University of Southern California Dissertations and Theses

Neural crest-derived cranial pericytes and their dysfunction in disease

(USC Thesis Other)

Neural crest-derived cranial pericytes and their dysfunction in disease

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

NEURAL CREST-DERIVED CRANIAL PERICYTES AND THEIR DYSFUNCTION IN
DISEASE

By
Casey Griffin
Mentor: Ruchi Bajpai

A Dissertation Present to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(DEVELOPMENT, STEM CELLS, and REGENERATIVE MEDICINE BIOLOGY)

December 2019
2
Table of Contents
1. Acknowledgements 3
2. Abbreviations 5
3. List of Figures 7
4. Chapter One: Cranial Pericytes 10
5. Chapter Two: In vitro derivation of cranial pericytes via a neural 15
crest intermediate
6. Chapter Three: Familial AD pericytes have lifelong, structural, 52
functional, and survival defects
7. Chapter Four: Structural and functional defects in sAD pericytes 83
8. Chapter Five: Defects in cranial pericytes identified in CHARGE 107
syndrome
9. Materials and Methods 121
10. References 142

3

ACKNOWLEDGEMENTS

I would like to thank my mentor Dr. Ruchi Bajpai, for providing me with every
opportunity to become the best scientist I can be. She gave me the freedom to pursue
my own questions within my project, as well as learn any and all experimental
techniques that were of interest. She always supported me and my work and gave me
the confidence to present my work with pride. She always gave me the chance to
present my project at conferences, as well as attend workshops to improve my skills.
Mostly, she supported me as a scientist, student, and person, and sought out nothing
but success for me and my work.
Thank you to my committee members: Dr. Gage Crump (chair), Dr. Berislav
Zlokovic, and Dr. Justin Ichida. Their insight and support over the years have helped to
shape my project, as well as my career path. Their guidance has been of unparalleled
importance, and I would not have completed my dissertation without them. Thank you
again to Dr. Gage Crump and to Dr. Michael Payne for their financial support through
training grants. They supported me for 3 years on their grants, allowing me to continue
my studies without obstacles, as well as providing me opportunities to grow and learn.
Also, a special thanks to Dr. Amy Merrill and Dr. Francesca Mariani, who both
supported me and my work throughout my time at USC and provided guidance and
insight whenever I needed it.
I want to thank my lab mates, past and present, for their collaboration, their
support, their help, and their friendship over the years. Thank you, Jennifer Oki, for
helping me so much when I joined the lab, and for keeping me sane during my rotation.
4
Thank you, Annie Lynch, for your friendship, your support, and your comfort, especially
during my qualifying exam times. Thank you, Yuhan Sun, for keeping me smiling,
lending a helping hand, and being there for me when I felt lost. Thank you, Susan
Smith, for being a shoulder to cry on, a person to vent with, my baked goods supplier,
and my friend, along with all your technical assistance and general knowledge and
insight. Thank you, Joanna Salva, even though you were never my lab mate, for your
friendship, your support, and your keen editing eye. Thank you Kaivalya Shevade, for
everything; you’re my colleague, my friend, and my brother, and I wouldn’t have
survived without you.
Lastly, thank you to all my family and friends, both here and back east, who have
supported me, comforted me, encouraged me, and guided me during these long five
years. Thanks to the friends out here for making the west coast bearable and thank you
to everyone back east for keeping me motivated to graduate so I could return back east.
Thank you especially to my husband Eric Tracy, without whom I would probably have
died on my couch because I would’ve forgotten to eat and sleep while working. He kept
me sane, he kept me happy, he kept me motivated, and he never failed to support or
encourage me during the hardest times.

5
ABBREVIATIONS
044-cPC = 044 sAD patient iPSC-derived cPC
1
o
PC = primary pericytes
ANGPT1 = angiopoietin 1
AD = Alzheimer’s disease
ADAD = autosomal dominant Alzheimer’s disease
BBB = blood-brain barrier
CHD7-cPC = CHARGE patient iPSC-derived cPC
CNS = central nervous system
cPC = cranial pericytes
EC = endothelial cells
ECM = extracellular matrix
fAD = familial Alzheimer’s disease
hESC = human embryonic stem cells
iFF = female fibroblast-derived
iMF = male fibroblast-derived
iPSC = induced pluripotent stem cells
iPSEN1* = PSEN1
A431E
mutation
iPSEN1
corr
= PSEN1
A431E
CRISPR/Cas9 corrected mutation
Media
cm
= conditioned media
mesPC = mesodermal pericytes
MSC = mesenchymal stem cells
NCC = neural crest cells
6
PC = pericytes
PCM = pericyte compatible media
sAD = sporadic Alzheimer’s disease
SMC = smooth muscle cells

7
List of Figures
1. Figure 2.1. Comparison of human brain and placental PC 17
2. Figure 2.2. Comparison of mouse and human brain 1°PC 19
3. Figure 2.3. Comparison of mesPC and cPC 21
4. Figure 2.4. in vitro derivation of cPC 24
5. Figure 2.5. Transition of cells from NCC to cPC 26
6. Figure 2.6. cPC transcriptionally reflect pericytes 29
7. Figure 2.7. cPC generate extracellular exosomes 31
8. Figure 2.8. cPC secrete various ECM proteins 32
9. Figure 2.9. cPC exhibit contraction in response to stimuli 34
10. Figure 2.10. cPC exhibit random and directed migration 35
11. Figure 2.11. cPC migrate toward EC and influence EC organization 37
12. Figure 2.12. cPC induce EC-EC contacts 38
13. Figure 2.13. cPC reduce EC proliferation 40
14. Figure 2.14. cPC induce EC tube formation 41
15. Figure 2.15. cPC home to craniofacial microvessels in Xenopus tadpoles 44
16. Figure 2.16. cPC home to vasculature in mouse brain slices 45
17. Figure 2.17. cPC induce barrier properties in EC 47
18. Figure 2.18. Transferrin uptake across in vitro barriers 48
19. Figure 2.19. Ab efflux across in vitro barriers 49
20. Figure 3.1. Generation of iPSEN1*-cPC 55
21. Figure 3.2. Structural and survival defects in iPSEN1*-cPC 56
22. Figure 3.3. NG2 mislocalization with loss of PSEN1 58
8
23. Figure 3.4. iPSEN1*-cPC unable to properly reduce EC proliferation 61
24. Figure 3.5. Poor EC-EC junctions induced by iPSEN1*-cPC 62
25. Figure 3.6. iPSEN1*-cPC show defect in inducing tube formation 63
26. Figure 3.7. Blockage of paracellular transport partially defective in 64
iPSEN1*-cPC
27. Figure 3.8. Poor Ab clearance across barrier with iPSEN1*-cPC 66
28. Figure 3.9. Poor Transferrin uptake across barrier with iPSEN1*-cPC 67
29. Figure 3.10. Structural defects in neonatal TgF344 rat pericytes 68
30. Figure 3.11. Pericyte defects occur ab initio in transgenic rat model of 70
fAD in vivo
31. Figure 3.12. fAD pericytes show reduced expression of ANGPT1 71
32. Figure 3.13. Rescue of EC-EC contacts in iPSEN1*-cPC cultures with 74
ANGPT1
33. Figure 3.14. Tube formation defect in iPSEN1*-cPC partially rescued 76
with ANGPT1
34. Figure 3.15. Loss of ANGPT1 in pericytes leads to structural defects 77
35. Figure 3.16. TIE2 inhibitor phenocopies ANGPT1 knockdown/fAD 79
36. Figure 3.17. Recombinant ANGPT1 rescues structural defects 80
37. Figure 3.18. ANGPT1 rescues morphological defects 81
38. Table 4.1. Source of post mortem brain tissue samples for primary 86
pericyte isolation
39. Figure 4.1. Structural defects in sAD 1°PC 87
40. Figure 4.2. sAD 1°PC have defect in EC proliferation arrest 89
9
41. Figure 4.3. sAD 1°PC have a defect in inducing EC tube formation 90
42. Figure 4.4. sAD 1°PC are able to block paracellular transport across 91
a barrier
43. Figure 4.5. sAD 1°PC show defects in transcytosis across a barrier 92
44. Figure 4.6. In vitro-derived sAD-cPC show similar defects as sAD 1°PC 95
45. Figure 4.7. Various processes and cellular components downregulated 98
in sAD 1°PC
46. Figure 4.8. Reduced ANGPT1 expression in sAD cells 99
47. Figure 4.9. ANGPT1 rescues EC communication defect in sAD 1°PC 101
48. Figure 4.10. Exogenous ANGPT1 rescues sAD 1°PC structural defects 103
49. Figure 4.11. Exogenous ANGPT1 rescues morphology and migration 104
defects in sAD 1°PC
50. Figure 4.12. PTPRB inhibition phenocopies ANGPT1 rescue 105
51. Figure 5.1. Structural defects in CHARGE pericytes 110
52. Figure 5.2. Reduced NG2 expression with CHD7 deficiency 111
53. Figure 5.3. Survival defect in CHD7-cPC 113
54. Figure 5.4. Migration defect in CHD7-cPC 115
55. Figure 5.5. Defective induction of tube formation in CHD7-cPC 116
56. Figure 5.6. Rescue of pericyte defects in CHD7-cPC with procainamide 118

10
Chapter One
Cranial Pericytes

11
What are pericytes?
Pericytes are perivascular cells surrounding the endothelial cells (EC) of the
blood vessels throughout the body. These cells were first known as ‘Rouget cells’, after
Charles-Marie Rouget, who identified them as contractile cells surrounding small blood
vessels (Armulik et al., 2011). While the definition of pericytes has long been based off
of their periendothelial location, as Rouget first described them (Rouget, 1873), they
eventually were more specifically defined as cells embedded in the basement
membrane of blood vessels (Sims, 1986). In addition to their perivascular location,
pericytes are also defined by the type of vessel they surround, specifically microvessels
(Diaz-Flores et al., 2009).
There are multiple cell types related to pericytes that occupy the same
perivascular spaces. These includes vascular smooth muscle cells (VSMC) fibroblasts,
mesenchymal stem cells (MSC), and macrophages. Similarity between the location,
morphology, and function of these cells, not to mention the lack of specific molecular
markers for pericytes, makes studying and characterizing pericytes difficult beginning
with the difficulty in and of itself of defining these cells.

Developmental trajectory of pericytes
Pericytes of the body can be classified into two groups based on their
developmental lineage: mesoderm-derived and neural crest-derived. Mesoderm-derived
pericytes make up a majority of the pericytes of the body: in the gut, liver, lungs, and
coronary vessels among other tissues pericytes have all been mapped to a mesodermal
origin (Wilm et al., 2005; Asahina et al., 2011; Que et al., 2008; Cai et al., 2008; Zhou et
12
al., 2008). The pericytes of the central nervous system (CNS) and craniofacial region,
however, are neural crest-derived (Bergwerff et al., 1998; Etchevers et al., 2001; Korn
et al., 2002). Neural crest cells (NCC) are a population of cells that delaminates from
the neural tube during development and then migrates laterally before giving rise to
various derivatives. In the craniofacial region, NCC give rise to muscle, bone, cartilage,
melanocytes, neurons, and pericytes.
Although there has been no definitive study comparing mesoderm-derived
pericytes (mesPC) and NCC-derived pericytes, there are a few inherent differences
between these two populations. All the regions of the body known as barrier regions
have pericytes that are NCC-derived – blood-brain-barrier (BBB), blood-retina-barrier,
blood-thymus-barrier, blood-spinal cord-barrier (Armulik et al., 2011). Pericyte density
also differs between mesPC and NCC-derived pericytes. In the CNS, the endothelial
cell (EC) to PC ratio is between 1:1-3:1, with high coverage of vessel surface by
pericytes (Mathiisen et al., 2010). In tissues with mesPC, however, EC to PC ratio is
highly variable but significantly lower than in the CNS, with ratios as low as 100:1 in
skeletal muscle (Diaz-Flores et al., 2009). It is possible that there are transcriptional and
functional differences between these pericyte populations, but the historical difficulty in
defining and characterizing pericytes, plus the lack of primary cells due to low yield from
tissue isolation has made it hard to study pericytes in a complete and multi-pronged
way.

13
Pericytes at the blood-brain-barrier
The blood-brain-barrier (BBB) is a border made up of cells that separates the
environments of the brain and the blood. Endothelial cells, pericytes, and astrocyte end
feet are the main components of the BBB, along with the ECM deposited by these cells
that enhances their interactions. These components come together to function as a
single unit of the BBB called the neurovascular unit (Winkler et al., 2011). Interactions
between cells are mediated through various junctions – peg and socket contacts
between EC and pericytes, tight junctions between EC, adhesion plaques between EC
and pericytes, hemichannels and gap junctions allowing molecular transfers between
cells – as well as signaling across the basement membrane and via various integrin
molecules (Winkler et al., 2011).
Proper BBB formation, function, and maintenance is integral for proper brain
function. The BBB maintains the necessary chemical and nutritional environment for the
brain, a role required as the brain has no nutritional reserve and has limited
mechanisms for waste clearance (Quastel and Wheatley, 1932). Defects in BBB
formation and/or maintenance leads to toxic accumulations, inflammation, vascular
defects, and neuronal dysfunction (Quaegebeur et al., 2010). Therefore, the
components of the BBB are required in order to establish and preserve brain health.

Functions of cranial pericytes
Cranial pericytes have multiple functions involved with the regulation and
maintenance of the BBB. Such functions include regulation of junctions between EC,
regulation of transcytosis across the BBB, stabilizing vessels, secretion of ECM
14
components, regulation of capillary diameter and blood flow, and phagocytosis (Winkler
et al., 2011). Pericyte activity begins even before the BBB is fully functional, with
pericytes aiding in the formation of the developing BBB (Armulik et al., 2010; Bell et al.,
2010). In fact, studies in PDGFRb deficient models leading to pericyte deficiency
revealed that the presence of pericytes is necessary for proper formation of the BBB
(Daneman et al., 2010).
Communication with EC via signaling pathways at the BBB is critical for PC to
perform their necessary activities. ANGPT1/TIE2 signaling mediated via pericyte
secretion of ANGPT1 is important for the regulation of EC proliferation, angiogenesis,
and EC-EC tight junctions (Sweeney et al., 2016). EC reciprocally secrete PDGF-BB, a
growth factor that pericytes respond to and which leads to pericyte recruitment to EC,
regulation of PC proliferation, and overall regulation of pericyte number (Winkler et al.,
2010; Lindahl et al., 1997; Olson and Soriano, 2011; Birbrair et al., 2014). Bidirectional
TGFb signaling between pericytes and EC regulates proliferation of both cell types,
overall BBB maturation, and stabilization of the vascular structure (Van Geest et al.,
2010; Lebrin et al., 2005). Overall, the importance of pericytes at the BBB goes hand-in-
hand with their proximity to EC and the structure and function of the neurovascular unit
as a whole to regulate and maintain the BBB.

15
Chapter Two
In vitro derivation of cranial pericytes via a neural crest intermediate

16
INTRODUCTION
Cranial pericytes (cPC) are key players at the blood-brain-barrier (BBB); proper
functioning of these cells is required to maintain BBB function and integrity. Therefore,
to understand the BBB fully it is necessary to study cPC, both as an isolated cell type as
well as part of the neurovascular unit. However, pericytes have historically been a
difficult cell type to study. For many years, pericytes were predominately defined by
their location along small capillaries, making their isolation from primary tissue difficult.
There remain no known pericyte-specific genes, making identification of pericytes in
vivo difficult and generation of genetic models cumbersome.
Due to the difficulties surrounding the study of cranial pericytes, various
surrogates have been used when generating BBB models or studying the functions of
cranial pericytes in pathological conditions. There are no available human pericyte cell
lines, so to study human pericytes, primary cells must be isolated from human tissues.
Due to the limited resources available for obtaining primary brain tissues, many people
have substituted pericytes from other tissues in their studies of brain pericytes.
However, pericytes from other tissues may not properly recapitulate the functions of
brain pericytes in experimental settings. Analysis of primary human brain and placental
pericytes shows the transcriptomic differences between the two types of pericytes
(Figure 2.1A). These transcriptomic differences underly functional differences,
highlighted by the molecular functions and pathways that are up- or downregulated in
brain primary pericytes (1°PC) compared to placental pericytes (Figure 2.1B). This
highlights the uniqueness of brain pericytes compared to other pericytes, and the
drawbacks of using other pericytes as a surrogate for brain pericytes.
17

Figure 2.1. Comparison of human brain and placental PC. A. Volcano plot
comparing adult human brain 1°PC with human placental 1°PC. > 2-fold up- (green) and
down- (grey) regulated genes are indicated. Genes with average reads < 10 were
eliminated from analysis. B. PANTHER molecular functions and pathways of genes
upregulated in cranial 1°PC compared to placental 1°PC, and PANTHER pathways of
genes downregulated in cranial 1°PC compared to placental 1°PC, with p-values listed.

Human cranial 1 PC vs Placental PC
o
A. B.
18
Mouse models are a common and necessary tool for studying in vivo functions of cells
in both normal and pathological conditions. Recent work from the Betsholtz group has
isolated cells from mouse brains and used single-cell transcriptomics to identify cell-type
specific gene signatures (He et al., 2016; Vanlandewijck et al., 2018). When the
pericyte-specific genes (~260) identified in mouse were compared to human 1°PC from
different sources, roughly two-thirds of these genes were found to be expressed at
similar levels in both mouse and human PC. However, 36% of the genes from this
group were specific to the mouse pericytes. These genes had expression levels that
were undetectable or below detection in the human pericyte samples (Figure 2.2A).
When Gene Ontology (GO) analysis was performed, the mouse-specific genes revealed
biological processes such as nephron development and negative regulation of muscle
cell differentiation, which could indicate differences between species that connect to
why mice do not develop Alzheimer’s disease (AD) (Figure 2.2B). This analysis shows
how, while mouse studies can provide many insights due to their in vivo context, there
are aspects that are lost from studying cells of a different species. Thus, we still need
human cells in our studies to truly understand the complete function of pericytes in
normal and pathological conditions.
Multiple studies have looked to utilize human in vitro-derived pericytes, however
these protocols generate pericytes from mesodermal intermediates, reflecting the
developmental trajectory of systemic pericytes, not brain pericytes (Dar et al., 2012;
Kusuma et al., 2013; Oriova et al., 2013; Kim et al., 2016; Kumar et al., 2016; Xu et al.,
2017; Chin et al., 2018). The developmental process through which the cells normally

19

Figure 2.2. Comparison of mouse and human brain 1°PC. A. Heatmap showing
normalized expression level of 259 mouse PC specific genes in mouse 1°PC as well as
human 1°PC. Note that human samples were cultured in a dish and mouse samples
were directly sequence post isolation from the brain. B. GO molecular function terms of
genes expressed in mouse brain PC and human brain PC and GO biological process
terms of genes expressed in mouse brain PC but little or no expression in human brain
PC, with p-values listed.

