis from the PNVP begins in the early ganglionic eminence and continues in a ventrolateral-dorsomedial direction and ultimately reaches the dorsal telencephalon (Fig. 1) (Tata et al., 2015).
During embryonic brain development, the vascular network undergoes extensive growth to meet the oxygen and nutrient needs of the growing neural tissue. In response to angiogenic cues from the neural tissue, nascent vessels grow, anastomose with other vessels and lumenize to form an elaborate vessel network. Following birth angiogenesis subsides, correlating with
reductions in angiogenic signals. However vascular remodeling, which is the fine tuning of vessel sprouting and regression, occurs well into a month post birth and is an essential process to meet the metabolic demands of the maturing brain (Harb et al., 2013).
As the CNS vasculature grows and remodels, a vascular tree is established. In the mature CNS, larger diameter arteries fed by the internal carotid arteries sit on the surface of the brain and give rise to smaller diameter arterioles that penetrate the CNS tissue. Pre-capillary arterioles in the CNS give rise to the capillary beds where gas and nutrient occurs. Capillaries provide a direct connection from the arteries to the outflowing post-capillary venules and veins. During development, each segment of the vascular tree obtains specific properties, most notably the high coverage of smooth muscle cells around arteries and arterioles, which are needed to regulate blood flow. Smooth muscle cells are also found along veins and venules but in much lower densities. Establishing this hierarchy within the brain vasculature is essential for a functioning
20 vascular network and supporting brain growth and function (Daneman and Prat, 2015, Grant et al., 2019).
Regulatory signals that direct blood vessel growth in the CNS
Vascular endothelial growth factor (VEGF) and Wingless/Integrated (WNT) signaling pathways are essential for blood vessel growth in the CNS. During development, these
angiogenic cues are secreted from neural progenitors and newly-born neurons, acting on the vascular network and stimulating growth. VEGFA is secreted by hypoxic neural progenitors in the ventricular zone to create a VEGFA concentration gradient (Fig. 1). By acting on the VEGFR2-expressing endothelium, VEGFA specifically promotes a migratory phenotype restricted to the distal endothelial cell of the angiogenic sprout, the tip cell, which extends long filopodia and ultimately migrates toward the VEGFA gradient. While VEGFA has a pro-
migratory effect on tip cells, the proximal stalk cells in the angiogenic sprout respond to VEGFA by increasing their proliferative capacity. When tip cells are stimulated by VEGFA, this
upregulates Delta-like canonical ligand-4 (Dll4) that subsequently activates Notch signaling in the adjacent stalk cell (Fig. 2). Notch signaling restricts the ability of the stalk cell to respond to VEGFA by suppressing VEGFR2 expression, thus preventing acquisition of tip cell
characteristics (Jakobsson et al., 2010).
VEGFA regulates developmental angiogenesis throughout the entire embryo, including the CNS, however the WNT family of soluble ligands and receptors has been shown to be
uniquely required for CNS vascular development. Canonical WNT-β-catenin signaling works via secreted WNT ligands binding to Frizzled/LRP cell-surface receptors. This leads to stabilization of -catenin by sequestration and inhibition of the -catenin destruction complex. This complex
21 entails many proteins, with the main effectors being GSK3, Axin, APC, and CK1 all of which work together to bind and phosphorylate -catenin at Ser33/37 and Thr41 targeting it for proteosomal degradation. Inhibition of this complex, through WNT-mediated signaling events, allows for -catenin to accumulate and translocate to the nucleus where it acts as a
transcriptional effector binding and activating WNT transcriptional machinery which includes TCF/LEF protein complexes (Moon, 2005).
Endothelial cells in the brain, spinal cord and retina, unlike endothelial cells in other organs, show evidence of WNT transcriptional activity as determined by transgenic reporter mouse lines (TOP-Gal and BAT-Gal) (Liebner et al., 2008, Daneman et al., 2009, Phng et al., 2009) and high expression of genes involved in the WNT-β-catenin pathway (Daneman et al., 2009, Hupe et al., 2017). Deletion of β-catenin or inactivation of its transcriptional activity in endothelial cells leads to CNS–specific blood vessel defects, ranging from brain areas with no vasculature to blood vessels that are enlarged and hemorrhagic (Stenman et al., 2008, Daneman et al., 2009, Zhou et al., 2014). WNT7a/b double-knockouts die at ~E12.5 with severe vascular defects in the CNS that largely phenocopy what is observed with inactivation of -catenin in endothelial cells (Stenman et al., 2008, Daneman et al., 2009). These studies demonstrated that WNT7a and WNT7b are the key WNT ligands in CNS vascular development. In the spinal cord and hindbrain, WNT7a and WNT7b are expressed in the neuroepithelium (Fig. 1)(Osumi et al., 1997; Daneman et al., 2009). In the forebrain, expression begins in the neuroepithelium but is expressed at even higher levels by post-mitotic neurons in the cortical plate (Fig. 1)(Stenman et al., 2008; Daneman et al., 2009). In the postnatal mouse neocortex, HIF1α activity in
oligodendrocyte progenitors cells drives production of WNT7a/b for postnatal angiogenesis (Yuen et al., 2014). WNT-driven vascularization of the CNS is mediated by the WNT receptors,
22 LRP5 and 6 (Zhou et al., 2014). Although Frizzled6 is highly expressed by brain endothelial cells during development (Daneman et al., 2009), it is unclear which Frizzled receptor is required for brain vascularization. Interestingly, Frizzled4 and its ligand Norrin, are uniquely required for retinal angiogenesis (Wang et al., 2012), suggesting Frizzled isoforms may elicit organ/region- specific signaling and developmental needs in regards to angiogenesis. Transduction of WNT7a/b binding to the Frizzled/LRP receptors is achieved through the G-coupled protein receptor, Gpr124 and its co-factor, Reck (Zhou and Nathans, 2014, Cho et al., 2017), a signaling event that is required for CNS vascularization (Fig. 2).
