PLACENTAL EXTRACELLULAR VESICLES IN MURINE PREGNANCY 

 
B(cid:92) 
 

Sea(cid:81) La(cid:80)-Vie(cid:81) Ng(cid:88)(cid:92)e(cid:81) 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

A DISSERTATION 

 
 

S(cid:88)b(cid:80)i(cid:87)(cid:87)ed (cid:87)(cid:82) 

Michiga(cid:81) S(cid:87)a(cid:87)e U(cid:81)i(cid:89)e(cid:85)(cid:86)i(cid:87)(cid:92) 

i(cid:81) (cid:83)a(cid:85)(cid:87)ia(cid:79) f(cid:88)(cid:79)fi(cid:79)(cid:79)(cid:80)e(cid:81)(cid:87) (cid:82)f (cid:87)he (cid:85)e(cid:84)(cid:88)i(cid:85)e(cid:80)e(cid:81)(cid:87)(cid:86)  

f(cid:82)(cid:85) (cid:87)he deg(cid:85)ee (cid:82)f 

 

Ce(cid:79)(cid:79) a(cid:81)d M(cid:82)(cid:79)ec(cid:88)(cid:79)a(cid:85) Bi(cid:82)(cid:79)(cid:82)g(cid:92)(cid:177)E(cid:81)(cid:89)i(cid:85)(cid:82)(cid:81)(cid:80)e(cid:81)(cid:87)a(cid:79) T(cid:82)(cid:91)ic(cid:82)(cid:79)(cid:82)g(cid:92)(cid:177)D(cid:82)c(cid:87)(cid:82)(cid:85) (cid:82)f Phi(cid:79)(cid:82)(cid:86)(cid:82)(cid:83)h(cid:92) 

 

2020 

ABSTRACT

PLACENTAL EXTRACELLULAR VESICLE TRAFFICKING IN MURINE PREGNANCY

By

Sean Lam-Vien Nguyen

The placenta is a critical extrafetal disc-shaped organ responsible for the growth and development
of the fetus during pregnancy. Anatomically, the placenta is at the interface of maternal and fetal
tissues and provides a unique opportunity for cellular communication between the mother and
fetus. Extracellular vesicles (EVs) are (40-150nm) membrane-enclosed nanostructures that contain
RNAs, proteins, and lipids that travel to different parts of the body and serve as a form of intercellular
communication. During pregnancy, EVs are released in high quantities from the placenta and have
been postulated to target multiple maternal cell types, including those of the vascular and immune
systems. However, most studies on pregnancy-associated EVs have used clinical samples and in
vitro models; to date, few studies have taken advantage of murine models in which pregnancy
can be precisely timed and manipulated. The placenta secretes copious amounts of exosomes
into maternal blood, but the trafficking of placental EVs in vivo and mechanisms mediate their
trafficking remains understudied. In this dissertation, we develop a computational software package
tidyNano to process EV quantification data across murine gestation and characterize alterations
in EV concentration during healthy pregnancy and during inflammation-associated preterm birth.
Next, we demonstrate specific, preferential trafficking of placental EVs to interstitial macrophages
in maternal lungs as well as Kupffer cells in the liver, with outer membrane integrins 
3
1, 
5
1,
and 
V
3 mediating trafficking to these tissues. Finally, we establish an in vivo murine model
for identifying fetal cells and placental EVs and propose a framework for studying maternal-fetal
interactions without experimental manipulation. Collectively, the work described in this dissertation
provides a computational framework to analyze nanoparticle data, identifies how placental EVs
target specific maternal tissues, and provides a novel mouse model to study placental EV trafficking
in vivo.

Copyright by
SEAN LAM-VIEN NGUYEN
2020

ACKNOWLEDGEMENTS

First and foremost, thank you to my thesis advisor Dr. Margaret Petroff for your mentorship and
guidance. These past few years in your lab have been transformative in my development as a
scientist and researcher. Your eagerness to support my research project ideas, encouragement to
attend conferences, learn programming, and allowing me to help run the Frontiers in Reproduction
course have been some of my fondest memories during my time in graduate school. I have truly
enjoyed my time in your lab.

I want to thank my guidance committee (Dr. Asgerally Fazleabas, Dr. Karl Olson, Dr. Cheryl
Rockwell, Dr. Karen Racicot, and Dr. Amy Ralston) for their thoughtful discussion during
my committee meetings. Your suggestions helped me improve my experimental approaches and
trained me to become a better scientist. I also would like to thank the environmental and integrative
toxicological sciences program and the integrative pharmacological sciences training program for
funding part of my graduate education and extracurricular development.

Thank you to my lab mates, past and present, for your continuous support throughout my
journey in the lab. Jacob Greenberg and Benjamin Collaer, thank you for helping me troubleshoot
all my experiments. Your contributions and findings helped establish the groundwork for all of
my experiments in my dissertation. Geoffrey Grzesiak and Sarika Kshirsagar, thank you for being
instrumental for making the lab run smoothly and being there for me during my journey’s highs and
lows. Cole McCutcheon, few individuals can understand the pain and joy of researching vesicles.
Thank you for your relentless optimism; your enthusiasm made the lab environment very enjoyable.
Dr. Soo Hyun Ahn, thank you for your help and insight in developing my project, especially towards
the latter half of my dissertation. I will always remember all those long days working on the big
experiments and the fruitful discussions about research and our experience working in the lab.

I want to thank my family for their support throughout my life, including pursuing a graduate
degree. Mom, dad, and Justin, you have been there for me through the thick and thin. I’m glad to
say that my graduate education journey is finally coming to a close! Finally, thank you, Chelsea

iv

Everson J.D., for being a source of stability and calm throughout my undergraduate and graduate
education. Thank you for putting up with me when I was tired and frustrated with my research and
celebrating the big successes along the way. Your words of encouragement throughout my journey
have been immensely beneficial to my graduate success.

v

TABLE OF CONTENTS

.

.

.

.

.

.

ix

1.2.1

1.3 EXTRACELLULAR VESICLE BIOLOGY OVERVIEW . . . . . . . . . . . . .

LIST OF FIGURES . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CHAPTER 1 GENERAL INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . .
1
1
1.1 PLACENTA OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2
1.2 PLACENTA STRUCTURE AND ANATOMY . . . . . . . . . . . . . . . . . . .
5
Immune Adaptations During Pregnancy . . . . . . . . . . . . . . . . . . .
6
.
9
1.3.1 EV function - physiology and disease . . . . . . . . . . . . . . . . . . . .
1.3.2 EV content
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3.3 EV trafficking and cellular uptake . . . . . . . . . . . . . . . . . . . . . . 11
1.3.4 EV biogenesis
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3.5 Methods of EV isolation, purification, and validation . . . . . . . . . . . . 13
1.4 MODELS OF STUDYING PLACENTAL EVS . . . . . . . . . . . . . . . . . . . 15
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
pEVs from blood .
pEVs from primary tissue . . . . . . . . . . . . . . . . . . . . . . . . . . 16
pEVs from cell lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.5 PLACENTAL EV OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.6 DISSERTATION OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

1.4.1
1.4.2
1.4.3

.

.

.

.

.
.

.
.

. .

.
INTRODUCTION .

CHAPTER 2 QUANTIFYING MURINE PLACENTAL EXTRACELLULAR VESI-
CLES ACROSS GESTATION AND IN PRETERM BIRTH DATA WITH
TIDYNANO: A COMPUTATIONAL FRAMEWORK FOR ANALYZ-
ING AND VISUALIZING NANOPARTICLE DATA IN R . . . . . . . . . . 21
2.1 ABSTRACT .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.3 MATERIALS AND METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.3.1 Mice and treatments
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.3.2 Exosome isolation and validation . . . . . . . . . . . . . . . . . . . . . . 25
2.3.3 NanoSight analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3.4
. . . . . . . . . . . . . . . . . . . . . . . 25
2.3.5 Data import and cleaning . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3.6 Data summarization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
. . . . . . . . . . . . . . . . . . 27
2.3.7 Data visualization and statistical analysis
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.4.1 Data import and cleaning . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.4.2 Data summarization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.3 Visualization . .
. 31
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.4.4
Statistical analysis
2.4.5 Effects of gestation day and inflammation on circulating EV concentrations
34
2.5 DISCUSSION .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.6 ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

tidyNano software development

2.4 RESULTS . .

.

.

.

.

.

.

.

.

.

.

.

.

vi

CHAPTER 3

.

.
.

.
.

. .

.
INTRODUCTION .

INTEGRINS MEDIATE PLACENTAL EXTRACELLULAR VESICLE
TRAFFICKING TO MATERNAL LUNG AND LIVER IN VIVO . . . . . . 39
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.1 ABSTRACT .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.2
3.3 MATERIALS AND METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.3.1 Animal experiments
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.3.2 Extracellular vesicle isolation . . . . . . . . . . . . . . . . . . . . . . . . 43
3.3.3
Fluorescent EV labeling . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.3.4 Nanoparticle tracking analysis . . . . . . . . . . . . . . . . . . . . . . . . 44
3.3.5 Western blot analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.3.6 Transmission electron microscopy . . . . . . . . . . . . . . . . . . . . . . 45
Proteinase K and inhibitory peptide treatment . . . . . . . . . . . . . . . . 45
3.3.7
3.3.8
Fluorescence/immunofluorescence microscopy and quantification . . . . . 45
3.3.9 Tissue clearing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3.10 Flow cytometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3.11 Whole organ imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.3.12 Statistical analysis
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Pregnant plasma EVs traffic to the lung . . . . . . . . . . . . . . . . . . . 48
Placental EVs traffic to lung interstitial macrophages in vivo. . . . . . . . . 51
. 54
Integrin 
3
1 allows trafficking of pEVs to the lung interstitium. . . . . . . 57
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

3.4.1
3.4.2
3.4.3 Outer membrane proteins influence pEV trafficking to the lung and liver.
3.4.4

3.5 DISCUSSION .

3.4 RESULTS . .

.

.

.

.

.

.

.

.

.

.

.

.

.

.
.

.
.

. .

. .

CHAPTER 4 A REPORTER MODEL FOR IN VIVO TRACKING OF PLACENTAL

.
INTRODUCTION .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

EXTRACELLULAR VESICLES IN MURINE PREGNANCY . . . . . . . 63
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.1 ABSTRACT .
4.2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.3 MATERIALS AND METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
. 66
4.3.1 Mouse studies . .
Placental EV isolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.3.2
. 67
4.3.3 Western blot
4.3.4 Tissue processing, immunostaining, and confocal microscopy . . . . . . . 68
4.3.5 mT/mG bmDC explant culture . . . . . . . . . . . . . . . . . . . . . . . . 69
4.3.6
Flow cytometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.3.7 Recombination genotyping . . . . . . . . . . . . . . . . . . . . . . . . . . 70
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Identification of placental EVs during pregnancy . . . . . . . . . . . . .
. 70
Fetal EV trafficking to maternal lung in vivo . . . . . . . . . . . . . . . . . 74
. . . . . . . . . . . . . . . 76
In vitro placental EV model of recombination
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
CHAPTER 5 GENERAL DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.1 SUMMARY OF FINDINGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.2 SCIENTIFIC CONTRIBUTION . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

4.5 DISCUSSION .

4.4 RESULTS . .

4.4.1
4.4.2
4.4.3

.

.

.

.

.

.

.

.

.

.

.

vii

5.3 POSSIBLE FUNCTIONS OF PLACENTAL EXOSOMES . . . . . . . . . . . . . 86
5.4 CONCLUDING REMARKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
. .
CHAPTER 2 SUPPORTING FIGURES . . . . . . . . . . . . . . . 90
CHAPTER 3 SUPPORTING FIGURES . . . . . . . . . . . . . . . 97
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

APPENDICES .
APPENDIX A
APPENDIX B

BIBLIOGRAPHY . .

.

.

.

.

.

.

.

.

.

.

.

.

viii

LIST OF FIGURES

Figure 1.1: Comparative placentation of human and murine pregnancy . . . . . . . . . . .

Figure 1.2: Overview of extracellular vesicle generation . . . . . . . . . . . . . . . . . . .

2

7

Figure 2.1: Schema of tidyNano framework . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Figure 2.2: Example workflow of tidyNano for analysis of NTA data.

. . . . . . . . . . . . 28

Figure 2.3: Data import and reformatting with nanoimport() and nanotidy().

. . . . . . . . 30

Figure 2.4: Multiparameter summary statistics and visualization.

. . . . . . . . . . . . .

. 31

Figure 2.5:

Interactive data manipulation and visualization with shinySIGHT web application. 33

Figure 2.6: Calculation of extracellular vesicle counts and statistics with nanocount(),

nanoShapiro() and nanoTukey(). . . . . . . . . . . . . . . . . . . . . . . . . . . 34

Figure 2.7: Peripheral exosome concentration of GD16.5 mice treated with 10
g LPS.

. . . 35

Figure 3.1: Pregnant plasma EVs traffic to murine lungs and liver in vivo. . . . . . . . . . . 49

Figure 3.2: Placental EV trafficking in vivo.

. . . . . . . . . . . . . . . . . . . . . . . . . 51

Figure 3.3: Flow cytometric analysis of pEV trafficking in murine lung. . . . . . . . . . . . 52

Figure 3.4: Placental EV Localization in Liver. . . . . . . . . . . . . . . . . . . . . . . . . 54

Figure 3.5:

Integrins mediate placental EV localization to murine lung and liver.

. . . . . . 55

Figure 3.6:

Integrin 
3
1 mediates placental EV trafficking to the lung. . . . . . . . . . . . 57

Figure 3.7: Localization of placental EVs in other organs.

. . . . . . . . . . . . . . . . . . 58

Figure 4.1: Cre-mT/mG Model of Recombination.

. . . . . . . . . . . . . . . . . . . . . . 72

Figure 4.2: CremT/mG Model to Study Fetal EVs in pregnancy.

. . . . . . . . . . . . . . . 73

Figure 4.3:

In Vivo Localization of mGFP in CMV-Cre Mated Maternal mT/mG Lung . . . 75

Figure 4.4:

In Vitro Transfer of Placental Cre to Recipient Cells. . . . . . . . . . . . . . . . 77

ix

Figure 4.5: Validation of Placental EV mediated genomic recombination.

. . . . . . . . . 78

Figure 5.1: Thesis Graphical Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

Figure A.1: Transmission electron microscopy of isolated exosomes. . . . . . . . . . . . . . 90

Figure A.2: Polystyrene bead data analysis.

. . . . . . . . . . . . . . . . . . . . . . . . . . 91

Figure A.3: NTA data import with nanotidy(). . . . . . . . . . . . . . . . . . . . . . . . . . 92

Figure A.4: Schema of experimental design. . . . . . . . . . . . . . . . . . . . . . . . . . . 92

Figure A.5: Visualization of samples with technical replicate data.

. . . . . . . . . . . . .

. 93

Figure A.6: Visualization of samples with nanolyze().

. . . . . . . . . . . . . . . . . . . . 94

Figure A.7: Placental mass across gestation. . . . . . . . . . . . . . . . . . . . . . . . . . . 95

Figure A.8: Sample NTA summary PDF file.

. . . . . . . . . . . . . . . . . . . . . . . . . 96

Figure B.1: Isolation and validation of plasma EVs.

. . . . . . . . . . . . . . . . . . . . . 97

Figure B.2: Validation of placental explant culture.

. . . . . . . . . . . . . . . . . . . . . 98

Figure B.3: Proteinase K inhibits pEV localization to lung interstitial macrophages. . . . . . 99

x

CHAPTER 1

GENERAL INTRODUCTION

1.1 PLACENTA OVERVIEW

The placenta is a critical extrafetal disc-shaped organ responsible for the growth and develop-
ment of the fetus during pregnancy. The placenta is attached to the decidual layer of the maternal
uterus and connects to the fetus by the umbilical cord, which joins placental vasculature to the fetal
circulation. The development of the placenta occurs in early gestation and serves as the physical
barrier between maternal and fetal blood before it is expelled during parturition [1, 2]. The placenta
functions as the primary organ for gas and nutrient exchange between the mother and fetus, and
maintains uterine quiescence during pregnancy through the synthesis and secretion of high levels
of progesterone [2, 3, 4, 5, 6]. In addition, the placenta establishes an immunological separation
between mother and fetal tissues, and mediates maternal tolerance to paternally inherited fetal
antigens [7, 8]. One mechanism by which the placenta can modulate the maternal immune system
is through the release of nano-sized, membrane enclosed vesicles known as extracellular vesicles.
All mammalian species harbor a placenta during development, but placental structures differ
among species based on the gross shape as well as the degree of anatomical arrangement of
fetal and maternal blood. Placentation of mice and humans share similarities in their placental
anatomy and function in that trophoblast cells are in direct contact with maternal blood and decidual
tissue (Figure 1.1 A, D) [9, 10]. Maternal blood carries oxygen, water, electrolytes, hormones,
antibodies, drugs, and in some instances, viruses that can pass through the trophoblast to enter
fetal circulation [11, 3, 12, 13]. Thus, placental venous blood is oxygen and nutrient-rich, while
arterial blood contains carbon dioxide, urea, waste products, and hormones and is returned to the
maternal circulation. The placenta also sheds fetal cells and extracellular vesicles which can travel
and influence the mother. This dissertation identifies the cellular targets of placental extracellular
vesicles and proposes a framework to study maternal-fetal interactions in vivo.

1

Figure 1.1: Comparative placentation of human and murine pregnancy. A. Pregnant human
uterus. The square indicates the location of the placenta. B. Overview of the human placenta. The
chorionic villi are surrounded by maternal blood, which is supplied by the maternal spiral arteries.
C. Cross-sectional view of a villous trophoblast. The fetal cells of the placenta include endothelial
cells, villous cytotrophoblasts (vCTB), and two layers of multinucleated syncytiotrophoblast cells
(SynT). D. Pregnant mouse uterus containing 5 fetuses in both uterine horns. The square indicates
the location of an individual placenta. E. Overview of mouse placenta, junctional zone (JZ),
trophoblast giant cell (TGC). F. Cross-sectional view of the junctional zone. Maternal blood is
separated by fetal mononuclear trophoblasts (MNT), two layers of syncytiotrophoblast, a layer of
mononuclear trophoblast cells, and fetal endothelial cells.

1.2 PLACENTA STRUCTURE AND ANATOMY

The human placenta consists of two portions, extravillous trophoblast (EVT) and the chorionic
villi. Trophoblast cells of both arise from the outer trophectoderm layer of the blastocyst, which
implants into the maternal decidua within the first week following conception [9]. Trophoblast cells
differentiate to become mononuclear cytotrophoblasts, which in turn become either multinucleated
syncytiotrophoblasts or extravillous trophoblast cells.
In humans, extravillous trophoblast form
anchoring column structures that connect the placenta to the decidua [10], and further invade
deeply into the decidual tissue and superficial myometrium.

2

A major function of extravillous trophoblast cells is to establish maternal-placental circulation
from the uterine spiral arteries [14]. These cells displace spiral artery endothelial cells, causing
the vessels to widen and increase blood flow to the uterus while also preventing contraction of the
arteries [15, 16]. EVTs also help protect the placenta and fetus from an adverse maternal immune
system response through the expression of HLA-G on their surface [14].
In mice, trophoblast
giant cells affect parallel events, although invasion is not as deep [10, 17]. The establishment of
maternal-placental circulation allows maternofetal gas and nutrient exchange, which occurs via the
outer syncytiotrophoblast layer that faces maternal blood. The syncytiotrophoblast layer also sheds
extracellular vesicles and cells directly into the blood and provides a mechanism for maternofetal
communication.

The second portion of the placenta consists of the chorionic villi, which are three-dimensional
tree-like villous structures that serve as the functional unit of maternal-fetal exchange of the placenta.
The villous structure consists of fetal endothelial cells, mesenchyme, and villous cytotrophoblast
cells, which continuously fuse to form the multinucleated syncytiotrophoblast that interfaces directly
with maternal blood (Figure 1.1 B, C) [18]. Gas and nutrient exchange occurs between maternal
blood and the outer syncytiotrophoblast layer. Gestation in humans is approximately 270 days with
definitive placental villous development first appearing on day 21.

The murine placenta is similar to the human placenta in the hemochorial arrangement between
the trophoblast and the maternal blood, but there are also significant differences in structure. The
murine placenta consists of three main regions: the first is the maternal decidua and vasculature
containing invasive trophoblast giant cells, the second is the junctional zone consisting of tro-
phoblast giant cells and spongiotrophoblast cells, and third is the labyrinth layer in which fetal and
maternal vascular exchange takes place (Figure 1.1 E) [11, 9, 4]. In mice, implantation occurs
at approximately gestation day (GD) 4.5, and at GD5.5, trophectoderm cells differentiate to form
trophoblast giant cells as well as the extraembryonic ectoderm and ectoplacental cone. Trophoblast
giant cells are responsible for producing steroid hormones and prolactin-related cytokines to help
the embryo implant into the maternal decidua [19]. Gestation in mice ranges from 19-20 days and

3

the formation of a functional placenta is established by gestational day (GD)10.5 and continues to
grow in size with increasing gestation [8]. The placenta is largest at GD14.5 in C57/BL6 mice and
gradually decreases in mass for the duration of pregnancy [20].

The junctional zone is the placental layer in mice that interfaces with the maternal decidua.
This layer spans from the trophoblast giant cells that line the maternal decidua to the placental
labyrinth. The spongiotrophoblast layer which lies between the trophoblast giant cell layer and the
labyrinth layer (Figure 1.1 E), is a compact network of non-syncytial cells that generates endocrine
hormones, maintains placental structural support, and allows for maternal venous blood to drain
from the labyrinth layer. Both the spongiotrophoblast and trophoblast giant cells produce placental
lactogens and vascular endothelial growth factor (VEGF), which promote growth of the fetus
and stimulate blood vessel development [10, 8]. After the establishment of the mature placenta,
glycogen trophoblast cells arise from the spongiotrophoblast layer at around GD12.5 and migrate
into the maternal decidua [15, 11]. These cells serve as a source of energy to the developing fetus
[18].

The placental labyrinth is the functional unit of the murine placenta in which maternal circu-
lation interfaces with fetal trophoblast cells, and nutrient/gas exchange occurs. The labyrinthin
interface has a complex network of interdigitated surfaces comprised of fetal endothelium, a bilayer
of multinucleated syncytiotrophoblast cells (SynT-I and SynT-II), and sinusoidal mononuclear tro-
phoblast cells that are in direct contact with maternal blood (Figure 1.1 E, F) [11]. Mononuclear
trophoblast cells overlay the syncytiotrophoblast cells in a discontinuous manner and express pla-
cental lactogen II, which is also expressed by trophoblast giant cells [21]. Placental lactogen II
facilitates fetal growth by promoting mobilization of maternal liver glycogen stores [22]. Maternal
blood enters the labyrinth layer via a central arterial canal, drains via venous channels within the
junctional zone and trophoblast giant cells [15]. Similar to humans, the murine syncytiotrophoblast
layers can produce and release extracellular vesicles directly into maternal circulation. Mice have
three trophoblast layers, two syncytiotrophoblast layers and a mononuclear trophoblast cell com-
pared to humans which have a single syncytiotrophoblast layer and 1-2 villous cytotrophoblast

4

layers, depending on stage of gestation.