B . A .
20
progress may be key to proper cell function; neonatal rat pericytes isolated from brain
and lung both express pericyte markers NG2 and PDGFRb, but only brain pericytes
express TFAP2a, reflective of their NCC origin (Figure 2.3A). When both cell types
were subjected to in vitro BBB modeling using a transwell system, brain pericytes were
more capable of both Ab efflux and transferrin uptake, two functions essential for
pericytes at the BBB (Figure 2.3B, C). To further analyze mesodermally-derived
pericytes (mesPC) compared to cPC, whole transcriptome correlation analysis was
performed between normal human brain PC and two mesPC lines independently
derived from human iPSCs, PC1 and PC2 (Kumar et al., 2017). The mesPC lines
clustered with mesenchymal stem cells (PC1) or smooth muscle cells (PC2) but were
separate from the cluster of brain pericytes (Figure 2.3D). This analysis highlights the
differences between mesPC and cPC, showing how both primary mesPC and in vitro-
derived mesPC fail to properly recapitulate cPC both functionally and transcriptionally.
The various analyses above show how there are no good substitutions for
studying human cPC, highlighting the necessity for access to human pericytes. To get
at the need for large numbers of cPC to study on structural, transcriptomic, and
functional levels, we developed a method of in vitro cPC derivation that follows the
developmental trajectory of these cells through a neural crest intermediate. We
analyzed the in vitro-derived cPC on multiple levels to highlight the robustness of our
protocol, as well as begin to define the ‘gold standard’ of brain pericytes.

21

Figure 2.3. Comparison of mesPC and cPC. A. IF showing similar NG2 expression
but differential expression of TFAP2a on wild type neonatal rat brain (NCC-derived) and
lung (mesoderm-derived) pericytes. Scale bar represents 10 µm. B. Schematic of
experimental setup to test clearance of Ab across transwells with or without conditioned
media, green circle = Ab(42)-488. Images of EC on transwells after 1hr of incubation with
Ab(42)-488 and culture with either rat brain or lung pericyte conditioned media. Scale bar
D . A .
B .
C i . C i i .
22
represents 10 µm. C. (i) Quantitation of clearance of Ab(42)-488 measured at 24hr and
48hr over 3 independent biological replicates. Error bars represent standard error of
means, *** is p<0.0001, ** is p<0.005 from unpaired student’s t-test. (ii) Quantitation of
uptake of Transferrin-TexasRed measured at 6hr and 24hr over 3 independent
biological replicates. Error bars represent standard error of means, *** is p<0.0001, * is
p<0.014 from unpaired student’s t-test. D. Spearman’s correlation coefficient matrix
showing hierarchical clustering of iPSC-derived and primary pericytes. Red shades
represent high correlation. Zoomed in section shows divergence of maturing PC1 and 2
from primary brain pericytes.

23
RESULTS AND DISCUSSION
In vitro derivation and characterization of cPC via NCCs
To generate cPC in vitro, we began with hESCs or iPSCs and generated NCC
via the previously established and highly characterized protocol (Bajpai et al., 2010;
Rada-Iglesias et al., 2012). Pure NCCs were collected by washing off neuroectodermal
spheres with PBS, and NCCs were confirmed to be triple positive for SOX9, P75, and
TFAP2a. NCCs were then plated in one of three pericyte compatible media (PCM1-3)
and pushed to cranial pericyte fate (Figure 2.4A, H). Compatible media was identified
as both capable of maintaining 1°PC in culture as well as inducing pericyte fate in NCCs
(Figure 2.4C, D, F). Along with identifying correct media conditions, it was necessary to
verify that NCCs were the optimal cell type to generate cPC; therefore, other cells types
were tested in PCM1-3, but only early NCC were found to be optimal for cPC generation
(Figure 2.4B, E, G). Together this established the necessary conditions of the protocol:
1) the optimal starting cells – immature cranial NCCs, and 2) the optimal medium – non
DMEM base medium, PDGF-BB, and low serum (Figure 2.4G).
As the cells transition from NCC to cPC, they go from NG2
-
PDGFRb
lo
to NG2
+

PDGFRb
hi
(Figure 2.5A, B). The cells undergo a morphological change, losing
circularity and appearing long and thin with extended processes, and retain this shape
for up to 8 months with continuous culture (Figure 2.5C, D). cPC are not static,
however, and their shape and processes change as the cells undergo random
movements (Figure 2.5E). This reflects the ability of pericytes to migrate and respond
to external cues, and accounts for morphological differences between cells in static
images.
24

A . B .
C .
D .
E .
F . G .
H .
25
Figure 2.4. In vitro derivation of cPC. A. Schematic of cPC generation from
pluripotent stem cells via an NCC intermediate. B. Schematic summary of conditions
tested for cPC generation. C. Media incompatible with pericyte growth and amplification:
phase image of live 1°PC (top) and IF (bottom) for NG2 and PDGFRb in fixed 1°PC
after 7 days culture in various media: company recommended PCM1, NCC medium
(cells did not live long enough for fixation and IF), Mesencult, DMEM with 10% FBS,
hESC medium, and smooth muscle cell medium. D. Left panels – media compatible
with pericyte growth and amplification: phase image of live 1°PC in PCM1 after 7 days
and IF for NG2 and PDGFRb in 1°PC grown for 7 days in PCM2 or PCM3. Right panels
– media compatible with pericyte induction from NCC: phase image of cPC induced
from NCC in PCM1 followed by IF for NG2 and PDGFRb in cPC induced from NCC for
14 days using PCM2 or PCM3. E. IF of NG2 and PDGFRb in various cell types after 14
days culture in PCM1: ESC-NCC, fetal fibroblasts, H9 ESCs, mature NCC, HUVECs,
and vascular smooth muscle cells. F. Percentage of NG2 positive cells obtained after 14
days culture of NCC in 5 different media conditions. Note viable cells obtained in PCM1
+ 10 nM sunitinib, a potent PDGFRb inhibitor, were all NG2 negative suggesting
PDGFRb function is essential for PC fate induction. G. Necessary factors needed for
generation of cPC in vitro. H. Comparative media composition chart for various media
tested.

26

E .
A .
B .
C .
D .
27
Figure 2.5. Transition of cells from NCC to cPC. A. Low magnification IF for NG2 and
PDGFRb in NCC and cPC compared with 1°PC. Scale bars represent 10 µm. B. High
magnification IF for NG2 and PDGFRb in ESC-cPC and 1°PC. C. Representative phase
images during NCC conversion to cPC fate. Scale bars represent 10 µm. D. Circularity
of cells determined during differentiation from NCC to cPC state and after continuous
culture for 8 months, with corresponding passage number. Circularity was measured on
N > 100 cells from 3 independent biological replicates per cell line. Error bars represent
standard error of means. E. Time lapse images of high density cPC expressing GFP.
White arrow points to a single cPC throughout the movie. Note rhythmic changes in
overall morphology in most cells accompanied with migration in random directions.
28
Morphological changes accompany transcriptional changes during cPC
differentiation. Comparison of bulk RNA-seq data from the starting NCC and the
differentiated cPC reflect a downregulation of many NCC-specific genes – such as
SOX9, SOX10, and MSX1 – and an upregulation of genes known to be expressed by
pericytes – such as NG2, ANGPT1, and CD13 (Figure 2.6A). Some NCC genes show
no change in expression between NCCs and cPCs – such as TWIST1 and SNAI2 – but
these genes reflect functions that are maintained from NCCs to cPCs. Transcriptional
analysis of cPCs allows for comparison between 1°PC as well as many related cell
types, to ensure the in vitro-derived cPC reflect the desired cell type. Based on
expression of known pericyte genes and lack of expression of key genes specific to
related cell types, it was determined that cPC reflect a brain pericyte transcriptome
(Figure 2.6B). Our method of cPC generation is able to robustly and repeatedly
generate cells that reflect the morphology and gene expression of brain pericytes.

29

Figure 2.6. cPC transcriptionally reflect pericytes. A. Examples of > 3-fold
downregulated (blue), highly expressed but not substantially altered (grey), and > 3-fold
upregulated (green) genes in cPC compared with NCC. B. Schematic of PC originating
from NCC or mesodermal cells and other related cell types: VSMC, MSC, EC,
oligodendrocytes, and astrocytes with distinctive features listed. Genes in red are not
known to be expressed in forebrain PC. Genes in green are expressed in cPC.

A .
B .
30
cPC function similarly to 1
o
PC
Aside from morphology and gene expression, we wanted to determine how
closely our cPC resemble 1°PC functionally, a necessity if cPC are to recapitulate 1°PC
experimentally. We performed GO analysis on the top 1400 upregulated genes in cPC
compared to NCC, and the top GO term was ‘extracellular exosomes’ (Figure 2.7A).
Pericytes participate in cell-cell signaling at the BBB with other pericytes, endothelial
cells (EC), and astrocytes, by using exosomes to secrete cargo to their neighboring
cells (Figure 2.7B). We found that cPC express both clatherin and caveolin, and
generate exosomes marked with CD81 (Figure 2.7C, D). Exosomes were found inside
the cells as well as secreted out of the cells, and high-magnification imaging showed
that specific cargo such as ANGPT1 colocalized with these exosomes, reflecting the
ability of the cPC to properly package and secrete signaling molecules (Figure 2.7E).
After ‘extracellular exosomes’, the next top GO term was ‘extracellular matrix
organization’ (Figure 2.7A). Pericytes contribute to the basement membrane that
surrounds the EC and PC of the BBB; they also produce and remodel the ECM as they
migrate during angiogenesis and wound healing. cPC are able to express and secrete
ECM proteins such as Fibronectin and Laminin in patterns similar to 1°PC (Figure 2.8).
cPC also express multiple integrins and proteoglycans that play a role in stabilizing,
maintaining, and remodeling the ECM of the pericytes as well as the basement
membrane of the BBB (Figure 2.6A and data not shown).
Pericytes utilize their ability to contract in order to regulate cerebral blood flow at
a local level by controlling the diameter of capillaries. It has been shown that pericytes
express a number of contraction proteins such as MYH1 and ACTA2 (Vanlandewijck et
31

Figure 2.7. cPC generate extracellular exosomes. A. The top GO terms (with p-
values) enriched with 1400 genes upregulated in cPC (> 2-fold up at p < 0.01, FDR <
0.5) compared to NCC. B. Schematic of exosome assembly and release in pericytes. C.
IF for Clatherin and Caveolin in cPC and 1°PC. D. IF for CD81 exosomal marker in cPC
and 1°PC. White box indicates zoomed in area. E. IF for ANGPT1 and CD81, with
zoomed images showing exosomes. White arrows indicate areas of ANGPT1 and CD81
colocalization. Scale bars represent 10 µm.

A .
B . C .
D .
E .
32

Figure 2.8. cPC secrete various ECM proteins. A. IF for Fibronectin ECM production
in cPC and 1°PC within 3 days of plating on a tissue culture dish. Scale bars represent
10 µm. B. IF of NG2 and Fibronectin and Fibronectin only channel in 1°PC and cPC to
highlight extracellular location of fibronectin. C. IF of laminin in 1°PC and cPC.

A .
B . C .
33
al., 2018). When plated in a defined medium, upon the addition of KCl, 1°PC and cPC
exhibited a slow but continuous contraction followed by an extended period of
expansion (Figure 2.9A-C). This was different from the normal morphological changes
the cells undergo without stimuli and is specific to the presence of potassium ions (K
+
),
as no contraction was seen in defined medium without KCl (Figure 2.9D), yet when KCl
was added, 100% of cells responded by contracting (Figure 2.9E). Upon transcriptional
analysis, it was found that the maxi K (big potassium) channel KCNMA1 was highly
expressed in cPC and 1°PC (> 5 FPKM) and was upregulated > 5-fold during the
transition of NCC to cPC (Figure 2.9F).
Along with stationary movement accompanying contraction, pericytes are mobile
cells with the ability to migrate with and without external cues. We first tested the ability
of the cPC to randomly migrate by performing scratch assays; cPC and 1°PC were both
able to fill up the scratch area by 48hrs, with similar dynamics of migration (Figure
2.10A-B). To test directed migration, cPC were cultured with affigel beads that were
either uncoated, coated with the PDGFRb inhibitor sunitinib malate, or coated with the
growth factor PDGF-BB (Figure 2.10C-D). cPC do not adhere to and show no directed
migration toward uncoated beads, but they will migrate away from beads soaked in
inhibitor and migrate toward and adhere to beads soaked in growth factor. This not only
shows the ability of the cPC to migrate directionally in response to a signal, but it also
shows how cPC can change their behavior and morphology in response to signals in
their environment. These aspects of cPC behavior and function show recapitulation of
cell-intrinsic functions of 1°PC.

34

Figure 2.9. cPC exhibit contraction in response to stimuli. A. Stills from a movie,
200 seconds apart, for 1°PC and ESC-cPC upon KCl addition at t=0. Red arrows
indicate the edges of a cell. B. Stills from a movie spanning 10 min of iPSC-cPC labeled
with GFP upon KCl addition at t=0. Scale bar represents 10 µm. C. Scatter plot shows
Dl (change in maximal length of each cell, in pixels) at time = 10 mins (corresponds to
maximum contraction in control cells). Dl was calculated for individual cells N > 50 from
at least 3 independent movies, and averages of cells in a field were plotted as a single
dot. ns = no statistical significance, ** is p < 0.0023 from one-way ANOVA. D. Stills from
a control movie (without KCl addition) of cPC expressing GFP to show entire
morphology. Note no change in morphology in medium without KCl. E. Percent of cells
responsive to KCl, N > 100 from 3 independent biological replicates. F. Expression level
of 110 K-channels in cPC, 1°PC, and NCC. Boxed area is zoomed in to show KCNMA1
level.

A . B . C .
D . E . F .
35

Figure 2.10. cPC exhibit random and directed migration. A. Scratch assay showing
migration of cPC into the scratch area at two time points. B. Bar graph shows percent
area covered by infiltrating PC across 3 independent biological replicates. Error bars
represent standard error of means, ns = no statistical significance, ** is p < 0.0012, * is
p < 0.014 from unpaired student’s t-test. C. Interaction of GFP expressing pericytes in
suspension with non-adherent affigel blue beads soaked in control medium without
additives, with inhibitor (Sunitinib malate), or with growth factor (human PDGF-BB).
cPCs do not adhere to (n=3/200), and avoid beads soaked in Sunitinib (dotted arrows)
(n=0/200); cPCs adhere to and are attracted to beads soaked in PDGF-BB (white arrow;
n=147/200). Zoom image shows cell processes spreading over the surface of PDGF-BB
soaked bead. D. Representative image of cPCs in a monolayer attracted to PDGF-BB
soaked beads (n=37/40 beads in three independent cPCs derived from different iPSC
lines) but not to control beads (n=0/40 beads).

A . B .
C . D .
36
cPC interact with EC and affect EC functions
Key to modeling the BBB in vitro and studying the role of pericytes in pathological
conditions is being able to study their interaction with EC. Therefore, in vitro-derived
cPC must interact with EC in a similar manner as 1°PC to be used as a tool for studying
brain pericytes. To test general interactions between EC and cPC, we plated co-cultures
of the cells at high and low density, as well as EC cultures with cPC conditioned media
to gauge changes in cell behaviors (Figure 2.11). At low density, live imaging captured
EC and cPC migrating toward each other, as well as cPC extending processes to
interact with nearby EC (Figure 2.11A). At high density, the addition of cPC to EC
cultures induced organization of EC into clusters surrounded by cPC (Figure 2.11B).
Even conditioned media alone was able to influence EC behavior, causing neighboring
EC to generate ‘zipper-like’ interactions with each other (Figure 2.11C), showing that
cues from cPC are able to induce changes in EC behavior. These EC-EC interactions
are key to maintaining BBB integrity, and pericytes are responsible for maintaining EC-
EC junctions. cPC are able to induce expression of junctional proteins in EC in
cocultures, with alignment between EC (Figure 2.12), in a manner similar to 1°PC, and
to similar extents.
Along with PC-EC interactions related to BBB function and integrity, pericytes
interact with EC during angiogenesis and blood vessel remodeling to control the growth,
branching, and stability of the vasculature. One method of mediating this control is the
ability of pericytes to arrest EC proliferation. In culture, EC are highly proliferative cells
(98% ± 2.5 Ki67
+
). With the addition of 1°PC to the culture, EC proliferation is reduced

37

Figure 2.11. cPC migrate toward EC and influence EC organization. A. Time lapse
images of a representative movie showing sustained interaction of GFP expressing cPC
with mCherry expressing EC (n=17/17) and directional migration of cPC toward EC
(n=21/23). B. Initiation of aggregation with addition of cPC to EC cultures. Fluorescently
labeled cPC are in green and pseudo-colored EC are in pink. C. IF of PECAM on EC
showing upregulation of PECAM, its junctional localization, and induction of EC-EC
contacts (white arrows) within 6hr of addition of cPC conditioned media to EC. Scale
bars represent 10 µm.

A .
B . C .
38

Figure 2.12. cPC induce EC-EC contacts. A. IF for g-catenin on EC cultured alone or
with 1°PC or cPC. B. IF for Occludin on EC culture alone or with cPC. Scale bars
represent 10 µm.

A . B .
39
3-fold (32.5% ± 5.7 Ki67
+
); this reduction is recapitulated with the addition of both cPC
(28.8% ± 4.4 Ki67
+
) and cPC conditioned media (30.4% ± 7 Ki67
+
) (Figure 2.13). To
test the overall ability to induce EC tube formation, we performed two types of tube
forming assays. In a thick matrix of growth-factor-reduced geltrex, EC alone were only
able to form small tubular structures that did not connect (Figure 2.14Ai, B). The
addition of cPC or 1°PC increased the overall area covered by tube structures (Figure
2.14B), and allowed for the formation of tube structures and lattices with branching
points (Figure 2.14F), with cPC and 1°PC lining up along the EC, preferentially at
branching points (Figure 2.14Aii-iii). In the presence of fibrin gel beads, EC alone fail to
generate tubes (Figure 2.14E); however, cPC are able to induce tube structures that
can form a lumen (Figure 2.14C-D). These tubes form in 3-dimensional structures, and
as with the geltrex assay, cPC line up along the EC. cPC are able to influence EC tube
formation by inducing growth and increasing stability.

40

Figure 2.13. cPC reduce EC proliferation. A. Quantitation of percentage of Ki67 and
PECAM double positive cells. Counts were done of maximum projection confocal
images of N > 100 cells from at least 3 independent biological replicates. Error bars
represent standard error of means, *** is p<0.0001 from one-way ANOVA. B. IF for
PECAM and Ki67 on EC cultured alone, with pericytes, or with pericyte conditioned
media. White arrows indicate junctional localization. Scale bar represents 10 µm.