Currently, it is not clear how WNT-β-catenin signaling regulates CNS angiogenesis.
Gene expression analysis of brain endothelial cells lacking β-catenin did not detect significant dysregulation in cell cycle genes and reported increased expression of angiogenesis-related pathways in ECs from older mutant embryos (Hupe et al., 2017), suggesting that WNT-β-catenin may not stimulate vascular growth in the developing brain through traditional angiogenic
mechanisms. However, WNT7a can act as a chemoattractant to endothelial cells in vitro
(Daneman et al., 2009), indicating that like VEGFA, neural-derived WNTs may stimulate growth of vessels into the CNS by regulating endothelial cell motility. WNT-β-catenin signaling, also similar to VEGFA signaling, induces expression of the Notch receptor, Dll4. However this may be limited to early embryonic vascular developmental events (Fig. 2)(Corada et al., 2010).
Furthermore, WNT-β-catenin may induce angiogenesis by regulating the expression of other key transcription factors like Sox17. Sox17 is a target gene of WNT--catenin signaling and is capable of inducing VEGFR2 expression in growing vessels therefore acting as a positive regulator of VEGFA signaling in endothelial cells (Ye et al., 2009, Kim et al., 2016b). Recent RNA-sequencing studies from Sox17-deficient endothelial cells during embryonic development
23 revealed Sox17 as a positive regulator of the WNT--catenin pathway in the vasculature (Corada et al., 2018). Sox17 knockdown in cultured endothelial cells reduces proliferation (Lee et al., 2015). Furthermore, conditional deletion of Sox17 in the endothelium results in reduced
vascularization in the hindbrain and retina, whereas overexpression of Sox17 in endothelial cells results in aberrant angiogenesis in these regions (Lee et al., 2014). These studies demonstrate that Sox17 is crucial for CNS vascular growth (Fig. 2) however strictly regulating its expression appears to be just as important for proper vascularization. Interestingly, overactivation of WNT signaling in the developing embryonic vasculature, leading to aberrant Sox17 expression, is hypothesized to affect endothelial specialization specifically by up-regulating artery-specific genes and suppressing venous-specific genes in the endothelium (Corada et al., 2010). This suggests that, like Sox17, keeping endothelial-WNT signaling in check is likely important for brain vascular growth and maturation properties.
Retinoic acid (RA) has been recently implicated as an important regulator of brain vascular development by modulating VEGFA and WNT signaling in the developing neocortex.
RA signaling is initiated after vitamin A is brought into the cell and converted into bioactive forms of RA (all-trans RA and 9-cis RA). RA then binds and activates RA receptors (RARs) or retinoid X receptors (RXRs) that act as both receptors and transcription factors to control gene transcription (Al Tanoury et al., 2013). Foxc1 embryonic mutants, which lack meningeal-derived RA (Siegenthaler et al., 2009), display severe cerebrovascular hemorrhage (Mishra et al., 2016).
Studies in Foxc1 mutants showed this was due to defects in vascular growth into the developing neocortex and loss of WNT7a/b and VEGFA expression in this brain region. Furthermore, WNT antagonists such as Dkk1 and secreted frizzled related protein (Sfrp)-1, 2, 4 and 5 were found to be up-regulated in the meninges/PNVP of the Foxc1 mutants, likely impairing pial-derived
24 vessels from growing into the underlying neocortex. Maternal supplementation with RA rescued cerebrovascular defects and restored normal expression of VEGFA and WNT ligands and WNT signaling antagonists in Foxc1 mutants, implicating meningeal RA as a regulator of
cerebrovascular development (Mishra et al., 2016). This work shows that RA has an important, non-cell autonomous role in regulating vascular growth in the brain through modulation of VEGFA and WNT signaling (Fig. 2). Embryonic mutants null for RARα and RARγ have CNS developmental defects and display brain hemorrhaging in the cerebral hemispheres (Lohnes et al., 1994). However the role of RA signaling and endothelial-expressed RARs during CNS vascular development has yet to be investigated.