1.2.1

Immune Adaptations During Pregnancy

One of the fundamental functions of the immune system is to distinguish self from non-self and to
protect the body from foreign pathogens. Transplanted foreign tissue is recognized as non-self and
results in tissue rejection by the host immune system due to inflammation. The fetus can be likened
to semi-foreign tissue; it is antigenically distinct from the mother due to the paternal contribution
of major and minor histocompatibility complex (MHC) genes. Prevention of an adverse immune
response is therefore critical for pregnancy success and is mediated in part by the placenta. The
placenta is the barrier between fetal and maternal blood and tissues, which occurs in the chorionic
villi and decidua, respectively. The syncytiotrophoblast cells express abundant immunosuppressive
molecules such as FasL, PD-L1, which can prevent adverse maternal T cell responses [23], while
extravillous trophoblast cells express these and HLA-G, which additionally suppresses antigen
presentation [24, 14]. All trophoblast cells of the placenta have restricted class Ia and class II MHC
antigen expression, which prevents cell-mediated maternal immune responses. The placenta has
various mechanisms to develop without eliciting an adverse maternal immune response including
the lack of classical MHC antigen expression and immunomodulatory molecules. Immune cells
that localize to the uterus include macrophages, dendritic cells, uterine natural killer (NK) cells,
and regulatory T cells; conventional T and B cells are rare [25, 26]. Uterine NK cells are the
most abundant leukocyte population at the fetal-maternal interface and are distinct from peripheral
NK cells in their ability to assist with the remodeling of the maternal spiral arteries through
secreted chemokines such as interferon-gamma and interleukin 18 as well as their unique ability to
recognize the class Ib MHC molecules expressed by extravillous trophoblast cells [27, 28]. Uterine
dendritic cells (DC) are also present among the maternal decidual leukocyte population and aid in
embryo implantation [29]. Uterine DCs are able to act on both the innate and adaptive immune
system through classical antigen presentation and cross-presentation respectively. Macrophages
are another antigen-presenting cell population in the decidua that has been found to be essential for

5

pregnancy, however, these cells do not possess the ability to cross-present exogenous fetal antigen
[29, 30]. Macrophages in the decidua produce indoleamine 2,3-dioxygenase (IDO), which prevents
effector T cell activation by catabolizing tryptophan and contribute to maternal immune tolerance to
the fetus [31, 32, 23]. The placenta also secretes extracellular vesicles which can further influence
the maternal immune system and will be described in the next sections.

1.3 EXTRACELLULAR VESICLE BIOLOGY OVERVIEW

Extracellular vesicles (EVs) are lipid bilayer membrane-enclosed structures that contain pro-
teins, nucleic acids, and hormones of the cell from which they are derived [33, 34]. EVs are
produced by nearly all nucleated cells and are of great interest because of their ability to travel
within the body to act on distant cells and potentially influence cell function. Although the precise
classification of and nomenclature of EVs has not reached a uniform consensus, EVs can generally
be classified as ectosomes/microvesicles or exosomes based on their size and site of origin within
the cell [35, 36, 37]. Microvesicles (MVs) (also called ectosomes or shedding vesicles) are EVs
that range from 150-1000 nm and arise from the outward budding of the plasma membrane (Figure
1.2 A) Exosomes are EVs that range from approximately 40-150 nm and arise from the invagination
of the limiting membrane of multivesicular endosomes [38, 39]. This invagination results in the
formation of intraluminal vesicles, which become exosomes once the multivesicular endosome
fuses with the plasma membrane and intraluminal vesicles are released into the extracellular space
(Figure 1.2 B) [34]. Once exosomes are released, they can travel within the body to bind and interact
with other cells to induce their biological effect. The placental extracellular vesicles shed from
the placenta are particularly interesting as they contain fetal antigen which is readily detectable in
maternal blood [40].

6

Figure 1.2: Overview of extracellular vesicle generation. A. Ectosomes/microvesicles (MVs)
arise from the outward budding of the plasma membrane and can express membrane cargo (red lines)
that exists on the plasma membrane surface on their exterior. Since MVs bud outwards, they can
contain cytosolic cargo (blue filaments) within their membrane. B. Exosomes arise from the double
invagination process of the plasma membrane. The first step is the inward formation of the plasma
membrane which forms the early endosome. The early endosome can fuse with the endoplasmic
reticulum and Golgi (dotted blue arrows). Note that externally facing membrane cargo on the
plasma membrane surface faces internally in early and early and late sorting endosomes. The early
endosome gives rise to the late sorting endosome and by inward budding of the limiting membrane to
form intraluminal vesicles (ILV), which later become exosomes. This second invagination process
results in the membrane cargo to be exposed on the external surface of ILVs. During the process
of ILV formation, cellular cargo (orange triangles) from the Golgi can be encapsulated internally
within the ILV or endosomal contents could be released ( bi-directional orange arrows). The late
endosomes containing multiple ILVs are classified as multivesicular endosomes (MVE). MVEs
can fuse with the plasma membrane and exocytose the ILVs as exosomes into the extracellular
space. Alternatively, MVEs can fuse with the lysosome or autophagosome and the ILVs and their
components will be degraded (not shown) and recycled into the cytosol.

7

Another class of extracellular vesicles are apoptotic bodies or apoptotic extracellular vesicles
which range from 100-5,000nm and arise from membrane blebbing of cells undergoing programmed
cell death [41]. Dying cells release damage-associated molecular patterns (DAMPs) to signal to
other tissues to induce inflammation to surrounding tissues [42]. Apoptotic EVs, which can
act as DAMPs, contain signaling proteins such as intercellular adhesion molecule (ICAM-3) and
phosphatidylserine to induce phagocytosis by macrophages, as well as inflammatory mediators
such as interleukin1 alpha (IL-1
). A new class of extracellular vesicles called apoptotic exosome-
like vesicles were identified as mediators inducing proinflammatory IL-1
 cytokine response [43].
These vesicles resemble exosomes, with the exception that their biogenesis is completely dependent
on the sphingosine-1-phosphate (S1P) and S1P receptor pathway.

Exosomes were first identified and described as platelet dust following differential ultracen-
trifugation of blood plasma [44]. The platelet dust was enriched with phospholipids, exhibited
coagulant properties, and could be released from granules of platelets. The term exosome was first
proposed in 1987 to describe nano-sized extracellular vesicles that are released into the cell culture
medium from sheep reticulocytes [45]. A seminal study in 2007 discovered, for the first time,
that exosomes contained functional mRNA and miRNAs that could be translated in recipient cells
[46]. The authors treated a human mast cell line with exosomes from a mouse mast cell line and
demonstrated exosome-mediated in vitro translation of murine protein in recipient human cells.
This study was one of the first to establish a role for exosomal RNA as a mediator of extracellu-
lar communication. Since this discovery, the field of exosome biology has expanded drastically
to identify other potential mechanisms by which cells can communicate through the release of
extracellular vesicles. Within the context of pregnancy, the placenta secretes large quantities of
exosomes that are detectable in maternal blood [40, 20]. The function of extracellular vesicles and
how they influence recipient cells will be covered in the next section.

8

1.3.1 EV function - physiology and disease

The ability of extracellular vesicles to harbor cellular cargo from the cell from which they are
derived and travel to act on other cells has resulted in a large number of studies in a variety of
disciplines including cancer biology, immunology, neuroscience, and reproductive biology [47, 48,
49, 50]. The implications that cancer-derived exosomes have on cancer growth, metastasis, and
detection have been profound. Cancer exosomes are associated with an increase in serum exosome
concentration, and breast cancer derived-exosomes can transcriptionally silence recipient cells by
processing pre-miRNAs into mature miRNAs, thereby promoting tumorigenesis of target cells [51].
It has also been postulated that exosomes could be used as a biomarker for cancer detection. Breast
and other cancers are associated with an increase in serum exosomes, and in 2015, researchers
found that the exosomes containing cell surface proteoglycan glypican-1 could be detected in the
serum of human pancreatic patients, but not in non-cancerous serum, and further, levels of GPC1
in blood correlated with tumor size [52].

Additional roles for exosomes in cancer metastasis have been identified. Pancreatic cancer
exosomes contain migration inhibitory factor (MIF), which induced Kupffer cell expression of
transforming growth factor 
 (TGF-
) as well as upregulation of fibronectin in hepatic stellate
cells.
Increased fibronectin resulted in the recruitment of bone marrow-derived macrophages
which in turn promoted a niche for future metastasis of pancreatic cancer cells [53]. Additional
work from this group identified that localization of exosomes to tissues is mediated by specific
integrins, and that certain types of cancers have a unique integrin profile that is important for
establishing pre-metastatic niches by upregulating the expression of proinflammatory S100 genes
[54]. The authors of this study also confirmed the finding that tissue-specific exosomal integrin
expression was elevated in cancer patients.

Exosomes have roles in numerous other diseases, including diabetes and viral diseases. Exo-
somes from adipose tissue macrophages contain microRNAs (miR-155) that can modulate insulin
resistance in vivo and in vitro in mice [55]. Exosomes have also been implicated in vascular
dysfunction in diabetes, as serum exosomes from diabetic patients are enriched in arginase1, which

9

inhibits endothelial nitric oxide (NO) production [56]. In obese mice, the exosomal miRNA profile
is altered, and exosomal miRNA can induce glucose intolerance and adipose tissue inflammation
[57]. Finally, studies have also found that exosomes can transmit viruses such as hepatitis C virus
(HCV) between cells in vitro [58]. Alterations in placental exosome composition and content have
been implicated in complications in pregnancy such as preeclampsia or preterm birth [59, 60, 61].
Collectively these studies demonstrate how extracellular vesicles are biologically active and have
an important role in intercellular communication.

1.3.2 EV content

Extracellular vesicles exhibit their biological function on cells from the different cargo expressed
on the outer surface and molecular cargo within their membranes. The double invagination process
in exosome biogenesis results in cell plasma membrane surface proteins to be expressed on the
outer surface of intraluminal vesicles (Figure 1.2 B). Outer membrane proteins include tetraspanins
(CD9, CD63, CD81), cell surface antigen, integrin, and cell adhesion proteins, immunomodulatory
proteins, as well as surface proteoglycan proteins [24, 62, 63]. The contents within exosomes
include cytosolic and nuclear proteins, extracellular matrix proteins, microRNAs, mRNAs, amino
acids, and metabolites [63]. The loading of cargo into exosomes is thought to occur during the
second invagination step of multivesicular endosomes where proteins can be shuttled from the Golgi
and become encapsulated inside of an intraluminal vesicle (Figure 1.2 B) [64]. Exosomes were
thought to harbor DNA however, this may be attributed to inadequate purification of true exosomes
[39]. Recently, researchers used high-resolution density gradient fractionation and direct affinity
immunocapture of exosomes to definitively demonstrate that argonaute 1-4, glycolytic enzymes,
cytoskeletal proteins, and DNA are absent from exosomes [37]. The authors provide evidence
of active DNA release in small extracellular vesicles in a multivesicular endosome dependent but
exosome independent process. This distinction suggests that the differences in EV content can be
attributed to the method of EV isolation.

10

1.3.3 EV trafficking and cellular uptake

EVs can induce intracellular changes following their internalization by recipient cells, although the
precise mechanism by which their contents are delivered into recipients is not well characterized
[34]. EVs can be internalized by recipient cells by micropinocytosis or phagocytosis, which are
typically performed by antigen-presenting cells such as dendritic cells and macrophages [65].
Macrophages are critical for the clearance of exosomes in vivo: macrophage-depleted mice are
unable to clear intravenously injected exosomes derived from B16BL6 melanoma cells [66, 67].
EVs can also be internalized by recipient cells by means of clathrin-dependent endocytosis, which
involves a coordinated invagination of the plasma membrane assisted by clathrin complexes resulting
in encapsulation of EVs in an intracellular vesicle [68, 65]. The clathrin-coated intracellular vesicle
can then undergo clathrin un-coating and the intracellular vesicle can fuse with the recipient cell
endosome where EV content may be released.

Another mechanism by which EVs influence cell function is through EV protein-mediated
binding to cell surface receptor proteins. EVs expressing major histocompatibility complexes
(MHC) on their surface can activate T lymphocytes by binding to the T cell receptors [69]. This
mechanism provides an additional method of stimulating an immune response. EV protein can
also mediate their binding to extracellular matrix proteins expressed in tissues such as fibronectin
and laminin can interact with exosomal integrins and vice versa [70, 54]. In fact, EV proteins have
been implicated in determining localization of EVs in vivo as well as inducing changes in recipient
cells.

1.3.4 EV biogenesis

The biological function and contents of extracellular vesicles arise from their site of origin within
the cell. Exosomes arise from a double invagination process of the plasma membrane. The first
invagination step occurs when the plasma membrane forms an inward cup-shaped cavity and fuses
upon itself to enter the cytosol to become an early endosome. During this process, the outer
surface protein cargo on the plasma membrane faces towards the lumen of the early endosome

11

(Figure 1.2 B). It is also during this stage that the early endosome can fuse with the endoplasmic
reticulum or Golgi, thus providing a mechanism for endocytic contents to reach these organelles
for recycling [63]. The early endosome then initiates the second invagination step, giving rise to
intraluminal vesicles (ILVs), which are 40-100 nm in diameter [71] and will eventually become
exosomes [72]. During this process, the endosome membrane forms another inward cup-shaped
cavity causing the surface protein cargo to facet outward (Figure 1.2 B). It is also during this
step that cytosolic, Golgi, and endoplasmic reticulum cargo can be loaded inside the intraluminal
vesicles. The presence of ILVs in the late sorting endosome is indicative of their maturation to
multivesicular endosomes (MVE) [73]. Migration of the MVE toward the cell surface and fusion
with the PM results in the release of the exosomes (formerly ILVs) into the extracellular space.
Alternatively, the multivesicular endosome can fuse to the lysosome or autophagosome for the
degradation and recycling of endosomal content [63].

Several proteins mediate the endsosomal inward budding process are collectively known as
endosomal sorting complexes required for transport (ESCRT) proteins [74]. ESCRT proteins
operate in the cytosol and consist of four complexes: ESCRT-0, -I, -II, and -III, all of which
function in the remodeling and maturation of endosomes into multivesicular endosomes [75, 76].
Tumor suppressor gene 101 (TSG101) is an ESCRT-I subunit involved with exosome biogenesis
and is often used as a marker for exosomes. Apoptosis-linked gene 2-interacting protein X (ALIX)
is an ESCRT-III interacting protein and is used as a marker of late endosomes [75]. Exosomes
are enriched with tetraspanins which are protein-membrane scaffolds associated with exosome
production in multivesicular bodies. Among the most abundant tetraspanins expressed on exosomes
are CD9, CD63, and CD81 [76, 37, 71]. Recently, tetraspanin 6 has been found to regulate the fate of
exosomes to be secreted or degraded by the lysosome [77]. The small GTPases Rab27a and Rab27b
influence the release of exosomes from MVE, and are required for normal exosome secretion [78].
Other exosomal markers include the lipid raft protein flotillin-1, lysosomal-associated membrane
protein 1 (LAMP-1), and syntenin-1 [73, 79, 63]. Since these proteins are tightly associated with
exosome biogenesis, they serve as good markers for exosome validation [63].

12

1.3.5 Methods of EV isolation, purification, and validation

The small nature of EVs makes isolation difficult with traditional centrifugation techniques used for
cells. Methods to isolate and purify vesicles can be classified into three broad categories, each of
which have different advantages and disadvantages. The first method, now considered traditional,
is by differential centrifugation. The general protocol to isolate exosomes is to first pellet and
discard cells and debris by relatively slow centrifugation; this is followed by centrifugation of
the supernatant at 10,000 xg, which pellets microvesicles [37]. Exosomes are then pelleted by
ultracentrifugation (>100,000 xg) followed by a wash step in phosphate-buffered saline (PBS).
From this, exosomes can further be purified by ultracentrifugation on a density gradient, as they
settle at a density of 1.12-1.19 g/ml [24, 80, 37]. The advantage of ultracentrifugation is the
input volume capacity, which ranges from 12-38 ml depending on the model and make of the
ultracentrifuge and rotor. The disadvantages are the cost of obtaining the equipment and the time
commitment to isolate samples. Ultracentrifugation requires the careful balancing of rotors prior
to use and spin times last on the order of hours per run.

The second method is ultrafiltration and size exclusion chromatography. This method separates
EVs from biological fluids based on their molecular size and weight using various inert media, which
can include dextran, sepharose, or agarose polymers. Once a sample is loaded, the chromatography
column allows for EVs to elute at a particular fraction, free of other contaminating proteins [81,
82, 83]. The typical protocol involves concentrating biological fluids with a centrifugal membrane
with a high molecular weight cut off (∼100 kDa) to reduce the sample volume for loading on
to a chromatography column [82]. This step requires a typical benchtop centrifuge, takes less
than one hour, and exposes the sample to speeds less than 5000 xg. EVs can then be isolated
by loading concentrated samples onto the chromatography column and the EV enriched fraction
can be collected and be used for downstream analysis. The purified EVs can be concentrated
further by centrifugal concentration with a 10 kDa membrane cutoff. Advantages of size exclusion
chromatography are ease of separating samples within a short period of time as well as the ability to
elute different fractions from a sample similar to gradient centrifugation. The disadvantage of this

13

method is the cost associated purchasing chromatography columns and centrifugal concentrators.
The third method of isolation is precipitation or affinity binding of exosomes. Precipitation of
exosomes from plasma, serum, and cell culture medium is possible through the use of proprietary
polyethylene glycol (PEG) formulations [20]. The advantage of the precipitation techniques is
that they allow for a large number of samples to be quickly isolated using a standard benchtop
microcentrifuge capable of spinning at 10,000 xg. Affinity capture purification is a bead-based
technique that uses antibodies to bind exosomal markers (CD9, CD63, CD81) from biological
samples [37]. The advantage of this method is that it definitely isolates exosomes from a sample
however, the overall yield is much lower compared to other isolation methods. Recently there have
been new approaches to rapidly isolate EVs using microfluidics and acoustic waves to separate EVs
from whole blood [84]; however, it remains to be seen if this methodology can scale to be widely
adopted in the future. Another method of detecting exosomes is by flow cytometry, although
this often requires modifications to resolve exosomes since the resolution limit of conventional
cytometers is typically less than 400 microns [81, 85].

Following extracellular vesicle isolation, samples must be validated using three general methods
to ensure EV purity. The first method of EV validation is determining the EV diameter and
concentration within a sample. The standard method of measuring EVs is with nanoparticle tracking
analysis (NTA), which determines the concentration and size of particles in a solution based on
Brownian motion [86]. Other methods of quantifying the concentration and diameter of samples
in a sample are dynamic light scattering which works by measuring diffracted light patterns from
particles in solution or tunable resistive pulse sensing (TPRS), which quantifies particles passing
through a nanopore [87, 88]. The concentration of extracellular vesicles in biological samples can
range from 1061012 particles per millimeter; thus, raw data from nanoparticle quantification data
is often large and may benefit from software to simplify data analysis.

The second method of extracellular vesicle validation is transmission electron microscopy
(TEM). Exosomes will exhibit a cup-like morphology which is thought to be induced from the
dehydration step during sample preparation for TEM imaging [89]. Examination by transmission

14

electron microscopy allows for resolving individual exosomes and identifying other contaminating
debris or protein in the sample. Immunogold labeling is a technique for validating the presence
of a protein of interest on exosomes and is achieved by labeling exosomes with gold-conjugated
antibodies [40]. Other methods for visualizing exosomes include cryogenic electron microscopy
(cryo-EM), scanning electron microscopy (SEM), and atomic force microscopy (AFM).

The third method assessing extracellular vesicle purity is Western blot analysis for the detection
of EV proteins. Typical markers for exosomes include proteins associated with their biogenesis
such as TSG101, CD9, CD63, CD81, and ALIX [37, 63]. Additional proteins can be used to
identify samples of non-exosomal origin such as the calnexin chaperone protein associated with
the endoplasmic reticulum or argonaute proteins associated with non-coding RNA processing [24,
35].

1.4 MODELS OF STUDYING PLACENTAL EVS

Similar to human and mouse models of placentation, the study of placental EVs presents itself
with unique approaches that have advantages and disadvantages. Regardless of model species,
placental EVs can be isolated with three general approaches. The first is isolating placental
EVs from whole blood, plasma, or serum. The second approach is to isolate placental EVs
from conditioned medium from primary placental tissue explant culture or perfusion. The third
approach is to collect placental EVs from placental cell line conditioned medium. Each method
must overcome certain limitations such as purifying placental EVs from a heterogeneous mixture
of EVs or increasing the quantity of EVs obtained from primary tissue.

1.4.1 pEVs from blood

Isolation of placental EVs from blood is advantageous in that biological replicates are relatively
easy to obtain compared to other approaches. The volume of blood that can be collected from
humans (10 ml or more) is much greater than that of mice (1 ml), which allows for more pEVs
that can be used for analysis or experimentation. Whole blood typically must be processed to

15

separate out whole cells and thus most studies will centrifuge samples to obtain plasma or serum
which can be easily aliquoted and frozen for later use. Plasma and serum are conducive to most
methods for isolating EVs but the challenge of working with pregnant blood is the identification
of placenta-specific EVs from non-placental EVs. One method of purifying placental EVs from
plasma or serum EVs is to use affinity bead capture with a human placental alkaline phosphatase
(hPLAP) antibody [80, 85].

1.4.2 pEVs from primary tissue

One of the main methods of isolating placental EVs is to culture placental tissue ex vivo or maintain
placental tissue viability through a perfusion-based system. The advantage of using primary tissue
for isolation of placental EVs is that it excludes the possibility of contaminating EVs from non-
placental tissues and that it most mimics what occurs in vivo relative to other methods for placental
EV isolation. An alternative approach is to dissociate the placenta into a single cell suspension and
purify cytotrophoblast cells which can then be analyzed for downstream analysis [24].

1.4.3 pEVs from cell lines

Another widely adopted method for studying placental EVs is to use cell lines derived from
placental tissues or to differentiate trophoblast stem cells. Common trophoblast cell lines derived
from choriocarcinomas include BeWo, JAR, and JEG-3 [90, 91, 92]. Other cell lines derived from
first trimester placenta include HTR-8/SVneo and Swann 71 cells [93]. Trophoblast stem cells from
human and mouse tissues can be cultured and differentiated to syncytial-like cells [94]. Synthetic
biology and organoid models have been developed to better model the placental environment in
vitro. In 2018, Turco and colleagues developed trophoblast organoids from human first-trimester
placentas. These organoids resembled human placental villous structures in vivo and had the ability
to differentiate to HLA-G+ extravillous trophoblasts and invade 3D Matrigel in culture [95].

16

1.5 PLACENTAL EV OVERVIEW

Early studies on placental EVs were performed with human placental tissues and co-culture
methods to understand the potential roles of placental EVs in normal placental physiology and pre-
eclampsia. One of the early studies on placental extracellular vesicles identified syncytiotrophoblast
microvesicles induced lymphocyte proliferation by inhibiting the expression of interleukin-2 (IL-2)
receptor [96].
In 1999, von Dadelszen and colleagues identified that the culture medium from
syncytiotrophoblast microvillous membranes (STBMs) cultured with endothelial cells resulted in
the activation of blood leukocytes in vitro [97, 65]. Placental explant culture from preeclamptic
women induced neutrophils to produce superoxide radicals in vitro, although the mechanism
by which this occurred was not identified [98]. Coculture of preeclamptic syncytiotrophoblast
microvillous membrane (STBM) tissue with endothelial cells did not yield drastic changes in
microarray gene expression analysis after 4,12, and 24 hours suggesting that STBM alone does not
account for the altered endothelial function in preeclampsia [99].