A . B .
41

Figure 2.14. cPC induce EC tube formation. A. Images of tube/lattice formation in EC
cultured on growth-factor-reduced geltrex matrix. EC are pseudocolored brown (i-ii) or
uncolored (iii) and fluorescently tagged cPC are in green. White arrowheads indicate
cPC alongside EC. B. Tube area in pixels squared was measured to quantify area
covered with cellular structures over 5 random fields in each of 3 independent biological
replicates. Error bars represent standard error of means, ns = no statistical significance,
*** is p<0.0001, ** is p<0.0014 from unpaired student’s t-test. C. (i) 3D projection of live
images of tubes radiating out from beads and forming networks with EC (magenta,
representing mCherry expression) and ESC-cPC (green, representing GFP expression)
embedded in fibrin gel. White arrowhead indicates cPC lining EC tubes. (ii) Single
confocal section of cPC induced tubes in fibrin gels. White asterisk indicates lumen in
confocal stack and white arrow shows cPC lining EC tubes. D. Phase images showing
similar tubes radiating out from beads with EC (uncolored) and iPSC-cPC (expressing
GFP in green) embedded in fibrin gel. Arrows point to luminized tubes. Dotted arrows
Ai.
Aiii. B.
Ci. Cii.
D. E. F.
Aii.
42
point to cPC lined tube at a different plane. E. Phase image showing beads with EC
only, with a lack of luminized tubes. F. Quantification of tube branches per unit area.
Counts were done over 5 random fields in each of 3 independent biological replicates
for each condition. Error bars represent standard error of means. Scale bars represent
10 µm.

43
cPC home to vasculature in vivo
To see if cPC have the ability to respond to the in vivo pericyte perivascular
niche, we turned to injections of cPC in Xenopus tadpoles. To empty the pericyte
niches, the tadpoles were treated with the PDGFRb inhibitor Sunitinib malate, which
kills the endogenous pericytes and causes leakiness of vessels without inducing any
gross hemorrhaging (Figure 2.15A). After treatment with sunitinib followed by 24hr
withdrawal, cPC that were fluorescently tagged with mCherry were injected
intracardially, and tadpoles were followed up to 2 weeks post-injection. cPC were found
in all tadpoles that were treated with sunitinib and injected with cells, and the cPC were
always found along the outside of the smaller vessels, preferentially in the CNS and the
craniofacial regions (Figure 2.15B-C). No cPC were ever found around larger vessels
that would normally be surrounded by VSMC, and the cPC that were found along the
small vessels showed a preference for locations at branching points (Figure 2.15C).
Similar to Xenopus, when cPC were added to mouse brain slices in culture, the cPC
homed to the vasculature, specifically microvessels with diameters < 10 µm (Figure
2.16). As in the case of the in vitro tube assays as well as the Xenopus tadpoles, the
cPC preferentially homed to junction points along the vasculature. This is consistent
with the role of pericytes stabilizing blood vessels in vivo.

44

Figure 2.15. cPC home to craniofacial microvessels in Xenopus tadpoles. A.
Normal and sunitinib treated Xenopus embryos with intracardial injection of 488-
dextran. White arrowhead indicates dextran leakage into surrounding tissue. Scale bars
represent 10 µm. B. Live images of anesthetized tadpoles that received 100 cPC
intracardially. (Top) White arrows point to large vessels devoid of human PC. Zoomed
image of boxed area shows region of the neural fold. mCherry fluorescence of cPC
outside small blood vessels in the CNS are seen in red. Red arrows point to mCherry
labelled cells on microcapillaries. (Bottom) Live image of tissue surrounding the eye, 10
days post injection, where blood vessels are visualized with 488-dextran and iMF-cPC
are labeled in red. Scale bars represent 10 µm. C. Chart showing quantification of cPC
location post injection in live Xenopus embryos.

A . B .
C .
45

Figure 2.16. cPC home to vasculature in mouse brain slices. A. Image of top
surface of mouse brain slice at 36hrs after slice-culture with 1000 iPSC-cPC (green)
deposited at t=0, showing homing of cPC to blood vessels with a diameter < 10 µm.
Blood vessels visualized with lectin in red. B. Zoomed image of a single GFP
expressing iPSC-cPC in close association with capillary (red). C. Chart showing
quantification of cPC location post culture on mouse brain slices.

A . B .
C .
46
cPC induce barrier functions in EC
Pericytes help regulate the transport of molecules across the BBB two-fold: first
pericytes reduce paracellular transport between cells of the BBB by regulating EC-EC
junctions and maintaining their integrity; second pericytes participate in transcytosis
across the BBB by aiding in cargo movement to and across EC. To test paracellular
transport across an in vitro BBB model, EC were plated on transwells and cultured with
and without cPC conditioned media (Figure 2.17A). When dextrans of various sizes
were added to the one side of the barrier, cPC were able to effectively reduce
paracellular transport compared to EC alone (Figure 2.17B, C). This is due to the ability
of cPC to induce EC-EC contacts, allowing for the production of a barrier with tight
junctions that can block transport between cells (Figure 2.17D). To test transcytosis
across the barrier, we examined the ability of cPC to aid in both Transferrin uptake and
Ab efflux. Transferrin uptake represents the ability for cPC to help EC take nutrients
from the blood and transport it into the brain. In the transwell system, cPC and 1°PC
promote proficient Transferrin uptake, well over the ability of EC alone (Figure 2.18A-
B). When visualized, we can see that conditioned media from cPC and 1°PC induced
EC to take up Transferrin into the cell, excluded from the nucleus (Figure 2.18C-D).
Transport at the BBB occurs in the opposite direction as well, such as the case with the
efflux of Ab that could build up in the brain out into circulation. In the transwell system,
as with Transferrin, cPC and 1°PC both induce efflux into the blood compartment more
efficiently than EC alone (Figure 2.19A-B). Visualization again shows that EC alone are
not able to take up Ab into the cells, but the addition of cPC or 1°PC conditioned media
causes EC to take up Ab into the cells, excluded from the nucleus, as the protein is
47

Figure 2.17. cPC induce barrier properties in EC. A. Schematic of experimental
setup to test paracellular transport in response to conditioned media, green circle = 10
kDa dextran-568 or 70 kDa dextran-488. B-C. Paracellular transport of 10 kDa dextran-
568 (B) and 70 kDa dextran-488 (C) measured at 12hr in various conditions over 3
independent biological replicates. Error bars represent standard error of means, *** is
p<0.0001, ** is p<0.0061 and p<0.0089 from one-way ANOVA. D. Images of transwells
showing EC coverage of pores with and without conditioned media. Scale bars
represent 10 µm.

* * *
A . B . C . D .
48

Figure 2.18. Transferrin uptake across in vitro barriers. A. Schematic of
experimental setup to test uptake of transferrin across transwell with or without
conditioned media, red circle = Transferrin-TexasRed. B. Uptake of Transferrin
measured at 6hr and 24hr over 3 independent biological replicates. Error bars represent
standard error of means, *** is p<0.0001 from unpaired student’s t-test. C. Schematic of
experimental setup to visualize transferrin transcytosis through GFP expressing EC,
against gravity. EC layered outside the well, hanging in the outer chamber that
represents the vessel lumen; abluminal/brain side is represented by the inner chamber
and contains cPC or conditioned medium, red circle = Transferrin-TexasRed. Images of
bottoms of transwells after 24hr incubation with Transferrin-TexasRed in outer chamber.
White arrows point to intracellular transferrin (overlap between green and red), seen as
rings as it is excluded from the nucleus. Dotted arrow indicates control EC with little or
no intracellular transferrin. D. Zoomed in image of Transferrin within an EC on the
transwell. Note the ring pattern of the molecules. n = nucleus.

A .
C .
D .
B .
49

Figure 2.19. Ab efflux across in vitro barriers. A. Schematic of experimental setup to
test clearance of Ab42 across transwell with or without conditioned media, green circle =
Ab42-488. B. Clearance of Ab42 measured at 24hr and 48hr in various conditions over 3
independent biological replicates. Error bars represent standard error of means, ** is
p<0.008 and p<0.006 from unpaired student’s t-test. C. Schematic of experimental
setup to visualize Ab42 trafficking across EC with hanging drop EC culture with or
without conditioned media, green circle = Ab42-488. Live images of the bottom of
transwells with EC 12hrs after incubation with Ab, showing the difference in Ab uptake
with and without conditioned media. D. Zoomed in images of Ab within an EC on the
transwells. n = nucleus.

A . B .
C . D .
50
transported across the barrier. cPC are able to block paracellular transport and induce
transcytosis across a barrier to the same extent as 1°PC, showing that cPC can be a
good surrogate for 1°PC experimentally.

CONCLUSION
The work in this chapter has shown how cPC can be generated in vitro and used
to model brain pericytes in experimental settings. Our method is robust and
reproducible, generating similar cPC across differentiations and starting ESC/iPSC
lines. This allows for unlimited numbers of cPC to be generated, both for experimental
usage as well as potential therapeutics. The ability for the cPC to home to vasculature
in vivo, as well as recapitulate primary pericyte functions, provides a cellular source of
normal pericytes that could be used for transplants in patients suffering from diseases in
which cranial pericytes are affected.
Since we began our work, two other groups have published protocols for in vitro-
derivation of brain pericytes via NCC (Faal et al., 2019; Stebbins et al., 2019). While
some aspects of the characterization of the cells is similar between all 3 protocols, our
system and our characterization remain both unique and important for the field. We
have shown a robustness across 14 lines, backed by RNA-seq data as well as
immunofluorescence. Our functional characterization is much more extensive than
either other group, taking into account as many aspects of pericyte function as we
could, attempting to not only characterize the cells derived in our protocol, but also
characterizing cranial pericytes overall, establishing a much-needed standard in the
field. We also show how our protocol can be used to interrogate pericytes in disease
51
settings, work which is described in the following chapters. Overall, our work has not
only established a robust protocol that generates pericytes of the BBB, but it also
provides context for a thorough description of BBB pericytes on multiple levels.

52
Chapter Three
Familial AD pericytes have lifelong structural, functional, and survival defects

53
INTRODUCTION
Alzheimer’s disease is the leading cause of dementia, accounting for as much as
80% of documented dementia cases (alz.org). An age-related disease, AD is clinically
described as a neurodegenerative disease leading to neuronal loss, cognitive
impairment, and eventually death. The majority of AD cases are known as sporadic AD
(sAD), and result from mostly unknown causes, potentially a combination of genetics
and environmental factors. About 5% of AD is early-onset, also known as familial AD
(fAD) or autosomal dominant AD (ADAD). fAD is caused by known genetic mutations
and occurs in patients as early as in their 30s; there is also faster progression of the
disease in fAD compared to sAD, with the disease lasting 6-7 years before death as
opposed to 7-10 years.
Experimental models of fAD have been developed in order to study the disease
genetically, as well as elucidate potential treatments. These models – which most
commonly involve mutations in APP or PSEN1/2 – have revealed vascular
dysregulation and capillary defects associated with AD, and these defects have been
seen in living AD patients, potentially preceding other pathological events such as
dementia in these patients (Arvanitakis et al., 2016; Kisler et al., 2017; Yew et al., 2017;
Montagne et al., 2015). In experimental models of pericyte loss, along with increased
BBB leakiness there was an association with cognitive decline and amyloid
accumulation (Sengillo et al., 2013; Winkler et al., 2014; Bell et al., 2010; Dalkara et al.,
2011). Despite these connections and associations, there has been no work done on
the specific defects of the pericytes in AD, and the molecular mechanisms underlying
vascular dysfunction are still unknown. Therefore, we look to characterize pericyte-
54
specific defects in AD using in vitro-derived cranial pericytes (cPC) from a patient with
fAD to see how pericytes, vascular dysfunction, and AD connect.

RESULTS AND DISCUSSION
Generation of in vitro-cPC from AD iPSCs reveal pericyte defects
To examine pericytes in the context of Alzheimer’s disease, we generated cPC in
vitro from iPSCs following our previously characterized method (Figure 3.1A) (Griffin
and Bajpai, in review) using a neural crest (NCC) intermediate. iPSCs derived from a
patient with the fAD PSEN1
A431E
mutation (iPSEN1*) were used to generate AD cPC. As
an isogenic control, CRISPR/Cas9 was used to correct the mutation (iPSEN1
corr
) in the
AD patient iPSCs; additional controls include iMF- and iFF-cPC (Figure 3.1B). PSEN1*-
iPSC were able to generate equal numbers of non-defective NCC as control iPSCs,
which was expected as patients do not have developmental defects in NCC or NCC-
derived tissues. iPSEN1*-NCC were triple positive for SOX9, P75, and TFAP2a, and
were NG2
-
and expressed low levels of PDGFRb (Figure 3.1D and data not shown).
There was no defect in cPC generation, as equal numbers of iPSEN1*-cPC and control
cPC were obtained that were all positive for PDGFRb, CD13, and CD105, and
expressed NG2 in similar proportions (iFF-cPC: 76.25 ± 7.23%, iPSEN1*-cPC: 71.03 ±
10.53%, iPSEN1
corr
-cPC: 75.68 ± 10.88%) (Figure 3.1C, E).
While iPSEN1*-cPC were generated, they showed structural and survival
defects. Overall morphology was altered in iPSEN1*-cPC, with reduced cell length and
increased numbers of shorter processes (Figure 3.2A-B). This morphological change
was not seen in iPSEN1
corr
-cPC, suggesting that the mutation itself is responsible for
55

Figure 3.1. Generation of iPSEN1*-cPC. A. Schematic of cPC generation in vitro from
ESC/iPSC via a neural crest intermediate. Below each cell type are listed key proteins
that are present or absent at each stage. B. Schematic showing genotypes of the cPC
derived via a neural crest intermediate in normal control iPSC (iFF/iMF), iPSC carrying
the genetic mutation PSEN1
A431E
(iPSEN1*), or isogenic mutation corrected control
iPSC (iPSEN1
corr
). C. IF of pericyte markers NG2 and PDGFRb in 1°PC, iPSC-cPC,
iPSEN1*-cPC, and iPSEN1
corr
-cPC. Scale bar represents 10 µm. D. IF of pericyte
markers NG2 and PDGFRb in NCC. Scale bar represents 10 µm. E. Quantitation of
PDGFRb, CD13, and CD105 (left), and NG2 (right) expression in iFF-cPC, iPSEN1*-
cPC, and iPSEN1
corr
-cPC. Counts were done of high-resolution confocal stacks of N >
100 cells from at least 3 independent biological replicates for each of 3 cPC derivations
from each line. Error bars represent standard error of means, ns = no statistical
significance from one-way ANOVA.
A .
B .
C .
D .
E .
56

Figure 3.2. Structural and survival defects in iPSEN1*-cPC. A. Representative
confocal images of iFF-cPC, iPSEN1*-cPC, and iPSEN1
corr
-cPC expressing mCherry or
labeled with WGA, highlighting the structural defect in iPSEN1*-cPC. Scale bars
represent 100 µm. B. Violin plot showing quantitation of cell morphology, highlighting
differences between iPSEN1*-cPC and controls. *** is p < 0.0001 from one-way
ANOVA. C. Quantitation of maximum passage number reached during long-term culture
of tested cell lines and clones from N = 3 independent derivations. Error bars represent
standard error of means, *** is p < 0.0001 from unpaired student’s t-test.

A .
B . C .
57
this defect. iPSEN1*-cPC also exhibited a decrease in long-term survival in culture.
While iFF-cPC/iMF-cPC and iPSEN1
corr
-cPC are able to survive up to 70 passages in
culture, iPSEN1*-cPC experience a sudden crash in population around passage 20
(Figure 3.2C).
Even though iPSEN1*-cPC expressed NG2 at similar levels to controls, there
was a defect in the localization of NG2 within the iPSEN1*-cPC. iMF-cPC show
asymmetric localization of NG2, with restriction to the edges of the cell along the cell
membrane, which is consistent with the secretion of NG2 protein to the cell membrane
and into the ECM; iPSEN1*-cPC, however, exhibited symmetrical NG2 distribution, with
expression of the protein diffuse throughout the cell (Figure 3.3A-B). Proper NG2
localization is restored in iPSEN1
corr
-cPC, suggesting that the mutation itself is
causative of the protein mislocalization. To further verify that this defect is due to the
dysfunction of the PSEN1/2 g-secretase protein complex, we used shRNA-mediated
gene knockdown to reduce expression of either the PSEN1 or the PSEN2 gene in wild
type cells. With loss of PSEN1/2 expression, the cells showed symmetric and diffuse
NG2 expression, consistent with iPSEN1*-cPCs (Figure 3.3C). Inhibition of g-secretase
function with a small molecule inhibitor, DAPT, also mimicked the NG2 localization
defect (Figure 3.3D), again connecting proper localization of NG2 with the function of
PSEN1/2.

iPSEN1*-cPC show defects in EC communication
To see if any of the vasculature defects associated with AD could be due to
improper PC-EC communication, we tested iPSEN1*-cPC in cocultures with wild type
58

Figure 3.3. NG2 mislocalization with loss of PSEN1. A. IF of NG2 and PDGFRb with
zoom of NG2 showing differences in NG2 localization. White arrowheads indicate broad
expression of NG2 in iPSEN1*-cPC in contrast to asymmetric localization in iPSEN1
corr
-
cPC indicated with double-headed arrows. Scale bars represent 10 µm. B. Quantitation
of NG2 localization across tested cell types. Counts were done of high-resolution
confocal stacks of N > 100 cells from at least 3 independent biological replicates for
each of 3 cPC derivations from each clone and/or line. Error bars represent standard
A . B .
C . D .
59
error of means, *** is p < 0.0001 from one-way ANOVA. C. IF of NG2 and PDGFRb of
single cPC cell with shRNA mediated knockdown of PSEN1 (shPSEN1.1) or PSEN2
(shPSEN2.1), or a nontargeting control shRNA (shCTRL), grown in culture with
knockdown for > 10 days. White arrows indicate cells with shRNA expression, indicated
in brown below; white arrowheads indicate NG2 expression. Scale bars represent 10
µm. D. 3D projection of NG2 expression in iFF-cPC with DMSO control (top) or DAPT
(bottom) treatment. Scale bar represents 10 µm.
60
human EC. PC control EC growth and vasculature structure by suppressing EC
proliferation. While control cPC suppressed EC proliferation by ~75%, iPSEN1*-cPC
were only able to reduce proliferation by ~25% (Figure 3.4). iPSEN1*-cPC were also
less efficient at inducing EC-EC tight junctions; they were able to induce junction protein
expression above EC cultures without pericytes, but the junctions were not continuous,
and expression was not as strong as with control cPC or iPSEN1
corr
-cPC (Figure 3.5).
Pericytes function to guide EC during angiogenesis, establishing the vasculature
network as well as stabilizing the blood vessels. In a 3D fibrin gel bead assay, iPSEN1*-
cPC cultured with EC were able to induce small tube structures, while iPSEN1
corr
-cPC
were able to induce structures with long tubes and multiple branching points like the
wild-type (Figure 3.6A). In a 2D matrix assay, iPSEN1*-cPC with EC grew tubular
structures, but failed to make extensive lattices with branching points and had bulbous
ends (Figure 3.6A-C). This defect was resolved in iPSEN1
corr
-cPC, which formed
extensive tube and lattice structures with stabilized branching points (Figure 3.6B-C).
Pericytes at the BBB are responsible for regulation of molecule transport across
the barrier. To test the ability to block paracellular transport between EC in a barrier, we
utilized a transwell assay system as an in vitro barrier (Figure 3.7A). iPSEN1*-cPC
were able to effectively block paracellular transport of various dextrans, with no
statistically significant difference between iPSEN1*-cPC and controls (Figure 3.7B).
However, when a tight barrier was induced first, and then transport of Evan’s Blue was
measured, iPSEN1*-cPC showed a small but significant defect in blocking transport
across intercellular space (Figure 3.7C). To test pericyte regulation of transcytosis
through EC at the barrier, we measured Ab clearance and Transferrin uptake across the
61

Figure 3.4. iPSEN1*-cPC unable to properly reduce EC proliferation. A. IF for Ki67
on EC when co-cultured with iPSEN1*-cPC or iPSEN1
corr
-cPC showing proliferation
status. Scale bar represents 10 µm. B. Quantification using maximum projection
confocal images of N > 100 cells from at least 3 independent biological replicates from
each cell line. Error bars represent standard error of means, *** is p < 0.0001, ** is p <
0.0027 from unpaired student’s t-test.