Maturation of the CNS vasculature – properties important for vascular stability
Soon after entering the neural tissue, blood vessels acquire many properties important for vessel stability. Of particular importance are enrichment of junctional complexes and abluminal coverage by pericytes, both of these properties being crucial for vascular stability and preventing hemorrhage (Stenman et al., 2008, Daneman et al., 2010b, Lindahl et al., 1997, Winkler et al., 2011). Junctional complexes are detected as vessels invade into the brain, forming between endothelial membranes to establish strong cell-cell connections through the formation of tight junctions (TJs) and adherens junctions (AJs) that seal and stabilize the growing vasculature respectively (Daneman et al., 2010b). Enrichment of these junctional complexes is thought to be driven by endothelial-WNT signaling (Liebner et al., 2008, Zhou et al., 2014, Zhou and Nathans, 2014) to create a nearly impermeable and continuous vascular network in the CNS (Fig. 3).
Pericytes are a subtype of mural cells that ensheath the abluminal surface of the capillary endothelium and are found at higher densities in the brain vasculature as compared to the
25 vasculature in other organs (Mathiisen et al., 2010). They build and share an extracellular matrix with the vasculature and form both structural and signaling connections with the endothelium (Winkler et al., 2011, Armulik et al., 2011). Platelet-derived growth factor B (PDGFB) is a mural cell mitogen and chemoattractant that is secreted by the endothelium. PDGFB is specifically concentrated at the tip cell and retained in the newly formed extracellular matrix around the vasculature. Concentrated PDGFB at the growing tip of the vessel attracts PDGF-receptor (PDGFR)-expressing pericytes (Lindblom et al., 2003). This drives pericyte recruitment along the vasculature through PDGFB/PDGFR-stimulated migration, proliferation and survival of pericytes (Fig. 3). Embryonic mouse mutants null for Pdgfb and Pdgfrb or defective in extracellular retention of PDGFB (Pdgfbret) lack pericytes embryo-wide. Loss of pericyte coverage in the CNS vasculature does not impact vessel growth or branching however there are significantly more endothelial cells per vessel. The resulting vessel hyperplasia increases susceptibility to brain hemorrhage (Lindahl et al., 1997, Hellstrom et al., 2001). Thus pericyte investment through PDGFB/PDGFR signaling during brain vascular growth is essential for vascular stability, potentially by limiting endothelial cell proliferation. PDGFB expression in the developing brain vasculature is especially high (Hupe et al., 2017). However the signaling pathways that drive the high density of pericytes and PDGFB expression in the CNS vasculature in unknown. Interestingly, overactivation of WNT signaling in CNS tumor endothelial cells increases PDGFB expression and mural cell coverage, including pericytes (Reis et al., 2012).
This suggests that endothelial-WNT signaling could be one pathway responsible for the high density of pericytes during CNS blood vessel development (Fig. 3). Importantly, limiting pericyte recruitment and PDGFB/PDGFR signaling also appears to be crucial for vascular
26 integrity. Studies using a mouse model of retinopathy of prematurity found that over-activation of PDGFB/PDGFR signaling results in excessive pericyte coverage in the retina and promotes the formation of neovascular tufts, which are indicative of endothelial cell over-proliferation.
This resulted in a weak vascular network that was susceptible to hemorrhage (Dubrac et al., 2018). Additionally, pericyte differentiation was impaired and vessel dilation was observed in the brains of mice where PDGFRβ was constitutively active during embryonic development (Olson and Soriano, 2011). These studies point to the importance of controlling
PDGFB/PDGFR signaling and pericyte recruitment to establish a stable vascular network within the CNS.
Exploring the role of RA in WNT-mediated blood vessel growth and maturation
Part I of this Chapter highlights the cellular events of the growing brain vasculature, angiogenic pathways important for CNS vascularization, and maturation properties that permit vascular stability. Much work has focused on signaling pathways that drive these processes however very little is known about how these cellular events, signaling pathways and maturation properties are controlled to build a stable vascular network. In Chapters II and III, I investigate how RA signaling regulates WNT-mediated vascular growth and maturation properties like pericytes that are needed for vascular stability (Fig. 5A). Specifically, Chapter II reveals that RA regulates WNT signaling in both non-cell and cell autonomous manners to regulate CNS
vascular growth and control endothelial-Sox17 expression, ultimately establishing a healthy brain vascular network. Chapter III extends this work where I explore the underlying mechanism through which RA regulates endothelial-WNT signaling and its role in controlling properties that are acquired by the growing brain vasculature, in particular pericyte recruitment.
27 Part II: Acquisition of blood-brain barrier properties by the growing CNS vasculature Blood-brain barrier properties – a developmental perspective
As soon as a vessel is formed in the brain, it acquires BBB properties to protect the CNS from potentially damaging contents within the blood. The BBB is a physical barrier created by brain endothelial cells through a progressive acquisition of unique properties, most notably enrichment of junctional complexes, transporters, and pericytes. Furthermore, the brain
endothelium has reduced transcytosis and fenestrations/pores, properties that are found in leakier vessels within non-CNS organs. Through these BBB properties, the brain vasculature controls the passage of molecules from circulation into the CNS.