In 2006, it was demonstrated that placental exosomes could be isolated from the plasma of
pregnant women by using human alkaline phosphatase (hPLAP) antibody [100]. The authors of
this study identified the expression of Fas ligand (FasL) and programmed death ligand-1 (PD-L1)
on placental exosomes and incubation of T cells with placental exosomes resulted in a reduction
of CD3-zeta and JAK3 expression, and an increase in SOCS-2, suggesting that placental exosomes
could have immunoregulatory properties. Other studies found that human placental exosomes
expressed MHC class I chain (MIC)-related proteins A and B, and caused the downregulation of
the natural killer (NK) cell receptor NKG2D on peripheral blood mononuclear cells similar to the
mechanism by which some tumors evade the immune system [101]. Our lab has demonstrated that
exosomes from first trimester and term placental explants express PD-L1 [24]; soluble programmed
cell death ligand 1 (sPD-L1) is detectable in serum of pregnant women and increases across
gestation, however, the authors did not determine whether sPD-L1 was present in EVs [102]. It
was also found that in vitro culture of human trophoblast cells secrete sPD-L1 and that it can
induce macrophage polarization toward an M2 phenotype associated with healing and decreasing

17

inflammation. Functional studies on placental exosomes identified the presence of Fas ligand
and TNF-related apoptosis-inducing ligand (TRAIL), providing additional evidence for the role of
placental exosomes in immunomodulation [103, 72]. Additionally, human trophoblast extracellular
vesicles contain placental specific microRNAs that prevent viral replication by inducing autophagy
[50]. Rice et al. found that human term placentas treated with EVs from macrophages but not
monocyte EVs induced placental production of IL-1
,1
, 6, 8, 10, and TNF 
 proinflammatory
cytokines, however, the authors did not report on the effect macrophage or monocyte EV treatment
on placental EV production [104].

Much of what is known about placental EVs have used human samples due to the relative ease
of obtaining blood and term placental tissue in a clinical setting. Ethical considerations make
experimental manipulation with human samples in vivo limited, thus making murine models of
studying placental EVs a more attractive option to understand placental EV physiology. Another
advantage of using mouse models is the ability to use transgenic strains for genetic manipulation. For
example, fetal microchimerism was identified using transgenic sires expressing green fluorescent
protein (GFP) mated to wildtype (non-fluorescent) dams resulting in pups and placentas expressing
GFP. GFP-positive fetal cells were identified in the maternal lung during late gestation and the
microchimeric cells undetectable after delivery [105].

Most experimental models of placental EV biology focused on the effect of EVs on recipient
cells in vitro. In 2017, Tong and colleagues purified and fluorescently labeled human placental
microvesicles and intravenously treated non-pregnant (NP) and pregnant (GD12.5) mice. The
authors identified localization of human MVs by whole body intravital imaging system (IVIS) in
the lungs and liver after 2 minutes, 30 minutes, and 24 hours in both NP and GD12.5 mice [106].
It also appears the route of administration may influence trafficking of EVs. Late gestation plasma
EVs injected intraperitoneally into recipient pregnant mice localized to the maternal-fetal tract and
intrauterine fetal compartments and induced preterm birth [107]. This group also identified intra-
amniotic injection of fluorescently labeled EVs from human epithelial amnion cells into pregnant
mice resulted in localization of fluorescence in maternal blood, uterus, and kidneys [108]. Although

18

the literature has identified evidence placental EV localization to organs in vivo, the specific cellular
targets have not been identified. Furthermore, evidence of specific targeting of placental exosomes
to highly vascularized organs such as the lung, liver, and uterus have not been well characterized.

1.6 DISSERTATION OVERVIEW

Extracellular vesicles play an important role in intercellular communication and are implicated
in influencing cellular responses in vivo. The placenta secretes extracellular vesicles into the
maternal blood. Much of the literature on placental EVs has focused on human samples and cell-
based approaches which have enhanced our understanding of placental EVs. Human plasma EVs
increase across gestation [80], however, a complete understanding of how plasma EVs change during
pregnancy has not been well characterized. Additionally, computational software to systemically
process and analyze nanoparticle quantification data is lacking. Placental EVs influence immune
and endothelial function in vitro, however, the extent that these findings actually occur in vivo
remain unknown [109]. One limitation of in vivo exosome research is the identification of specific
cells that take up exosomes from ones that do not [110].

The overall objective of this dissertation is to: 1) Quantify circulating maternal extracellular
vesicles in normal and abnormal pregnancy. 2) Identify maternal immune cell populations targeted
by murine placental EVs in vivo. 3) Develop a model for studying placental EV interactions
in vivo and ex vivo.
In Chapter 2, we develop a computational software package tidyNano to
process EV quantification data across increasing periods of gestation and characterize alterations
in EV concentration during healthy pregnancy and during inflammation-associated preterm birth.
In Chapter 3, we demonstrate specific, preferential trafficking of placental EVs to interstitial
macrophages in maternal lungs as well as Kupffer cells in the liver, with outer membrane integrins

3
1, 
5
1, and 
V
3 mediating placental EV trafficking to these tissues.
In Chapter 4, we
establish a model for identifying fetal cells and placental EVs in vivo and propose a framework for
studying maternal-fetal interactions without experimental manipulation. Finally, in Chapter 5, we
conclude the dissertation with a general discussion with a summary of the research findings, the

19

scientific contribution to the field, and future directions.

20

QUANTIFYING MURINE PLACENTAL EXTRACELLULAR VESICLES ACROSS

GESTATION AND IN PRETERM BIRTH DATA WITH TIDYNANO: A

COMPUTATIONAL FRAMEWORK FOR ANALYZING AND VISUALIZING

CHAPTER 2

NANOPARTICLE DATA IN R

Sean L. Nguyen1,2, Jacob W. Greenberg3, Hao Wang4, Benjamin W. Collaer5, Jianrong Wang4,
Margaret G. Petroff1,3,6

1Cell and Molecular Biology Program, Michigan State University, East Lansing, MI, USA
2Institute for Integrative Toxicology, Michigan State University, East Lansing, MI, USA
3Department of Microbiology and Molecular Genetics, Michigan State University, East Lansing,
MI, USA
4Department of Computational Mathematics, Science and Engineering, Michigan State
University, East Lansing, MI, USA
5Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing,
MI, USA
6Department of Pathobiology Diagnostic Investigation, Michigan State University, East Lansing,
MI, USA

Originally published in PLoS ONE :

2019 June 18 14(6): e0218270.

https://doi.org/10.1371/journal.pone.0218270

This research was supported by NIH grant R21HD091429, MSU AgBioResearch (NIFA

MICL02447), and MSU funds (MGP). SN was supported by the Integrative Pharmacological
Sciences Training Program grant T32GM092715. The funders had no role in study design, data

collection, and analysis, decision to publish, or preparation of the manuscript.

21

2.1 ABSTRACT

Extracellular vesicles (EVs) are increasingly recognized as important mediators of intercellular
communication that carry protein, lipids, and nucleic acids via the circulation to target cells
whereupon they mediate physiological changes. In pregnancy, EVs are released in high quantities
from the placenta and have been postulated to target multiple cell types, including those of the
vascular and immune systems. However, most studies of pregnancy-associated EVs have used
clinical samples and in vitro models; to date, few studies have taken advantage of murine models
in which pregnancy can be precisely timed and manipulated.
In this study, we used a murine
model to determine whether the quantity of EVs is altered during healthy pregnancy and during
inflammation-associated preterm birth. To facilitate data analysis, we developed a novel software
package, tidyNano, an R package that provides functions to import, clean, and quickly summarize
raw data generated by the nanoparticle tracking device, NanoSight (Malvern Panalytical). We
also developed shinySIGHT, a Shiny web application that allows for interactive exploration and
visualization of EV data. In mice, EV concentration in blood increased linearly across pregnancy,
with significant rises at GD14.5 and 17.5 relative to EV concentrations in nonpregnant females.
Additionally, lipopolysaccharide treatment resulted in a significant reduction in circulating EV
concentrations relative to vehicle-treated controls at GD16.5 within 4 hours. Use of tidyNano
facilitated rapid analysis of EV data; importantly, this package provides a straightforward framework
by which diverse types of large datasets can be simply and efficiently analyzed, is freely available
under the MIT license, and is hosted on GitHub (https://nguyens7.github.io/tidyNano/). Our data
highlight the utility of the mouse as a model of EV biology in pregnancy, and suggest that placental
dysfunction is associated with reduced circulating EVs.

2.2 INTRODUCTION

Extracellular vesicles (EVs) encompass a broad class of membrane-enclosed structures secreted
by cells, and are classified based on their size and subcellular origin. Exosomes, which range from
40-150 nm, arise from the inward budding of late endosomes, and microvesicles, which range from

22

100-1000 nm, develop as a result of outward budding of the plasma membrane. Because EVs have
the capacity to induce physiological responses in recipient/target cells, they are of immense interest
for many life science disciplines including immunology, cancer biology, and medicine [24, 40, 52,
111]. Molecular contents of EVs, which include lipids, proteins, and nucleic acids, serve as the basis
for intercellular communication [72]. Indeed, EVs have been shown to be effective and important
in mediating processes including antigen cross-presentation [112], establishing local niches for
metastasis of cancer cells [54], delivery of microRNAs for suppression of gene expression in target
tissues [113], and even transfer and subsequent translation of mRNAs into target cells [110]. Within
the context of pregnancy, the conceptus-derived placenta secretes copious amounts of EVs that are
detectable in maternal blood as early as the first trimester in women [114, 115, 80]. Humans and
mice share in common the hemochorial anatomic arrangement of the placenta, in which maternal
and fetal circulations are separated by only 2 to 4 cellular layers. Importantly, in both species, the
embryo-derived trophoblast is suffused with maternal blood during much of pregnancy, serving as
a surrogate endothelium across which maternal nutrients and fetal waste products are exchanged.
This intimacy also allows for the trophoblast to shed large amounts of EV directly into the maternal
circulation. Studies in women have suggested that placental EVs contribute to critical processes
in pregnancy, including vascular development of the maternal-fetal interface and establishment of
maternal immune tolerance to the antigenically foreign fetus [24, 116]. Further, because placental
EVs circulate in maternal peripheral blood, they can serve to provide a noninvasive liquid biopsy
indicating fetal health in utero; similar concepts are applied to overall health as a potential diagnostic
for cancer [52, 54]. The similarities in placentation between humans and mice highlight the
utility of the murine model to understand the function of pregnancy-associated EVs; surprisingly,
however, only a few recent studies [107, 94] have explored the pregnant mouse as a model for
placental EV function, and no studies have yet explored the response of pregnancy-associated
EV to inflammation-induced preterm birth. Interest in EVs has burgeoned for the above reasons;
however, their isolation and analysis has posed a number of technical challenges. Because of their
small size, direct observation and quantification of isolated EVs require specialized equipment

23

that generates large amounts of data. One method of measuring the size and concentration of
nanoparticles is nanoparticle tracking analysis (NTA; NanoSight, Malvern Instruments, USA),
which makes use of a microfluidic system and laser together with a camera and software to track
individual particles, measuring their size and concentration on the basis of Brownian motion [86,
117]. However, while NanoSight provides an effective means of measuring EVs, experiments with
more than one condition require the use of independent spreadsheet software for extraction of raw
counts, treatment information, statistical analysis, and graphical representation. Further, because
this approach necessitates direct manipulation of raw data by the user, it is susceptible to user-
or software-introduced error [118]. Analysis of large number of samples would benefit from a
computational approach but requires experiment-specific scripts, which are not easily adaptable to
complex experimental designs or conducive for reproducible research [119, 120]. In this study, we
aimed to establish baseline quantities of murine EVs across pregnancy, and further to determine the
effects of inflammation on pregnancy-associated EVs in a model of preterm birth. In addition, we
sought to develop a computational framework to standardize the process of NTA analysis including
data import, organization, visualization, and statistical analysis.

2.3 MATERIALS AND METHODS

2.3.1 Mice and treatments

All experiments were approved by Michigan State University Institutional Animal Care and Use
Committee protocol: 04/18-050-01. Mice (C57BL/6; Jackson Laboratory, Bar Harbor, ME, USA)
were anesthetized by 3% isoflurane and euthanized by cervical dislocation and subsequent bilateral
pneumothorax. Mice were housed in temperature-controlled environments in a 12-hour light/dark
cycle with standard diet and water available ad libitum. Timed matings were performed in 6-8 week
old mating pairs, and the presence of a vaginal plug was designated as gestational day (GD) 0.5.
For preterm birth experiments, mated GD16.5 females were injected with 10
g LPS (Salmonella
enterica, Sigma-Aldrich, St. Louis, MO, USA) in 100
l PBS or 100
l PBS (vehicle control) i.p.
and sacrificed 4 h later. Following anesthesia with isofluorane and oxygen (2L/min), whole blood

24

was harvested via cardiac puncture using 1.2 ml S-Monovette EDTA-containing tubes fitted with a
22-gauge needle (Sarstedt, Newton, NC, USA). Plasma was isolated by two rounds of centrifugation
at 2000 xg for 15 minutes at 4°C and frozen at -80°C.

2.3.2 Exosome isolation and validation

Frozen plasma samples were thawed on ice and exosomes were harvested using Total Exosome
Isolation reagent (Thermo Fisher, Burlingame, CA, USA) for plasma following the manufacturers
instructions. Samples were resuspended in 25-50 
l of phosphate buffered saline (PBS) [Corning,
Manassas, VA, USA]. Plasma exosomes were examined by transmission electron microscopy as
previously described [40] (Figure A.1). Briefly, pelleted exosomes were resuspended with 4%
paraformaldehyde in PBS, loaded on formvar-carbon coated grids and counterstained with 2.5%
glutaraldehyde and 0.1% uranyl acetate in PBS, and imaged on a JEOL100 CXII (JEOL, Peabody,
MA, USA).

2.3.3 NanoSight analysis

Nanoparticle tracking analysis was performed on a NanoSight NS300 (Malvern Panalytical, West-
borough, MA, USA) equipped with a 488 nm excitation laser and an automated syringe sampler.
Plasma EV samples were diluted 1:125-1:500 in PBS and loaded into 1 ml syringes with the flow
rate set to 50 and the operator blinded to sample identity. Diluted samples were measured by two
separate injections, each by three 30-second videos. Measurements were recorded at camera level
11 and detection threshold of 4. Experiment summary CSV files generated by NTA software v3.2
were used for computational analysis and development of software.

2.3.4

tidyNano software development

Figure 2.1 summarizes the core functions of tidyNano, which serve to simplify importation of raw
NanoSight data into a data frame suitable for rapid computational analysis in the R environment.
To this end, tidyNano rearranges data into a tidy format in which observations are represented

25

Figure 2.1: Schema of tidyNano framework.
tidyNano (purple) is designed to facilitate the
process of importing and formatting data into a tidy format, such that the data are compatible with
existing visualization and statistical packages (cyan).

in rows and variables in a single column, rearranging the data from the canonical wide format
(provided by NTA software), to the long format that is characteristic of tidy data (Figure 2.2, 2.3 A,
and 2.3 B) [121], and conducive for manipulation with the dplyr package [122], visualization with
the ggpplot2 package [123], and computation with base R statistics [124] (Figure 2.1). tidyNano
also parses out sample information and performs back calculations to account for sample dilution
during NTA measurement. Once in this tidy format, tidyNano provides functions to summarize
data and calculate statistics that are immediately suitable for further analysis visualization with
ggplot2 and/or shinySIGHT, as described below (Figure 2.1). Functions within tidyNano make
use of dplyr and tidyr packages to transform and aggregate data, allowing each tidyNano output to
be compatible with a wide variety of R packages and suitable for recent changes in visualization
paradigms [122, 123, 125, 126].

2.3.5 Data import and cleaning

TidyNano provides nanoimport() and nanocombine() functions to import individual NTA experi-
ment files or combine multiple experiment files into a single data frame within the R environment.
Both functions determine NTA software versions and import raw tabular data, and can accom-
modate user-specified NTA parameters including bin widths and particle ranges. The functions
automatically determine the number of samples in an NTA experiment, extract sample names and
particle counts for each sample, and return a single data frame. To do this, the nanotidy() function

26

uses the gather() function from the tidyr package to convert the data from wide to long format,
i.e., the condensing of individual sample columns into key-value pairs [122]. The function then
splits the sample column into multiple columns based on user-specified names, and performs back
calculations to determine the true count of particles in the original sample. Finally, the column
headers are converted to the correct class (e.g., categorical factor or numeric) (Figure 2.2).

2.3.6 Data summarization

The nanolyze() function calculates summary statistics by group and generates a data frame that
includes the number of samples within groups, mean, standard deviation, and standard error.
Further, it calculates a wide variety of statistics such as technical and biological replicate means or
differences between other experimental conditions. Nanolyze() contains arguments for specifying
prefixes to summary statistic columns which allows for the function to be used sequentially to
average replicates such as technical, biological, and group/treatment replicates. Each iteration of
Nanolyze returns a visualizable tidy data frame (Figure 2.2). The nanocount() function allows for
rapid calculation of the total sum of particles within groups of samples and can be combined with
existing functions such as filter() to subset data.

2.3.7 Data visualization and statistical analysis

We also developed shinySIGHT, a web application built within the R shiny framework that allows
the user to upload, interact with, and visualize the NanoSight data without needing to program.
shinySIGHT facilitates dynamic exploration of NTA data and can provide an understanding of
particle size distribution and concentration with user-adjustable sliders to specify particle size
ranges (Figure 2.2). shinySIGHT is available within the tidyNano package and can be run locally
by using the shinySIGHT() command.

27

Figure 2.2: Example workflow of tidyNano for analysis of NTA data. Core functions of
tidyNano (violet) are to facilitate extraction, formatting and aggregation of NTA data. Following
data import, tidyNano functions can be easily visualized using existing packages such as ggplot2
or with the interactive web application shinySIGHT.

28

2.4 RESULTS

To demonstrate the utility of the package, we used tidyNano to analyze three NanoSight
datasets. The first set consisted NanoSight data from polystyrene fluorescent and non-fluorescent
bead standards (Figure A.2) and a second set consisted of peripheral exosomes from C57BL/6
female mice across six time points of pregnancy. The third dataset consisted of peripheral exosome
data from a lipopolysaccharide (LPS) model of preterm birth in GD16.5 C57 BL/6 mice.

2.4.1 Data import and cleaning

To demonstrate the nanotidy() function, we imported raw NTA data from a dataset of murine plasma
exosomes from a preterm birth model of pregnancy. We used nanoimport() to import NTA data
into R which worked by extracting column headers and individual particle counts and combined
them into a large data frame (Figure A.3). If desired, NTA data can also be cleaned manually by
the user in spreadsheet software such as Excel, as was done with the murine pregnancy time course
exosome dataset (Figure 2.3 A). Next, nanotidy() was used to reshape the data into a key-value pairs
such that the individual samples and values were condensed into single columns resulting in the
conversion from wide to long format in an intermediate step (Figure 2.3 B). Nanotidy() then parsed
the Sample column into individual (user-specified) columns and performed back-calculations to
obtain a True_count column containing the corrected count values (Figure 2.3 C). The output of
nanotidy() resulted in a tidy data frame that could be summarized, visualized or manipulated for
further analysis (Figures 2.1 and 2.2).

2.4.2 Data summarization

To study the concentration of peripheral exosomes in murine pregnancy, we sampled blood from
6 experimental groups (nonpregnant, four time points of gestation corresponding to different
embryonic developmental stages, and one postpartum) (n = 5-7/group). From each experimental
group, three technical replicates, and two injection replicates were measured, resulting in 222

29

Figure 2.3: Data import and reformatting with nanoimport() and nanotidy(). A. Output
from nanoimport() or manually-cleaned NTA data. B. Intermediate step of nanotidy() function
which converts data from wide to long format, generating tidy data that can be easily filtered to
isolate individual sample values. C. Finalized output from nanotidy() function which separates the
Sample column into user-specified columns. D. Representative visualization of technical replicates
of nanoimport() output which includes multiple sample injections from one sample with ggplot2,
lines represent technical replicate measurements.

distinct measurements (Figure A.4). Nanolyze() was used to average the technical and injection
replicates within each biological replicate, resulting in a data frame that was suitable for plotting
(Figure A.5 and Figure A.6). To demonstrate sequential use of nanolyze(), we calculated averages
of technical replicates from two separate injections of the same dilution (Figure A.5, Figure 2.4
A). Nanolyze() was then repeated sequentially to determine the mean exosome concentration
between biological replicates using data in Figure 2.4 A as input (Figure 2.4 B). This was possible
due to the name argument, which allows custom naming of the column headers by users, and
param_var argument, which specifies the column on which to perform the calculation. We then
used nanocount() (data in Figure 2.4 A) to determine the total concentration of particles within

30

each biological replicate, resulting in a data frame that could be plotted.

Figure 2.4: Multiparameter summary statistics and visualization. A. Output from nanolyze()
which calculates mean, standard deviation, and standard error within specified groups. Correspond-
ing line graph of exosome concentration as a function of size, split by gestational age, different
colored lines within each group represent a single biological replicate. B. Summarization of bio-
logical replicate data from nanolyze() of 2.3(A) data. Corresponding line graph depicting mean
exosome concentration of biological replicates as a function of size, grey area surrounding lines
represent standard error of the mean.

2.4.3 Visualization

Each tidyNano function described above returns a data frame conducive for flexible downstream
graphical representation. In the examples above, we created all plots from nanolyzed data using
ggplot2. Further, we developed an R Shiny web application, shinySIGHT, that facilitates NTA
data visualization, interactive exploration, and interpretation, and may be particularly practical for

31

users who are less experienced in computer programming. To this end, after importing and tidying
the murine exosome time course data with nanoimport() and nanotidy(), respectively, we used
the nanosave() function to create a .Rds file. We then used the shinySIGHT() function to launch
shinySIGHT, and imported the .Rds file. This allowed interactive visualization of NTA data from
the preterm birth data without needing to be explicitly coded (Figure 2.5). A detailed vignette
outlining how to use shinySIGHT is available at https://nguyens7.github.io/tidyNano/
articles/tidyNano.html#tidynano-vignette. If desired, technical and injection replicates
( Figure 2.A.5) and/or individual samples (Figure 3D) may be visualized using shinySIGHT (not
shown). Summarization steps from each time point across gestation was examined by line graph
using ggplot2, and demonstrated mean exosome concentration of each animal split by gestational
time point (Figure 2.4 A) as well as mean concentration of biological replicates (Figure 2.4 B). After
plotting mean concentration of individual replicates with nanocount(), total exosome concentration
was plotted to determine the variations in exosome concentrations across gestation (Figure 2.6 B).

32

Figure 2.5: Interactive data manipulation and visualization with shinySIGHT web applica-
tion. shinySIGHT allows users to upload tidyNano data to visualize and manipulate data using
a graphical user interface. shinySIGHT automatically generates plots from user uploaded data as
well as displays the underlying data frames that make up the visualizations.

2.4.4 Statistical analysis

We used nanoShapiro() to determine the normality of the murine exosome dataset, which returns a
data frame with suggestions for the appropriate statistical test (Figure 2.6 C). Next, we used existing
base R statistical packages [124] to run an ANOVA on our samples to compare total exosome count
across gestational days using data in Figure 2.6 A. The result of the ANOVA suggested that there
were significant differences in exosome concentrations across gestational age (p < 0.001) (Figure
2.6 D). We then utilized the nanoTukey() function to run a Tukey post-hoc test, which returned a
tidy data frame of pair-wise comparisons of exosome concentrations (Figure 2.6 E).

33

Figure 2.6: Calculation of extracellular vesicle counts and statistics with nanocount(),
nanoShapiro() and nanoTukey(). A. nanocount() function determines the total concentration
of particles. B. Boxplot of murine peripheral exosomes across pregnancy, each point represents
an individual biological replicate. C. nanoShapiro() function determines Gaussian distribution of
data. D. ANOVA output generated from nanocount of data frame in A. E. nanoTukey() output of
pair-wise Tukey post hoc analysis comparing extracellular vesicle concentration across gestation.