A . B .
62

Figure 3.5. Poor EC-EC junctions induced by iPSEN1*-cPC. A. IF for g-catenin on
ECs when co-cultured with iPSEN1*-cPC or iPSEN1
corr
-cPC with zoomed images below
and arrows indicating cell-cell contacts. Scale bar represents 10 µm. B. Quantification
using maximum projection confocal images of N > 100 cells from at least 3 independent
biological replicates. Cumulative signal calculated on N > 5 images with identical
thresholding, without saturated pixels. Error bars represent standard error of means, ***
is p < 0.0001 from unpaired student’s t-test.

A . B .
63

Figure 3.6. iPSEN1*-cPC show defect in inducing tube formation. A. (Top) Confocal
images showing fewer and shorter lumen filled tubules growing in fibrin gel from beads
coated with iPSEN1*-cPC and mCherry expressing EC (shown in magenta). White
arrowheads indicate lumen filled tubules; white arrows point to branch junctions.
(Bottom) Bright field images of endothelial tubes/lattices formed by EC when co-cultured
with iPSEN1*-cPC or iPSEN1
corr
-cPC. White arrow points to branch junction; red arrow
points to ‘stubby ends’ found primarily in co-culture with mutant cells. B. Tube area in
pixels squared measured to quantify total length of tubes along with their width over 5
random fields in each of 3 independent biological replicates. C. Tube branch points
counted over 5 random fields in each of 3 independent biological replicates. Error bars
represent standard error of means, ns = no statistical significance, *** is p < 0.0001 from
unpaired student’s t-test.

A . B .
C .
64

Figure 3.7. Blockage of paracellular transport partially defective in iPSEN1*-cPC.
A. Schematic of experimental setup with sub-confluent cells on transwell to test
paracellular/intercellular transport in response to conditioned media, green circle = 70
kDa dextran-488. B. Movement of 70 kDa dextran measured at 12hr in various
conditions over 3 independent biological replicates. Error bars represent standard error
of means, ** is p < 0.0061 and p < 0.0089 from one-way ANOVA. C. Reduced Evan’s
Blue albumin (67 kDa) movement across transwell with high density EC, with or without
conditioned media measured across 3 independent biological replicates confirming
barrier establishment. Zoom in of graph shows difference in iPSEN1*-cPC paracellular
transport compared to controls. Error bars represent standard error of means, *** is p <
0.0001, * is p < 0.0202 from one-way ANOVA.
A . B . C .
65
transwells (Figure 3.8A, 3.9A). iPSEN1*-cPC showed a defect in Ab efflux that was
resolved with iPSEN1
corr
-cPC (Figure 3.8B). When visualized, EC on transwells showed
little retention of Ab after the duration of the clearance experiment; EC cultured with
iPSEN1*-cPC, however, showed excessive accumulation of Ab within the cells (Figure
3.8C). This phenotype points at a dual process for Ab transport across EC of the BBB,
and PSEN1 playing a role in both the uptake of Ab on the abluminal side, and the
transport of Ab on the luminal side, with perhaps a greater role in the second process,
leading to the Ab accumulation within the cells in the case of AD (Figure 3.8D). In the
case of Transferrin uptake, iPSEN1*-cPC exhibited a defect in uptake across the BBB,
while iPSEN1
corr
-cPC demonstrated similar uptake to controls (Figure 3.9B).

Pericyte defects occur ab initio in ADAD in vivo rat model
The in vitro-derived cPC represent a developmentally young cell, experiencing
roughly 5 weeks between ESC/iPSC state and matured cPC state. Therefore, we can
think of these cells as representing late development or early postnatal stages. AD is
considered an aging disorder, and even early-onset fAD does not clinically present itself
until one’s 30s at the earliest, yet we are seeing defects in young iPSEN1*-cPC. To see
if AD-related defects are noticeable at birth, we turned to an in vivo rat model of AD with
a truncated PSEN1 gene, TgF344 (Figure 3.10A). These rats show clinical symptoms
of AD in 6-month-old adult rats, such as amyloid plaques, neurofibrillary tangles,
neuronal loss, vascular defects, and cognitive behavioral deficits (Cohen et al., 2013).
We found that along with these symptoms, adult rats also showed signs of pericyte
defects, such as symmetric and diffuse NG2 localization, compared to asymmetric
66

Figure 3.8. Poor Ab clearance across barrier with iPSEN1*-cPC. A. Schematic of
experimental setup to test clearance of Ab across transwell with or without conditioned
media, green circle = Ab(42)-488. B. Clearance of Ab measured at 24hr and 48hr in
various conditions over 3 independent biological replicates. Error bars represent
standard error of means, ** is p < 0.0027 and p < 0.0024 from unpaired student’s t-test.
C. Images of EC monolayer on same transwells from B showing Ab(42)-488
accumulation in EC treated with iPSEN1*-cPC conditioned media. D. Schematic of two-
step process of Ab clearance by EC monolayer in response to normal PC and defects in
both steps in iPSEN1*-cPC.

A . B .
C .
D .
67

Figure 3.9. Poor Transferrin uptake across barrier with iPSEN1*-cPC. A. Schematic
of experimental setup to test Transferrin uptake in response to conditioned media, red
circle = Transferrin-TexasRed. B. Reduced uptake of Transferrin-TexasRed measured
at 6hr and 24hr in various conditions over 3 independent biological replicates. Error bars
represent standard error of means, *** is p < 0.0001, ** is p < 0.0031 from unpaired
student’s t-test.

A . B .
68

Figure 3.10. Structural defects in neonatal TgF344 rat pericytes. A. Schematic of
transgenic rat fAD model. B. IF of pericyte markers on 12-month-old TgF344 rat brain
pericytes. Red arrow points to abnormal distribution of NG2. Scale bars represent 10
µm. C. IF of pericyte markers in 5-day-old wild type and TgF344 rat brain pericytes from
two different litters. Red arrows point to abnormal distribution of NG2 seen in TgF344
rats; white arrow points to normal NG2 localization in controls. Scale bars represent 10
µm.
A . B .
C .
69
localization in age-matched controls (Figure 3.10B). To examine pericyte defects from
birth, we isolated pericytes from neonatal pups and evaluated these cells for the
presence of defects we identified in iPSEN1*-cPC. Like the adult rat pericytes and the
iPSEN1*-cPC, neonatal rat pericytes from the TgF344 animals showed NG2
mislocalization compared to littermate controls (Figure 3.10C). As in the case of
iPSEN1*-cPC, the neonatal rat pericytes also showed defects in their ability to
communicate with EC. Average tube length and area of coverage was reduced in
TgF344 pericytes in a 2D tube formation assay compared to littermate controls (Figure
3.11A). Transgenic rat pericytes were also less efficient than controls at inducing
expression of tight junction proteins in EC, and in generating EC-EC junctions (Figure
3.11B-C). This in vivo AD model corroborates our in vitro evidence that defects in AD
pericytes occur from the start of and throughout the lifespan of the pericytes.

ANGPT1 was downregulated in ADAD cPC and exogenous ANGPT1 rescued
pericyte defects
In order to further characterize AD pericyte defects, we performed RNA-seq to
compare normal iPSC-cPC and iPSEN1*-cPC (Figure 3.12A). Our analysis revealed
many downregulated pathways in iPSEN1*-cPC, specifically pathways associated with
the defects we observed such as anatomical structure morphogenesis, behavior, and
cell adhesion (Figure 3.12A). One gene that we found to be downregulated compared
to controls was ANGPT1, a gene encoding a secreted protein responsible for PC-EC
communication via secretion of ANGPT1 by pericytes that binds to the TIE2 receptor on
EC. To verify our RNA-seq data, we looked at ANGPT1 protein expression via
70

Figure 3.11. Pericyte defects occur ab initio in transgenic rat model of fAD in vivo.
A. Box and whiskers plot of average tube lengths of lattices/tubes formed by EC in 4
days when co-cultured with control or TgF344 rat pericytes over 4 technical replicates
per animal. B. IF for g-catenin on EC when co-cultured with control or TgF344 rat PC or
conditioned media. Scale bar represents 10 µm. C. Quantitation of EC-EC junctions
using high-resolution maximum projection confocal images of N > 100 cells from at least
3 independent biological replicates. Cumulative signal calculated on N > 5 images with
identical thresholding, without saturated pixels. Error bars represent standard error of
means, *** is p < 0.0001, * is p < 0.0332 from one-way ANOVA.

A .
B . C .
71

Figure 3.12. fAD pericytes show reduced expression of ANGPT1. A. (Top) cPC
derived from iPSC in a dish and differentially expressed genes comparing biological
replicates of iPSEN1*-cPC and normal cPC. (Bottom) GO terms of cellular components
and biological processes downregulated in fAD-cPC compared to normal cPC. B. IF
showing reduced ANGPT1 expression in iPSEN1*-cPC compared to iFF-cPC, iMF-cPC,
A . B .
C . D .
E .
72
and iPSEN1
corr
-cPC. Scale bars represent 10 µm. C. IF showing ANGPT1 expression in
cPC + shPSEN1 14 days post shRNA induction compared to control shRNA. The
shRNA expression cell is indicated with a double-headed arrow and a small filed is
zoomed in (I and ii) showing change in ANGPT1 expression with PSEN1-specific
shRNA expression. Scale bar represents 10 µm. D. Quantitation of ANGPT1 expression
using high-resolution maximum projection confocal images of N > 100 cells from at least
3 independent biological replicates. Cumulative signal calculated on N > 5 images with
identical thresholding, without saturated pixels. Error bars represent standard error of
means. E. Western Blot for ANGPT1 showing decreased level of protein in fAD
compared to control. GAPDH confirms equal protein loading.
73
immunofluorescence; control cPC and iPSEN1
corr
-cPC all expressed ANGPT1 protein,
but expression was severely downregulated in iPSEN1*-cPC (Figure 3.12B, D-E).
shRNA knockdown of PSEN1 in control cPC leading to a decrease in ANGPT1
expression recapitulated the pattern seen in iPSEN1*-cPC (Figure 3.12C-D).
To see if ANGPT1 loss was associated with the defects described above, we
added soluble ANGPT1 protein to cocultures of EC and iPSEN1*-cPC to see if any of
the defects were mitigated. In simple co-culture, EC exposed to iPSEN1*-cPC with
exogenous ANGPT1 showed an increase in overall EC-EC contacts, with more visible
cell-cell interactions (Figure 3.13A). Junctional protein expression also increased with
the addition of ANGPT1, both in expression level and integrity (Figure 3.13B-D).
ANGPT1 was also able to rescue tube formation defects in iPSEN1*-cPC cocultures.
With the addition of ANGPT1 to 2D tube formation assays, there was an increase in
overall tube area covered, as well as an increase in branching points, albeit not to the
level of controls (Figure 3.14). This indicates that ANGPT1 expression is required for
some EC communication functions of cPC, but not all the communication defects we
characterized are mediated via ANGPT1.
To determine if ANGPT1 loss is causative of any pericyte defects, we decided to
examine the effects of ANGPT1 loss in wild type cells. We first used targeted siRNA to
reduce expression of ANGPT1. Within a short time (16hrs), loss of ANGPT1 caused a
change in cell morphology, from an elongated cell to a shorter and squatter cell,
reminiscent of AD-cPC (Figure 3.15A, D). Along with loss of ANGPT1 expression with
the siRNA, we also saw a change in NG2 expression from asymmetric membranous
localization to symmetric diffuse localization (Figure 3.15B-C). Aside from targeting
74

Figure 3.13. Rescue of EC-EC contacts in iPSEN1*-cPC cultures with ANGPT1. A.
(Left) mCherry-labeled EC co-cultured with iPSEN1*-cPC conditioned media with and
without 50 ng/mL ANGPT1 showing EC-EC interactions. (Right) Surfaces of contact
between EC were manually isolated and show long stretched and multiple overlapping
contacts upon addition of ANGPT1. Scale bars represent 10 µm. B. IF for N-cadherin on
EC when co-cultured with iPSEN1*-cPC conditioned media with and without 50 ng/mL
ANGPT1. Scale bar represents 10 µm. C. Quantitation of N-cadherin signal using high-
resolution maximum projection confocal images of N > 100 cells from at least 3
independent biological replicates. Cumulative signal calculated on N > 5 images with
identical thresholding, without saturated pixels. D. Quantitation of g-catenin signal using
high-resolution maximum projection confocal images of N > 100 cells from at least 3
independent biological replicates. Cumulative signal calculated on N > 5 images with
A . B .
C . D .
75
identical thresholding, without saturated pixels. Error bars represent standard error of
means, *** is p < 0.0001 from unpaired student’s t-test.

76

Figure 3.14. Tube formation defect in iPSEN1*-cPC partially rescued with
ANGPT1. A. Images of lattices formed by EC in 2 days with or without 50 ng/mL
ANGPT1 when cultured independently or co-cultured with iPSEN1*-cPC in growth-
factor-reduced geltrex matrix. B. Tube area in pixels squared measured to quantify area
covered with cellular structures over 5 random fields in each of 3 independent biological
replicates. Error bars represent standard error of means, *** is p < 0.0001, ** is p <
0.0018 from unpaired student’s t-test. C. Tube branch points were counted over 5
random fields in each of 3 independent biological replicates. Error bars represent
standard error of means, *** is p < 0.0001 from unpaired student’s t-test.

A . B .
C .
77

Figure 3.15. Loss of ANGPT1 in pericytes leads to structural defects. A. Confocal
images of iFF-cPC expressing pgk-GFP at 3hr and 16hr after treatment with siANGPT1,
showing changes in cell morphology. B. IF of ANGPT1 in iFF-cPC treated with non-
targeting control siRNA or siANGPT1. C. IF of NG2 in iFF-cPC treated with non-
targeting control siRNA or siANGPT1. D. Violin plot showing quantitation of cell
morphology of 1°PC treated with siControl or siANGPT1 over a 24-36hr period.
A . B . C . D .
78
ANGPT1 directly, we also tested the effect of inhibiting the ANGPT1 signaling pathway,
using an inhibitor for the TIE2 receptor. Upon addition of the TIE2 inhibitor to wild type
cPC, there again was a shift in cell morphology from elongated to rounder and shorter
cells (Figure 3.16).
ANGPT1 is a secreted protein and is involved in cell-cell signaling between PC
and EC; however, cPC also express TIE2 receptor, and are therefore able to respond to
extracellular ANGPT1. To see if ANGPT1 could rescue any cell autonomous defects in
AD pericytes, we added exogenous protein to cultures of iPSEN1*-cPC. The addition of
ANGPT1 led to a shift toward more asymmetric and membranous localization of NG2
(Figure 3.17). There was also a change in cell morphology with time upon ANGPT1
addition, in both iPSEN1*-cPC and transgenic TgF344 neonatal rat pericytes (Figure
3.18). These data show that ANGPT1 is involved in multiple pathways associated with
proper pericyte function, and mutations in PSEN1 disrupt these functions via loss of
ANGPT1 expression.

CONCLUSION
Here we have shown how our previously described method of in vitro-generation
of cPC can be used to gain insight into cell type-specific defects in patients. In the case
of early-onset fAD, with patient iPSCs we have been able to define a set of defects seen
in the pericytes, including overall cell structure, cell survival, and cell communication.
We are able to support our findings with CRISPR/Cas9-corrected isogenic cell lines,
alternative methods of mutations (such as shRNAs and small molecules), and in vivo
recapitulation of the defects. Coupled with in vivo verification, our method has also
79

Figure 3.16. TIE2 inhibitor phenocopies ANGPT1 knockdown/fAD. A. Confocal
images of iFF-cPC either untreated (top) or treated with TIE2 inhibitor for 3hr (bottom)
showing changes in cell morphology. Cells are stained with WGA-488 for visualization.
B. Violin plot showing quantitation of cell morphology of iFF-cPC untreated or treated
with TKi over a 6hr period.

A .
B .
80

Figure 3.17. Recombinant ANGPT1 rescues structural defects. A. IF of NG2 on
iPSEN1*-cPC cultured with or without 50 ng/mL ANGPT1. B. Quantitation of NG2
expression and localization. Counts were done of high-resolution maximum projection
confocal images of N > 100 cells from at least 3 independent biological replicates. Error
bars represent standard error of means, ** is p < 0.0013 from unpaired student’s t-test.

A . B .
81

Figure 3.18. ANGPT1 rescues morphological defects. Still images from 36hr movie
showing structural rescue of cell morphology in iPSEN1*-cPC expressing mCherry upon
ANGPT1 addition. White arrows indicate elongated processes. Scale bars represent 10
µm.
82
revealed that these pericyte defects occur early in life, perhaps ab initio, in the patients,
preceding any clinical signs or symptoms of disease. This points to pericytes as a
potential target for preventative therapeutics, as they may be the first cell type to
experience defects associated with AD.
By scrutinizing pericytes in such an extensive manner, we have also begun to
tease apart some of the molecular mechanisms behind AD-related dysfunction. We
have identified ANGPT1 as a key regulator of some pericyte functions that become
dysfunctional in the AD setting. ANGPT1 has previously been implicated in EC
angiogenesis, vessel architecture, and vessel stability (Thurston et al., 1999; Uemura et
al., 2002), but the role of ANGPT1 specifically in pericyte function, as well as a
connection between ANGPT1 and AD has not previously been established. Therefore,
our work identifies new therapeutic avenues for AD, as well as reiterates the importance
of studying specific cell types with rigor to identify the true story in pathological settings.