Studies investigating when the BBB seals during development found that the
permeability of horseradish peroxidase (HRP) subsided around E15 in mice (Risau et al., 1986a).
However, in the early postnatal rodent many molecules that are excluded from the adult brain are still capable of crossing the BBB (Johanson, 1980, Fabian and Hulsebosch, 1989). This suggests that there is a gradual maturation in the BBB during brain development, likely due to a
progressive tightening of junctional complexes and loss of leaky properties. This is supported by ultrastructural studies in late embryonic and postnatal rat that correlate the disappearance of fenestrations and interjunctional clefts (spaces between endothelial cells without components of junctions) with decreasing permeability to HRP that plateau in the 3rd postnatal week (Stewart and Hayakawa, 1994). A separate developmental study utilizing freeze fracture to visualize tight junctions between endothelial cells in rat cortex observed a dramatic increase in the density and complexity of TJs in the last three days of fetal development (E18-P1) (Kniesel et al., 1996).
These studies demonstrate that the BBB matures during embryonic and postnatal periods likely through the gradual acquisition of BBB properties.
28 Enrichment of barrier properties in brain endothelial cells during development
Both AJs and TJs are important for inducing BBB function by stabilizing and sealing the endothelium. AJs are found on the abluminal (brain) side of the endothelium and maintain the connections between neighboring endothelial membranes. Strong AJ connections facilitate TJ formation on the luminal (blood) side to “zip” up the endothelium and restrict free movement of molecules from the blood to the brain (Dejana et al., 2009). Expression of TJ proteins, like Claudin5, Occludin, and Zona Occludin-1 (ZO-1), can be found as early as E12.5 in the brain vasculature (Daneman et al., 2010b). AJ proteins, like VE-cadherin, are constitutively expressed in the vasculature and are essential for embryonic angiogenesis (Gory-Faure et al., 1999).
Catenins are AJ adaptor proteins and like the TJ adaptor protein, ZO-1, facilitate connections with the actin cytoskeleton (Fig. 3). The actin cytoskeleton is an important factor in maintaining junctional connections by eliciting cell shape and essentially pushing the endothelial membranes together to maintain cell-cell contacts (Hawkins and Davis, 2005). Junctional complexes and the cytoskeleton work together to stabilize the endothelium and restrict the movement of substances into the brain.
Since TJs seal the endothelium, transport of CNS waste and nutrients essential for proper neural function, is strictly regulated by influx and efflux transporters that are specific to the brain endothelium. The glucose transporter Glut1, which is enriched in brain endothelial cells, is present as early as E11 in embryonic mouse brain vasculature (Fig. 3)(Daneman et al., 2009).
Studies employing translating ribosome affinity purification (TRAP) RNA-sequencing to identify gene transcripts that are actively being translated in embryonic brain endothelial cells found that transporter genes (for the transport of water, ions and amino acids) are highly
expressed in the brain endothelium at E14.5 (Hupe et al., 2017). In some cases, the expression of
29 a specific transporter is limited to the luminal or abluminal side of the endothelial cells thus conferring a polarity to the endothelial cell that ensures proper directional movement of molecules and proteins (Abbott et al., 2010).
Astrocytes are a glial cell type that extend polarized processes called endfeet to ensheath the mature brain vasculature. The close proximity of the astrocytic endfeet to the abluminal side of the endothelial cell wall and their BBB inducing affect in endothelial cell co-cultures have led to the speculation that astrocytes participate in BBB function. However astrocytes do not appear until around birth in mice (Molofsky and Deneen, 2015), well after a functioning BBB is in place (Ben-Zvi et al., 2014). This indicates that astrocytes can induce barrier properties but are not required for prenatal BBB development. Recent studies by Kubotera et al., tested whether astrocytes participate in BBB maintenance by using laser-ablation approaches to remove
astrocytes or their processes and endfeet in the adult mouse. They found that BBB function was not compromised in either instance providing evidence that astrocytes do not participate in BBB maintenance (Kubotera et al., 2019). That said, there is evidence of roles for astrocytes in regulating water transport across the endothelium and neurovascular coupling, a process in
which astrocytes regulate blood flow in response to neuronal activity (Daneman and Prat, 2015).
Suppression of peripheral endothelial properties in the developing brain vasculature
Pores or fenestrations in endothelial membranes are characteristic of blood vessels in organs that require the rapid exchange of molecules. Early during brain development, around E11, fenestrations can be found within the immature developing brain vasculature. As the vasculature matures, fenestrations, identified through EM and the expression of the
plasmalemma vesicle associated protein (PLVAP), decline by E13 in rodents (Fig. 3)(Stewart
30 and Hayakawa, 1994, Daneman et al., 2010b). This forces the transport of macromolecules through either transporters or highly regulated transcytosis processes.
Transcytosis is a process in which larger macromolecules can be transported in and out of tissues through endothelial cells and is a cellular process that is suppressed in brain endothelial cells. Alternatively, passage of macromolecules is highly regulated via either receptor-mediated transcytosis (RMT) or absorptive-mediate transcytosis (AMT). RMT is essential for transport of iron through the transferrin receptor and lipoproteins through the Apolipoprotein E receptor 2.