2.4.5 Effects of gestation day and inflammation on circulating EV concentrations

Development of the nanoTidy package enabled rapid analysis of the changes in circulating EVs
across gestation (Figure 2.6 B) and following LPS treatment (Figure 2.7). Maternal plasma EV
rose linearly with advancing gestational age, peaking at GD14.5 and corresponding with placental
mass (Figure A.7). EV concentrations were significantly higher (p < 0.05) between non-pregnant
and GD14.5, non-pregnant and GD17.5, and trends (p < 0.1) towards an increase between GD5.5
and GD14.5. Likewise, there was a trend towards reduction in circulating EVs between GD14.5
and 1 day postpartum. Further, EV concentrations at GD5.5 and 10.5 were intermediate between
those of nonpregnancy and GD14.5 and 17.5. Administration of 10
g LPS i.p. in GD16.5 mice

34

resulted in wholesale fetal death as indicated by hemorrhagic implantation sites, as early as 4 hours
post-treatment and indicative of impending preterm delivery. This was associated with a rapid and
significant decrease in circulating EV concentration (p < 0.0017) (Figure 2.7).

Figure 2.7: Plasma samples were collected 4 h following i.p.
saline (PBS) or 10
g LPS. Each point represents a biological replicate. Welchs t-test.

injection of phosphate buffered

2.5 DISCUSSION

In this study, we quantified the murine plasma EVs across normal gestation, as well as in a model
of inflammation-induced preterm birth. The principle findings were that plasma EVs concentrations
increased with advancing gestational age in mice in comparison to nonpregnant females, and that
LPS-induced fetal loss was associated with a striking reduction in circulating EV. Additionally, we
report a novel pipeline in the R platform for rapid exploration and visualization of data generated

35

from NanoSight analysis of EVs. In pregnant females, plasma EV concentrations rose progressively
and were highest during the latter stages of gestation, reaching a maximum at GD14-17. This finding
is consistent with both a recent report of EV across gestation in mice and with reports that EVs
are highest during the third trimester of human pregnancy [107, 127]. Additionally, mean plasma
EV concentrations at GD5.5 and 10.5 rose to intermediate concentrations levels between those
of nonpregnancy and GD14.5, reminiscent of increases in circulating EV observed during the
first trimester of human pregnancy [127]. This may reflect heightened release of maternal, rather
than placental, EVs, as neither the first trimester of human pregnancy nor the first half of murine
pregnancy is characterized by established maternal blood flow to the placenta. Factors regulating
an early pregnancy-associated increase in EV are unknown; hormones from the conceptus and
embryo are already very high at these stages, and may induce systemic possibly vascular release of
maternal EVs. Early rises in circulating EV during pregnancy may reflect a mechanism that controls
systemic changes in the mother in preparation for her own dramatic physiological adaptations to
pregnancy and/or her accommodation of rapid fetal growth. EV concentrations tended (p = 0.076)
to drop rapidly at one day postpartum from the highest values at GD14. These data are also in
line with rapid clearance of placental EVs from the circulation of women following birth [114].
Changes in exosome content during late gestation have been suggested to contribute to parturition in
mice, due in particular to a progressive increase in EVs carrying proinflammatory mediators [107].
Alterations in the EV microRNA profile towards late gestation were also observed in women, and
further changes occurred in preterm birth [128]. Interestingly, we observed a striking reduction
in circulating EV concentrations following LPS administration into pregnant dams, modeling
infection-induced preterm birth. This reduction in peripheral exosomes may reflect acute placental
dysfunction and thus placental EV secretion. In humans, peripheral EVs increase in pre-eclampsia,
however no quantitative differences were observed between normal term and spontaneous preterm
birth [128, 61, 129]. Time of sample collection relative to onset of preterm birth symptoms as well
as variable etiology of preterm birth in human clinical samples may explain differences observed
in our study. It also remains possible, and even likely, that inflammation-induced preterm birth in

36

mice alters EV cargo as in women. Ongoing studies are addressing this possibility.

In addition to our experimental findings, we developed a computational framework, tidyNano,
in R that facilitates analysis of data generated by NanoSight. Following sample measurement, the
NanoSight software generates a PDF report that includes a line graph depicting size distribution
and particle concentration, as well as summary statistics including mean, median, and mode size
of the EVs (Figure A.8). For downstream analysis, including combinatorial analysis of multiple
datasets within a single experiment, users can access the raw NanoSight data, which includes
particle counts and size, as well as user-defined acquisition parameters. However, experiments
typically consist of multiple conditions, and require the use of independent spreadsheet software
(e.g., Microsoft Excel). Although spreadsheets can facilitate this by providing intuitive processing
through a point-and-click interface, the requirement for users to directly interact with raw data is
a major disadvantage as it is tedious and increases the chances for user-introduced error [118].
Further, Excel itself has been known to inadvertently change underlying raw data (e.g., the gene
name OCT4 can be converted to the date 1004), which can also become problematic for downstream
analysis [118, 130, 131]. Spreadsheet software is also limited to user-specific pipelines, cannot be
easily expanded to accommodate multiparameter experiments, and the work and time required to
analyze data typically increases linearly with added data.

To develop software that avoids these pitfalls, we took advantage of a recent paradigm in
computational data analysis known as data tidying, which has become increasingly popular due
to its consistent format that is amenable to rapid data exploration and analysis [121]. Tidy data
is organized into a format such that observations are represented in rows and variables in a single
column, which allows for subsetting, statistical calculation, and visualization of the data. Several
popular packages that accept tidy data have been developed, and allow for efficient rearrangement,
manipulation, programming, and visualization of data [122, 123, 125, 126]. TidyNano employs a
similar strategy, allowing users to quickly import and transform nanoparticle data into a tidy format.
We showed here that this software was able to efficiently process and analyze large sets of biological
data. We also created an R Shiny web application, shinySIGHT, for further exploration of the data

37

using a graphical user interface that can facilitate the process of inspecting and visualizing data
without requiring the user to know how to code.

In summary, the present study reveals changes in EV concentrations across gestation and in a
model of inflammation-induced preterm birth, and describes a novel software package that provides
a framework for analyzing NTA data by performing functions that address the key components
of data analysis:
import and cleaning, summarization, visualization, and statistical calculation.
TidyNano, which will be updated to accommodate future NanoSight models and NTA software, is
an open source package developed in R, hosted on GitHub (https://github.com/nguyens7/
tidyNano), and is freely available under the MIT license. Data, supporting documentation, and
vignettes can be accessed at the package website (https://nguyens7.github.io/tidyNano/).
TidyNano summarization functions are general purpose and can be adapted to analyze other tidy
data, including non-nanoparticle data.

2.6 ACKNOWLEDGEMENTS

The authors thank Alicia Withrow for transmission electron microscopy analysis assistance,
and Gregory Burns and Thomas Spencer of the University of Missouri, Wei-Ting Hung and Lane
Christenson of The University of Kansas Medical Center, and Neva Kandzija and Manu Vatish
of The University of Oxford for validation and feedback on the tidyNano package. The authors
also acknowledge Soo Hyun Ahn and Sarika Kshirsagar for helpful discussions in designing and
implementing the software.

38

CHAPTER 3

INTEGRINS MEDIATE PLACENTAL EXTRACELLULAR VESICLE TRAFFICKING

TO MATERNAL LUNG AND LIVER IN VIVO

Sean L. Nguyen1,2, Soo Hyun Ahn3, Benjamin W. Collaer4, Jacob W. Greenberg5, Dalen W.
Agnew3, Ripla Arora6, Margaret G. Petroff1,3,5

1Cell and Molecular Biology Program, Michigan State University, East Lansing, MI, USA
2Institute for Integrative Toxicology, Michigan State University, East Lansing, MI, USA
3Department of Pathobiology Diagnostic Investigation, Michigan State University, East Lansing,
MI, USA 4Department of Biochemistry and Molecular Biology, Michigan State University, East
Lansing, MI, USA
5Department of Microbiology and Molecular Genetics, Michigan State University, East Lansing,
MI, USA
6Department of Obstetrics, Gynecology, and Reproductive Biology, Institute for Quantitative
Health Science and Engineering, College of Human Medicine, Michigan State University, East
Lansing, MI, USA

This research was supported by NIH grant R21HD091429, MSU AgBioResearch (NIFA

MICL02447), and MSU funds (MGP). SN was supported by the Integrative Pharmacological
Sciences Training Program grant T32GM092715. The funders had no role in study design, data

collection, and analysis, decision to publish, or preparation of the manuscript.

39

3.1 ABSTRACT

Placental extracellular vesicles (EVs) are 40-150 nm membrane-enclosed structures containing
RNAs, proteins, and lipids that exhibit immunomodulatory effects during pregnancy. Placental
EVs are abundant in maternal plasma; however, the small nature of EVs makes studying their effects
on recipient cells in vivo challenging. Other studies have found that placental EVs can localize to
the lung and liver, but the specific cell types that they impact within these tissues have not been
identified. Here, we aimed to identify the specific maternal cells targeted by placental EVs, as
well as to elucidate the mechanisms by which they traffic to these tissues and cells. We found that
localization to the lung was specific to pregnancy-associated EVs, as plasma EVs obtained from
pregnant mice, but not those obtained from non-pregnant mice plasma, preferentially localized to
this tissue following administration. Additionally, we found that placental EVs were specifically
associated with interstitial macrophages in lungs and Kupffer cells in the liver. Treatment of EVs
with proteinase K ablated localization of the lung and liver in vivo, indicating that proteins on the
surface of EVs direct tissue localization. Integrins on placental EVs appeared to be selectively
expressed: purified placental EVs express integrins 
3, 
5, 
1, and 
3, whereas integrins 
V or

6 were not detected, despite expression of all of these integrins in the placenta. Pre-incubation
of EVs with HYD-1 or RGD inhibitory peptides reduced the localization of EVs to the lung in
vivo, suggesting functional roles for integrins in tissue localization. In summary, we demonstrate
specific, preferential trafficking of placental EVs to interstitial macrophages in maternal lungs as
well as Kupffer cells in the liver, with outer membrane integrins 
3
1 and 
5
1 integrins mediating
placental EV trafficking to these tissues.

3.2 INTRODUCTION

During pregnancy, the placenta is responsible for the development and protection of the fetus,
and in doing so, actively communicates with the maternal cellular environment. Hemochorial
placentation in mouse and human species is characterized by direct contact of fetal trophoblast
tissue with maternal blood circulation. The multinucleated syncytiotrophoblast is the outermost

40

layer of the human placenta and releases extracellular vesicles including exosomes, microvesicles,
apoptotic bodies, and syncytial nuclear aggregates into the maternal blood [100, 80, 127]. In fact,
fetal syncytial nuclear aggregates and trophoblast cells are detectable in the maternal lung tissue
after pregnancy [132, 133].

Exosomes are 40-150 nm membrane-enclosed structures formed by inward budding of the
limiting membrane of endosomes. Exosomes can carry RNAs, proteins, and lipids to distant
cells and tissues wherein they mediate biological effects [46, 54, 110]. Placental EVs are readily
detectable in maternal plasma EVs [80], and we previously demonstrated that total plasma EVs
increase with advancing gestational age in mice, peaking at gestation day (GD) 14.5 [20]. Placental
EVs have potential immunomodulatory properties, as they express paternally inherited fetal antigen,
placenta-specific microRNAs that inhibit viral replication, and programmed death-ligand 1 (PD-
L1/B7-H1) [40, 50, 24, 102]. Placental EVs have also been implicated in pregnancy complications
such as preterm birth and preeclampsia [134, 129].

Much of our understanding of placental EV function results from studying how they affect
recipient cells in vitro. However, it remains to be seen if these effects occur in vivo during
pregnancy. In vivo studies have found that placental EVs localize to various organs, depending
on the mode of administration and possibly source of EVs. Intravenously (i.v.) injected human
placental EVs localize to the lung and liver in mice, whereas intraperitoneally (i.p.) injected murine
plasma EVs localize to the uterus, fetal membranes, and the placenta [109, 106, 107].

Questions remain, however, regarding the specific cellular targets of placental EVs in vivo
as well as the mechanisms influencing their trafficking.
Integrins are transmembrane proteins
expressed on the surface of cells that facilitate binding to the extracellular matrix (ECM) and
induce changes in cellular signaling [135]. Cancers that metastasize to different organs release EVs
that express unique integrin signatures. Integrins on the surface of tumor-derived EVs mediate
the binding to distant organs and establish a premetastatic niche [54]. In this study, we identify
the cellular targets of placental EVs in vivo using labeled placental EVs, and identify a role for
outer membrane integrins in localization to the lung and liver. We identify the cell types to which

41

placental EVs localize in the lung and liver and the role of pEV integrins in trafficking. We
also demonstrate that blocking of integrins with inhibitory peptides ablates localization in vivo.
Collectively, these results suggest that maternal lung is a specific target of placental EVs during
pregnancy.

3.3 MATERIALS AND METHODS

3.3.1 Animal experiments

All animal experiments were approved by the Michigan State University Institutional Animal Care
and Use Committee (protocol no. 201800176). Wildtype C57BL/6J (stock no: 000664, were
purchased from the Jackson Laboratory (Bar Harbor, ME) and housed in temperature-controlled,
12-hour light/dark cycle rooms with food and water available ad libitum. All mice were aged 6-12
weeks and timed matings were performed. The presence of a vaginal plug was used as a sign of
mating and designated gestational day (GD) 0.5. Plasma was isolated as previously described [9];
briefly, mice were deeply anesthetized with 3% isoflurane infused with oxygen (2L/min) and whole
blood was collected by cardiac puncture with 22 gauge needles into 1.2 ml S-Monovette EDTA-
containing blood collection tubes (Sarstedt, Newton, NC). Mice were immediately euthanized
thereafter by cervical dislocation and bilateral pneumothorax. Plasma was collected after two
rounds of centrifugation at 2000 xg for 15 minutes at 4°C before being frozen at -80°C for future
analysis.

For cell culture, EV-free media was made with Roswell Park Memorial Institute (RPMI) 1640
(Gibco, Waltham, MA), 50
M 
-mercaptoethanol (LifeTechnologies, Carlsbad, CA), 100
g/ml
Penicillin/Streptomycin (Gibco, Waltham, MA), 1mM Sodium pyruvate (Gibco, Waltham, MA)
with 10% fetal bovine serum and was ultracentrifuged in an SW28.1 rotor (Beckman Coulter,
Brea, CA) at 100,000 xg for 18 hrs at 4°C and filter sterilized through a 0.22
m stericup vacuum
filter (Millipore). Placental explant culture was performed as previously described [10]; briefly,
individual GD14.5-16.5 placentas were isolated cut into four pieces and placed in 74
m mesh
15mm net well inserts (Corning, Corning, NY) in a 12 well plate filled with 3 ml of EV-depleted

42

media for 18 hours at 37 °C, 5% CO2. Tissue viability was assessed by hematoxylin and eosin
staining using a double blinded scoring of placental tissue necrosis (Leica Biosystems, Buffalo
Grove, IL). Following explant culture, media was collected and centrifuged twice at 500 xg for 15
minutes and 2000 xg for 15 minutes to pellet cells and cellular debris respectively. Media was filter
sterilized through a 0.22
m PVDF membrane steriflip filter (cat # SE1M179M6, EMD Millipore,
Billerica, MA) and concentrated using 20 ml 100 kDa molecular weight cut off Vivaspin centrifugal
ultrafiltration columns (cat # VS2042, Sartorius, Stonehouse, UK) at 3320 xg to a volume of 500
l
before being frozen at -80°C for future analysis.

3.3.2 Extracellular vesicle isolation

Extracellular vesicles were isolated using qEVsingle (50 - 100
l samples) or qEVoriginal (100 -
500
l samples) size exclusion chromatography columns (Izon Sciences, Medford, MA) following
manufacturer’s instructions. Briefly, thawed plasma or explant media samples were loaded onto
columns equilibrated with phosphate-buffered saline (Corning, Manassas, VA) and fractions 6-8
or 7-9 (200
l for qEVsingle, 500
l for qEVoriginal columns respectively) were collected and
concentrated with Amicon Ultra 4 10kDa centrifugal filters(cat # UFC801096, EMD Millipore,
Darmstadt, Germany) to 50
l for further analysis.

3.3.3 Fluorescent EV labeling

Purified EVs were labeled with PKH26 lipophilic red fluorescent dye (Sigma-Aldrich, St. Louis,
MO) using manufacturer instructions, followed by a sucrose cushion purification protocol with
modifications to separate excess unbound dye/micelles from labeled EVs [136]. Briefly, 100
l
purified EVs or PBS (dye only control) were diluted with 400
l Diluent C and were pipetted
into ultracentrifuge tubes (Beckman Coulter, Brea, CA) containing 3
l PKH26 dye and 497
l
Diluent C before being gently mixed and incubated for 3 minutes. The labeling reaction was
stopped with the addition of 1 ml of 10% BSA in PBS followed by 2.25 ml EV-depleted media.
Next, 750
l of 0.971M sucrose was gently pipetted beneath the 3 ml of EV/BSA/dye solution

43

and labeled EVs were pelleted by ultracentrifugation at 4°C for 1.5 h at 150,000 xg in SW50.1
rotors (Beckman Coulter, Brea, CA). The supernatant was discarded and the fraction beneath
the sucrose cushion (labeled EVs) was washed with PBS and concentrated with Amicon Ultra
4 10kDa centrifugal filters as described above. For whole organ imaging, purified EVs were
labeled with 0.246mM (DiIC18(7) (1,1’-Dioctadecyl-3,3,3’,3’-Tetramethylindotricarbocyanine Io-
dide)) DiR (Thermo Fisher,Rockford, IL) for 3 minutes and was ultracentrifuged on a sucrose
cushion as described above. After centrifugation, the label vesicles were loaded on IZON single
columns and fractions 6-8 were collected and concentrated with 10kDa concentrators as described
above and used for downstream analysis.

3.3.4 Nanoparticle tracking analysis

EVs were quantified using NanoSight NS300 (Malvern Panalytical, Westborough, MA) as described
previously with modifications [20]. Briefly, concentrated samples were diluted 1:500-2000 in PBS
and loaded in a 1 ml syringe at a flow rate of 50. For each sample, five 20-second videos were
recorded at camera level 12 and a detection threshold of 4 before data raw data were analyzed with
the tidyNano R package [20].

3.3.5 Western blot analysis

For protein analysis, 20ul of each fraction SEC fraction was heat-denatured at 95°C for 15 minutes
with RIPA buffer containing DTT. Denatured samples were run on 4-20% SDS-PAGE gels (Bio-
Rad, Irvine, CA) and samples were transferred onto PVDF Amersham Protran 0.45
m nitrocellu-
lose membranes (GE Healthcare, Chicago, IL). Ponceau S stain (Sigma-Aldrich, St. Louis, MO)
was used for visualizing protein before being washed with tris buffered saline (TBS) with 0.02%
tweenr 20 (Acros, Morris Plains, NJ). Blots were blocked with superblock solution (ThermoFisher,
Rockland, IL) for 30 minutes before overnight incubation with primary antibodies (Supplemental
Table 1) at 4°C followed by three washes with TBST and incubation with appropriate secondary
horseradish peroxidase antibody for 1 h at room temperature. Blots were incubated with Super-

44

Signal pico plus chemiluminescent substrate (ThermoFisher, Rockford, IL) for 5 minutes before
development on film or digital imager (ThermoFisher, Rockford, IL).

3.3.6 Transmission electron microscopy

For transmission electron microscopy, concentrated EVs were fixed with 4% paraformaldehyde in
PBS and placed on formvar-carbon coated grids and counterstained with 2.5% glutaraldehyde and
0.1% uranyl acetate in PBS. Negative TEM was performed on a JEOL100CXII (JEOL, Peabody,
MA).

3.3.7 Proteinase K and inhibitory peptide treatment

For proteinase K treatment, placental EVs were treated with 10
g/ml proteinase K (Life Technolo-
gies, Carlsbad, CA) in PBS and PBS only for the control treatment. Both samples were incubated for
10 minutes at 37 °C before heat inactivation at 65 °C for 10 minutes followed by addition of 100
M
phenylmethylsulfonyl fluoride (PMSF) protease inhibitor (Sigma-Aldrich, St. Louis, MO). For
inhibitory peptide experiments, RGD peptide (A8052-5MG, Sigma-Aldrich, St. Louis, MO) was
commercially purchased and HYD-1 (KIKMVISWKG) was synthesized (Genscript,Piscataway,
NJ). All reagents were 95% or greater purity and resuspended in deionized water, aliquoted, and
frozen. Labeled pEVs were incubated with 0.6 
M of inhibitory peptide and incubated at 37 °C
for 30 minutes before intravenous administration into recipient mice.

3.3.8 Fluorescence/immunofluorescence microscopy and quantification

Tissues were fixed in 4% paraformaldehyde (PFA) [Sigma-Aldrich, St. Louis, MO] in PBS and
30% sucrose in PBS at 4°C overnight before being flash-frozen with isopentane and liquid nitrogen
in Tissue Tek O.C.T embedding compound (Sakura Finetek, Torrance, CA). For fluorescence
microscopy, 5 
m tissue cryosections were mounted with DAPI nuclear stain (Vector, Burlingame,
CA) and imaged on a Nikon Eclipse Ti epifluorescent microscope fitted with a 20x NA 0.5 plan
fluor objective using Nikon NIS-Elements AR 4.40 software. For PKH26 quantification, five

45

random fields were imaged and individual color channels processed using ImageJ. PKH26 and
nuclei counts were computationally quantified using a custom pipeline developed in CellProfiler
3.0 software [137]. For immunofluorescence microscopy, slides were stained with BioLegend
antibodies and mounted with mounting medium with DAPI.

3.3.9 Tissue clearing

Fixed organs were embedded in 2% agarose in PBS and 200
m sections were cut on a Leica
VT1200 S vibratome (Leica, Buffalo Grove, IL) and placed in 12-well plates and permeabilized
with PBS with 0.05% tween 20 and blocked with SuperBlock solution at RT for 4-12 hrs. The
samples were then stained with primary or directly conjugated antibodies at 37 °C in an orbital
shaker for 24 hours before the addition of DAPI (10
g/ml) and incubated for another 24 hours.
The tissues were washed with PBS with 0.05% tween 20 on an orbital shaker at RT for 12 hrs.
Secondary antibody was incubated overnight at 37 °C and samples were washed for 12 hrs. Samples
were then immersed in 2-3 ml of CUBIC1 solution and incubated overnight at 37 ° before being
washed briefly with PBS before immersed with 2-3 ml of CUBIC2 solution for 12hrs [138, 139].

3.3.10 Flow cytometry

Lungs were minced with spring-loaded surgical scissors (FST) in digestion buffer consisting of
RPMI1640, 0.05% W/V liberase TH (Roche), 100U/ml DNase1 (Sigma-Aldrich, St. Louis, MO)
for 1 hr at 37 °C. Larger tissue pieces were further dispersed by extruding undigested tissue pieces
The digestion was quenched with albumin buffer (5% bovine serum albumin (Sigma-Aldrich, St.
Louis, MO) in PBS), pelleted at 500 xg for 5 minutes before treatment with ACK lysis buffer
(150mM NH4, 10mM KHCO3, 0.1mM Na2EDTA, in distilled water) and resuspended with flow
staining buffer (FSB) (2% fetal bovine serum, 0.01% sodium azide in PBS). Cells were plated in
a U bottom 96 well plate (1E6cells/well) and treated Live/Dead Yellow 1:1000 (LifeTechnologies,
Carlsbad, CA) for 10 minutes on ice followed by FC block 1:200/well (BioLegend, San Diego,
CA) and stained with antibody cocktails for 30 minutes on ice. Cells were washed twice and fixed

46

with 1% PFA in flow staining buffer. Spectral flow cytometry was performed on an Aurora (Cytek
BioSciences, Fremont, CA) with eUltracomp compensation beads (LifeTechnologies, Carlsbad,
CA) used for single color controls. 100-200 thousand events were collected for each well and
analysis was done on Kaluza 2.1 (Beckman Coulter, Brea, CA). Flow cytometry t-distributed
stochastic neighbor embedding (t-SNE) analysis was performed using the Cytofkit R package [140]
with default settings.