83
Chapter Four
Structural and functional defects in sAD pericytes

84
INTRODUCTION
Alzheimer’s disease (AD) is a form of dementia associated with neuronal loss,
cognitive decline, and eventually death; it is considered an aging disorder, with age of
the patient being the greatest risk factor for developing disease. The majority of AD
cases are referred to as sporadic AD (sAD), that is, the patient did not carry any known
heritable genetic mutations associated with AD. For some sAD cases the cause can be
determined, either by the patient’s APOE status, or the presence of a de novo mutation.
However, many sAD cases are considered to have no known cause and may even be
due to a combination of genetics and environmental factors. Whatever the initial cause,
all patients with sAD end up experiencing the same devastating impact of the disease.
While sAD and early-onset familial AD (fAD) are considered almost to be two
distinct diseases, the end-stage of both is represented by the same set of clinical
pathologies – amyloid plaques, neurofibrillary tangles, neuronal loss, and cognitive
decline. Notwithstanding the different initial causes of disease or the differences in age-
of-onset and progression, there is a similarity in the way the brain is insulted, and
therefore a similarity in the defects between patients. Therefore, it might be the case
that pericytes at the BBB are similarly defective in fAD and sAD, despite the mutations
or genetic and epigenetic insults to the cells being so different. From this conclusion, it
could be drawn that there is an overall pericyte defect signature specific to AD that can
be identified and treated in patients regardless of the type of AD they suffer from, or the
causes of their disease. It is therefore important to study pericytes in the context of sAD
to determine if there are similarities in pericyte defects across AD, or if defects in
pericytes differ due to specific mutations or insults.
85

RESULTS AND DISCUSSION
sAD primary pericytes show structural defects
To begin assessing sAD pericytes, we obtained three different patient primary
pericytes isolated post-mortem (Table 4.1) all who had sAD due to an unknown cause.
For comparison we also obtained two different control patients’ primary pericytes,
isolated post-mortem, both who did not have any signs of AD or dementia. To begin
characterization of these cells, we looked at expression of NG2 and PDGFRb, two
proteins expressed in pericytes. While PDGFRb expression in all 3 sAD 1°PC was
equal to that of both wild type 1°PC, NG2 expression in the sAD 1°PC was symmetric
and diffuse as opposed to asymmetric in controls (Figure 4.1A-B). Overall cell
morphology was also altered in the sAD 1°PC; instead of thin and elongated cells seen
in the controls, sAD 1°PC were rounder and shorter, with smaller processes (Figure
4.1C-D). These structural defects were consistent across all 3 different patients and
were similar to the structural defects we previously found in in vitro-derived fAD
pericytes (Chapter 3).

sAD primary pericytes show defective communication with EC
Communication between EC and PC is paramount to maintaining the integrity of
the BBB. To determine the ability of sAD pericytes to properly communicate with EC, we
co-cultured EC with sad 1°PC. Pericytes normally signal to EC to arrest EC proliferation.
86

Table 4.1. Source of post mortem brain tissue samples used for primary pericyte
isolation. Table of information on patient sources of post mortem brain tissue samples
used for primary pericyte isolation. Primary criteria for patient selection was late onset
AD.
87

Figure 4.1. Structural defects in sAD 1°PC. A. IF of pericyte markers on two control
1°PC. B. IF of pericyte markers on three sAD 1°PC. White arrowheads indicate broad
NG2 expression in sAD PC. Scale bars represent 10 µm. C. Confocal images of control
1°PC expressing mCherry. D. Confocal images of sAD 1°PC expressing mCherry.
Scale bar represents 100 µm.
A . B .
C . D .
88
When cocultured with EC, sAD 1°PC were unable to suppress EC proliferation, causing
a majority of the EC in the culture to remain Ki67
+
(Figure 4.2). PC-EC communication
is also imperative during angiogenesis and blood vessel remodeling, as PC signal to EC
to control the vessel architecture, as well as stabilize vessels. In 3D tube formation
assays, sAD 1°PC cultured with EC on fibrin gel beads were unable to produce
extensive tube structures, capable only of making small tubes with no branching points
(Figure 4.3A). Likewise, in a 2D tube formation assay, sAD 1°PC cocultured with EC
showed reduced tube length, area of coverage, and branching points (Figure 4.3B-C).
Proper BBB architecture lends itself to proper BBB function, so it goes that
defects in vessel structure would be accompanied with defects in function. To test
barrier function of sAD pericytes, we used a transwell system to mimic an in vitro BBB.
In terms of paracellular transport, or transport between cells, sAD 1°PC were as efficient
as controls in blocking transport of dextrans across the EC barrier (Figure 4.4). To
examine transcytosis through EC, we looked at the ability of sAD 1°PC to aid EC in Ab
clearance and transferrin uptake. When a tight EC barrier is induced before adding
pericytes, the addition of sAD 1°PC reveals a small but statistically significant defect in
blocking Evan’s Blue paracellular transport (Figure 4.5A-B). Measuring the capability of
the cells to promote Ab clearance or transferrin uptake reveals a more significant defect,
with sAD 1°PC much less capable than control 1°PC to promote transcytosis in either
direction (Figure 4.5C-D). When the EC on the barrier are imaged, EC cultured with
sAD 1°PC show excessive accumulation of Ab within the cells, a phenotype that is not
seen in EC cultured with control 1°PC (Figure 4.5E). This points at pericytes not only
communicating with EC at the BBB in regard to uptake of molecules, but also in the
89

Figure 4.2. sAD 1°PC have defect in EC proliferation arrest. A. IF for Ki67 and
PECAM of EC co-cultured with sAD 1°PC or control 1°PC. Representative example of
all 3 patient lines. Scale bar represents 10 µm. B. Quantitation of Ki67 expression done
using maximum projection confocal images of N > 100 cells from at least 3 independent
biological replicates. Error bars represent standard error of means, *** is p < 0.0001
from unpaired student’s t-test.

A . B .
90

Figure 4.3. sAD 1°PC have a defect in inducing EC tube formation. A. Confocal
images showing fewer and shorter lumen filled tubules growing in fibrin gel from beads
coated with sAD 1°PC and mCherry-expressing EC (shown in magenta) compared to
normal control 1°PC in identical conditions. White arrow indicates normal tube with
branch points; red arrows indicate short tubes with sAD 1°PC. B. Phase contrast image
showing lack of networks and ‘stubby ends’ (red arrow) found in EC co-culture with sAD
1°PC, compared to EC with control 1°PC on growth-factor-reduced geltrex. C. (Left)
Tube area in pixels squared measured to quantify area covered with cellular structures
in tube assay on growth-factor-reduced geltrex, over 5 random fields in each of 3
independent biological replicates. Error bars represent standard error of means, * is p <
0.013, *** is p < 0.0001 from unpaired student’s t-test. (Right) Tube branch points were
counted over 5 random fields in each of 3 independent biological replicates. Error bars
represent standard error of means, *** is p < 0.0001 from unpaired student’s t-test.
A . B .
C .
91

Figure 4.4. sAD 1°PC are able to block paracellular transport across a barrier. A.
Schematic of experimental setup to test paracellular transport in response to
conditioned media, blue circle = 70 kDa dextran-488. B. Paracellular transport of 70 kDa
dextran-488 measured 30 mins post addition in various conditions over 3 independent
biological replicates. Error bars represent standard error of means, ns = no statistical
significance from one-way ANOVA.

A . B .
92

Figure 4.5. sAD 1°PC show defects in transcytosis across a barrier. A. Schematic
of experimental setup of high-density EC to test Transferrin uptake and clearance of Ab
in response to conditioned media, green circle = Ab(42)-488. B. Reduction of Evan’s Blue
transfer confirms minimal free diffusion via intercellular spaces prior to addition of
biomolecules for transcellular transport assays. Inset shows magnified view of the small
but significant further reduction in Evan’s blue in pericyte conditioned medium as
compared to EC alone, indicating strengthening of the pre-formed barrier by PC. Error
bars represent standard error of means, * is p < 0.0202 from one-way ANOVA. C.
Decreased uptake of Transferrin-TexasRed in sAD 1°PC measured at 6hr and 24hr in
various conditions over 3 independent biological replicates. Error bars represent
standard error of means, *** is p < 0.0001, * is p < 0.0440 from one-way ANOVA and
unpaired student’s t-test. D. Decreased clearance of Ab(42)-488 measured at 24hr and
48hr in various conditions over 3 independent biological replicates. Error bars represent
A . B . C .
D . E .
93
standard error of means, *** is p < 0.0001, ** is p < 0.0061 from one-way ANOVA. E.
Confocal images of EC on transwells after 3 days of incubation with Ab(42)-488 and
conditions media from normal or sAD 1°PC. Ab(42)-488 accumulation inside cells with
sAD conditioned media is shown in green. Insets zoom in to a single cell to show
accumulation around the nucleus.
94
release of molecules from the cells, a step that appears defective in the context of sAD.
These barrier defects again show consistency across sAD patient cells and similarity
with previously identified defects in fAD pericytes.

In vitro-derived sAD-cPC show similar yet milder defects
While all the defects noted above are consistent between fAD and sAD pericytes,
the fAD characterization was done in in vitro-derived cPC, while the sAD
characterization was done in 1°PC. To better compare the two distinct types of AD, we
obtained fibroblasts from one of our sAD patients, generated iPSCs from these
fibroblasts, and generated in vitro-derived cPC (044-cPC) (Figure 4.6A). These cPC not
only provide a better direct comparison with our work in fAD-cPC, but they also
represent a developmentally younger version of pericytes from the 044 patient, allowing
for comparison between the young 044-cPC and the old 044 1°PC.
When we generated 044-cPC and looked at the cell population expression of
pericyte markers, we saw that 044-cPC showed a milder phenotype than the 044 1°PC
(Figure 4.6B-C). While almost all NG2
+
044 1°PC showed symmetric NG2 expression,
only half of the NG2
+
044-cPC had symmetric expression and the other half had normal
asymmetric expression. This could be due to the fact that 044-cPC are younger cells
than the 044 1°PC, so their defect is not as severe. It is possible that there are two
populations of pericytes in this patient, and the normal asymmetric NG2 population dies
out over time; or the pericytes in this patient could transition with time from asymmetric
NG2 expression to symmetric expression, and the younger cells show that not all of the
pericytes have yet made this transition. Despite this milder phenotype in NG2
95

Figure 4.6. In vitro-derived sAD-cPC show similar defects as sAD 1°PC. A.
Schematic showing generation of sAD-patient cPC. B. Quantitation of NG2 expression
and localization. Counts were done of maximum projection confocal images of N > 100
cells from at least 3 independent biological replicates. Error bars represent standard
error of means, *** is p < 0.0001, ** is p < 0.0031, * is p < 0.0413 from unpaired
student’s t-test. C. IF of pericyte markers on 044 sAD-cPC with zoomed in images of
mixed population of NG2 localization. Scale bars represent 10 µm. D. Quantitation of g-
catenin on ECs when co-cultured with two different clones of 044 sAD-cPC or sex-
matched controls, quantified using high-resolution maximum projection confocal images
of N > 100 cells from at least 3 independent biological replicates. Cumulative signal
calculated on N > 5 images with identical thresholding, without saturated pixels. Error
bars represent standard error of means, *** is p < 0.0001 from unpaired student’s t-test.
A .
B .
C .
D . E . F .
96
E. Tube area in pixels squared was measured to quantify area covered with
lattices/tubes formed by EC in 2 days when co-cultured with 044 sAD-cPC or controls
on growth-factor-reduced geltrex matrix over 5 random fields in each of 3 independent
biological replicates. Error bars represent standard error of means, *** is p < 0.0001
from unpaired student’s t-test. F. Tube branch points were counted over 5 random fields
in each of 3 independent biological replicates. Error bars represent standard error of
means, *** is p < 0.0001 from unpaired student’s t-test.
97
localization, 044-cPC show a defect in inducing EC junction proteins and in inducing
tube formation that is of a similar magnitude as 044 1°PC (Figure 4.6D-F). This shows
that although some defects may be milder in young cells, defects in pericytes still
present themselves very early in AD.

sAD pericytes show reduction in ANGPT1
To further examine the defects in sAD 1°PC, we compared bulk RNA-seq data
from the 044-patient sAD 1°PC with control 1°PC. When the differentially expressed
genes were put into GO analysis, we found many processes and cellular components
that were downregulated in sAD 1°PC that aligned with our defects: cell junction,
membrane, and cell-cell signaling among others (Figure 4.7). Among the many
differentially expressed genes, we chose to follow up on ANGPT1 due to its importance
in signaling between PC and EC at the BBB, and because of our previous findings with
fAD and ANGPT1 (Chapter 3). To verify the RNA-seq data showing that ANGPT1
mRNA is decreased in sAD 1°PC (data not shown), we performed immunofluorescence
and western blot analysis; our results showed clear ANGPT1 expression in control
1°PC, and reduced or absent expression in all patient sAD 1°PC (Figure 4.8A-C). We
also saw the same reduction in the 044-cPC (Figure 4.8D), confirming the loss of
ANGPT1 across all AD samples.

sAD pericyte defects can be rescued with exogenous ANGPT1
To determine if there is a connection between the defects in sAD 1°PC and loss
of ANGPT1, we looked to see if ANGPT1 could rescue any of the previously described
98

Figure 4.7. Various processes and cellular components downregulated in sAD
1°PC. A. 1°PC from normal and sAD patients and differentially expressed genes
comparing biological replicates of disease and normal 1°PC. B. GO terms of cellular
components and biological processes downregulat4ed in sAD 1°PC compared to
normal 1°PC.

A .
B .
99

Figure 4.8. Reduced ANGPT1 expression in sAD cells. A. IF showing ANGPT1
expression on two independent samples of normal 1°PC versus three samples of sAD
1°PC. Scale bars represent 10 µm. B. Western Blot for ANGPT1 showing decreased
level of protein in sAD 1°PC compared to control; GAPDH confirms equal protein
loading. C. Graph showing approximately 4-fold reduction in average ANGPT1 protein
between sAD 1°PC and control 1°PC. Error bars represent standard deviation between
patients, *** is p < 0.0001 from unpaired student’s t-test. D. IF of ANGPT1 on two
different clones of 044 sAD-cPC. White box indicates area of zoom.
A .
B .
C .
D .
100
phenotypes. In terms of PC-EC communication, when soluble ANGPT1 protein is added
to cocultures, PC-induced tight junction protein expression in EC increases (Figure
4.9A). ANGPT1 is also able to rescue defects in tube formation, including total area of
coverage and number of branching points, albeit branching points were not rescued to
the level of controls (Figure 4.9B-D). Looking at cell autonomous defects, we added
exogenous ANGPT1 to cultures of sAD 1°PC. We saw a shift in NG2 expression in sAD
1°PC from mostly symmetric to majority asymmetric (Figure 4.10). Overall cell
morphology and random movement was also affected by addition of ANGPT1; cells
shifted from shorter, circular cells with many processes to longer, thinner cells with
fewer processes, and the longer cells showed more locomotion in random directions
than the defective cells (Figure 4.11). These signs of rescue are not simply due to the
presence of ANGPT1 protein but the activation of ANGPT1 signaling – targeting other
members of the pathway results in rescue as well. PTPRB is a protein responsible for
dephosphorylating the TIE2 receptor after ANGPT1 is bound to it, resulting in a switch
from active signaling to inactive signaling (Figure 4.12A). Using an siRNA targeting
PTPRB, resulting in loss of protein and therefore continuous activation of ANGPT1
signaling leads to a rescue in sAD 1°PC cell morphology (Figure 4.12B-C). This opens
doors for potential therapeutics, showing that multiple members of a signaling pathway
can be targeted to overcome disease-related defects.

CONCLUSION
We have shown here that sAD pericytes have a set of structural and functional
defects that is consistent across patients despite unknown causes of disease. These
101

Figure 4.9. ANGPT1 rescues EC communication defect in sAD 1°PC. A.
Quantitation of cumulative g-catenin on EC when co-cultured with sAD 1°PC conditioned
media with and without 50 ng/mL ANGPT1, quantified using high-resolution maximum
projection confocal images of N > 100 cells from at least 3 independent biological
replicates. Cumulative signal calculated on N > 5 images with identical thresholding,
without saturated pixels. Error bars represent standard error of means, *** is p < 0.0001
from unpaired student’s t-test. B. Images of lattices formed by EC in 2 days with or
without 50 ng/mL ANGPT1 when cultured independently or co-cultured with sAD 1°PC
in growth-factor-reduced geltrex matrix. C. Tube area in pixels squared was measured
to quantify area covered with cellular structures, over 5 random fields in each of 3
independent biological replicates. Error bars represent standard error of means, *** is p
< 0.0001 from unpaired student’s t-test. D. Tube branch points were counted over 5
A .
B .
C .
D .
102
random fields in each of 3 independent biological replicates. Error bars represent
standard error of means, *** is p < 0.0001 from unpaired student’s t-test.

103

Figure 4.10. Exogenous ANGPT1 rescues sAD 1°PC structural defects. A. IF of
NG2 on sAD 1°PC when cultured with or without 50 ng/mL ANGPT1. White arrows
indicate cells with asymmetric NG2 localization. B. Quantitation of NG2 expression and
localization in sAD 1°PC with and without 50 ng/mL ANGPT1. Counts were done of
high-resolution maximum projection confocal images of N > 100 cells from at least 3
independent biological replicates. Error bars represent standard error of means, *** is p
< 0.0001 from unpaired student’s t-test.

A .
B .
104

Figure 4.11. Exogenous ANGPT1 rescues morphology and migration defects in
sAD 1°PC. A. (Top) Still images from a movie of characteristic 044 sAD 1°PC at
intervals up to 6hr. White arrows point to multiple projections of a single sAD-PC at t=0.
Note minimal change in overall morphology over the 6hr period. (Bottom) Rapid and
dramatic change in morphology of both ‘paintbrush-like’ (white star) and ‘stellate-like’
(white arrows) sAD 1°PC upon ANGPT1 addition, to acquire a more elongated shape
with active directional migration reminiscent of normal PC. B. Still images from 8hr
movies with migratory tracks of individual pericytes (shades of yellow and green) and
minimally displaced cells (white tracks) shown. Note rescue of directional migration in
majority of sAD 1°PC treated with one acute dose of ANGPT1.

A .
B .
105

Figure 4.12. PTPRB inhibition phenocopies ANGPT1 rescue. A. Schematic of Tie2
receptor in active state (binding of ANGPT1 and phosphorylation) and inactive state
(dephosphorylation by PTPRB). B. Confocal images of sAD 1°PC treated with siControl
(left) or siPTPRB (right) showing difference in cell morphology. C. Violin plot showing
quantitation of cell morphology of sAD 1°PC treated with siPTPRB or siControl.
A . B . C .
106
defects were seen in patient 1°PC as well as in vitro-derived cPC leading to two
conclusions: 1) our previously described protocol for in vitro-generation of cPC is able to
mimic disease-specific defects seen in primary cells, and 2) the defects described in
post-mortem sAD 1°PC were similar to those of the developmentally young in vitro-
cPC, suggesting that these pericyte defects exist in the patients long before any other
signs and symptoms of disease.
This multi-prong interrogation of sAD pericytes has also revealed the importance
of ANGPT1 in pericyte structure and function. Our previous work identified ANGPT1 as
being a key player in pericyte function based on defects seen in fAD-cPC; our work here
has supported this work in sAD 1°PC and cPC, showing the similarities between fAD
and sAD pericyte defects and the similarities between the molecular mechanisms
behind these defects. Both types of AD show a reduction of ANGPT1, and gain of
ANGPT1 is able to rescue many of the pericyte defects described, suggesting that
although fAD and sAD are different types of AD, there may be a similarity across AD in
how disease progresses regardless of cause of disease or age-of-onset of disease.