AMT is important for the transportation of cationized proteins through the endothelium (Abbott et al., 2010). Pericytes actively suppress transcytosis in the brain vasculature (Fig. 3)(Armulik et al., 2010, Daneman et al., 2010b) most likely to ensure transport of macromolecules occurs through the regulated processes of RMT and AMT. Although junctional disorganization is observed in Pdgfrb mutants (Daneman et al., 2010b), increased transcytosis in Pdgfb, Pdgfbret, and Pdgfrb mutants is thought to underlie BBB leakage (Armulik et al., 2010, Daneman et al., 2010b). The pericyte derived signal(s) that suppress transcytosis in the endothelium has not been identified. Mfsd2a is a CNS endothelial cell enriched protein that, when deleted, causes
increased transcytosis in brain endothelium and BBB leakiness in embryonic and adult mice.
Pericytes stimulate expression of Mfsd2a in endothelial cells and expression of Mfsd2a is reduced in mice that lack pericytes, indicating that Mfsd2a is a key target of a pericyte-derived signal(s) needed for barrier development and maintenance (Fig. 3)(Ben-Zvi et al., 2014). It is important to note that recent live imaging studies of the BBB using two-photon microscopy and laser ablation of individual pericytes did not induce paracellular (junctional defect) or
transcellular (increased transcytosis) mediated BBB leakage in adult animals (Berthiaume et al., 2018). These studies suggest that there may be a threshold in BBB maintenance by pericytes or
31 they are not a major factor in inhibiting transcytosis and facilitating junctional stability in the adult brain.
Regulatory signals that drive blood-brain barrier properties during development – a focus on BBB junctions and transporters
Early work using mouse-chick and chick-quail chimera experiments provided evidence that neural tissue drives the acquisition of BBB characteristics in endothelial cells. When avascular mouse or quail brain was transplanted into chick chorio-allantoic membrane or coelomic cavity, it induced expression of BBB-specific proteins and the formation of TJs on endothelial cells in the non-neural tissue (Stewart and Wiley, 1981, Risau et al., 1986b).
Subsequent studies have identified several neural-derived factors that are important for BBB development.
Neural-derived WNT ligands and endothelial WNT-β-catenin signaling is vital for BBB development. Treatment of brain endothelial cells with a canonical WNT ligand, WNT3a, leads to upregulation of junctional proteins and the appearance of TJs between cultured brain
endothelial cells (Liebner et al., 2008, Daneman et al., 2009). The BBB-inducing effect of WNT on the developing brain vasculature is dependent on canonical WNT signaling and requires Gpr124 and β-catenin transcriptional activity (Zhou et al., 2014, Zhou and Nathans, 2014, Cho et al., 2017). Loss of endothelial-Sox17 in embryonic mice and -catenin in postnatal animals revealed an increase in PLVAP expression (Liebner et al., 2008, Corada et al., 2018) indicating that endothelial WNT--catenin-Sox17 signaling is required to reduce fenestrations in immature brain blood vessels (Fig. 3). These and other studies also found a downregulation in the TJ proteins Claudin-3/5, combined with a loss of BBB integrity in postnatal animals lacking
32 endothelial -catenin transcriptional activity or expression (Fig. 3)(Liebner et al., 2008, Zhou et al., 2014). Furthermore, embryos with the conditional deletion of β-catenin in endothelial cells fail to upregulate the BBB-enriched transporter Glut-1 (Fig. 3) but express other junctional proteins like ZO-1 and VE-cadherin (Stenman et al., 2008, Daneman et al., 2009). This suggests that WNT signaling may regulate specific barrier properties within the brain vasculature. A more complete picture of the BBB properties downstream of WNT-β-catenin in the brain endothelium of wild-type and mutants with endothelial specific knock-out of -catenin supported the findings in these earlier studies using TRAP-RNA sequencing. They found that loss of -catenin in brain endothelial cells of developing embryos resulted in an attenuation of vascular maturation
processes. Specifically they found reductions in brain endothelial transporters and Claudin-5 expression in -catenin deficient endothelial cells. This study also identified the transcription factors Foxf2 and ZIC3 that act downstream of WNT--catenin signaling to induce the expression of a variety of BBB transcripts indicative of vessel maturation (Hupe et al., 2017).
Collectively these studies indicate that canonical WNT-β-catenin signaling in the CNS vasculature is critical for both the initiation and maturation of the BBB, likely through transcriptional activation of BBB-specific junctional and transporter genes.
RA is also implicated in induction of BBB properties, specifically by inducing junctional gene expression. Astrocytes are predicted to be a major source of RA as they express the retinoic acid synthesis protein, RALDH, in human brain tissue (Mizee et al., 2013). In culture
experiments, astrocyte-derived RA induced barrier properties in brain endothelial cells. Further studies showed that embryonic exposure to a pan-RAR inhibitor impaired BBB development and reduced ZO-1 expression in the brain vasculature (Mizee et al., 2013, Mizee et al., 2014).