3.3.11 Whole organ imaging

To quantify localization in whole organs, 2.5 x 109 GD14.5 DiR labeled pEVs were injected
intravenously into NP female mice. After 24 hours, mice were euthanized and the lung, liver,
spleen, thymus, brain, uterus, spleen, lymph node, kidney, and heart were harvested and fixed in
4% PFA in PBS for four hours before being stored in PBS. Whole organs were imaged on a LiCor
odyssey infrared imager on manual scan mode (21 micron) under automatic acquisition settings
(LiCor BioSciences, Lincoln, NE). Raw tifs were analyzed using Fiji and converted to 8 bit before
mean intensity was measured in each tissue [141].

3.3.12 Statistical analysis

For boxplots, the middle line represents median value, upper and lower box regions correspond to
third and first quartiles (75th and 25th percentiles) and whiskers represent 1.5 times the interquartile
range. All plots and analyses were generated in R v4.0. All raw data, analysis, and scripts for
generation of figures are available on GitHub. Fiji/imageJ macros and CellProfiler pipelines
are available on GitHub. Data were subjected to a Shapiro normality test before selecting the
appropriate parametric or non-parametric statistical test.

47

3.4 RESULTS

3.4.1 Pregnant plasma EVs traffic to the lung

Placental EVs are continuously released into maternal circulation and are readily detected in
maternal plasma in mice and humans [80, 114, 85, 142]. Tong et al. demonstrated the localization
of human placental microvesicles in murine lung [109]. To determine if plasma EV trafficking to
the lung was specific, EVs were isolated from plasma of pregnant or non pregnant female mice by
size exclusion chromatography, allowing purification of EVs without the presence of contaminating
proteins that appear in later elution fractions [82]. We observed a 2.4-fold increase in the total
number of vesicles in plasma from gestation day (GD)14.5 mice compared to plasma of nonpregnant
mice (P = 0.015) (Figure B.1 A). EVs from both GD14.5 and nonpregnant mice ranged from 50-
150 nm in diameter, which is typical of exosomes [37]. Western blot revealed presence of the
exosomal marker TSG101 in chromatography fractions 7-9, while Ponceau S staining revealed the
presence of other proteins in later fractions, confirming purification of EVs (Figure B.1 B). Plasma
EVs isolated by size exclusion chromatography exhibited typical spherical morphology of plasma
exosomes by transmission electron microscopy (Figure B.1 C) [20].

48

Figure 3.1: Pregnant plasma EVs traffic to murine lungs and liver in vivo. A. Plasma EVs
(2.5x1010) isolated from nonpregnant or GD14.5 mice were labeled with PKH26 and administered
i.v. into NP mice. Mice were sacrificed after 30 minutes and lung and liver tissue was analyzed
by epifluorescence microscopy. B. Representative fluorescence microscopy of lung from a mouse
treated with nonpregnant or GD14.5 plasma EVs (Arrowheads, GD 14.5 plasma EVs). C. Fluo-
rescence quantification of plasma EVs isolated from nonpregnant (NP) or GD14.5 pregnant mice
in nonpregnant recipient lung tissue. D. Representative fluorescence microscopy of liver from a
nonpregnant mouse treated with plasma EVs from nonpregnant or GD14.5 pregnant mice EVs (Ar-
rowheads, plasma EVs). E. Fluorescence quantification of plasma EVs isolated from nonpregnant
or GD14.5 pregnant mice in recipient liver tissue. Dots represent biological replicates, Wilcoxon
test.

49

We next labeled the purified EVs from nonpregnant and pregnant gestational day (GD) 14.5
mice with PKH26 red fluorescent dye, and administered the labeled EVs intravenously (i.v.) into
nonpregnant recipient females. Lung and liver tissue were harvested after 30 minutes, and PKH26-
labeled EVs were quantified by CellProfiler using epifluorescence microscopy (Figure 3.1 A).
Notably, only transferred EVs from pregnant mice were readily detectable in recipient female lungs;
those purified from nonpregnant mice were scarce or undetectable in recipient tissues (Figure 3.2
B). PKH26 signal in the lung tissue was significantly elevated in mice injected with GD14.5 plasma
EVs compared to those injected with EVs from nonpregnant mice (Figure 3.1 C). Interestingly, in
a limited number of samples, the same trend did not appear to be true for EV localization to the
liver (Figure 3.1 D, E). The localization of plasma EVs from pregnant mice but not nonpregnant
mice suggests a preferential pregnancy-specific EV localization to lung tissue.

50

Figure 3.2: Placental EV trafficking in vivo. A. Schematic diagram of placental EV (pEV) isola-
tion from explant culture and size exclusion chromatography. B. Transmission electron microscopy
of EVs obtained from explant culture of GD14.5 placentas, representative random field from four
independent experiments. C. Experimental overview of i.v. injection of 2.5x1010 PKH26 labeled
GD14.5 placental EVs into mice; lung and liver were analyzed 30 minutes after treatment. D.
Representative flow cytometry t-SNE analysis of murine lung cells after treatment of mice with
GD14.5 placental EVs. Dots represent individual cells, purple dotted line represents CD45+ cells,
blue dotted line represents cells positive for PKH26. E. Immunofluorescence microscopy of lungs
from mice treated with GD14.5 placental EVs. Arrows, PKH26 pEV foci. Representative images
of random fields chosen from three independent experiments.

3.4.2 Placental EVs traffic to lung interstitial macrophages in vivo.

We next tested the hypothesis that placental EVs preferentially target immune cells in the lung. To
isolate placental EVs we cultured GD14.5 murine placental explants and collected EVs from the
supernatant by ultrafiltration and size exclusion chromatography (Figure B.2 A). We first tested
explant viability as well as quality and quantity of EVs from explants cultured under various

51

conditions. We found no significant differences in placental histology necrosis scoring between
freshly isolated placentas and placentas cultured for 8 or 24 hours at low or ambient oxygen levels
(Figure B.2 A). We observed the highest yield of EVs from placentas cultured for 18-24 hours at 37
°C, and therefore used these conditions for pEV isolation. EVs obtained in this manner exhibited
typical cup-like morphology by TEM analysis and were 50-150 nm in diameter by NanoSight
analysis (Figure 3.2 B and Figure B.2 B).

Figure 3.3: Flow cytometric analysis of pEV trafficking in murine lung. A. Adapted gating
strategy [143] to identify murine lung cells colozlizing with PKH26 labeled pEVs. Representative
flow cytometry plots identifying macrophage (M0), alveolar macrophage (AM), and interstitial
macrophages (iM) from murine lung treated with GD14.5 placental EVs. B. Autofluorescence
signature of macrophage, alveolar macrophage, and interstitial macrophage populations.

Following isolation, placental EVs were labeled with PKH26 and administered i.v. into non-
pregnant recipient mice. Recipient tissues were harvested after 30 minutes for downstream analysis
by flow cytometry and immunofluorescence microscopy. To identify cells in the lung that took up

52

PKH26 by flow cytometry, we used t-distributed stochastic neighbor embedding (t-SNE) analysis
as a dimensionality reduction technique to cluster related individual cells based on marker intensity
[140]. This approach facilitated the identification of PKH26+ cells with other lung cells from
NP mice treated with PKH26-labeled placental EVs.
t-SNE analysis readily identified CD45+
cells (Figure 3.2 D, purple dashed line) and could be used to further identify populations of cells
positive for PKH26 EVs (Figure 3.2 D, blue dotted line). Cells positive for PKH26 were also
CD45+, MHCIIhi/lo, F4/80+, autofluorescence (AF) -, CD11clo, CD64+, all characteristic markers
of interstitial macrophages, which are distinct from alveolar macrophages (Figure 3.2 D) [140, 144,
145, 146]. Using a complementary forward-gating strategy together with spectral flow cytometry,
we found that PKH26+ placental EVs localize predominantly with interstitial macrophages and not
alveolar macrophages (Figure 3.3 A, B) [143]. We confirmed the localization of placental EVs in
the lung by immunofluorescence microscopy, detecting PKH26+ EVs with LYVE1+ and CD68+
interstitial macrophages (Figure 3.2 E). Confocal microscopy of liver tissue from the same animals
revealed localization of placental EVS within CD31 endothelial cells and F4/80+ Kupffer cells
(Figure 3.4 A, B).

53

Figure 3.4: Placental EV Localization in Liver. Immunofluorescence confocal microscopy pEV
localization with A. CD31 endothelial cells. B. F4/80 macrophages. Representative images from
three independent experiments.

3.4.3 Outer membrane proteins influence pEV trafficking to the lung and liver.

Adhesion molecules such as integrins (ITGs) mediate the trafficking of EVs to specific organs
[54]. We profiled the placenta and placental EVs for integrins by Western blot and determined if
the treatment of EVs with or without proteinase K would remove integrin expression on the outer
surface. We titrated proteinase K treatment and found that 10
g/ml was sufficient to eliminate pEV
ITG 
1 expression but did not eliminate the exosomal marker CD9 since the antibody binds to the
cytoplasmic moiety of EVs. Proteinase K treatment did not change the morphology of placental
EVs by TEM (Figure 3.5 A,B) [37]. Placental EVs expressed 
3, 
5, 
V, 
1, and 
3, whereas
integrin 
6 was not detected in placental EVs, despite their expression in the placenta (Figure 3.5
C). As expected, proteinase K treatment removed outer membrane-associated integrins. (Figure
3.5 C).

54

Figure 3.5: Integrins mediate placental EV localization to murine lung and liver. A. West-
ern blot of placenta (plac) and placental EVs treated with proteinase K, representative of three
independent experiments. B. Representative transmission electron microscopy of EVs treated with
proteinase K. C. Western Blot analysis of integrin expression in murine placenta, placental EVs,
and placental EVs treated with proteinase K, three independent experiments. D. Representative
fluorescence microscopy images of murine lung treated with GD14.5 placental EV after 30 min-
utes. E. Fluorescence microscopy quantification of EV positive cells in lungs treated with placental
EVs after 30 minutes. F. Representative fluorescence microscopy images of the liver treated with
placental EVs after 30 minutes. G. Fluorescence microscopy quantification of EV-positive cells in
liver treated with placental EVs after 30 minutes. White arrows represent placental EV foci, points
represent biological replicates.

Next, we determined whether pEV localization to lung and liver tissue was mediated by outer
membrane proteins. After enzymatically digesting EVs with proteinase K, we labeled them with
PKH26, and administered them i.v. to NP mice. We readily detected PKH26-labeled EVs in both
lung and liver in mice treated with placental EVs, but not with EVs treated with proteinase K
(Figure 3.5 D, F). We then quantified PKH26 labeled placental EVs and found that proteinase K
treatment significantly reduced localization in both tissues. To account for the possible degradation

55

by proteinase K treatment we also administered 10-fold more EVs and still observed a reduction in
localization (Figure 3.5 E, G). Interestingly, treatment with proteinase K to remove outer membrane
proteins significantly reduced the localization of EVs to interstitial macrophages (Figure B.3 A).
To test the hypothesis that integrins mediate pEV targeting to the lung in vivo, we pre-incubated
placental EVs with RGD and HYD-1 peptides, which block binding of integrins 
5
1 and 
3
1
respectively to their substrates [147, 148], prior to i.v. administration into nonpregnant recipient
females. Mice were sacrificed after 30 minutes, and lung and liver were analyzed by fluorescence
microscopy (Figure 3.5 D, F). Although proteinase K ablated localization of placental EVs to both
lung and liver (Figure 3.5 D and F, upper panels), placental EVs remained detectable in both tissues
after pretreatment with integrin-blocking peptides (Figure 3.5 D and F, lower panels). In the lung,
neither peptide significantly reduced the total numbers of placental EVs (Figure 3.5 E). In the liver,
however, RGD pretreatment resulted in a significant reduction of placental EV localization (Figure
3.5 G).

Although the total numbers of placental EVs that localized to the lung were not reduced by
HYD-1, we noticed that the tissue distribution of the vesicles was altered, with EVs appearing to
remain within vessels (Figure 3.5 D, lower left panel). To better characterize the effect of HYD-
1 treatment on pEV localization to the lung, 200-micron lung sections were cleared by CUBIC
and imaged by 3D confocal microscopy [149]. Strikingly, we saw localization of placental EVs
exclusively within the vasculature, suggesting that they were blocked from entering interstitium
(Figure 3.6 A).

56

Figure 3.6: Integrin 
3
1 mediates placental EV trafficking to the lung. A. Representative
CUBIC-cleared lung treated with placental EVs after 30 minutes, three independent experiments.
B. Near-infrared (NIR) whole lung imaging of murine lung treated with labeled placental EVs after
24 hours, representative of three independent experiments. C. Quantification of NIR pEV positive
area from whole lung treated with placental EVs, a.u., arbitrary units, points represent biological
replicates.

3.4.4

Integrin 
3
1 allows trafficking of pEVs to the lung interstitium.

Since HYD-1 pretreatment appeared to inhibit migration placental EVs to lung interstitium, we
asked whether the peptide impacted pEV localization to lung tissue after 24 hours. We used
Li-Cor whole organ imaging as a high throughput method to screen for the presence of placental
EVs following treatment with inhibitory peptides. We labeled EVs with near-infrared (NiR)
dye, administered them i.v.
into nonpregnant recipients, and quantified localization by whole-
organ imaging of the lung (Figure 3.6 B). Untreated EVs remained detectable after 24 hours, but
pretreatment with HYD-1 significantly decreased their presence in the lung (Figure 3.6 C). RGD
pretreatment reduced localization to the lung but the result was not statistically significant.

57

Figure 3.7: Localization of placental EVs in other organs. Representative whole organ NIR
imaging of mice treated with pEVs after 24 hrs, representative of three independent experiments.

Other studies suggested that placental EVs localize to the uterus and impact uterine physiology
during pregnancy [107, 142]. Therefore, we imaged other organs including, the brain, thymus,
heart, kidney, paraaortic lymph node, spleen, gastrointestinal tract, uterus, and ovaries following
injection of NIR-labeled placental EVs. After 24 hours, we did not observe any significant changes
in NIR signal relative to the vehicle control treatments suggesting that placental EVs do not traffic
to these tissues in vivo (Figure 3.7).

3.5 DISCUSSION

Since the discovery of EVs and their effects on distant cells, research interest in the role of
placental EVs during pregnancy has greatly increased [46, 35, 150, 151, 152]. The total quantity
and concentration of EVs in maternal plasma rise across gestation, with the placenta contributing

58

significantly to this increase [20, 40]. While a number of effects of placental EVs have been
suggested, few studies have attempted to quantify their biodistribution in vivo.
In this study,
we show that placental EVs traffic to the lung and the liver when administered i.v. Further, we
demonstrate that trafficking to the lung appears to be specific to placental EVs, as only plasma EVs
isolated from pregnant dams, but not those from non-pregnant mice, localized to the lung. This
suggests that localization to the lung is specific to the placenta and pregnancy, and not merely due
to the high degree of vascularization of the lung or due to direct routing of venous blood to this
tissue.

Our results are consistent with those found by Tong et al. [109, 106], who found that human
placental EVs administered into mice also localized to the lung. A limitation to the approach by Tong
et al. was the administration of human placental EVs into immune competent mice representing In
our study, we additionally identified the cellular targets of placental EVs in this tissue as well as in
the lung. Using multiparameter flow cytometry and both targeted and untargeted analyses, together
with immunofluorescence microscopy, we found that murine placental EVs specifically colocalized
to lung interstitial macrophages and Kupffer cells of the liver. This result is not surprising as
macrophages are highly phagocytic, and aligns with data published by Atay et al. that identified
human trophoblast EVs were readily taken up by monocytes and macrophages in culture and induce
pro-inflammatory cytokine production [70, 153].

Another important function of macrophages is antigen presentation. Maternal T cells are
made aware of fetal antigen exclusively through indirect antigen presentation by maternal antigen-
presenting cells [154], which may include macrophages. Maternal CD8 and CD4 T cells that
recognize fetal antigen do not mount an adverse immune response to fetal antigen, even when
artificially stimulated with high concentrations of adjuvant [154], suggesting that antigen presenting
cells in pregnancy convey powerful tolerogenic signals. We and others have postulated that EVs may
be the source of these tolerogenic signals, possibly through cargo that include potent suppressors
such as PDL1 and FasL [24, 62, 155, 101, 103, 156]. Intriguingly, maternal recognition of fetal
alloantigens occurs as early as post conception, and can be identified as physically associated

59

with follicular dendritic cells in the spleen and lymph nodes [154, 157]. The placenta, through
emission of free and/or EV-associated antigen, has been presumed to be the source [158, 157];
indeed, paternally-inherited ovalbumin is secreted in several hundred microgram quantities daily
into maternal circulation in mice, and is expressed on the outer surface of exosomes [157, 40].
Surprisingly, our current results show minimal or no localization of pEVs to the maternal spleen
and lymph nodes, but rather to lung interstitial macrophages and Kupffer cells of the liver. Lung
interstitial macrophages are a class of tissue-resident macrophages that arise from monocytes and
play a role in lung homeostasis [144, 145, 146]. These cells have antigen-presenting capabilities and
interact with resident dendritic cells to present antigenic peptides [145, 144]. As such, future work
may examine the role of pEVs in inducing interstitial macrophages antigen presentation, thereby
providing a possible mechanism for fetal placental antigen exposure to the maternal immune system.
The placenta shares many properties with cancer cells, in that fetal trophoblast cells undergo
epithelial to mesenchymal transition, invade into uterine tissue in early placental development,
and express a number of tumor-associated proteins [40, 62]. EVs can serve as biomarkers for
early detection of cancer, similarly an area of active investigation in pregnancy research [52, 159,
160, 161, 162, 163]. Additionally, adhesion and extracellular matrix proteins associated with EVs
appear to play an important role in tumor metastasis.
In a murine xenograft model of cancer,
EVs derived from human metastatic tumors specifically traffic to the lung and liver of nude mice,
serving to establish a niche for future metastasis to these tissues, with integrins playing a major role
in specific localization [164, 52, 54, 53]. We found that placental integrin expression did not fully
recapitulate integrin expression in placental EVs. Integrins 
3, 
V, 
5, 
1, and 
3 were expressed
in both placenta and EVs, while 
6 was found only in the placenta and not in EVs. This suggests
a preferential selection of protein from cells into EVs during their biogenesis. Interestingly, we
found the presence of both heavy and light chains in integrins 
3 and 
V in the placenta but
the light chains were expressed at higher levels. Proteinase K treatment removed placental EV
integrin expression and abrogated localization to the lung and liver tissues; this effect was also
Interestingly, EV expression of CD9 was preserved after
observed in interstitial macrophages.

60

proteinase K treatment suggesting that the antibody used may bind the interior domain of EVs.
Further, we found that pre-incubation of EVs with RGD peptide, which blocks ITG 
5
1 binding
to its receptor, fibronectin, abrogated placental EV to the liver, which is rich in fibronectin [53, 165,
166]. In addition, HYD-1, which blocks ITG 
3
1 binding to the basement membrane glycoprotein
laminin (LN)-5, blocked entry of placental EVs into the lung parenchyma [167, 148, 168, 169]. The
pulmonary basement membrane, which contains LN-5, underlies endothelial cells and surrounds
the interstitial matrix [170]. Our 3-dimensional imaging of placental EVs pre-treated with HYD-1
confirms that ITG 
3
1 is necessary for penetrating the basement membrane and accessing the
interstitium (Figure 3.6 A).

The above results highlight possible parallel roles for integrins on placental EVs. Trophoblast
cells use integrins to bind the endometrial epithelium during implantation, and a role for blastocyst
EVs, presumably arising from the trophectoderm, has been postulated.
It is entirely possible
that these trophectoderm EVs play a role in establishing a niche for blastocyst implantation.
Additionally, it is interesting that when trophoblast cells turn malignant in cases of choriocarcinoma,
the lung is the primary site for metastasis. Placental EV trafficking to the lung could play an
unfortunate role parallel to those found for other metastatic tumors.

The route of administration, cellular source, and quantity of EVs are important factors for
biodistribution studies [171]. B16-F10 melanoma EVs injected into mice i.v. localize to the liver
and spleen within 24 hours [82]. These authors also found that localization of EVs to the lung
depended on method of isolation: treatment with EVs isolated by ultracentrifugation resulted in EV
localization of EVs to the lung, but EVs isolated by ultrafiltration or size exclusion chromatography
did not. Using size exclusion chromatography, we found that placental EVs rapidly localize to the
lung and liver and remain detectable after 24 hours.

A limitation of the current study is that a bolus injection of purified EVs, as performed here,
does not recapitulate the physiological process of continuous EV release during pregnancy. We
sought to address this caveat in two ways: first, we used an EV quantity for injection rather than
protein, and further, mimicked quantities found during pregnancy (Chapter 2) [20]. Additionally,

61

we administered EVs intravenously into the tail vein, as it more closely mimics the hematogenous
route that EVs travel during pregnancy. Maternal uterine venous blood flows to the iliac vein and
the inferior vena cava before entering the heart for oxygenation and recirculation throughout the
body [172, 173].
Intravenous administration of placental EVs allows for immediate circulation
to all organs. Notably, we observed localization of pregnant plasma in the lungs of recipient
animals but did not observe localization if plasma EVs from nonpregnant animals suggesting that
the localization to the lung is specific. Additionally, we found that intravenous injection of 25
billion placental EVs did not localize to uterine, renal, gastrointestinal, brain, heart, or spleen
tissues after 24 hours. Interestingly, our results did not agree with another study that administered
murine plasma EVs intraperitoneally [107]. Murine late-term plasma-derived EVs administered
intraperitoneally into mid-gestation mice localized to the uterus and placenta, and further, induced
preterm birth. The authors of that study concluded that these EVs must play a role in parturition;
however, given the lack of localization of EVs to the uterus when delivered intravenously in our
study, it seems unlikely that placental EVs are the cause of their findings.

Collectively, we have established that placental EVs preferentially traffic to maternal pulmonary
interstitial macrophages, and Kupffer cells in vivo. We demonstrate that integrin 
3
1 is necessary
for localization to the lung interstitium and integrin 
5
1is necessary for localization to the liver.
We also identify that the intravenous administration of pEVs do not target other peripheral organs
by whole-organ imaging. Future work may seek to identify how placental EVs influence maternal
interstitial macrophages and Kupffer cells, as well as the physiology of the lung and liver, during
pregnancy.

62

CHAPTER 4

A REPORTER MODEL FOR IN VIVO TRACKING OF PLACENTAL

EXTRACELLULAR VESICLES IN MURINE PREGNANCY

Sean L. Nguyen1,2, Soo Hyun Ahn3, Jacob W. Greenberg4, Benjamin W. Collaer5, Margaret G.
Petroff1,3,4

1Cell and Molecular Biology Program, Michigan State University, East Lansing, MI, USA
2Institute for Integrative Toxicology, Michigan State University, East Lansing, MI, USA
3Department of Pathobiology Diagnostic Investigation, Michigan State University, East Lansing,
MI, USA
4Department of Microbiology and Molecular Genetics, Michigan State University, East Lansing,
MI, USA
5Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing,
MI, USA

This research was supported by NIH grant R21HD091429, MSU AgBioResearch (NIFA

MICL02447), and MSU funds (MGP). SN was supported by the Integrative Pharmacological
Sciences Training Program grant T32GM092715. The funders had no role in study design, data

collection, and analysis, decision to publish, or preparation of the manuscript.