107
Chapter Five
Defects in cranial pericytes identified in CHARGE syndrome
108
INTRODUCTION
CHARGE syndrome is a developmental disorder characterized by a set of
congenital anomalies affecting multiple tissues. Among the tissues affected are neural
crest cell (NCC) derivatives – including cranial bones and cartilage, outflow tract of the
heart, and the inner ear – due to defects in NCC formation during development (Siebert
et al., 1985). Mutations in the CHD7 gene cause CHARGE syndrome; CHD7 encodes a
chromodomain helicase that functions as a nucleosome remodeling factor (Vissers et
al., 2004; Bouazoune and Kingston, 2012). Loss of CHD7 has been shown to result in
NCC formation, specification, and migration defects (Bajpai et al., 2010; Asad et al.,
2016), resulting in classification of CHARGE syndrome as a neurocristopathy (Aramaki
et al., 2007).
Many clinical aspects of CHARGE syndrome have previously been
characterized; however, no previous work has connected any defects in pericyte
number, survival, or function to CHARGE syndrome. As most NCC-derived tissues are
affected in CHARGE syndrome, it suggests that NCC-derived pericytes may also show
defects due to NCC-related dysfunction. To examine the structure and function of
pericytes in CHARGE syndrome, we obtained iPSC from two CHARGE patients with
two different mutations in CHD7: CHD7
1701X
and CHD7
1036X
. By generating in vitro-
derived cranial pericytes (cPC) from these patient iPSC, we can characterize pericytes
in the context of CHARGE syndrome and determine if pericytes of the BBB share NCC-
related defects in these patients.

109
RESULTS AND DISCUSSION
cPC from CHARGE patients show structural and survival defects
To begin characterization of pericytes in CHARGE syndrome, cPC were
generated from two different patient iPSCs via a neural crest intermediate. As expected,
patient iPSCs generated fewer NCCs and exhibited delayed differentiation compared to
wild type iPSCs (Bajpai et al., 2010; Rada-Iglesias et al., 2012). Once equal numbers
were obtained, NCC were induced to cPC fate (Griffin and Bajpai, in review). Upon cPC
differentiation, the first noticeable difference between the samples is the morphological
anomalies in CHARGE-cPC (Figure 5.1). Compared to the thin, elongated cells of
controls and CHD7
corr
-cPC, CHD7
1701X
-cPC and CHD7
1036X
-cPC are shorter and boxier.
This morphological defect does not change with time, indicating it is not a feature of
delayed differentiation, but an overall defect in cellular structure.
Upon examination of pericyte marker expression, both CHD7-cPC showed
defects in NG2 expression. Expression levels of NG2 was reduced in both CHD7-cPC,
along with the overall percentage of NG2
+
cells (Figure 5.2A-C). We were able to
recapitulate both the morphological and the NG2 expression defects in wild type cPC
upon addition of an shRNA targeting CHD7 for knockdown (Figure 5.2D-F). Reduced
levels of CHD7 expression led to a short, square morphology and reduced NG2
expression, indicating the defects in CHD7-cPC are due to mutations.
Extended culture of CHD7
1701X
-cPC and CHD7
1036X
-cPC revealed a long-term
survival defect. CHD7-cPC grow at a slower rate than controls, and with time there is a
reduction in proliferation, eventually resulting in senescence of the population between
110

Figure 5.1. Structural defects in CHARGE pericytes. A. Phase images of CHD7
1701X
-
cPC, ESC-cPC, and CHD7
corr
-cPC showing morphological changes with mutation in
CHD7. Scale bars represent 10 µm. B. Confocal images of two CHARGE patient cPCs
and corrected control marked with cell surface dye highlighting cell morphology. Scale
bars represent 10 µm.

A . B .
111

Figure 5.2. Reduced NG2 expression with CHD7 deficiency. A. IF for NG2,
PDGFRb, CD13, ANGPT1, and ENDOGLIN in CHD7
1701X
-cPC, CHD7
1036X
-cPC, and
CHD7
corr
-cPC. White arrows indicate NG2 expression. Scale bars represent 10 µm. B.
Quantitation of NG2 expression. Counts were done of high-resolution maximum
projection confocal images of N > 100 cells from at least 3 independent biological
replicates derived at different times. Error bars represent standard error of means, ns =
no statistical significance, *** is p < 0.0001 from unpaired student’s t-test. C.
Quantitation of cumulative NG2 signal. Counts were done of high-resolution maximum
A . B .
C .
D .
E .
F .
112
projection confocal images of N > 100 cells from at least 3 biological replicates. Error
bars represent standard error of means, *** is p < 0.0001 from unpaired student’s t-test.
D. Phase images of normal cPC and two clones of cPC with shCHD7. White arrows
indicate cells with no shRNA expression; red arrows indicate cells with shCHD7
expression and subsequent rectangular shape. Scale bars represent 10 µm. E. IF of
NG2 in cPC expressing shCHD7. Red arrows indicate shRNA expressing cells; white
arrows indicate cells not expressing shRNA. F. Quantitation of percent cells with normal
or irregular cell morphology. Counts were done of high-resolution phase images of N >
100 cells from at least 3 independent biological replicates. Error bars represent
standard error of means.

113

Figure 5.3. Survival defect in CHD7-cPC. Quantitation of maximum passage number
reach during long-term culture of tested cell lines from N = 3 independent derivations.
Error bars represent standard error of means.
114
passages 25-30 (Figure 5.3). This survival defect is absent in CHD7
corr
-cPC, indicating
a connection between proper CHD7 expression levels and cell survival.

CHARGE cPC have defects in migration
Defects in cell migration have previously been described in CHARGE syndrome
(Okuno et al., 2017). Migration is a normal function of pericytes, as they migrate during
vascular development as well as in response to signals to regulate and maintain the
BBB. Therefore, we tested migration in CHARGE patient cPC to determine if there were
any defects. We previously saw that migration defects are not inherent to diseased
pericytes, as there were no defects in AD-cPC migration in a scratch assay (Figure
5.4A). When CHD7-cPC were assessed via a scratch assay, we saw that CHD7
1701X
-
cPC were unable to completely fill in the scratch area, and CHD7
1036X
-cPC were unable
to move past the scratch barrier at all (Figure 5.4B-C). These results add pericytes to
the collection of tissues and cell-types that experience migration defects in the context
of CHARGE syndrome.

CHARGE cPC have defects in communication with EC
Along with cell autonomous functions, we examined the ability of CHD7-cPC to
communicate with endothelial cells (EC), as PC-EC interaction is the key feature of BBB
integrity. We assessed the ability of CHD7-cPC to induce EC tube formation in a 2D
lattice/tube formation assay on growth-factor-reduced geltrex matrix, where EC alone
generate minimal tubes but wild type cPC are able to induce extensive tube structures
115

Figure 5.4. Migration defect in CHD7-cPC. A. Quantitation of percent of scratch area
filled in 48hr after scratch across cell types of various diseases, compared to controls,
across 3 independent biological replicates. Error bars represent standard error of
means, ** is p < 0.001 from one-way ANOVA. B. Schematic of scratch assay and
regions of imaging. C. Zoomed in live phase images of control (d), disease (a, c, e), and
corrected control (b, f) cPC 48hr after scratch in a scratch assay.

A . B .
C .
116

Figure 5.5. Defective induction of tube formation in CHD7-cPC. A. Images of
tubes/lattices formed by EC after 2 days on growth-factor-reduced geltrex with EC only
and cPC controls. White arrowhead points towards branch junctions; open arrowhead
identifies small structures unable to make tubes. B. Images of tubes/lattices formed by
EC after 2 days on growth-factor-reduced geltrex with CHD7-cPC. White arrowheads
point towards branch junctions; white asterisks indicate thin tubes.
A .
B .
117
with multiple branching points (Figure 5.5A). Both CHD7
1701X
-cPC and CHD7
1036X
-cPC
are able to induce the formation of tube structures above EC only with proper
branching, their structures were thinner and longer than controls (Figure 5.5B). There
were also regions of EC unable to make lattice structures, suggesting defects in
inducing tube formation, as well as stability of the tubes.

Rescue of CHARGE cPC with procainamide
As CHARGE syndrome is a developmental disorder affecting patients from birth,
much work has been done to relieve these children of the burden of this disorder. A
recent study has identified a number of small molecules with the potential to reverse
disease phenotypes based on rescue in zebrafish models (Asad and Sachidanandan,
2019). One small molecule identified was procainamide, a DNMT1 (DNA
methyltransferase 1) inhibitor, was able to partially rescue craniofacial defects in
cartilage, neurons, and Schwann cells. Therefore, we decided to test the ability of
procainamide to rescue pericyte structure and function defects in CHD7-cPC.
To observe the effect of procainamide on pericytes, we first simply cultured
CHD7
1701X
-cPC and CHD7
1036X
-cPC with procainamide to see if there were any
phenotypic changes. We observed an increase in NG2 expression with the addition of
procainamide in both patient lines (Figure 5.6A), with no changes in NG2 expression in
controls. NG2 expression also appeared properly localized in CHD7
1701X
-cPC, a
property that was not quantifiable without procainamide due to the low levels of
expression. Along with structure, we tested the ability of procainamide to rescue
pericyte functional defects. We performed scratch assays on CHD7
1701X
-cPC with
118

Figure 5.6. Rescue of pericyte defects in CHD7-cPC with procainamide. A. IF for
NG2 and PDGFRb in control and CHARGE cPC with and without 1 µM procainamide.
B. Quantitation of percent of scratch area filled at 0hr and 48hr post scratch with 0 nM,
500 nM, or 1 µM procainamide, across 3 independent biological replicates. Error bars
represent standard error of means.
A . B .
119
procainamide added to the cultures, and we observed improved migration ability at
multiple drug concentrations (Figure 5.6B). Procainamide improved the distance across
the scratch covered, as well as the overall coverage of the scratch area, to levels
comparable to controls. These results support the previous in vivo study showing
potential therapeutic applications of procainamide for CHARGE patients.

CONCLUSION
While CHARGE syndrome is a rare developmental disorder, it has a devastating
impact on patients and their families. It is imperative to understand the mechanisms
behind the disease and the causes/effects of specific defects in order to develop
successful treatments. We’ve shown here that NCC-derived pericytes generated from
CHARGE patient iPSCs have structural and functional defects that have the potential to
impair BBB function in vivo. This has implications for patients, as there has been no
previous description of pericyte defects, so this work can serve as an indication for
patients to have MRIs to test for BBB leakage. Providing information on and testing for
pericyte and BBB defects can help patients adopt preventative plans to improve their
overall wellbeing as well as prevent lasting damage.
We have also provided further support for procainamide as a treatment for
defects seen in CHARGE syndrome. Our work shows that multiple aspects of pericyte
structure and function can be ameliorated by procainamide, adding to the list of cell
types and features that have seen improvement with this treatment. While procainamide
does not represent a preventative or curative treatment, it could potentially provide
abatement of symptoms, leading to improved quality of life in patients. Decreasing the
120
burden of the disease will always benefit the patients, even if the disease is not cured.
The fact that pericytes are not the first cell type to show rescue with procainamide is
also encouraging, suggesting procainamide as a therapeutic option could be helpful for
a variety of symptoms, limiting the financial and medical burdens a drug could add to
the patients. Overall, our study on pericyte defects in CHARGE syndrome provides
patients with a rare disorder some insight into relief, as well as provides additional
support for already promising therapeutic avenues.

121
Materials and Methods
122
Experimental Model and Subject Details

Special thank you to the following: Valerie Hennes and Justin Ichida for providing iPSCs
with PSEN1 mutations; Abhay Sagare and Berislav Zlokovic for providing primary
patient pericytes; Brian Leung and Terrance Town for providing TgF344 rats and control
littermates; and William Stallcup for providing NG2 antibody.

Frogs
Wild type Xenopus laevis mature females were used to collect eggs. After fertilization,
embryos were staged using the Nieuwkoop and Faber staging method (Nieuwkoop and
Faber, 1994). Embryos were allocated into control and experimental groups with no bias
toward weight, size, sex, or any other feature in accordance to institutional IACUC
approved protocols.

Rats
We used the rat genetic model of AD known as TgF344 (Cohen et al., 2013). This
ADAD model is driven by bi-cistronic expression of two human AD-predisposing
mutations: the Swedish APP mutation that can form aggregation prone Ab (APP
swe
) and
a dominant negative truncated PSEN1 (PSEN1
DE9
). At 6 months of age, adult rats show
amyloid plaques, neurofibrillary tangles, neuroinflammation, neuronal loss,
cerebrovascular pathology, and behavioral deficits. All experiments with neonatal
transgenic rats were carried out with littermate controls.

123
Cell lines and primary cultures
Cells were grown at 37
o
C with 5% CO2.

Pluripotent stem cell lines:
All pluripotent stem cell lines described here were tested negative for mycoplasma
every three months (Lonza, LT07-318), and were found to have normal karyotype at
onset of experiments (CHLA cytogenetics core).
1. H9 = wildtype human embryonic stem cell line with normal karyotype from
Wicell (WA09)
2. iFF = human induced pluripotent stem cell line derived from normal female
fibroblasts with non-integrating reprogramming vectors
3. iMF = human induced pluripotent stem cell line derived from normal male
fibroblasts with non-integrating reprogramming vectors
4. iPSEN1* = human induced pluripotent stem cell line derived from male fAD
patient fibroblasts carrying PSEN1
A431E
mutation with non-integrating
reprogramming vectors
5. iPSEN1
corr
= PSEN1
A431E
mutation corrected with CRISPR-Cas9 based
genome editing to generate isogenic, control, human induced pluripotent stem
cells
6. CHD7
1071X
= human induced pluripotent stem cell line derived from CHARGE
patient fibroblasts carrying CHD7
1071X
mutation with non-integrating
reprogramming vectors
124
7. CHD7
1036X
= human induced pluripotent stem cell line derived from CHARGE
patient fibroblasts carrying CHD7
1036X
mutation with non-integrating
reprogramming vectors
8. CHD7
corr
= CHD7
1071X
mutation corrected with CRISPR-Cas9 based genome
editing to generate isogenic, control, human induced pluripotent stem cells
9. 044-iPSC = human induced pluripotent stem cell line derived from sAD
patient #044 fibroblasts with non-integrating reprogramming vectors
10. iPSEN1
homo
= human induced pluripotent stem cell line derived from male fAD
patient fibroblasts carrying PSEN1
A431E
mutation on both alleles of PSEN1
with non-integrating reprogramming vectors
11. iPSEN1
trunc
= human induced pluripotent stem cell line derived from male fAD
patient fibroblasts carrying a truncation of the PSEN1 gene at Lys109 with
non-integrating reprogramming vectors

iPSCs were generated using non-integrating episomal vectors and were PCR tested for
loss of inducible factor plasmids. Genotype of mutant and corrected lines were tested
via PCR followed by Sanger sequencing.

The mutation in iPSEN1* is A431E in exon 12, caused by a single nucleotide
substitution of an adenine for a cytosine, leading to amino acid 431 being changed from
alanine to glutamic acid. This mutation is known as the Jalisco mutation (rs63750083)
due to its origin in a single founder in Jalisco, Mexico. This mutation was corrected in
iPSEN1
corr
by switching the adenine back to a cytosine, as well as making a change in
125
the PAM located at base pair 86,667 from a cytosine to a thymine to prevent cutting
after insertion of the correction.

All pluripotent stem cell lines tested positive for OCT4, NANOG, E-Cadherin, and SOX2,
and retained the potential for multi-lineage differentiation.

All experiments characterizing cPC were performed with H9-cPC, iFF-cPC, and iMF-
cPC, using 1
o
PC (age 20) as controls. Representative images of a single control or
single disease line are shown in figures when results were similar across the various
lines.

Method Details

Cell Culture and Differentiation
ESC/iPSC culture
ESCs and iPSCs were grown and expanded on growth-factor reduced geltrex- (Gibco,
#A1413302) coated dishes in mTeSR medium (Stem Cell Technologies, Cat#05850) to
80% confluence. Media was changed daily. Cells were split every 5-7 days with
accutase (Stem Cell Technologies, #07920) (Bajpai, Lesperance, Kim, & Terskikh,
2008) and plated at a 1:6 diluted on geltrex-coated plates.
NCC differentiation
For NCC differentiation, ESCs/iPSCs were collected using collagenase IV (Gibco,
#17104-019), rinsed with PBS to remove all traces of enzyme, and transferred to
126
serum-free NCC medium, as described in Bajpai et al., 2010. Media was changed every
three days. All clusters of neuroectodermal spheres were washed off on day 10 with
PBS, and the remaining adherent NCC were used for cPC generation.
cPC generation
Adherent NCCs were collected on Day 10 by treatment with accutase. A small
proportion of the cells were plated in wells for quality control of NCC derivation,
including assessing morphology, migration ability, and presence of multiple
immunological markers expressed by NCC. The remaining NCC were replated in TC
dishes in PC induction medium (usually PCM1 or 5% serum containing medium with
PDGF-BB for 14 days as noted). Cells were split every 2-3 days during the 14-day
induction process with 0.05% trypsin-EDTA. cPCs were maintained in PCM1, 2, or 3
with regular medium changes every 3-4 days. Cells cultured in PCM1 were followed for
up to at least 8 months, using 0.05% trypsin-EDTA to split every 4-8 days. cPCs were
checked for maintenance of PC markers every 10 passages.

Identification of Pericyte Incompatible Media
Primary pericytes were split and transferred into the following media conditions: NCC
medium, MesenCult (Stem Cell Technologies, #05449), DMEM with 10% FBS, mTeSR
ESC medium, or Smooth Muscle Cell Medium. Cells were cultured for 7 days, live
imaged, then fixed in fresh 4% PFA followed by immunofluorescence for pericytes
markers. The media was determined incompatible based on the cell morphology, cell
survival, and immunofluorescence.

127
Identification of Pericyte Compatible Media
Primary pericytes were cultured in the company’s suggested medium (ScienCell,
#1201), as well as two recipes for pericyte media. Cells were cultured for 7 days, live
imaged, then fixed in fresh 4% PFA and tested for presence of pericyte markers. To
verify that these media were pericyte compatible, iFF-NCC were collected and
transferred into all 3 media. The cells were cultured and monitored for cell morphology
changes for 14 days, then fixed in fresh 4% PFA and tested for the presence of pericyte
markers. All 3 media were determined to be pericyte compatible based on cell
morphology, cell survival, successful induction of pericytes, and successful
maintenance of pericytes (triple positive for NG2, PDGFRb, and CD13).

Identification of Optimal Starting Cells for Generation of Pericytes
The following cell types were cultured in their respective media for 7 days: NCCs, fetal
fibroblasts, ESCs, mature NCCs (cranial mesenchyme), HUVECs, and VSMCs. After 7
days, cells were split, collected, and transferred to pericyte medium (PCM1). The cells
were cultured and monitored for cell morphology changes for 14 days, then fixed in
fresh 4% PFA and tested for the presence of pericyte markers. Cells were considered
capable of generating pericytes based on cell morphology, cell survival, and
immunofluorescence results after 14 days in pericyte medium.