33 Pharmacological concentrations (>5M) of all-trans RA increases ZO-1 and VE-cadherin
expression and improves BBB tightness in cultured human brain endothelial cells. This effect is hypothesized to be mediated by RAR (Mizee et al., 2013). The pharmacological effect of all- trans RA is also observed in induced pluripotent stem cell-derived brain endothelial cells (Lippmann et al., 2014, Katt et al., 2016). These studies demonstrate that pharmacological concentrations of all-trans RA is a valuable tool to strengthen barrier properties by inducing junctional protein expression in BBB culture models. However the pharmacological
concentrations of all-trans RA used in these studies are considerably above what is thought to be a physiological range of all-trans RA (2-600nM)(Napoli et al., 1991). High concentrations of all- trans RA have been shown to isomerize to 9-cis RA (Urbach and Rando, 1994). This raises the possibility that high concentrations of all-trans RA may also act via 9-cis RA related pathways like PPAR/RXR or LXR/RXR (Szanto et al., 2004). Even though these studies suggest that RA may play an important role BBB development, the physiological role of RA in BBB
development is unclear (Fig. 3). Moreover, it is unknown how the pharmacological
concentrations of all-trans RA are working to induce junctional gene expression in BBB culture models.
Investigating the role of RA in the acquisition of blood-brain barrier properties
In this section, I discussed the acquisition of BBB properties by the developing brain endothelium and how those properties work together to create an impermeable vascular network in the brain. I also reviewed signals, like WNT and RA signaling, that are implicated in inducing BBB properties. In Chapter IV I test if RA plays a physiological role in BBB development using
34 RA deficient embryos (Rdh10 mutants). I also investigate how pharmacological concentrations of all-trans RA used in BBB culture models work to induce junctional gene expression (Fig. 5B).
Part III - Breakdown of the blood-brain barrier in viral encephalitis
Impact of blood-brain barrier leakage in disease – emphasis on viral infection in the CNS Deterioration of BBB properties resulting in BBB dysfunction is observed in stroke, neurodegenerative diseases, epilepsy, and other encephalopathies such as encephalitis in the CNS. Many diseases share similarities in how BBB properties are eroded. For example, reductions in vascular junctions, acceleration in immune cell infiltration (an event that is generally suppressed by a healthy BBB), and pericyte loss is observed in almost all of these diseases. This leads to reduced oxygen/nutrient delivery, leakage of serum constituents, and acceleration of immune cell infiltration. These phenomenon are thought to contribute to poor tissue oxygenation, edema, heightened inflammatory responses and disruption of neuronal activity, all hallmarks of many CNS diseases. Even though BBB leakage and loss of barrier properties is commonly observed, the mechanisms that initiate BBB breakdown are not fully understood (Hawkins and Davis, 2005, Obermeier et al., 2013, Engelhardt and Liebner, 2014, Liebner et al., 2018). Efforts to uncover these mechanisms are well underway in diseases such as Alzheimer’s where a PubMed search for “Blood-brain barrier disruption Alzheimer’s disease”
yields 211 articles. A similar search for BBB leakage in viral causes of encephalitis, a major cause of encephalitis in humans, yields only 41 articles and many of them fall short of identifying mechanisms inducing BBB leakage.
The major characteristics of viral infection in the CNS include edema, increased immune infiltration, and poor neuronal activity – all potential consequences of poor BBB integrity.
35 Uncovering the mechanisms of BBB breakdown in viral encephalitis could ultimately provide new therapies to improve treatment in patients.
Disease pathogenesis and blood-brain barrier leakage in viral encephalitis
Mouse models of viral encephalitis have yielded important knowledge regarding viral pathogenesis in humans. In general, viral infection in the CNS results in poor neuronal activity and cell death in the infected cells. Cell death and viral antigens initiate an immune response, which involves activation of microglia and astrocytes, along with infiltration of neutrophils, macrophages and T cells. This accelerates the disease due to increases in cell death-inducing cytokines and bystander engulfment of healthy cells by phagocytes (Dahm et al., 2016).
Disruption to BBB integrity has been observed in human cases of HIV-related encephalitis (Dallasta et al., 1999, Rahimy et al., 2017) and is speculated to occur in other viral causes of encephalitis in humans caused by Japanese encephalitis (JEV), West Nile (WNV), La Crosse (LACV), Alphaviruses and Herpes simplex viruses type 1 and 2 (HSV1/2).
Exposure to encephalitis-causing viruses such as JEV, WNV, LACV and Alphaviruses can occur through arthropod vectors, like mosquitos. Following exposure, viral entry into the CNS occurs through several mechanisms. These mechanisms include 1) direct infection of the brain endothelial cells, 2) infection and subsequent migration of immune cells into the CNS, or 3) retrograde axonal transport of viral particles from peripheral nerves (Dahm et al., 2016).