63

4.1 ABSTRACT

Extracellular vesicles (EVs) are (40-150 nm) membrane-enclosed nanostructures that contain
RNAs, proteins, and lipids that travel to different parts of the body and serve as a form of intercellular
communication. Using human and mouse models, we previously established that the placenta
secretes many EVs directly into the maternal blood during pregnancy. Further, we found that
placental EVs can be taken up by macrophages in organs distant to the placenta, most notably
the liver and lung. However, previously employed methods for identifying the cells targeted by
EVs require exogenous labeling and adoptive transfer of a bolus of purified EVs, which does
not necessarily recapitulate what occurs physiologically. Additionally, exogenous administration
of labeled EVs does not reveal whether cells received EV cargo. In this study, we employed a
Cre-loxP system in which the placenta and fetus constitutively express Cre recombinase, as well
as a maternal reporter gene by which placentally-derived EVs can be identified. When released
endogenously from the placenta during pregnancy, EVs expressing Cre recombinase are taken up
in reporter cells which undergo recombination. Here, we establish a model for identifying fetal
cells and placental EVs in vivo and propose a framework for studying maternal-fetal interactions
without the need to administer endogenously labeled EVs. We demonstrate that placentas release
EVs that are directly detectable in maternal lung tissue and provide in vitro evidence of placental
Cre-mediated recombination of recipient reporter bone marrow dendritic cells.

4.2 INTRODUCTION

The placenta is the critical organ that is responsible for the growth and development of the fetus
during pregnancy. Placental trophoblast cells in mice and humans are in direct contact with ma-
ternal decidual tissue and blood throughout pregnancy, and the multinucleated syncytiotrophoblast
layer secretes copious amounts of extracellular vesicles (EVs), which are readily detectable in
maternal blood [80]. Exosomes are 40-150 nm membrane-enclosed vesicles that arise from the
invagination of multivesicular endosomes [71]. Exosomes contain molecular cargo in the form of
RNAs, proteins, and lipids of the cell from which they are derived, and can induce physiological

64

changes in recipient cells both in vitro [46, 174, 106, 175] and in vivo [176, 110, 47, 59]. Placental
EVs (pEVs) have potential immunomodulatory properties due to their expression of paternally
inherited antigens, as well as programmed death ligand-1 (PD-L1/B7-H1), FasL, and other im-
munosuppresors on their surface [40, 24]. Placental EVs have also been implicated in activating
peripheral blood leukocytes and endothelial cells in preeclampsia [97, 134, 175, 150].

While many functions of placental EVs have been characterized using in vitro methods, fewer
studies have elucidated their roles in vivo. One way of studying placental EV function in vivo
is to label them with a lipophilic dye and administer them into the biological system of choice.
This approach can provide valuable insight into identifying placental EV cellular targets in vivo.
As seen in Chapter 3, we demonstrated that placental EVs use integrins on the outer membrane
surface to mediate trafficking to the interstitial macrophages in the lung and Kupffer cells in the
liver. However, a limitation of this approach is that there is variation in localization depending on
route of administration: labeled plasma EVs from pregnant mice injected intraperitoneally (i.p.)
can localize to maternal reproductive tract tissues and fetal compartments [107], while intravenous
injection reveals localization almost exclusively to the liver and lung, with no localization to uterine
tissues (Chapter 3) [106, 109]. In addition, recapitulation of continuous release of EVs, as the
placenta does, is a significant technical challenge, and single bolus injections do not necessarily
reflect what occurs in vivo.

In this chapter, we propose an in vivo model that reveals the true trafficking patterns of placental
EVs during pregnancy. This model uses the transgenic Cre/loxP reporter system [177, 178], wherein
cells undergo a permanent genetic change in the presence of Cre recombinase. Prior studies have
used this system to show that EVs can deliver Cre recombinase mRNA and/or protein to reporter
cells and tissues, thereby inducing recombination and expression of a reporter gene and demarcating
the cells that internalize EVs from those that did not [176, 47, 110, 59]. Using this approach, we
studied maternal-fetal interactions during unmanipulated pregnancy in vivo. This model allows
for the identification of fetal placental EV interactions with recipient maternal cells and provides a
paradigm for studying placental EV function in vivo.

65

4.3 MATERIALS AND METHODS

4.3.1 Mouse studies

All animal studies were approved by the Michigan State University Institutional Animal Care
& Use Committee (IACUC), and experiments were performed according to the approved animal
use form (AUF) protocol #201800176. B6.129(Cg)-Gt(ROSA)26Sortm4(ACTB-tdTomato,-EGFP)Luo/J
(mT/mG) mice (Stock No. 007676) were used as dams and B6.C-Tg(CMV-Cre)1Cgn/J (CMV-Cre)
mice (Stock No. 006054) or WT C57 BL/6J (Stock No. 000664) were used as sires and purchased
from Jackson Laboratories (Bar Harbor, ME). Mice were maintained as homozygotes and were
housed in a controlled environment with a 12:12 light-dark cycle and food and water available
ad libitum. For breeding studies, CMV-Cre or WT sires were paired with mT/mG dams, and
the presence of the copulatory plug was designated gestational day 0.5. Mice were euthanized
by cervical dislocation followed by bilateral pneumothorax and organs and were harvested for
experimental analysis.

4.3.2 Placental EV isolation

Placental EVs were isolated from GD14.5 placental explant culture as described previously with
modifications [40]. Briefly, placentas from GD14.5 mice were cut into four equal pieces and
placed inside12-well culture plates fitted with 74 
m membrane Netwell inserts (CLS3479-48EA,
Corning, Manassas, VA) filled with 3 ml of EV culture medium consisting of RPMI 1640 (Gibco,
Waltham, MA) supplemented with 1x antibiotic antimycotic solution (15240096 Gibco, Waltham,
MA) 1x Sodium Pyruvate, 1x 
 mercaptoethanol, and 10% heat-inactivated, EV-depleted fetal
bovine serum (FBS) (A2720801, Gibco, Waltham, MA). Placental explants were cultured for 18
hrs, and media was collected and centrifuged twice at 500 xg at room temperature (RT) for 15
minutes to pellet cells and debris. The supernatant was collected and centrifuged twice at 2000 xg
RT to further pellet smaller debris. The supernatant was collected and filter sterilized through a
0.22 
m PVDF filter (SCGP00525, EMD Millipore, Burlington, MA) to remove EVs larger than

66

220 nm. Placental explant conditioned medium (∼25 ml/prep) was concentrated with centrifugal
columns with 100 kDa cut off (VS2042, Sartorius, Bohemia, NY) centrifuged at 3220 xg to ∼50

l before being frozen at -80°C for downstream analysis. For placental EV isolation, concentrated
conditioned medium was loaded onto IZON qEVoriginal 35 nm size exclusion chromatography
columns (SP5, IZON Biosciences, Cambridge, MA) according to manufacturer’s instructions. EVs
were eluted in the fractions 5-7 with contaminating non-EV proteins eluted in later fractions [82].
The eluted EV-fractions were then concentrated in 4 kDa centrifugal columns (UFC801024, EMD
Millipore, Burlington, MA) at 3220 xg to 100 
l and used for downstream analysis.

4.3.3 Western blot

Placental tissue samples were flash-frozen in liquid nitrogen and stored at -80°C for downstream
analysis. Tissue samples were homogenized in RIPA buffer supplemented with 2 
g/ml aprotinin,
5
g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride (PMSF). Protein samples were quan-
tified using a Qubit 3 fluorometer (Life Technologies, Carlsbad, MA). For Western blot, samples
were incubated with 6X loading buffer containing dithiothreitol (DTT) and heat-denatured at 98°C
for 15 minutes and were loaded (10 
g/lane) on precast 4-20% SDS-PAGE gels (4568093, Bio-Rad,
Hercules, CA) with dual standard plus ladder (4568093, Bio-Rad, Hercules, CA). Samples were
run for 35 minutes at 200V, and protein was transferred onto nitrocellulose membrane for 1 hr at
100V. Membranes were blocked for 20 minutes in superblock solution at RT before being incubated
with 1:5000 rabbit anti-EGFP (TA150032, Origene, Rockville, MD) or 1:1000 rabbit anti-TSG101
(ab30871, AbCam, Cambridge, MA) in tris buffered saline (TBS) with 3% skim milk overnight at
4°C. Membranes were washed three times for 5 minutes in TBS with 0.05% Tween (TBS-T) at RT
and incubated with 1:2000 horseradish peroxidase (HRP) conjugated goat anti-rabbit secondary
antibody (7074S, Cell Signaling Technologies, Danvers, MA) for 1hr at RT. Membranes were
washed three times with TBS-T for 5 min and incubated with chemiluminescent substrate (34580,
Thermo Fisher, Waltham, MA) for 5 min before being imaged on a digital iBright membrane imager
(Thermo Fisher, Waltham, MA).

67

4.3.4 Tissue processing, immunostaining, and confocal microscopy

All tissues were fixed in 4% paraformaldehyde (Sigma, St Louis, MO) in phosphate buffered saline
(PBS) pH7.4 (Corning, Manassas, VA). For fetomaternal interface imaging, fixed uteroplacental
tissues from CMV-Cre;mT/G and WT;mT/mG fetuses were embedded, and 2% agarose and 200
m
slices were cut by vibratome (Leica, Richmond, IL). Vibratome sections were cleared using a
modified CUBIC protocol, as described in Chapter 3 [179, 180, 181, 182]. Briefly, tissue sections
were cleared with CUBIC-1 (25% urea, 25% Quadrol, 15% Triton X-100 in water) solution with
10
g/ml 4,6-diamidino-2-phenylindole (DAPI) overnight at 37 °C, washed three times with PBS for
1 hr at RT. Sections were then cleared with CUBIC-2 (25% urea, 50% sucrose, 10% triethanolamine)
solution overnight and mounted directly on glass slides with CUBIC-2 solution and No.0 coverslips
(260300, Ted Pella, Redding, CA) were placed on top of sections and sealed with clear nail polish
[139]. For immunostaining, fixed lung samples were placed in 30% (W/V) sucrose solution in PBS
overnight before being embedded in optimal cutting tissue (O.C.T.) compound (Sakura Finetek,
Torrance, CA) and flash frozen inside 2-propanol cooled with liquid nitrogen. Frozen embedded
tissues were cut (7
m slices) onto charged glass slides by cryostat. Cut sections were outlined
with a hydrophobic PAP-Pen (008899, Thermo Fisher, Waltham, MA) and blocked with a PBS
solution supplemented with 10% superblock and 3% goat serum for 20 min at RT. Sections were
then incubated with 1:1000 rabbit anti-EGFP primary antibody (TA150032, Origene, Rockville,
MD) overnight at 4°C. Slides were washed three times with 0.05% Tween PBS (PBS-T) for 5 min,
and sections were incubated with 1:200 secondary goat anti-rabbit conjugated to Alexa fluor 647
(A27040, Thermo Fisher, Waltham, MA) at RT in a humid environment. Slides were washed
three times with 0.05% Tween PBS (PBS-T) for 5 min, and sections were incubated with 1:200
secondary goat anti-rabbit conjugated to Alexa fluor 647 (A27040, Thermo Fisher, Waltham,
MA) at RT in a humid environment. Slides were washed three times with PBS-T for 5 min and
mounted with mounting medium containing 4,6-diamidino-2-phenylindole (DAPI) (H-1200-10,
Vector, Burlingame, CA) and glass coverslips. CUBIC-cleared tissue, immunostained lung section,
and co-cultured dendritic cells were imaged on a Nikon A-1 confocal microscope fitted with a 20X

68

and 40X oil objectives lenses and images were processed using FIJI image analysis software [141].

4.3.5 mT/mG bmDC explant culture

To determine if placentas could induce recombination in vitro, we isolated bone marrow dendritic
cells from mT/mG mice. Bone marrow was harvested using a previously established protocol with
modifications [183]. Briefly, bone marrow was isolated from the femurs and tibias of mT/mG
females, and 10 million cells/well were seeded on a six-well plate with 4 ml of complete culture
medium as described above with 20ng/ml murine GM-CSF (576302, BioLegend, San Diego, CA)
and was cultured at 37 °C. On the second day, half the medium was removed, and 2 ml of new
complete culture medium with 40ng/ml GM-CSF was added and cultured at 37 °C. The culture
medium was removed entirely on day three and replaced with fresh medium containing 20ng/ml
GM-CSF and was cultured for an additional three days before downstream analysis. For explant
culture, 100,000 dendritic cells were seeded in 12 well plates, and (colorless) CMV-Cre or WT
placentas were co-cultured overnight in Netwell inserts as described above. Following overnight
culture, the placentas were removed, and dendritic cells were cultured for an additional five days
to allow for sufficient time for recombination [20]. 500 
l of new EV free media was added every
other day to maintain cell viability. For microscopy experiments, cells were seeded on to coverslips
before placental co-culture. Cells were then fixed with 4% PFA and stained with 1:5000 rabbit
anti-EGFP (Origene) and anti-rabbit AF647 (10
g/ml). Coverslips were mounted with mounting
medium with DAPI (Vector) on glass slides and were imaged on an A1 Confocal microscope.

4.3.6 Flow cytometry

To characterize bmDC expression of mTomato or mGFP fluorescence, co-cultured dendritic cells
were removed from culture plates with cell scrapers and stained with live/dead yellow viability
stain (L34959, Thermo Fisher, Waltham, MA) for 15 min on ice. Cells were then blocked with
TruStain FcX block (Biolegend, San Diego, CA) for 10 min and stained with anti-CD45, CD11c,
and MHC -II antibodies (BioLegend, San Diego, CA) for 30min on ice. Cells were then washed

69

three times with flow staining buffer made of 1x PBS supplemented with 5% FBS and 2mM
sodium azide. Single color controls were made with individual antibodies along with eUltracomp
beads (Thermo Fisher, Waltham, MA), and spleen cells from mTomato and mGFP mice were
used as endogenous fluorescence controls. Stained cells were analyzed by an Aurora spectral flow
cytometer (CytekBio, Fremont, CA), and data were analyzed by Kaluza flow cytometry software
(Beckman Coulter, Pasadena, CA).

4.3.7 Recombination genotyping

To detect the presence of genomic recombination in co-cultured mT/mG dendritic cells, DNA was
isolated with a Quick DNA/RNA miniprep kit (D7005, Zymo, Irvine, CA) following the manufac-
turer’s instructions. DNA was amplified using a modified two-step polymerase chain reaction proto-
col targeting the unrecombined mTomato locus and recombined mGFP locus from mT/mG reporter
mice [184]. The mTomato locus was amplified with mT-F 5’-GCAACGTGCTGGTTATTGTG-3’
and mT-R 5’-TGATGACCTCCTCTCCCTTG-3’ primers yielding a 200bp amplicon. The mGFP
locus was amplified with mG-F 5’-GTTCGGCTTCTGGCGTGT-3’ and mG-R 5’-TGCTCACGGA
TCCTACCTTC-3’ primers yielding a 376 amplicon. Genomic DNA from dendritic cells was
amplified for 35 cycles, and 3
l of the PCR product was used as the template input for a second
round of PCR. Genomic tail DNA from WT, mTomato, and mGFP mice were used as positive and
negative controls for both PCR reactions. PCR products were run on 1.5% agarose gel with 100bp
ladder (New England Biolabs, Cambridge, MA) and visualized on an iBright digital gel imager
(Thermo Fisher, Waltham, MA).

4.4 RESULTS

4.4.1

Identification of placental EVs during pregnancy

We previously demonstrated that GD14.5 placental EVs injected intravenously (i.v.) traffic pri-
marily to lung and liver tissue in vivo. However, a single bolus injection of purified EVs does
not necessarily recapitulate the sustained release of EVs from the placenta, and does not reflect

70

the sustained quantities of placental EVs present in maternal circulation across all of pregnancy
(Chapter 3). To circumvent this caveat, we developed a model in which fetal EVs could be detected
within the context of normal pregnancy, using mT/mG transgenic female mice as dams. These
mice express the mT/mG reporter construct within the ROSA26 locus under the control of the
chicken actin promoter and the cytomegalovirus (CMV) enhancer (pCA), such that all cells consti-
tutively express the membrane-targeted red fluorescent protein, tandem dimer Tomato (mTomato),
in the absence of Cre recombinase (Figure 4.1) [185]. In the presence of Cre recombinase, the
maternal mTomato locus is excised, resulting in the expression of membrane-targeted enhanced
green fluorescent protein (EGFP; mGFP) (Figure 4.1). Homozygous mT/mG dams were mated
with CMV-Cre sires in which an X-linked Cre recombinase is expressed under the control of the
cytomegalovirus promoter. In this system, all fetuses inherit maternal mT/mG and female fetuses
inherit the CMV-Cre locus, and thus express mGFP in all tissues including the placenta; male em-
bryos inherit the WT locus and continue to express the mTomato red fluorescence protein (Figure
4.2 A).

71

Figure 4.1: Cre-mT/mG Model of Recombination. Representative mT/mG reporter construct in
which all unrecombined mTomato red fluorescence on the surface in the absence of Cre recombi-
nase. In the presence of Cre recombinase, the mT/mG reporter construct will undergo Cre-loxP
mediated excision of the mTomato locus such that the membrane-targeted enhanced green fluores-
cent protein will be expressed. Arrow represents transcriptional direction. Triangles represent loxP
sites in which Cre-mediated recombination occurs. Red octagon represents a stop codon sequence.

Using this system, we characterized placentas and purified pEVs at gestational day (GD) 14.5
expression for EGFP. EGFP was readily detectable by Western blot analysis in both placentas and
pEVs from CMV-Cre;mT/mG but not WT;mT/mG fetuses (Figure 4.2 B); both CMV-Cre;mT/mG
and mT;mT/mG placentas and pEV samples expressed the control protein, TSG101, which is
associated with multivesicular body genesis and enriched in exosomes (Figure 4.2 B) [71]. We then
confirmed the distinct expression of fetal mGFP and maternal mTomato expression at the GD14.5

72

maternal-fetal interface by confocal microscopy (Figure 4.2 C). Placental expression of mGFP was
prominent against the background of maternal mTomato in the decidua. We were additionally
able to visualize the glycogen trophoblast cells that had invaded the decidua [11]. These results
confirmed that female fetuses inherited paternal Cre recombinase and underwent Cre-mediated
recombination, and further, that placental EVs expressed the recombined mGFP protein (Figure
4.2). Next, we asked whether fetal mGFP+ EVs within plasma of the dams in this model carried
EGFP in the lung.

Figure 4.2: CremT/mG Model to Study Fetal EVs in pregnancy. A. An mT/mG female mated
to a CMV-Cre (X-linked) male gives rise to female pups ubiquitously expressing mGFP and male
pups expressing mTomato. B. Western blot of GD14.5 placenta (Plac) and placental EVs (pEVs)
from mT/mG females mated with CMV-Cre males. C. GD14.5 uterine and placental interface, fetal
mGFP-expressing placental cells are readily distinguishable from maternal mTomato-expressing
decidual cells; Dec, decidua; Sp, spongiotrophoblast; glycogen trophoblast cell, GTC. The dashed
line demarcates the fetomaternal interface. Representative image of three independent experiments.

73

4.4.2 Fetal EV trafficking to maternal lung in vivo

After confirming the expression of mGFP in CMV-Cre;mT/mG placentas, and pEVs, we asked
whether fetally-derived mGFP expressing pEVs trafficked to maternal lungs, as they do when
injected intravenously. We collected lung from GD14.5 CMV-Cre mated mT/mG females and
analyzed them by confocal microscopy. We were able to readily detect mGFP fluorescence in
maternal tissues (Figure 4.3 A). Most of the signal observed was punctate and associated with
maternal mTomato-expressing cells (Figure 4.3 A), similar to what was observed when placental
EVs were administered intravenously (Chapter 3).
Interestingly, these foci were less than 10
microns in diameter and not associated with DAPI nuclear stain, thus suggesting that they were not
recombined, mGFP-expressing cells (Figure 4.3 A). However, we also observed mGFP expression
that was clearly membrane-associated, surrounding distinct DAPI-stained nuclei (Figure 4.3 B, C).
These cells did not express mTomato, clearly indicating that these cells had undergone Cre-mediated
recombination. mGFP-associated foci and cells were observed only in mT/mG dams mated to CMV-
Cre males, and not in those mated to WT males. Thus our mT/mG reporter model of pregnancy
demonstrates how fetal EVs traffic to the maternal lung within the context of unmanipulated normal
pregnancy.

Recombined mGFP-positive cells in the lungs could represent either fetal cells that trafficked
from the fetus/placenta to the maternal lungs (microchimerism) [105, 186, 187], or they could be
maternal cells that underwent Cre-mediated recombination as a result of EVs carrying Cre mRNA
or protein to recipient cells (Figure 4.3 C) [110, 188, 176].

74

Figure 4.3: In Vivo Localization of mGFP in CMV-Cre Mated Maternal mT/mG Lung A.
Confocal image of punctate mGFP-positive foci localization in maternal GD14.5 lung. B. Confocal
localization of mG+, recombined cell in GD14.5 lung. C. Confocal image of mGFP+ recombined
cell in GD14.5 lung. Representative of five biological replicates. White arrows; punctate mGFP-
positive foci. Yellow arrows; mGFP-positive (mT-negative), recombined cells.

75

4.4.3

In vitro placental EV model of recombination

To determine if placental EVs could induce Cre-mediated recombination of reporter target cells,
we developed a model in which mT/mG bone marrow-derived dendritic cells (DC) are co-cultured
with GD14.5 (colorless) CMV-Cre placentas or WT placental explants (Figure 4.4 A). Explants
were separated from the dendritic cells by a 70-micron insert. In this way, EVs released into the
medium could taken up by underlying dendritic cells; Cre-mediated recombination of the dendritic
cells was used as an endpoint. After 18 hours, placentas were removed from the co-cultures, and
dendritic cells were cultured for an additional five days to allow sufficient time for maximal Cre
recombination to occur [185, 110, 188] and were subsequently screened for recombined mGFP
positive dendritic cells (Figure 4.4 A). Flow cytometric analysis revealed no changes in mTomato
expression in mT/mG dendritic cells treated with WT placentas (Figure 4.4 B). However, mT/mG
dendritic cells treated with CMV-Cre placentas resulted in a shift of mTomato-positive to mTomato-
negative and mGFP negative population (17.3%) as well as an increase in the proportion of mGFP
positive cells (0.53%) (Figure 4.4 B). Confocal microscopy revealed the presence of cell-associated
and non-cell-associated mGFP, which also colocalized with mTomato (Figure 4.4 E). To confirm
this observation, the same cells were also stained with an anti-GFP antibody. In dendritic cells
co-cultured with CMV-Cre placentas, but not those cultured with WT placentas, cells positive
for mGFP and noncellular mGFP foci stained positive for anti-GFP antibody (Figure 4.5 C-E).
Interestingly, the morphology of dendritic cells co-cultured with CMV-Cre placentas revealed large
clusters of cells that were not observed in dendritic cells co-cultured with WT placentas (Figure
4.5 C-E). Both the flow cytometry and confocal microscopy data provide evidence of Cre-mediated
recombination of mT/mG dendritic cells as control treatments did not result in mGFP expression.

76

Figure 4.4: In Vitro Transfer of Placental Cre to Recipient Cells. A. mT/mG bmDCs were
cultured with (colorless) GD14.5 CMV-Cre placentas for 24 hrs and cultured for an additional five
days. B. Representative flow cytometry plots of mT/mG bmDCs treated with CMV-Cre or WT
placentas and cultured for five days. C. Representative confocal microscopy of reporter bmDCs
co-cultured with CMV-Cre or WT placentas. Arrows, mG foci. D. mT/mG bmDCs treated with WT
placentas. E. mT/mG bmDCs treated with CMV-Cre placentas. Representative of three independent
experiments.

To further confirm the above results, we isolated the DNA from the treated mT/mG dendritic
cells and performed PCR with primers specific for the unrecombined and recombined mTomato
and mGFP loci (Figure 4.5 A) [184]. To increase the sensitivity of detecting the recombined
mGFP locus, we performed two rounds of PCR on co-cultured bmDC DNA. We observed the
mGFP amplicons in DNA isolated from mT/mG dendritic cells treated with CMV-Cre but not in
that of DNA from WT placentas indicating that dendritic cells underwent genomic recombination
(Figure 4.5 B). Our results demonstrate a proof of concept model for identifying recipient cells that

77

internalize placenta EVs and provides an additional mechanism for studying how placental EVs
induce biological effects on recipient cells.