Cell Immunofluorescence
Cells were plated on prewashed German glass coverslips coated with fibronectin and
grown to 60-80% confluence (usually 2 days after plating). They were then fixed using
128
fresh 4% paraformaldehyde. Cells were washed with PBS and placed in PBTx blocking
solution (1% BSA, 0.1% Triton-X in PBS) for 1 hour at room temperature. Primary
antibodies were brought to the desired dilution (1:100 to 1:1000) in PBTx and incubated
with cells overnight at 4
o
C. Afterward, the cells were washed with PBTx. Fluorophore
conjugated secondary antibodies diluted in PBTx (1:1000) were incubated with cells for
45 minutes in the dark, at room temperature. Cells were washed again with PBS, then
DAPI (0.1 mg/mL in PBS) was added for 3-5 minutes, before final washing with PBS.
Coverslips were mounted on slides using 20% glycerol and imaged using a Leica SP8
confocal microscope. Images of five random fields were taken for each sample, and all
IFs with the same antibodies but different cell lines were done in parallel for
comparative analysis. Images in figures are representative.

Paracellular Transport on Transwells
All transwell experiments were performed based on Takata et al. 2013, with
modifications. Transwell plates (0.4 μm pore size, Corning Cat#3450) were coated with
fibronectin at 4
o
C overnight. HBMECs were plated at 1.5 x 10
5
cells per cm
3
and grown
for 2 days. Then the media of the transwell insert was changed to a 1:1 mix of
endothelial cell medium (ScienCell, Cat#1001) with either prewarmed regular pericyte
media (EC only controls) or pericyte conditioned media (media taken from a plate
growing pericytes for at least 3 days, media
cm
) for 12 hours. Then, 5 μg 10 kDa or 70
kDa Dextran-488 in endothelial cell media was added to the outer chamber. At short (15
min, 30 min, 1 hr, 2 hr) or long (24 hr, 48 hr) time-intervals, 100 μl was taken from the
inner chamber and replaced with prewarmed 1:1 mix EC medium with PCM1 or
129
media
cm
. The removed volume was placed in a 96-well plate and measured on a plate
reader (Victor
3
V*, Perkinelmer #1420-040) (Ex(λ) 485 nm; Em(λ) 530 nm). Average
fluorescence values in RFUs were calculated and plotted for each sample, timepoint,
and replicate experiment, with error bars = 1 standard error of deviation calculated on
individual averages and their independent standard deviation. At least three biological
replicate experiments were performed for each cell line. Media
cm
from all cell lines was
collected from dishes with equal number of cells and was collected immediately before
use.

Aβ Clearance with Transwells
Transwell plates were coated with fibronectin at 4
o
C overnight. HBMEC were plated at
1.5 x 10
5
cells per cm
3
and grown for 4-5 days until a tight barrier was formed in
endothelial cell media. Media was then changed to 1:1 mix of endothelial cell media
(ECM) with either (i) regular pericyte media (EC only) or (ii) conditioned pericyte media
(media
cm
) v/v for 12 hours. Barrier tightness was tested by adding 4% Evan’s Blue to
either the inner chamber and determining flow through to the other chamber over 2 hr,
or vice versa. Samples from the opposite chamber of Evan’s Blue addition were taken
every 15 minutes and measured on a SpectraMax iD3 Multi-Mode Microplate reader
(Molecular Devices) for absorbance at 630 nm. The transwells were then moved to a
new plate, rinsed gently with PBS, and media was changed (1:1 ECM to PCM1 or
media
cm
). This media mix also contained Aβ42-488 to a final concentration of 440 nM
and was added to the inner chamber representing the abluminal side. Samples were
taken from the outer chamber at 2, 24, and 48 hours and measured on a plate reader
130
(Victor
3
V*, Perkinelmer #1420-040) (Ex(λ) 485 nm; Em(λ) 530 nm). Average
fluorescence in RFUs across 3 replicate experiments was calculated and plotted for the
different cell types and time points, with error bars = 1 standard error of deviation
calculated on individual averages and their independent standard deviation. At least
three biological replicate experiments were performed for each cell type and/or
condition. Images of cells on transwells were taken by placing transwells with live cells
onto glass-bottom dishes with minimal medium and imaged on a Leica SP8 confocal
microscope.

Transferrin Uptake with Transwells
Transwell plates were coated with fibronectin at 4
o
C overnight. HBMEC were plated at
1.5 x 10
5
cells per cm
3
and grown for 4-5 days until a tight barrier was formed in
endothelial cell media. Media was then changed to 1:1 mix of endothelial cell media
(ECM) with either (i) regular pericyte media (EC only) or (ii) conditioned pericyte media
(media
cm
) v/v for 12 hours. Barrier tightness was tested by adding 4% Evan’s Blue to
either the inner chamber and determining flow through to the outer chamber over 2 hr,
or vice versa. Samples from the opposite chamber of Evan’s Blue addition were taken
every 15 minutes and measured on a SpectraMax iD3 Multi-Mode Microplate reader
(Molecular Devices) for absorbance at 630 nm. The transwells were then moved to a
new plate, rinsed gently with PBS, and media was changed (1:1 ECM to PCM1 or
media
cm
). ECM containing Transferrin-TexasRed to a final concentration of 100nM and
was added to the outer chamber representing the luminal side. Samples were taken
from the inner chamber at 2, 6, and 24 hours and measured on a plate reader (Victor
3

131
V*, Perkinelmer #1420-040) (Ex(λ) 596 nm; Em(λ) 615 nm). Average fluorescence in
RFUs across 3 replicate experiments was calculated and plotted for the different cell
types and time points, with error bars = 1 standard error of deviation calculated on
individual averages and their independent standard deviation. At least three biological
replicate experiments were performed for each cell type and/or condition. Images of
cells on transwells were taken by placing transwells with live cells onto glass-bottom
dishes with minimal medium and imaged on a Leica SP8 confocal microscope.

RNA-seq
Cells were collected using 0.05% trypsin-EDTA (cPCs and 1
o
PCs) or accutase (NCCs),
spun down (900 rpm for 3 minutes), washed with PBS, spun again, and put into Trizol
(ThermoFisher, #15596026). RNA was isolated using 0.2 mL chloroform per 1 mL
Trizol, washed with isopropanol and 70% ethanol, and resuspended in RNAse-free
water. Samples were run on a 2% agarose gel to check for successful RNA isolation
and quantified with the Qubit RNA HS assay kit (ThermoFisher, #Q32852). Libraries for
sequencing were made using the KAPA stranded mRNA-seq kit (KAPA Biosystems)
with capture beads for poly-A RNA isolation starting from 1 µg total RNA. Samples were
quantified, and library quality checked and sequenced using HiSeq2000 (DNA Link
USA, Inc) to obtain at least 20 million reads per sample.

RNA Analysis
RNA-seq analysis was done using Partek flow suite and USC high performance
computational nodes. Fastq sequencing files obtained from DNA link were quality
132
checked, trimmed to remove index sequences, and then mapped to the genome and
transcriptome of the hg19 build of the human genome using star aligner at standard
settings. > 85% of reads uniquely mapped to the human genome in each sample. Fastq
files downloaded from encode and/or gene expression omnibus were also processed
simultaneously using the identical settings in a batch analysis mode. The absolute
transcript abundance was quantified using Partek E/M annotation model. The transcript
levels were normalized either to the upper quartile or to a set of predetermined
housekeeping genes. Subsequent results were not significantly different if either of the
two datasets were used for further analysis. All transcripts with a geometric mean <1
were filtered out. Complete table with raw read counts and normalized FPKM values is
provided via the GEO link GSE104141. Post normalization variation among a selected
group of samples was estimated by visualizing principal components. PCA in Partek
Flow is done using correlation matrix to find the components. All the features are
standardized to a mean of 0 and standard deviation of 1. The results are presented as a
scatterplot, with each dot on the plot being a sample, while the axes represent the PCs
and the axes values correspond to the respective PC values. By default, the first three
PCs are shown on the X-, Y-, and Z-axis respectively, with the information content of an
individual PC in the parenthesis. Pairwise comparisons of biological replicate datasets
were then done using built in differential expression tools. The datasets were further
filtered to remove transcripts with an average value of less than 10 read counts at
standard settings or otherwise noted in the text. All analysis was done with assistance
from Dr. Ming Li from the Norris medical library bioinformatics center.

133
Calculating Cell Circularity
Circularity was calculated as a ratio of two perpendicular lengths of the cell, where the
first length is the longest measured length and the second is its perpendicular diameter.
Thus 1 = perfect circle. Cells from at least five random images from three biological
replicate experiments per cell type per time point were measured. The average
circularity per time point for each representative experiment was plotted with error bars
= 1 standard error of deviation calculated on individual averages and their independent
standard deviation.

Cell Counts
Immunofluorescence was quantified by taking confocal images of at least 5 random
fields per sample, with at least three biological replicates per cell line per IF. For
comparative analysis, all IFs for different cell lines were performed simultaneously in
identical conditions, and all images were collected at the same settings of exposure
collection and threshold. Total cell number was counted with the help of Fiji tools using
nuclear staining for DAPI and expression of specific proteins was counted using the
corresponding channel at a set threshold. The average of the 5 fields for each replicate
experiment was plotted with error bars = 1 standard error of deviation calculated on
individual averages and their independent standard deviation.

Generation and Infection of Virus
To make pgk-gfp or pgk-mCherry viruses, mid-log phase 293T cells were used,
passaged every 3-4 days and grown in DMEM with 10% FBS. 293Ts were co-
134
transfected with lentiviral vector along with VSVG expression vector and gagpol
expression vector (pMGD2 and pCMV R8.74). After 12-16 hours, the media was
discarded and replaced with serum-free Ultraculture media (Lonza, #12-725F) with 2X
glutamine for virus harvesting. Viral supernatant was collected the next day and
concentrated by spinning at 15,000 rpm at 4
o
C overnight in an ultracentrifuge. Virus was
resuspended in remaining media by shaking on ice for 2-12 hours. To infect cPCs or
1
o
PCs, cells were collected with 0.05% trypsin-EDTA and placed in 200 μl media. 50-
150 μl of viral supernatant was added along with polybrene to a final concentration of 10
μg/mL. The cells were incubated at 37
o
C for 45 minutes, then transferred to a new
plate.

Scratch Assays
Cells were plated and grown to >75% confluence. A cell scraper was used to make a
thin scratch along the diameter of the well. The edge of the scratch was marked on the
outside of the plate and cells were imaged every 24 hr starting immediately after the
scratch (at 0 hr) until the scratch was filled in or up to 72 hr, whichever occurred first.
Area of the scratch was measured, and the amount of the area filled in with cells was
calculated at each time point using Fiji tools for surface area calculation. Are of the
scratch was set at 100% and percentage of this area filled in at each timepoint was
calculated. At least 3 biological replicate experiments were performed for each cell line
and/or condition, and averages of replicates were plotted with error bars = 1 standard
error of deviation calculated on individual averages and their independent standard
135
deviation. Phase images in figures are of live cells at stated time point and are
representative images of the biological replicates.

Contractility Assay
Experiments were based on previous work done by Chow et al., 2007. Cells were either
sparsely plated in order to see separate individual cells or grown to >60% confluence to
assess larger numbers of cells. At time of assay, media was changed to HBSS (with
calcium and magnesium) for 5 minutes and cells were imaged to note the resting state.
Media was then changed to HBSS + 75 mM KCl for 2 minutes, then returned to HBSS.
Cells were imaged every 20 seconds for 10 minutes. The experiment was repeated with
at least three wells of cells of each cell type. Number of cells in each field that changed
their total extension by at least 10% of their original maximal length were counted as
positively contracting in response to KCl. For control, cPCs in HBSS with no addition of
KCl were imaged every 20 seconds for 10 minutes to establish baseline fluctuations in
length. To calculate amount of contraction, 50 cells across multiple fields were
measured along the longest axis at 0 seconds and again at 600 seconds using Fiji. The
change in size was calculated by taking the difference between the length at the two
timepoints, and averages of single fields were then plotted using Prism, with error bars
= 1 standard error of deviation calculated on individual averages and their independent
standard deviation.

136
EC-PC Co-culturing
Human Brain Microvascular Endothelial cells (HBMECs) were grown on fibronectin-
coated plates using endothelial cell medium. To co-culture with pericytes, ECs and
cPCs or 1
o
PCs were plated on fibronectin coated plates or glass coverslips at a ratio of
2:1 (EC to PC) in endothelial cell medium. For comparison, cocultures were also grown
in pericyte medium only or v/v endothelial cell medium and pericyte medium.

Quantifying Cumulative Fluorescence Signal
Maximum projection confocal images of cells were input into Fiji and thresholded using
identical settings, excluding saturated pixels. The images were then analyzed for
integrated density, calculating the product of the area of pixels and the mean intensity
value of these pixels. Averages of these values across multiple fields and replicates
were plotted as cumulative fluorescence signal, analyzing differences in area of
fluorescence and intensity of fluorescence across experimental samples.

Tube Assay on Geltrex
HBMEC were plated at 1x10
4
cells per cm
3
on a matrix of 12X growth-factor-reduced
geltrex and incubated at 37
o
C for 30 minutes. Pericytes were then added to the ECs in
a PC-EC ratio or 1:1 or 1:3. In EC only samples, no PC were added. All cell
combinations were grown in ECM only. Cultures were grown for 5 days and were
imaged daily using a Leica DMi8 inverted fluorescent microscope. Tube area was
calculated by measuring the area of a field covered with cells using Fiji automated
surface area determination tools. Tube branching points were calculated by counting
137
the number of branching points in a field of 0.5 cm
2
. Five random fields per sample were
imaged, and at least three biological replicate experiments were performed for each cell
type and/or condition. For tube area and branching plots, error bars = 1 standard error
of deviation calculated on individual averages and their independent standard deviation.

Frogs
Adult female frogs were super-ovulated to induce egg laying the night before, eggs were
collected and fertilized, and embryos were grown until tadpole stage (stage 40). Half of
the tadpoles were kept as untreated controls while the other half were moved to ⅓ MBS
with 10 nM sunitinib malate for 48 hours. Tadpoles were tested for blood vessel leakage
using a pinch test (tails were pinched gently with forceps and examined for bleeding). Of
the tadpoles showing effects of the drug treatment, half were kept as uninjected controls
while the other half were anesthetized and injected directly into the heart with 10
3
pgk-
mCherry H9-cPCs or pgk-mCherry iMF-cPCs. An equal number of the untreated
controls were also injected with cells in the same manner as untreated injected controls.
All tadpoles were kept for up to 2 weeks post injection, with twice daily changes of
tadpole medium and live imaged periodically. Before imaging, embryos were
anesthetized with MS222 and injected directly into the heart/aorta with 1 μl of 50 ng/mL
70 kDa Dextran-488 using a fine pulled glass capillary needle to visualize the blood
vessels. Experiments were repeated at least 3 times for both cell types from individual
batches of embryos.

138
Rat PC Isolation
Neonatal (5 day old) or adult rat brains were obtained, washed thoroughly with PBS,
and dissociated using a dounce homogenizer. Cell solutions were then incubated with
collagenase for 30 min at 37°C. Cells were spun down at 1200 rpm for 5 min, further
dissociated by hand with a pipette, spun and washed with PBS, and placed in pericyte
medium for culture. Cells were cultured and passaged with 0.05% trypsin-EDTA until
uniform populations were obtained (roughly 2 weeks of culture). Cells were analyzed for
pericyte identity using immunofluorescence for pericyte markers, cell morphology, and
cell function.

Western Blot
Cells were grown for 4 days, changing media once every 24 hours. After 4 days, cells
were lysed using RIPA buffer (50 mM Tris HCl pH 8, 150 mM NaCl, 1% NP-40, 0.5%
Sodium Deoxycholate, 0.1% SDS). Lysed cells were spun at 14,000 x g for 15 min, and
total protein was quantified using the Micro BCA kit (Pierce, #23235). Equal amounts of
protein were loaded and run on an acrylamide SDS gel. Gels were transferred onto
blotting paper, blocked with blocking buffer (Roche, #11921673001) for 15 min at 37°C,
and probed with primary antibody overnight at 4°C. Blots were washed with TBST,
probed with secondary antibody for 45 min at room temperature, washed with TBST,
and developed with Lumi-light substrate (Roche, #12015200001). GAPDH was used as
a loading control. At least two biological replicate experiments were performed for each
cell type.

139
Culture with ANGPT1
All assays assessing ANGPT1 rescue were set up as explained above. Controls of iFF-
cPC with and without ANGPT1 and iPSEN1*-cPC and/or 044 1°PC were always used.
The experimental group was cultured the same as the controls in the assays, but with
50 ng/mL ANGPT1 added to the medium daily. Cells were grown for at least 2 days and
up to 6 days, with ANGPT1 added every 24hr to the medium in the experimental group.
Experiments were assessed at the end timepoint as described above.

ANGPT1 Knockdown with siRNA and Inhibitor
Controls of iFF-cPC untreated and treated with non-targeting control siRNA were
always used. The experimental group was cultured the same as the controls, but with 1
µM siANGPT1 or 250 nM Tie2 Kinase Inhibitor. siRNA and inhibitor were added every 6
or 12hr for the duration of the experiment. Cells were then fixed, and IF was performed.

Cell Tracking
Movies of live cells were taken using the Leica SP8 confocal microscope. After imaging,
the Leica software was used to make motility tracks for each individual cell for the
extent of the movie.

Quantification and Statistical Analysis

Statistical details can be found in the figures, figure legends, and results section. To
compare two values, we used unpaired student’s t-test, and to compare three or more
140
we used one-way ANOVA, via Prism with standard settings. For all cell staining
quantifications, N > 100, where N = single cells, and a minimum of 5 random fields from
one experiment from a minimum of three replicate experiments were used to obtain N.
For quantification of area in scratch assays and tube assays, a minimum of 5 replicates
(scratch assays) or 5 fields (tube assays) from at least three replicate experiments was
used to calculate area. For quantification of contractility, N > 50, where N = single cell,
and a minimum of 4 repeats of the experiment were used. Error bars were drawn for all
graphs as ± 1 standard error of deviation calculated on individual averages and their
independent standard deviation. Significance was defined as p-value ≤ 0.05, as
calculated via unpaired student’s t-test or one-way ANOVA. No data points were
excluded from any calculations or graphs. Images shown in all figures are
representative images of each experiment, and all replicates and repeat experiments
were included in calculations.