Reactivation of latent HSV in the temporal lobe or peripheral ganglia and subsequent retrograde axonal transport is hypothesized to be a cause of HSV-induced encephalitis in patients (Gnann and Whitley, 2017). Regardless of entry routes, loss of BBB integrity has been observed in
36 mouse models of Zika, JEV, WNV, LACV, and Alphavirus-induced encephalitis (Daniels and Klein, 2015, Leibrand et al., 2017).
Alterations in blood-brain barrier properties during CNS viral encephalitis
TJs have been the main focus for how BBB leakage may occur in mouse models of viral encephalitis. The TJ proteins Claudin-5, Occludin, and ZO-1 are reported to be reduced within the brain vasculature of mice infected with neurotropic viruses and this correlates with BBB leakage (Chai et al., 2014, Bleau et al., 2015, Li et al., 2015, Kim et al., 2016a). In contrast, other studies have reported that TJ protein expression is unaltered but localization at the cell borders may be disrupted during infection (Daniels et al., 2014, Miner et al., 2015, Cain et al., 2017). The latter studies have implicated the cytoskeleton as a major component of regulating TJ protein stability and localization at cell junctions. The prevailing model is that Rac and Rho-GTPases have opposing functions in regulating BBB integrity, with the former being barrier protective and the latter barrier disruptive. For example, activation of innate immune responses by pathogen-associated molecular patterns (PAMPs) can influence barrier integrity in brain endothelial cells including altering localization of TJ proteins at the endothelial cell borders.
Daniels et al., found that Interferon- receptors (IFNR) are barrier protective and attenuate the barrier disrupting effect of the PAMPs. The authors found that IFNR-mediated signaling activates Rac1-GTPases, proteins that are responsible for strengthening cell-cell connections by stabilizing the actin cytoskeleton (Fig. 4). These studies suggest that Type I IFN signaling mediated by IFNR is barrier protective during CNS viral infection. Interestingly, a number of studies are emerging indicating that Type II IFN signaling mediated by IFN is barrier disruptive (Chai et al., 2014, Li et al., 2015, Daniels and Klein, 2015).
37 Major pathways involved in blood-brain barrier breakdown in viral encephalitis
Activation of IFN signaling is essential to control viral infections. IFN signaling is initiated by IFN ligands, such as IFN and IFN that bind to and activate IFNR, or IFN that bind to and activate IFNGR. This leads to the phosphorylation and heterodimerization of
STAT1/2 or dimerization of STAT1 for Type I or Type II IFN signaling respectively. Type I IFN signaling is thought to function early in viral infection and is responsible for eliciting anti-viral responses to limit viral replication and activating innate immune responses. Type II IFN signaling, mediated by IFN, is also important for anti-viral responses but is thought to be a strong inducer of adaptive immune responses by enhancing antigen presentation to activate T cells and create memory T cells. Although IFN signaling is important for immune and anti-viral responses, all cell types have IFN receptors and are capable of responding to IFN ligands, in this way IFNs can induce a variety of cellular responses during disease (Platanias, 2005, Lee and Ashkar, 2018). These include the BBB protective effects of astrocyte-mediated IFN/IFNAR- signaling (Daniels et al., 2014) and the hypothesized action of IFN/IFNGR-signaling in reducing BBB properties in brain endothelial cells (Chai et al., 2014, Li et al., 2015). Studies investigating IFN-mediated effects on brain endothelial cells during viral infection used brain lysate from infected mice to induce barrier disruption in cultured brain endothelial cells. They found that neutralization of IFN using antibodies directed at IFN improves TJ protein expression and barrier properties in these experiments. Furthermore, neutralization of IFN
improves BBB leakage in mice infected with Japanese encephalitis and rabies viruses (Chai et al., 2014, Li et al., 2015). These studies suggest that Type II IFN signaling, induced by IFN, disrupts the BBB during CNS viral infection (Fig. 5). However the direct action of IFN on brain
38 endothelial cells has yet to be tested and could have multiple effects on barrier properties in addition to affecting TJ protein expression. For example, IFN activation of Rho-associated protein kinase (ROCK) is hypothesized to initiate cytoskeletal rearrangements, forcing the internalization of TJs in epithelial cells (Utech et al., 2005). It’s possible a similar effect is observed in brain endothelial cells (Fig. 5). Additionally, a majority of these studies have focused solely on TJs. The effect of CNS viral infection on other BBB properties like AJs, pericytes and transcytosis, are unknown.
Investigating the breakdown of the blood-brain barrier in viral encephalitis
In this chapter, I introduced how BBB leakage impacts disease and how barrier properties are altered during disease. I focused on how BBB leakage is known to occur in cases of viral- induced encephalitis but unlike other CNS diseases, very little is known about the mechanisms and pathways underlying BBB leakage. To address this, in Chapter V, I conducted studies to understand the widespread effect of viral infection in the CNS on BBB properties in an
experimental mouse model of viral encephalitis (Fig. 5C). During these studies, I identify IFN- mediated signaling as a significant contributor to BBB leakage and use cell culture studies to understand how mechanistically, IFN reduces barrier properties in brain endothelial cells.