Figure 4.5: Validation of Placental EV mediated genomic recombination. A. Schematic dia-
gram of forward (mT-F/mG-F) and reverse (mT-R/mG-R) primers for identification of unrecom-
bined mTomato and recombined mGFP. PCR amplification of genomic DNA with these primers
results in differences in PCR products for non-recombined and recombined cells (200bp and 376bp,
respectively). B. Genotyping electrophoresis gel of mGFP or mTomato loci from genomic DNA
from purified wild type (WT), mTomato (mT), mTomato (mT) and mGFP tail DNA, or dendritic
cells treated with WT placenta or CMV-Cre placenta. No template control, NTC. Representative
gel image of five independent experiments.

4.5 DISCUSSION

Placental EVs can impact the function of a wide range of cells including macrophages, endothe-
lial cells, smooth muscle, bone, and dendritic cells. Many biological effects of EVs demonstrated
in vivo can be confirmed in vivo by adoptive transfer of EVs. However, little is unknown about the
bona fide in vivo targets of placental EVs, as no studies to date use unmanipulated experimental
systems that can track endogenous release and trafficking of EVs vivo [189, 151, 50, 190]. Here,
we developed a model for studying placental EVs in maternal tissues in vivo without the need to
isolate, label and administer them. In this system, fetal tissues and EVs derived from them express
mGFP as a reporter that allows us to determine their in vivo tissue and cellular targets. Putative

78

EVs could be detected as punctate foci within the maternal lung during pregnancy. Their small
size, punctate distribution, and their localization to the maternal lung mirrored our prior findings
of intravenous injection of exogenously derived placental EVs (Chapter 3) [106, 109]. We propose
that this model is most representative date of continuous placental EV release and trafficking, and
that it will allow further studies on the effects of EVs on maternal physiology in vivo [40, 20].

Microchimerism may also be influenced by surface receptors present on fetal cells. EVs from
metastatic tumors traffic to specific organs through the expression of integrins, and establish a
metastatic niche [53, 54]. In chapter 3, we demonstrated that placental EVs expressing 
3
1, 
5
1
and 
V
3 integrins and localize to maternal lung tissue, and our results here align with the idea
that EVs may prime the maternal lung environment for microchimeric fetal cells. Consistent with
this idea is the fact that when trophoblast cells become cancerous, a prime location for metastasis
is the lung. Further studies will clarify the function of placental EVs on maternal pulmonary
function, as well as their possible role in promoting microchimerism or metastasis in pregnancy
and choriocarcinoma, respectively.

We also identified whole cells in the lung that had recombined to express mGFP but not
mTomato. Recombined mGFP cells in the maternal lung may arise from two mechanisms: fetal
microchimerism or Cre-mediated recombination of maternal lung cells. Fetal microchimerism -
fetal cells that traffic into maternal tissues was first described with the presence of fetal trophoblast
cells residing in the maternal lung over 125 years ago [133], and has been confirmed by numerous
studies since [105, 191, 132, 192]. Fetal microchimerism is persistent: male fetal progenitor cells
are detectable in maternal blood after as many as 27 years postpartum [186]. In contrast, murine
models of microchimerism suggest that fetal nucleated cells are rarely detectable in maternal blood
[193]. Murine models of pregnancy using wild-type dams together with GFP-expressing sires
confirmed persistent microchimerism of fetal cells in the maternal lung, although similar to our
results, but not readily in the liver [105, 194]. Tissue specificity of microchimerism may depend on
the conditions, however, since ethanol-induced injury promoted fetal microchimerism to the liver
and kidney in a rat model [191].

79

A second explanation for the presence of recombined mGFP-expressing cells in the maternal
lung may arise from a novel method in which EVs contain Cre recombinase RNA or protein ,
and target cells that express a Cre reporter internalize these EVs undergo recombination, which
results in phenotypic conversion. Thus, in our system, maternal lung cells may have internalized
Cre RNA- or protein-containing EVs from the fetal-placental unit, and subsequently undergone
recombination to exhibit green fluorescence. Preliminary experiments identified the presence of
Cre mRNA in placentas and placental EVs, experiments or currently in progress verify these
results. To examine whether this was possible in principle, we used an explant culture model in
which CMV-Cre placentas shed vesicles, and appeared to have induced recombination of reporter
mT/mG-derived dendritic cells. Future work may seek to identify Cre-recombinase in EVs and
use adoptive embryo transfer models of colorless CMV-Cre embryos; alternatively, intravenous
injection of colorless CMV-Cre placental EVs into recipient mT/mG dams. This approach will rule
out the possibility of microchimerism explaining the presence of mGFP-expressing cells in this
model, as mGFP expression could only result from maternal recombination.

Collectively we demonstrate a new model for further expanding the field of understanding
placental EV interactions in vivo and provide a framework for visualizing maternal-fetal interactions
without the use of antibody labeling. Further, this system offers advantages to traditional methods
of understanding placental EV kinetics in vivo as purified EVs do not need to be purified and
administered to recipient animals. Future studies will identify the maternal cell types that colocalize
with placental EVs and determine how placental EVs induce changes in recipient cells in vivo.

80

CHAPTER 5

GENERAL DISCUSSION

5.1 SUMMARY OF FINDINGS

Increasing evidence has identified roles for extracellular vesicles in cell-to-cell communication,
immunomodulation, signal pathway activation, and diagnostic use as a biomarker for disease
detection [46, 70, 52, 110]. Within the context of pregnancy, the placenta secretes copious amounts
of extracellular vesicles into the maternal circulation. Placenta EVs are enriched for microRNAs,
proteins such as fibronectin, Fas ligand, TRAIL, and PD-L1 [50, 94, 70, 24]. While the putative
functions of placenta extracellular vesicles in vitro have been described in the literature, the extent
to which these effects occur in vivo is not well characterized. Major gaps in knowledge in our
understanding placental EV function during pregnancy include 1) identification of placental EV
cellular targets and mechanisms mediating their binding and 2) computational methods to analyze
nanoparticle/EV quantification data and an experimental method to study placental EV function
within a normal physiological context. To address these gaps in knowledge, we used a murine
model of pregnancy to study placental EVs and implemented fluorescent labeling techniques to
identify their localization in vivo as well as transgenic mice to identify fetally derived placenta EVs
and cells (Figure 5.1).

81

Figure 5.1: Thesis Graphical Overview. Chapter 2 quantified EVs across gestation, chapter
3 identified the cellular localization of placental EVs, and chapter 4 established a model of for
studying placental EVs in vivo.

In chapter 2, we quantified maternal murine plasma EV concentrations across normal gestation
by nanoparticle tracking analysis and identified that plasma EVs increased with advancing with
increasing gestational age peaking at GD14.5. To address constraints in analyzing nanoparticle
tracking analysis data, we developed a computational software package tidyNano that facilitates
importing, processing, analyzing, and visualizing raw nanoparticle tracking analysis data. We then
used tidyNano to quantify maternal plasma EV concentration during normal gestation and in an
inflammation-induced model of preterm birth. Our lab has previously demonstrated that placental
EVs are detectable in the maternal plasma at GD15.5 and 17.5 in mice [40]. The results from this
chapter identified the significant increases in maternal plasma EV concentration at GD14.5 and 17.5,
which corresponded to placental mass, suggesting that the placenta contributes to total maternal
plasma EV quantities. We also identified a significant reduction in plasma EV concentration when
mice were administered lipopolysaccharide-induced preterm birth which may be attributed to acute

82

placental dysfunction resulting in a decrease in placental EV output. Future work may seek to
characterize placental EV output in response to insults in utero.

After identifying the gestational plasma EV profile in mice, we sought to identify the localization
of placental EVs in vivo. In chapter 3, we found that intravenous injection of EVs isolated from
pregnant mice into recipient nonpregnant mice localized to the lung, and surprisingly, did not
observe this localization when mice were injected with EVs from nonpregnant mice. This finding
suggests that pregnant EV localization to the lung is specific and not merely a result of high
vascularization in the lung. We also identified lung interstitial macrophages and liver Kupffer cells
as targets of placental EVs. Both interstitial macrophages and Kupffer cells are tissue-resident
macrophage populations capable of phagocytosis as well as antigen presentation. Future studies
may determine if placental EVs can be a source of fetal antigen to both macrophage cell types. In
addition to identifying the cellular localization of placental EVs in vivo, we determined that outer
membrane proteins were required for trafficking, as proteinase K treatment abrogated this effect.
Further, incubation of placental EVs with inhibitory peptides identified integrins 
3
1, 
5
1,
and 
V
3 as mediators for binding to laminin expressed in the lung and fibronectin expressed
in the liver [167, 195, 166, 147, 54]. Additionally, whole-organ imaging revealed that placental
EVs administered intravenously do not appear to localize to the spleen, kidney, heart, brain,
gastrointestinal tract, uterus, or lymph nodes. Placental EV protein expression may be altered
in placental pathologies such as preeclampsia or preterm birth and thus may influence integrin
expression. Future work may analyze placental EV protein expression in response to placental
insults. Supporting this idea, treatment of human term placental explant culture with environmental
pollutants polybrominated diphenyl ethers (PBDEs) and bisphenol A (BPA) did not alter the size,
quantity, morphology of pEVs secreted into the media but changed the pEV protein composition
phenotype of cellular injury [196].

In chapter 4, we established a model for studying placental EVs in vivo without the requirement
for experimental manipulation. In our system, we were able to identify fetal expression of mGFP
in placentas including placental EVs and demonstrated that fetal GFP foci localized to maternal

83

tissues, confirming our findings in Chapter 2. Interestingly, we also identified recombined mGFP-
positive/mTomato-negative cells in the maternal lung. The presence of recombined cells in the
lung could arise from two possibilities: 1) microchimerism of fetal recombined cells of presumed
placental origin migrated to maternal lung tissue or 2) EV mediated Cre recombination of maternal
lung cells. Multiple research groups have implemented EV-mediated Cre recombination reporter
systems with human cell culture lines and demonstrated EVs could induce recombination in re-
cipient cell reporters [188, 110, 176, 59]. Thus, to test the possibility of placental EV mediated
recombination, we co-cultured reporter mT/mG bone marrow derived dendritic cells with colorless
CMV-Cre placentas. We identified cells that express mGFP but also expressed mTomato and
confirmed our results by genotyping PCR. Future work may seek to identify the mechanism and
source of Cre in our model system. Preliminary experiments identified the presence of Cre mRNA
in placental EVs, but further experimentation is required to rule out the possibility for potential Cre
recombinase protein associated with EVs.

A limitation of our EV reporter model is that the CMV-Cre transgene was expressed on the X
chromosome, and only female fetuses in our model express mGFP and Cre. Thus, the number of
Cre-expressing mGFP positive EVs was derived from only ∼50% of the fetuses. To overcome this,
we attempted use of another strain as the sire that expresses Cre recombinase under the chicken
Beta-actin promoter, reasoning that all pups would inherit and express Cre. Interestingly, we did
not observe maternal recombined cells in the lungs of mT/mG dams mated to Beta-actin Cre
sires. Future work may seek to expand on this model by treating reporter mT/mG bone marrow-
derived dendritic cells with purified colorless CMV-Cre placental EVs to determine if EVs alone
can induce recombination. However, it was previously reported that only adding purified EVs
containing Cre mRNA was insufficient to induce recombination in reporter cells as recombination
only occurred with Cre cell co-culture [110]. In a separate model of studying EV recombination in
brain recipient cells, it was discovered that EV mediated Cre recombination was rare in the brain,
but was significantly increased during inflammation-induced injury [176]. Thus, it may be possible
that placental EV mediated maternal recombination may increase during inflammatory conditions

84

such as in preterm birth.

5.2 SCIENTIFIC CONTRIBUTION

General approaches to identify the localization of EVs in vivo include transfecting host cells to
produce EVs with reporter molecules such as luciferase or biotin the surface of EVs for detection in
vivo [197]. Intravenous injection of these modified EVs from human embryonic kidney cells (HEK)
into athymic nude mice resulted in the localization of EVs in the spleen and liver. Interestingly,
the authors found that the EVs were eliminated by hepatic and renal routes within six hours.
We demonstrated in chapter 3 that placental EVs specifically traffic to lung and liver tissues
and remain in these tissues after 24 hours. We also did not observe localization in the spleen,
heart, brain, uterus, kidney, lymph nodes, or gastrointestinal tract. EV isolation method also
influences localization to different tissues in mice. Injection of B16 F10 melanoma EVs isolated
by ultrafiltration and size exclusion chromatography will localize to liver and spleen compared to
EVs isolated by ultracentrifugation which localizes to lung, liver, and spleen [82]. Interestingly
our method of placental EV isolation uses ultrafiltration and size exclusion chromatography and
we were able to detect localization to the lung. Wiklander and colleagues identified in vivo
localization in mice differed based on the source of EVs and the route of administration [171].
In Chapter 3 we demonstrated that intravenous administration of pregnant plasma EVs resulted
in specific localization to the lung. Our results differed from other EV studies that administered
plasma EVs intraperitoneally which identified localization in the uterine and fetal compartments
[21]. We chose to administer EVs intravenously as placental EVs are readily detected in maternal
blood in both mice and humans [80, 40]. The rationale for intraperitoneal administration of plasma
EVs is difficult to reconcile given that plasma EVs would normally exist in the peritoneal cavity
during normal pregnancy. We did not detect the localization of placental EVs in the uterine and
fetal compartments when they were administered intravenously. Our work in chapter 4 provides
further evidence that the localization of placental EVs traffics to the lung during pregnancy.

The small nature of extracellular vesicles makes studying their function challenging when

85

compared to cell-based methods. To isolate placental EVs, we used conditioned medium from
placental explant cultures which resulted in a large quantity of EVs compared to in vitro trophoblast
cell culture. A limitation of this method is that it is not possible to determine the specific cellular
identity producing the EVs in the conditioned medium. Alternative approaches may be to use
cell type-specific promoters to drive expression to the syncytiotrophoblast cells. Gcm1 is a gene
responsible for syncytiotrophoblast formation in mice and may be a suitable driver for identifying
placental EVs derived from the syncytiotrophoblast cells [21]. A cell line-based approach is another
method of isolating placental EVs, however, most trophoblast cell lines are of human origin which
may not be conducive for working with transgenic murine models of pregnancy due to the xenograft
incompatibility. An alternative approach is to isolate murine trophoblast stem cells and differentiate
them to syncytial-like cells which would address the pEV purity and viability from pEVs derived
from explant culture [94]. A limitation of this approach as with other cell line approaches is that
the overall EV yield is lower. A limitation with EV trafficking studies is the requirement to label
EVs with fluorescent dyes to identify where they go in a biological system and this approach does
not. Alternative strategies to labeling exosomes directly is using a pH-dependent fluorescently
labeled tetraspanin reporter to directly label EVs as they are released [198, 199]. This method
takes advantage of the slightly acidic environment (∼pH5.5) of the multivesicular body to quench
the optical reporter CD63-pHluorin. When the multivesicular endosome fuses with the plasma
membrane, the exosomes become immediately detectable in the neutral (∼pH7.4) extracellular
space by fluorescence microscopy. Future work may implement this approach to generate a mouse
model in which only EVs are fluorescently labeled and are only produced by syncytiotrophoblast
cells.
In this way, the presence of mGFP foci would only arise from EVs and not from fetal
microchimerism.

5.3 POSSIBLE FUNCTIONS OF PLACENTAL EXOSOMES

This dissertation provides evidence for specific trafficking of pregnancy associated exosomes
to the lung in non-pregnant recipient animals. More specifically, placental exosomes traffic to

86

lung interstitial macrophages in vivo and removal of exosome outer membrane proteins reduced
localization to the lung including interstitial macrophages. It remains to be seen if these integrins
are required for trafficking to pulmonary interstitial macrophages. Placental exosome trafficking
to the lung appears to be specific, however, how the exosomes influence pulmonary physiology in
pregnancy remain to be seen. Interstitial macrophages possess antigen presentation capability and
may be a mechanism by which the maternal immune system may be primed for paternally inherited
antigen expressed on the surface of placental exosomes [145, 40]. Another possible function of
placental exosomes on maternal lung cells is inducing inflammatory response. Exosomal integrins
from human cancer cells induced the upregulation of pro-inflammatory S100 genes including
S100A4, -A6, -A10, -A11, -A13 and -A16 in human bronchial epithelial cells but this effect was not
observed with exosomes lacking integrins 
4 or 
5 [54]. Interestingly, S100 genes are associated
with cancer metastasis and induction of these genes may prime organs to become a pre-metastatic
niche for tumor cell migration. Microchimerism of fetal trophoblast cells to maternal lung tissue
resembles tumor metastasis to distant organ sites and the work in this dissertation has identified
how placental exosomes share similar integrins as tumor exosomes. Future work may determine
how placental exosomes influence the upregulation of pro-inflammatory gene expression in whole
lung or in vitro lung cell line approaches. Pregnant women are often more susceptible to exhibiting
a more severe immune response to pulmonary infections such as influenza virus or even severe
acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and this effect may be due to the presence
of placental exosomes [200, 201]. The evolutionary advantage of placental exosomes trafficking
to the lung is not readily clear however, it is notable that these exosomes appear to interact
with interstitial macrophages. In addition antigen presentation capabilities, these resident tissue
macrophages exhibit a tolerizing phenotype and maintain lung homeostasis which may serve to
counteract some pro-inflammatory effects from placental exosomes [145]. Circulating plasma
exosome concentration changes throughout pregnancy (chapter 2) and recent evidence suggests
that plasma exosome profile changes from a noninflammatory to pro-inflammatory profile towards
the later have of pregnancy. Future work may also examine the effect of placental exosomes from

87

different gestation periods and their localization in vivo.

5.4 CONCLUDING REMARKS

In this dissertation, we have contributed to the understanding of placental extracellular traffick-
ing dynamics using two separate approaches: 1) Intravenous injection of placental EVs identified
targeting to lung interstitial macrophages and Kupffer cells. We also identified that placental EV
trafficking to lung and liver tissue was specific and that integrins mediate their localization. 2)
A transgenic model to identify fetal placental EV trafficking in vivo without the requirement for
experimental manipulation. This model demonstrated that placental EVs could functionally alter
recipient cells ex vivo. Collectively, the work described in this dissertation further expands on
our understanding of the mechanism by which placental EVs interact with the maternal cells and
provides a computational framework for quantifying EVs as well as an experimental mouse model
to examine fetomaternal EV interactions (Figure 5.1).

88

APPENDICES

89

APPENDIX A

CHAPTER 2 SUPPORTING FIGURES

Figure A.1: Transmission electron microscopy of isolated exosomes. Representative electron
micrographs of plasma exosomes isolated by Total Exosome Isolation reagent of A. non-pregnant
and B.C. GD14.5 exosomes. Scale bar represents 100 nm.

90

Figure A.2: Polystyrene bead data analysis. Sample workflow of importing data into R using
tidyNano functions.

91

Figure A.3: NTA data import with nanotidy(). Raw count data from NTA .csv files can be
extracted and imported into the R environment with the nanotidy() function.

Figure A.4: Schema of experimental design. Each plasma exosome sample was diluted, separated
into two separate syringes, and was measured by NanoSight through recording of three 30-second
videos.

92

Figure A.5: Visualization of samples with technical replicate data. Faceted plot of plasma
exosome size and particle concentration of all samples (n = 76) in experimental study where each
sample was tested twice with three technical replicates.

93

Figure A.6: Visualization of samples with nanolyze(). Each summary output of the nanolyze()
aggregation function can be visualized. A. Plot of sample mean particle concentration from three
technical replicate measurements (n = 76). B. Plot of mean injection data from two syringe injection
measurements (n = 76).

94

Figure A.7: Placental mass across gestation. Placental mass across gestation in WT mated
C57B/6 mice. Points represent individual placentas.

95

Figure A.8: Sample NTA summary PDF file. Representative file containing sample acquisition
data following nanoparticle tracking analysis measurement.

96

APPENDIX B

CHAPTER 3 SUPPORTING FIGURES

Figure B.1: Isolation and validation of plasma EVs. A. Representative NanoSight size distri-
bution of plasma EVs from non-pregnant (NP) and gestational day (GD) 14.5 mice obtained by
size exclusion chromatography. Line represents mean concentration; shaded area represents SEM
of five technical replicates. B. Western blot analysis of plasma EVs from a GD14.5 mouse. Lanes
represent size exclusion chromatography fractions; WP, whole plasma; red box indicates fractions
used for in vivo experiments; arrow indicates TSG101 protein band. Representative data of three
independent experiments. C. Transmission electron microscopy of plasma EVs from a nonpregnant
mouse (representative image of n = 3).

97

Figure B.2: Validation of placental explant culture. A. Hematoxylin and eosin (H&E) blinded
histological analysis of placentas cultured in ambient and low (8% O2) oxygen levels. B. Repre-
sentative histogram of placental EVs analyzed by NanoSight NTA. /textbfC. Western blot analysis
of plasma EVs, lanes represent SEC fractions, red box indicates SEC fractions used for in vivo
experiments.

98

Figure B.3: Proteinase K inhibits pEV localization to lung interstitial macrophages. Flow
cytometric quantification of pEV localization to interstitial macrophages in lung, points represent
biological replicates. Welch
s T-test.

99

BIBLIOGRAPHY

100

BIBLIOGRAPHY

[1]

J A Whitsett. “Specializations in plasma membranes of the human placenta”. en. In: J.
Pediatr. 96.3 Pt 2 (Mar. 1980), pp. 600–603.

[2] Mario Sideri et al. “The Ultrastructural Basis of the Nutritional Transfer: Evidence of
Different Patterns in the Plasma Membranes of the Multilayered Placental Barrier”. In:
Fetal Nutrition, Metabolism, and Immunology: The Role of the Placenta. Ed. by Richard K
Miller and Henry A Thiede. Boston, MA: Springer US, 1984, pp. 15–25.

[3] Michelle Desforges and Colin P Sibley. “Placental nutrient supply and fetal growth”. en.

In: Int. J. Dev. Biol. 54.2-3 (2010), pp. 377–390.

[4] Eddy Solomon et al. “Major mouse placental compartments revealed by diffusion-weighted
MRI, contrast-enhanced MRI, and fluorescence imaging”. en. In: Proc. Natl. Acad. Sci. U.
S. A. 111.28 (July 2014), pp. 10353–10358.

[5] A I Csapo et al. “The significance of the human corpus luteum in pregnancy maintenance.
I. Preliminary studies”. en. In: Am. J. Obstet. Gynecol. 112.8 (Apr. 1972), pp. 1061–1067.

[6] G Thordarson et al. “Interaction of mouse placental lactogens and androgens in regulating
progesterone release in cultured mouse luteal cells”. en. In: Endocrinology 138.8 (Aug.
1997), pp. 3236–3241.

[7] P B Medawar. “Some immunological and endocrinological problems raised by the evolution

of viviparity in vertebrates”. In: 1953, pp. 320–338.

[8] A Malassiné, J L Frendo, and D Evain-Brion. “A comparison of placental development and
endocrine functions between the human and mouse model”. en. In: Hum. Reprod. Update
9.6 (Nov. 2003), pp. 531–539.

[9]

J C Cross, Z Werb, and S J Fisher. “Implantation and the placenta: key pieces of the
development puzzle”. en. In: Science 266.5190 (Dec. 1994), pp. 1508–1518.

[10]

J Rossant and J C Cross. “Placental development: lessons from mouse mutants”. en. In:
Nat. Rev. Genet. 2.7 (July 2001), pp. 538–548.