Data and Software Availability

Data
Raw and analyzed data from this paper: GSE104141

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE104141
Zhang et al. 2017 raw data: GSE93511
https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE93511
He et al. 2016 raw data: GSE75668
https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE75668
141
Placental Pericyte raw data from ENCODE: GSE78604
https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE78604

Software
Partek Flow for analysis of raw sequencing data files
http://www.partek.com/partekflow
Fiji for biological image analysis
http://fiji.sc/
Prism for graphing and statistical analysis
https://www.graphpad.com/scientific-software/prism/
Homer for sequencing analysis
http://homer.ucsd.edu/homer/

142
REFERENCES
1. Aramaki M, Kimura T, Udaka T, Kosaki R, Mitsuhashi T, Okada Y, Takahashi T,
Kosaki K. (2007). Embryonic expression profile of chicken CHD7, the ortholog of
the causative gene for CHARGE syndrome. Birth Defects Res A Clin Mol Teratol,
79(1), 50-57.
2. Armulik A, Genove G, Mae M, Nisancicoglu MH, Wallgard E, Niaudet C, He L,
Norlin J, Lindblom P, Strittmatter K, et al. (2010). Pericytes regulate the blood-
brain barrier. Nature, 468, 557-561.
3. Armulik A, Genove G, Betsholtz C. (2011). Pericytes: developmental,
physiological, and pathological perspectives, problems, and promises. Dev Cell,
21, 193-215.
4. Arvanitakis Z, Capuano AW, Leurgans SE, Bennett DA, Schneider JA. (2016).
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.
5. Asad Z, Pandey A, Babu A, Sun Y, Shevade K, Kapoor S, Ullah I, Ranjan S,
Scaria V, Bajpai R, Sachidanandan C. (2016). Rescue of neural crest-derived
phenotypes in a zebrafish CHARGE model by Sox10 downregulation. Hum Mol
Gen, 25, 3539-3554.
6. Asad Z, Sachidanandan C. (2019). Chemical screens in a zebrafish model of
CHARGE syndrome identifies small molecules that ameliorate disease-like
phenotypes in embryo. Eur J Med Gen, doi.org/10.1016/j.ejmg.2019.04.018.
143
7. Asahina K, Zhou B, Pu Wu, Tsukamoto H. (2011). Septum transversum-derived
mesothelium gives rise to hepatic stellate cells and perivascular mesenchymal
cells in developing mouse liver. Hepatology, 53(3), 983-995.
8. Bajpai R, Lesperance J, Kim M, Terskikh AV. (2008). Efficient propagation of
single cell accutase-dissociated human embryonic stem cells. Mol Reprod Dev,
75(5), 818-827.
9. Bajpai R, Chen DA, Rada-Iglesias A, Zhang J, Xiong Y, Helms J, Chang CP,
Zhao Y, Swigut T,Wysocka J. (2010). CHD7 cooperates with PBAF to control
multipotent neural crest formation. Nature, 463(7283), 958-962.
10. Bell RD, Winkler EA, Sagare AP, Singh I, LaRue B, Deane R, Zlokovic BV.
(2010). Pericytes control key neurovascular functions and neuronal phenotype in
the adult brain and during brain aging. Neuron, 68, 409-427.
11. Bergwerff M, Verberne ME, DeRuiter MC, Poelmann RE, Gittenberger-de Groot
AC. (1998). Neural crest cell contribution to the developing circulatory system:
implications for vascular morphology? Circ Res, 82(2), 221-231.
12. Birbrair A, Zhang T, Files DC, Mannava S, Smith T, Wang ZM, Messi ML, Mintz
A, Delbono O. (2014). Type-1 pericytes accumulate after tissue injury and
produce collagen in an organ-dependent manner. Stem Cell Res Ther, 5, 122.
13. Bouazoune K, Kingston RE. (2012). Chromatin remodeling by the CHD7 protein
is impaired by mutations that cause human developmental disorders. Proc Natl
Acad Sci USA, 109(47), 19238-19243.
144
14. Cai CL, Martin JC, Sun Y, Cui L, Wang L, Ouyang K, Yang L, Bu L, Liang X,
Zhang X, Stallcup WB, Denton CP, McCulloch A, Chen J, Evans SM. (2008). A
myocardial lineage derives from Tbx18 epicardial cells. Nature, 454, 104-108.
15. Chin CJ, Li S, Corselli M, Casero D, Zhu Y, He CB, Hardy R, Peault B, Crooks
GM. (2018). Transcriptionally and functionally distinct mesenchymal
subpopulations are generated from human pluripotent stem cells. Stem Cell Rep,
10(2), 436-446.
16. Chow N, Bell RD, Deane R, Streb JW, Chen J, Brooks A, Van Nostrand W,
Miano JM, Zlokovic BV. (2007). Serum response factor and myocardin mediate
arterial hypercontractility and cerebral blood flow dysregulation in Alzheimer’s
phenotype. Proc Natl Acad Sci USA, 104(3), 823-828.
17. Cohen RM, Rezai-Zadeh K, Weitz TM, Rentsendorj A, Gate D, Spivak I, Bholat
Y, Vasilevko V, Glabe CG, Breunig JJ, et al. (2013). A transgenic Alzheimer rat
with plaques, tau pathology, behavioral impairment, oligomeric abeta, and frank
neuronal loss. J Neurosci, 33, 6245-6256.
18. Dalkara T, Gursoy-Ozdemir Y, Yemisci M. (2011). Brain microvascular pericytes
in health and disease. Acta Neuropathol, 122, 1-9.
19. Daneman R, Zhou L, Kebede AA, Barres BA. (2010). Pericytes are required for
blood-brain barrier integrity during embryogenesis. Nature, 468, 562-566.
20. Dar A, Domev H, Ben-Yosef O, Tzukerman M, Zeevi-Levin N, Novak A,
Germanguz I, Amit M, Itskovitz-Eldor J. (2012). Multipotent vasculogenic
pericytes from human pluripotent stem cells promote recovery of murine ischemic
limb. Circulation, 125, 87-99.
145
21. Diaz-Flores L, Gutierrez R, Madrid JF, Varela H, Valladares F, Acosta E, Martin-
Vasallo P, Diaz-Flores L Jr. (2009). Pericytes. Morphofunction, interactions and
pathology in a quiescent and activated mesenchymal cell niche. Histol
Histopathol, 24(7), 909-969.
22. Etchevers HC, Vincent C, Le Douarin NM, Couly GF. (2001). The cephalic neural
crest provides pericytes and smooth muscle cells to all blood vessels of the face
and forebrain. Development, 128, 1059-1068.
23. Faal T, Phan DTT, Davtyan H, Scarfone VM, Varady E, Blurton-Jones M,
Hughes CCW, Inlay MA. (2019). Induction of mesoderm and neural crest-derived
pericytes from human pluripotent stem cells to study blood-brain barrier
interactions. Stem Cell Reports, 12(3), doi: 10.1016/j.stemcr.2019.01.005.
24. Griffin C, Bajpai R. (2017). Neural crest-derived human cranial pericytes model
primary forebrain pericytes and predict disease-specific cranial vasculature
defects. SSRN, http://dx.doi.org/10.2139/ssrn.3189103.
25. He L, Vanlandewijck M, Raschperger E, Andaloussi Mae M, Jung B, Lebouvier T,
Ando K, Hofmann J, Keller A, Betsholtz C. (2016). Analysis of the brain mural cell
transcriptome. Sci Rep, 6, 35108.
26. Kim JM, Hong KS, Song WK, Bae D, Hwang IK, Kim JS, Chung HM. (2016).
Perivascular progenitor cells derived from human embryonic stem cells exhibit
functional characteristics of pericytes and improve the retinal vasculature in a
rodent model of diabetic retinopathy. Stem Cells Transl Med, 5, 1268-1276.
146
27. Kisler K, Nelson AR, Rege SV, Ramanathan A, Wang Y, Ahuja A, Lazic D, Tsai
PS, Zhao Z, Zhou Y, et al. (2017). Pericyte degeneration leads to neurovascular
uncoupling and limits oxygen supply to brain. Nat Neurosci, 20, 406-416.
28. Korn J, Christ B, Kurz H. (2002). Neuroectodermal origin of brain pericytes and
vascular smooth muscle cells. J Comp Neurol, 442, 78-88.
29. Kumar DK, Eimer WA, Tanzi RE, Moir RD. (2016). Alzheimer’s disease: the
potential therapeutic role of the natural antibiotic amyloid-beta peptide.
Neurodegener Dis Manag, 6, 345-348.
30. Kumar A, D’Souza SS, Moskvin QV, Toh H, Wang B, Zhang J, Swanson S, Guo
LW, Thomson JA, Slukvin II. (2017). Specification and diversification of pericytes
and smooth muscle cells from mesenchymoangioblasts. Cell Rep, 19, 1902-
1916.
31. Kusuma S, Shen YI, Hanjaya-Putra D, Mali P, Cheng L, Gerecht S. (2013). Self-
organized vascular networks from human pluripotent stem cells in a synthetic
matrix. Proc Natl Aca Sci USA, 110(31), 12601-12606.
32. Lebrin F, Deckers M, Bertolino P, Ten Dijke P. (2005). TGF-beta receptor
function in the endothelium. Cardiovasc Res, 65, 599-608.
33. Lindahl P, Johansson BR, Leveen P, Betsholtz C. (1997). Pericyte loss and
microaneurysm formation in PDGF-B-deficient mice. Science 227, 242-245.
34. Mathiisen TM, Lehre KP, Danbolt NC, Ottersen OP. (2010). The perivascular
astroglial sheath provides a complete covering of the brain microvessels: an
electron microscopic 3D reconstruction. Glia, 58(9), 1094-1103.
147
35. Montagne A, Barnes SR, Sweeney MD, Halliday MR, Sagare AP, Zhao Z, Toga
AW, Jacobs RE, Liu CY, Amezcua L, et al. (2015). Blood-brain barrier
breakdown in the aging human hippocampus. Neuron, 85, 296-302.
36. Nieuwkoop and Faber (1994). Normal Table of Xenopus laevis (Daudin). Garland
Publishing Inc., New York ISBN 0-8153-1896-0
37. Okuno H, Mihara FR, Ohta S, Fukuda K, Kurosawa K, Akamatsu W, Sanosaka
T, Kohyama J, Hayashi K, Nakajima K, Takahashi T, Wysocka J, Kosaki K,
Okano H. (2017). CHARGE syndrome modeling using patient-iPSCs reveals
defective migration of neural crest cells harboring CHD7 mutations. eLife, 6,
e21114.
38. Olson LE, Soriano P. (2011). PDGFRbeta signaling regulates mural cell plasticity
and inhibits fat development. Dev Cell, 20, 815-826.
39. Orlova VV, Drabsch Y, Freund C, Petrus-Reurer, van den Hil FE, Muenthaisong
S, Dijke PT, Mummery CL. (2013). Functionality of endothelial cells and pericytes
from human pluripotent stem cells demonstrated in cultured vascular plexus and
zebrafish xenografts. Arterioscler Thromb Vasc Biol, 34(1), 177-186.
40. Quaegebeur A, Segura I, Carmeliet P. (2010). Pericytes: blood-brain barrier
safeguards agains neurodegeneration? Neuron, 68(3), 321-323.
41. Quastel JH, Wheatley AH. (1932). Oxidations by the brain. Biochem J, 26, 725-
744.
42. Que J, Wilm B, Hasegawa H, Wang F, Bader D, Hogan BL. (2008). Mesothelium
contributes to vascular smooth muscle and mesenchyme during lung
development. Proc Natl Acad Sci USA, 105, 16626-16630.
148
43. Rada-Iglesias A, Bajpai R, Prescott S, Brugmann SA, Swigut T, Wysocka J.
(2012). Epigenomic annotation of enhancers predicts transcriptional regulators of
human neural crest. Cell Stem Cell, 11, 633-648.
44. Rouget CB. (1879). Sciences, 88, 916-918.
45. Sengillo JD, Winkler EA, Walker CT, Sullivan JS, Johnson M, Zlokovic BV.
(2013). Deficiency in mural vascular cells coincides with blood-brain barrier
disruption in Alzheimer’s disease. Brain Pathol, 23, 303-310.
46. Siebert JR, Graham JM, MacDonald C. (1985). Pathologic features of the
CHARGE association: support for involvement of the neural crest. Teratology,
31, 331-336.
47. Sims DE. (1986). The pericyte – a review. Tissue Cell, 18(2), 153-174.
48. Stebbins MJ, Gastfriend BD, Canfield SG, Lee MS, Richards D, Faubion MG, Li
WJ, Daneman R, Palecek SP, Shusta EV. (2019). Human pluripotent stem cell-
derived brain pericyte-like cells induce blood-brain barrier properties. Sci Adv, 5,
no. eaau7375.
49. Sweeney MD, Ayyadurai S, Zlokovic BV. (2016). Pericytes of the neurovascular
unit: key functions and signaling pathways. Nat Neurosci, 19, 771-783.
50. Thurston G, Suri C, Smith K, McClain J, Sato TN, Yancopoulos GD, McDonald
DM. (1999). Leakage-resistant blood vessels in mice transgenically
overexpressing angiopoietin-1. Science, 286, 2511-2514.
51. Uemura A, Ogawa M, Hirashima M, Fujiwara T, Koyama S, Takagi H, Honda Y,
Wiegand SJ, Yancopoulos GD, Nishikawa S. (2002). Recombinant angiopoietin-
149
1 restores higher-order architecture of growing blood vessels in mice in the
absence of mural cells. J Clin Invest, 110, 1619-1628.
52. Van Geest RJ, Klaassen I, Vogels IM, Van Noorden CJ, Schlingemann RO.
(2010). Differential TGF-beta signaling in retinal vascular cells: a role in diabetic
retinopathy? Invest Ophthalmol Vis Sci, 51, 1857-1865.
53. Vanlandewijck M, He L, Mae MA, Andrae J, Ando K, Del Gaudio F, Nahar K,
Lebouvier T, Lavina B, Gouveia L, et al. (2018). A molecular atlas of cell types
and zonation in the brain vasculature. Nature, 554, 475-480.
54. Vissers LE, van Ravenswaaij CM, Admiraal R, Hurst JA, de Vries BB, Janssen
IM, van der Vliet WA, Huys EH, de Jong PJ, Hamel BC, Schoenmakers EF,
Brunner HG, Veltman JA, van Kessel AG. (2004). Mutations in a new member of
the chromodomain gene family cause CHARGE syndrome. Nature Genetics, 36,
955-957.
55. Wilm B, Ipenberg A, Hastie ND, Burch JB, Bader DM. (2005). The serosal
mesothelium is a major source of smooth muscle cells of the gut vasculature.
Development, 132, 5317-5328.
56. Winkler EA, Bell RD, Zlokovic BV. (2010). Pericyte-specific expression of PDGF
beta receptor in mouse models with normal and deficient PDGF beta receptor
signaling. Mol Neurodegener, 5, 32.
57. Winkler EA, Bell RD, Zlokovic BV. (2011). Central nervous system pericytes in
health and disease. Nat Neurosci, 14, 1398-1405.
58. Winkler EA, Sagare AP, Zlokovic BV. (2014). The pericyte: a forgotten cell type
with important implications for Alzheimer’s disease? Brain Pathol, 24, 371-386.
150
59. Xu J. Gong T, Heng BC, Zhang CF. (2017). A systematic review: differentiation
of stem cells into functional pericytes. FASEB, 31(5).
60. Yew B, Nation DA, Alzheimer’s Disease Neuroimaging, I. (2017).
Cerebrovascular resistance: effects on cognitive decline, cortical atrophy, and
progression to dementia. Brain, 140, 1987-2001.
61. Zhou B, Ma Q, Rajagopal S, Wu SM, Domian I, Rivera-Feliciano J, Jiang D, von
Fise A, Ikeda S, Chien KR, Pu WT. (2008). Epicardial progenitors contribute to
the cardiomyocyte lineage in the developing heart. Nature, 454, 109-113.

Abstract (if available)

Abstract Forebrain pericytes are critical players in the blood-brain barrier (BBB). Defects in or loss of functional forebrain pericytes leads to compromised microvessel function and ultimately breakdown of the integrity of the BBB, causing leakage of toxins and pathogens into the brain and aggravating neuroinflammation. Defective blood vessels and leakiness of the BBB has been found to play a part in numerous neurodegenerative diseases, most notably Alzheimer's disease (AD), and tortuous vessels have been detected prior to onset of dementia in AD patients and carriers of AD risk alleles. Despite their importance, little is known about forebrain pericytes and what makes this population of pericytes both able to maintain the BBB and become prone to damages with aging and disease. Utilizing a mathod to generate in vitro-derived cranial pericytes, I have been able to identify a set of characteristics defining forebrain pericytes, as well as identify a set of defects inherent in pericytes in AD and the developmental disorder CHARGE syndrome. I take steps toward understanding the molecular mechanisms underlying the defects in AD, as well as examine the similarities and differences in the defects across types of AD and across AD versus CHARGE. This work will help the field of pericyte biology gain a better understanding of forebrain pericytes, as well as open the door for potential therapeutic avenues in multiple diseases.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Understanding the role of APP and DYRK1A in human brain pericytes

PDF

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

PDF

Blood-brain barrier pathophysiology in cognitive impairment and injury

PDF

The role of vascular dysfunction in cognitive impairment

PDF

Functional study of C9ORF72 and its implication in the pathogenesis of amyotrophic lateral sclerosis

PDF

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

PDF

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

PDF

Derivation, expansion and characterization of human hippocampal primordial cells from normal and diseased iPSCs

PDF

Identification of therapeutic targets for neurons and microglia in amyotrophic lateral sclerosis

PDF

Functional compensation between hematopoietic stem cell clones in vivo

PDF

Genetics and the environment: evaluating the role of noncoding RNA in autism spectrum disorder

PDF

The autism-associated gene SYNGAP1 regulates human cortical neurogenesis

PDF

A novel role for post-migratory neural crest cells in cardiac outflow tract alignment: insights into the molecular basis of concomitant congenital cardiac and cleft malformations

PDF

Elucidating the functional role of CHD7 associated nuclear PDH complex and other associated proteins on neural crest development

PDF

Signaling networks in complex brain disorders

PDF

Understanding human nephrogenesis and scaling synthesis of organoids facilitate modeling of kidney development and disease

PDF

An iPSC-based biomarker strategy to identify neuroregenerative responders to allopregnanolone

PDF

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

PDF

Opposing function of chd7 and phf6 in zebrafish craniofacial development

PDF

Transcriptomic maturation of developing human cone precursors in fetal and 3D hESC-derived tissues

Asset Metadata

Creator Griffin, Casey (author)

Core Title Neural crest-derived cranial pericytes and their dysfunction in disease

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Development, Stem Cells and Regenerative Medicine

Publication Date 12/04/2019

Defense Date 07/23/2019

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

Tag Alzheimer's disease,blood-brain barrier,CHARGE syndrome,endothelial cells,neural crest,OAI-PMH Harvest,pericyte,stem cell

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Crump, Gage (committee chair), Bajpai, Ruchi (committee member), Ichida, Justin (committee member), Zlokovic, Berislav (committee member)

Creator Email caseyg523@comcast.net,crgriffi@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-241853

Unique identifier UC11674671

Identifier etd-GriffinCas-7984.pdf (filename),usctheses-c89-241853 (legacy record id)

Legacy Identifier etd-GriffinCas-7984.pdf

Dmrecord 241853

Document Type Dissertation

Rights Griffin, Casey

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