39 Figure 1.1 – Blood vessel development and angiogenic ligand expression in the CNS: In the E9.5 mouse spinal cord, angiogenic sprouts from the perineural vascular plexus (PNVP) grow inward toward the ventricular lumen and establish an immature vascular plexus by E11.5 in response to the low-to-high WNT and VEGF gradients. Around E10 in mouse, radially oriented branches of
the PNVP have extended into the hindbrain and eventually form lateral connections with adjacent vessels by E11. The initial growth of vessels into the hindbrain relies on VEGF and WNT expressed the neuroepithelium lining the roof of the 4th ventricle. A steep, high-to-low gradient of VEGF at the midline (dark blue) is important for elaboration of the plexus through secondary branching and fusion. In the mouse forebrain, angiogenic sprouts from the PNVP at the pial surface penetrate the forebrain surface in a dorsal to ventral manner around E10.5; in the
level dorsal forebrain, the pial vessel may grow inward in response to Wnt7a/b expressed by postmitotic neurons in the forming cortical plate. Figure adapted from Chapter 43 in the 2nd edition of Comprehensive Developmental Neuroscience and is currently under review. Bonney
S., Mishra S., Pleasure S.J., Siegenthaler J.A. “Meninges and Vasculature”.
40 Figure 1.2 – Pathways involved in blood vessel growth in the CNS: Neuroepithelial-derived VEGFA acts on the VEGFR2-expressing endothelium to induce ingression of the perineural vascular plexus (PNVP) into the CNS. This occurs through VEGFR2-induction of delta-like canonical ligand-4 (Dll4) expression to promote a migratory phenotype in the tip cell of the angiogenic sprout. Dll4 activation of Notch signaling in stalk cells induces proliferation to promote blood vessel development in the CNS. WNT7a/b, which is specifically required for angiogenesis the CNS, is secreted from the growing neuroepithelium and post-mitotic neurons (purple) to stimulate blood vessel growth similar to VEGFA signaling potentially through Dll4- Notch signaling. Additionally, Sox17, a WNT target gene, has been shown to be involved in CNS angiogenesis potentially by mediating endothelial cell proliferation. WNT7a/b signaling in
the endothelium is mediated by Fzd, Lrp5/6, Gpr124 and Reck, all of which are required for blood vessel growth into the CNS. Retinoic acid (RA) likely derived from the meninges, has been shown to be positive regulator of endothelial-VEGFA and WNT7a/b signaling in a non-cell
autonomous manner. This figure is adapted and updated from Obermeier et al., 2013 Nature Medicine.
41 Figure 1.3 – Blood brain barrier properties and regulatory signals during development: Early
during brain vascular development, WNT-signaling plays a part in the acquisition and enrichment of junctional complexes. This includes tight junction proteins like Cldn5 (WNT
target gene), Ocln and ZO1. Furthermore, WNT-mediated stabilization of -catenin (-cat) localizes to and stabilizes adherens junctions through interactions with VE-cadherin (VE-cad).
Additionally PDGFB secreted by the growing endothelium, and potentially regulated by endothelial WNT signaling, attracts PDGFR-expressing pericytes. As the vasculature matures, WNT-signaling induces a variety of blood-brain barrier (BBB) properties like the expression of
brain endothelial transporters, such as Glut-1, and the suppression of PLVAP to reduce fenestrations. Additionally, pericytes suppress transcytosis by maintaining Mfsd2a expression in endothelial cells (ECs). Retinoic acid (RA) has also been implicated in regulating junctional gene
expression in BBB culture models however it is unclear if it plays a physiological role during development. Acquisition of these properties is crucial to establish a stable and healthy BBB
network in the CNS. This figure is adapted and updated from Obermeier et al., 2013 Nature Medicine.
42 Figure 1.4 – Blood-brain barrier breakdown in viral encephalitis: During viral infection in the CNS, Astrocyte-derived IFN/ activates IFNAR on brain endothelial cells (ECs). This induces
barrier protective mechanisms mediated by Rac-GTPase activity that stabilizes the actin cytoskeleton. This helps to facilitate cell shape and strengthen tight (TJ) and adherens junction (AJ) connections. It is hypothesized that later in disease, T cell secretion of IFN is heightened and induces barrier disruption, potentially by disrupting cell-cell contacts or reducing AJ and TJ
protein expression, however the underlying mechanism is unknown. This figure is adapted and from Daniels and Klein., 2015 PLOS Pathogens.
43 Figure 1.5 – Investigations of blood vessel growth, BBB properties and breakdown in disease:
(A) In Chapters II and III, I explore the role of retinoic acid in regulating WNT-mediated vascular growth and maturation of early brain vascular properties like junctions and PDGFB/PDGFR-mediated pericyte recruitment. (B) In Chapter IV, I investigate if retinoic acid plays a role in the acquisition of BBB properties. I also test how pharmacological concentrations
of all-trans retinoic may induce BBB properties in brain endothelial culture models. (C) In Chapter V, I perform a comprehensive investigation into how various BBB properties are disrupted and the signaling events involved to induce BBB leakage in an experimental mouse
model of viral encephalitis.