[11] Erica D Watson and James C Cross. “Development of structures and transport functions in

the mouse placenta”. en. In: Physiology 20 (June 2005), pp. 180–193.

[12]

Jonathan J Miner et al. “Zika Virus Infection during Pregnancy in Mice Causes Placental
Damage and Fetal Demise”. en. In: Cell 165.5 (May 2016), pp. 1081–1091.

101

[13] E Mourier et al. “Non-invasive evaluation of placental blood flow: lessons from animal

models”. en. In: Reproduction 153.3 (Mar. 2017), R85–R96.

[14] Henrieta Papuchova et al. “Three types of HLA-G+ extravillous trophoblasts that have
distinct immune regulatory properties”. en. In: Proc. Natl. Acad. Sci. U. S. A. (June 2020).

[15] S Lee Adamson et al. “Interactions between trophoblast cells and the maternal and fetal

circulation in the mouse placenta”. en. In: Dev. Biol. 250.2 (Oct. 2002), pp. 358–373.

[16] R Pijnenborg, L Vercruysse, and M Hanssens. “The uterine spiral arteries in human preg-

nancy: facts and controversies”. en. In: Placenta 27.9-10 (Sept. 2006), pp. 939–958.

[17] Koji Hayakawa et al. “Nucleosomes of polyploid trophoblast giant cells mostly consist of
histone variants and form a loose chromatin structure”. en. In: Sci. Rep. 8.1 (Apr. 2018),
p. 5811.

[18] Laura Woods, Vicente Perez-Garcia, and Myriam Hemberger. “Regulation of Placental
Development and Its Impact on Fetal Growth-New Insights From Mouse Models”. en. In:
Front. Endocrinol. 9 (Sept. 2018), p. 570.

[19] Dong Hu and James C Cross. “Development and function of trophoblast giant cells in the

rodent placenta”. en. In: Int. J. Dev. Biol. 54.2-3 (2010), pp. 341–354.

[20] Sean L Nguyen et al. “Quantifying murine placental extracellular vesicles across gestation
and in preterm birth data with tidyNano: A computational framework for analyzing and
visualizing nanoparticle data in R”. en. In: PLoS One 14.6 (June 2019), e0218270.

[21] David G Simmons and James C Cross. “Determinants of trophoblast lineage and cell
subtype specification in the mouse placenta”. en. In: Dev. Biol. 284.1 (Aug. 2005), pp. 12–
24.

[22] M M Shida et al. “Placental-specific expression from the mouse placental lactogen II gene

promoter”. en. In: Proc. Natl. Acad. Sci. U. S. A. 89.9 (May 1992), pp. 3864–3868.

[23] Stephanie E Ander, Michael S Diamond, and Carolyn B Coyne. “Immune responses at the

maternal-fetal interface”. en. In: Sci Immunol 4.31 (Jan. 2019).

[24] S K Kshirsagar et al. “Immunomodulatory molecules are released from the first trimester

and term placenta via exosomes”. en. In: Placenta 33.12 (Dec. 2012), pp. 982–990.

[25] Mary K Collins, Chin-Siean Tay, and Adrian Erlebacher. “Dendritic cell entrapment within
the pregnant uterus inhibits immune surveillance of the maternal/fetal interface in mice”.
en. In: J. Clin. Invest. 119.7 (July 2009), pp. 2062–2073.

102

[26] Ashley Moffett and Francesco Colucci. “Uterine NK cells: active regulators at the maternal-

fetal interface”. en. In: J. Clin. Invest. 124.5 (May 2014), pp. 1872–1879.

[27] Zofia Madeja et al. “Paternal MHC expression on mouse trophoblast affects uterine vas-
cularization and fetal growth”. en. In: Proc. Natl. Acad. Sci. U. S. A. 108.10 (Mar. 2011),
pp. 4012–4017.

[28] B Anne Croy et al. “Uterine natural killer cells: insights into their cellular and molecular

biology from mouse modelling”. en. In: Reproduction 126.2 (Aug. 2003), pp. 149–160.

[29] Vicki Plaks et al. “Uterine DCs are crucial for decidua formation during embryo implanta-

tion in mice”. en. In: J. Clin. Invest. 118.12 (Dec. 2008), pp. 3954–3965.

[30] Alison S Care et al. “Macrophages regulate corpus luteum development during embryo

implantation in mice”. en. In: J. Clin. Invest. 123.8 (Aug. 2013), pp. 3472–3487.

[31] A L Mellor et al. “Prevention of T cell-driven complement activation and inflammation by
tryptophan catabolism during pregnancy”. en. In: Nat. Immunol. 2.1 (Jan. 2001), pp. 64–68.

[32] David H Munn et al. “Potential regulatory function of human dendritic cells expressing

indoleamine 2,3-dioxygenase”. en. In: Science 297.5588 (Sept. 2002), pp. 1867–1870.

[33] Graça Raposo and Willem Stoorvogel. “Extracellular vesicles: exosomes, microvesicles,

and friends”. en. In: J. Cell Biol. 200.4 (Feb. 2013), pp. 373–383.

[34] Marina Colombo, Graça Raposo, and Clotilde Théry. “Biogenesis, secretion, and intercel-
lular interactions of exosomes and other extracellular vesicles”. en. In: Annu. Rev. Cell Dev.
Biol. 30 (Aug. 2014), pp. 255–289.

[35]

Jan Lötvall et al. “Minimal experimental requirements for definition of extracellular vesicles
and their functions: a position statement from the International Society for Extracellular
Vesicles”. en. In: J Extracell Vesicles 3 (Dec. 2014), p. 26913.

[36] Maria Yu Konoshenko et al. “Isolation of Extracellular Vesicles: General Methodologies

and Latest Trends”. en. In: Biomed Res. Int. 2018 (Jan. 2018), p. 8545347.

[37] Dennis K Jeppesen et al. “Reassessment of Exosome Composition”. en. In: Cell 177.2 (Apr.

2019), 428–445.e18.

[38] Stephen J Gould, Amy M Booth, and James E K Hildreth. “The Trojan exosome hypothesis”.

en. In: Proc. Natl. Acad. Sci. U. S. A. 100.19 (Sept. 2003), pp. 10592–10597.

[39] Raghu Kalluri. “The biology and function of exosomes in cancer”. en. In: J. Clin. Invest.

126.4 (Apr. 2016), pp. 1208–1215.

103

[40] Susmita Jasti et al. “Immune response to a model shared placenta/tumor-associated antigen
reduces cancer risk in parous mice”. en. In: Biol. Reprod. 96.1 (Jan. 2017), pp. 134–144.

[41] Sarah Caruso and Ivan K H Poon. “Apoptotic Cell-Derived Extracellular Vesicles: More

Than Just Debris”. en. In: Front. Immunol. 9 (June 2018), p. 1486.

[42] Pedro Morillas et al. “Inflammation and apoptosis in hypertension. Relevance of the extent

of target organ damage”. en. In: Rev. Esp. Cardiol. 65.9 (Sept. 2012), pp. 819–825.

[43] Soo Jeong Park et al. “Molecular mechanisms of biogenesis of apoptotic exosome-like
vesicles and their roles as damage-associated molecular patterns”. en. In: Proc. Natl. Acad.
Sci. U. S. A. 115.50 (Dec. 2018), E11721–E11730.

[44] P Wolf. “The nature and significance of platelet products in human plasma”. en. In: Br. J.

Haematol. 13.3 (May 1967), pp. 269–288.

[45] R M Johnstone et al. “Vesicle formation during reticulocyte maturation. Association of
plasma membrane activities with released vesicles (exosomes)”. en. In: J. Biol. Chem.
262.19 (July 1987), pp. 9412–9420.

[46] Hadi Valadi et al. “Exosome-mediated transfer of mRNAs and microRNAs is a novel
mechanism of genetic exchange between cells”. en. In: Nat. Cell Biol. 9.6 (June 2007),
pp. 654–659.

[47] Kirsten Ridder et al. “Extracellular vesicle-mediated transfer of functional RNA in the

tumor microenvironment”. en. In: Oncoimmunology 4.6 (June 2015), e1008371.

[48] Mauro Poggio et al. “Suppression of Exosomal PD-L1 Induces Systemic Anti-tumor Im-

munity and Memory”. en. In: Cell 177.2 (Apr. 2019), 414–427.e13.

[49] Nunzio Iraci et al. “Extracellular vesicles are independent metabolic units with asparaginase

activity”. en. In: Nat. Chem. Biol. 13.9 (Sept. 2017), pp. 951–955.

[50] Elizabeth Delorme-Axford et al. “Human placental trophoblasts confer viral resistance to
recipient cells”. en. In: Proc. Natl. Acad. Sci. U. S. A. 110.29 (July 2013), pp. 12048–12053.

[51] Sonia A Melo et al. “Cancer exosomes perform cell-independent microRNA biogenesis

and promote tumorigenesis”. en. In: Cancer Cell 26.5 (Nov. 2014), pp. 707–721.

[52] Sonia A Melo et al. “Glypican-1 identifies cancer exosomes and detects early pancreatic

cancer”. en. In: Nature 523.7559 (July 2015), pp. 177–182.

[53] Bruno Costa-Silva et al. “Pancreatic cancer exosomes initiate pre-metastatic niche formation

in the liver”. en. In: Nat. Cell Biol. 17.6 (June 2015), pp. 816–826.

104

[54] Ayuko Hoshino et al. “Tumour exosome integrins determine organotropic metastasis”. en.

In: Nature 527.7578 (Nov. 2015), pp. 329–335.

[55] Wei Ying et al. “Adipose Tissue Macrophage-Derived Exosomal miRNAs Can Modulate

In Vivo and In Vitro Insulin Sensitivity”. en. In: Cell 171.2 (Oct. 2017), 372–384.e12.

[56] Huina Zhang et al. “Serum exosomes mediate delivery of arginase 1 as a novel mechanism
for endothelial dysfunction in diabetes”. en. In: Proc. Natl. Acad. Sci. U. S. A. 115.29 (July
2018), E6927–E6936.

[57] Carlos Castaño et al. “Obesity-associated exosomal miRNAs modulate glucose and lipid
metabolism in mice”. en. In: Proc. Natl. Acad. Sci. U. S. A. 115.48 (Nov. 2018), pp. 12158–
12163.

[58] V Ramakrishnaiah et al. Exosome-mediated transmission of hepatitis C virus between

human hepatoma Huh7.5 cells. 2013.

[59] Samantha Sheller-Miller et al. “Cyclic-recombinase-reporter mouse model to determine
exosome communication and function during pregnancy”. en. In: Am. J. Obstet. Gynecol.
221.5 (Nov. 2019), 502.e1–502.e12.

[60] Mancy Tong, Vikki M Abrahams, and Lawrence W Chamley. “Immunological effects of

placental extracellular vesicles”. en. In: Immunol. Cell Biol. (Mar. 2018).

[61] Preenan Pillay et al. “Placental exosomes and pre-eclampsia: Maternal circulating levels in
normal pregnancies and, early and late onset pre-eclamptic pregnancies”. en. In: Placenta
46 (Oct. 2016), pp. 18–25.

[62] S M K Alam et al. “Trophoblast Glycoprotein (TPGB/5T4) in Human Placenta: Expression,
Regulation, and Presence in Extracellular Microvesicles and Exosomes”. en. In: Reprod.
Sci. 25.2 (Feb. 2018), pp. 185–197.

[63] Raghu Kalluri and Valerie S LeBleu. “The biology, function, and biomedical applications

of exosomes”. en. In: Science 367.6478 (Feb. 2020).

[64]

Joanna Kowal, Mercedes Tkach, and Clotilde Théry. “Biogenesis and secretion of exo-
somes”. en. In: Curr. Opin. Cell Biol. 29 (Aug. 2014), pp. 116–125.

[65] Laura Ann Mulcahy, Ryan Charles Pink, and David Raul Francisco Carter. “Routes and

mechanisms of extracellular vesicle uptake”. en. In: J Extracell Vesicles 3 (Aug. 2014).

[66] Yuki Takahashi et al. “Visualization and in vivo tracking of the exosomes of murine
melanoma B16-BL6 cells in mice after intravenous injection”. en. In: J. Biotechnol. 165.2
(May 2013), pp. 77–84.

105

[67] Takafumi Imai et al. “Macrophage-dependent clearance of systemically administered B16BL6-

derived exosomes from the blood circulation in mice”. en. In: J Extracell Vesicles 4 (Feb.
2015), p. 26238.

[68] Marko Kaksonen and Aurélien Roux. “Mechanisms of clathrin-mediated endocytosis”. en.

In: Nat. Rev. Mol. Cell Biol. 19.5 (May 2018), pp. 313–326.

[69] Mercedes Tkach et al. “Qualitative differences in T-cell activation by dendritic cell-derived

extracellular vesicle subtypes”. In: EMBO J. 36.20 (2017), pp. 3012–3028.

[70] Safinur Atay, Cicek Gercel-Taylor, and Douglas D Taylor. “Human trophoblast-derived
exosomal fibronectin induces pro-inflammatory IL-1
 production by macrophages”. en. In:
Am. J. Reprod. Immunol. 66.4 (Oct. 2011), pp. 259–269.

[71] Mathilde Mathieu et al. “Specificities of secretion and uptake of exosomes and other
extracellular vesicles for cell-to-cell communication”. en. In: Nat. Cell Biol. 21.1 (Jan.
2019), pp. 9–17.

[72] Guillaume van Niel, Gisela D’Angelo, and Graça Raposo. “Shedding light on the cell
biology of extracellular vesicles”. en. In: Nat. Rev. Mol. Cell Biol. 19.4 (Apr. 2018),
pp. 213–228.

[73] Clotilde Théry, Laurence Zitvogel, and Sebastian Amigorena. “Exosomes: composition,

biogenesis and function”. en. In: Nat. Rev. Immunol. 2.8 (Aug. 2002), pp. 569–579.

[74]

James H Hurley. “ESCRT complexes and the biogenesis of multivesicular bodies”. en. In:
Curr. Opin. Cell Biol. 20.1 (Feb. 2008), pp. 4–11.

[75] William Mike Henne, Harald Stenmark, and Scott D Emr. “Molecular mechanisms of the
membrane sculpting ESCRT pathway”. en. In: Cold Spring Harb. Perspect. Biol. 5.9 (Sept.
2013).

[76] Nina Pettersen Hessvik and Alicia Llorente. “Current knowledge on exosome biogenesis

and release”. en. In: Cell. Mol. Life Sci. 75.2 (Jan. 2018), pp. 193–208.

[77] Rania Ghossoub et al. “Tetraspanin-6 negatively regulates exosome production”. en. In:

Proc. Natl. Acad. Sci. U. S. A. 117.11 (Mar. 2020), pp. 5913–5922.

[78] Angélique Bobrie et al. “Rab27a supports exosome-dependent and -independent mecha-
nisms that modify the tumor microenvironment and can promote tumor progression”. en.
In: Cancer Res. 72.19 (Oct. 2012), pp. 4920–4930.

[79] Bong Hwan Sung et al. “Directional cell movement through tissues is controlled by exosome

secretion”. en. In: Nat. Commun. 6 (May 2015), p. 7164.

106

[80] Suchismita Sarker et al. “Placenta-derived exosomes continuously increase in maternal
circulation over the first trimester of pregnancy”. en. In: J. Transl. Med. 12 (Aug. 2014),
p. 204.

[81] Aizea Morales-Kastresana et al. “Labeling Extracellular Vesicles for Nanoscale Flow Cy-

tometry”. en. In: Sci. Rep. 7.1 (May 2017), p. 1878.

[82]

Joel Z Nordin et al. “Ultrafiltration with size-exclusion liquid chromatography for high yield
isolation of extracellular vesicles preserving intact biophysical and functional properties”.
en. In: Nanomedicine 11.4 (May 2015), pp. 879–883.

[83] Yong Qin Koh et al. “Exosome enrichment by ultracentrifugation and size exclusion chro-

matography”. en. In: Front. Biosci. 23 (Jan. 2018), pp. 865–874.

[84] Mengxi Wu et al. “Isolation of exosomes from whole blood by integrating acoustics and
microfluidics”. en. In: Proc. Natl. Acad. Sci. U. S. A. 114.40 (Oct. 2017), pp. 10584–10589.

[85] Terry K Morgan. “Cell- and size-specific analysis of placental extracellular vesicles in

maternal plasma and pre-eclampsia”. en. In: Transl. Res. 201 (Nov. 2018), pp. 40–48.

[86] Vasco Filipe, Andrea Hawe, and Wim Jiskoot. “Critical evaluation of Nanoparticle Track-
ing Analysis (NTA) by NanoSight for the measurement of nanoparticles and protein aggre-
gates”. en. In: Pharm. Res. 27.5 (May 2010), pp. 796–810.

[87] Patrcia M Carvalho et al. “Application of Light Scattering Techniques to Nanoparticle

Characterization and Development”. en. In: Front Chem 6 (June 2018), p. 237.

[88] Md Matiur Rahman et al. “Acidification effects on isolation of extracellular vesicles from

bovine milk”. en. In: PLoS One 14.9 (Sept. 2019), e0222613.

[89] Min Kyo Jung and Ji Young Mun. “Sample Preparation and Imaging of Exosomes by

Transmission Electron Microscopy”. en. In: J. Vis. Exp. 131 (Jan. 2018).

[90] R A Pattillo et al. “JAR CELL LINE-CONTINUOUS HUMAN MULTIHORMONE PRO-
DUCTION AND CONTROLS”. In: IN VITRO-JOURNAL OF THE TISSUE CULTURE
ASSOCIATION. Vol. 6. 1971, p. 398.

[91] B Wice et al. “Modulators of cyclic AMP metabolism induce syncytiotrophoblast formation

in vitro”. en. In: Exp. Cell Res. 186.2 (Feb. 1990), pp. 306–316.

[92] R R Adler, A K Ng, and N S Rote. “Monoclonal antiphosphatidylserine antibody inhibits
intercellular fusion of the choriocarcinoma line, JAR”. en. In: Biol. Reprod. 53.4 (Oct.
1995), pp. 905–910.

107

[93] C H Graham et al. “Establishment and characterization of first trimester human trophoblast

cells with extended lifespan”. en. In: Exp. Cell Res. 206.2 (June 1993), pp. 204–211.

[94] Adrianne L Stefanski et al. “Murine trophoblast-derived and pregnancy-associated exosome-
enriched extracellular vesicle microRNAs: Implications for placenta driven effects on ma-
ternal physiology”. en. In: PLoS One 14.2 (Feb. 2019), e0210675.

[95] Margherita Y Turco et al. “Trophoblast organoids as a model for maternal-fetal interactions

during human placentation”. en. In: Nature 564.7735 (Dec. 2018), pp. 263–267.

[96] G Thibault et al. “Inhibitory effect of human syncytiotrophoblast plasma membrane vesicles
on Jurkat cells activated by phorbol ester and calcium ionophore”. en. In: Cell. Immunol.
139.1 (Jan. 1992), pp. 259–267.

[97] P von Dadelszen, G Hurst, and C W Redman. “Supernatants from co-cultured endothelial
cells and syncytiotrophoblast microvillous membranes activate peripheral blood leukocytes
in vitro”. en. In: Hum. Reprod. 14.4 (Apr. 1999), pp. 919–924.

[98] Amal S Aly et al. “Neutrophils are stimulated by syncytiotrophoblast microvillous mem-
branes to generate superoxide radicals in women with preeclampsia”. en. In: Am. J. Obstet.
Gynecol. 190.1 (Jan. 2004), pp. 252–258.

[99] A M Hoegh et al. “Effect of syncytiotrophoblast microvillous membrane treatment on gene
expression in human umbilical vein endothelial cells”. en. In: BJOG 113.11 (Nov. 2006),
pp. 1270–1279.

[100] Anuradha Sabapatha, Cicek Gercel-Taylor, and Douglas D Taylor. “Specific isolation of
placenta-derived exosomes from the circulation of pregnant women and their immunoreg-
ulatory consequences”. en. In: Am. J. Reprod. Immunol. 56.5-6 (Nov. 2006), pp. 345–355.

[101] Lucia Mincheva-Nilsson et al. “Placenta-derived soluble MHC class I chain-related molecules
down-regulate NKG2D receptor on peripheral blood mononuclear cells during human preg-
nancy: a possible novel immune escape mechanism for fetal survival”. en. In: J. Immunol.
176.6 (Mar. 2006), pp. 3585–3592.

[102] Yong-Hong Zhang et al. “Trophoblast-secreted soluble-PD-L1 modulates macrophage po-

larization and function”. en. In: J. Leukoc. Biol. 17 (May 2020), p. 469.

[103] Ann-Christin Stenqvist et al. “Exosomes secreted by human placenta carry functional Fas
ligand and TRAIL molecules and convey apoptosis in activated immune cells, suggesting
exosome-mediated immune privilege of the fetus”. en. In: J. Immunol. 191.11 (Dec. 2013),
pp. 5515–5523.

[104] Thomas F Rice et al. “Macrophage- but not monocyte-derived extracellular vesicles induce

placental pro-inflammatory responses”. en. In: Placenta 69 (Sept. 2018), pp. 92–95.

108

[105] Kiarash Khosrotehrani et al. “Natural history of fetal cell microchimerism during and

following murine pregnancy”. en. In: J. Reprod. Immunol. 66.1 (June 2005), pp. 1–12.

[106] Mancy Tong et al. “Placental Nano-vesicles Target to Specific Organs and Modulate Vas-

cular Tone In Vivo”. en. In: Hum. Reprod. 32.11 (Nov. 2017), pp. 2188–2198.

[107] Samantha Sheller-Miller et al. “Exosomes Cause Preterm Birth in Mice: Evidence for

Paracrine Signaling in Pregnancy”. en. In: Sci. Rep. 9.1 (Jan. 2019), p. 608.

[108] Samantha Sheller-Miller et al. “Feto-Maternal Trafficking of Exosomes in Murine Preg-

nancy Models”. en. In: Front. Pharmacol. 7 (Nov. 2016), p. 432.

[109] Mancy Tong et al. “In vivo targets of human placental micro-vesicles vary with exposure

time and pregnancy”. en. In: Reproduction 153.6 (June 2017), pp. 835–845.

[110] Anoek Zomer et al. “In Vivo imaging reveals extracellular vesicle-mediated phenocopying

of metastatic behavior”. en. In: Cell 161.5 (May 2015), pp. 1046–1057.

[111] Fengmei Pi et al. “Nanoparticle orientation to control RNA loading and ligand display on
extracellular vesicles for cancer regression”. en. In: Nat. Nanotechnol. 13.1 (Jan. 2018),
pp. 82–89.

[112]

J Wolfers et al. “Tumor-derived exosomes are a source of shared tumor rejection antigens
for CTL cross-priming”. en. In: Nat. Med. 7.3 (Mar. 2001), pp. 297–303.

[113] Heedoo Lee et al. “Lung Epithelial Cell–Derived Microvesicles Regulate Macrophage
Migration via MicroRNA-17/221–Induced Integrin 
1 Recycling”. en. In: The Journal of
Immunology 199.4 (Aug. 2017), pp. 1453–1464.

[114] Murray D Mitchell et al. “Placental exosomes in normal and complicated pregnancy”. en.

In: Am. J. Obstet. Gynecol. 213.4 Suppl (Oct. 2015), S173–81.

[115]

Jennifer Southcombe et al. “The immunomodulatory role of syncytiotrophoblast microvesi-
cles”. en. In: PLoS One 6.5 (May 2011), e20245.

[116] Mercy PrabhuDas et al. “Immune mechanisms at the maternal-fetal interface: perspectives

and challenges”. en. In: Nat. Immunol. 16.4 (Apr. 2015), pp. 328–334.

[117] Rebecca A Dragovic et al. “Sizing and phenotyping of cellular vesicles using Nanoparticle

Tracking Analysis”. en. In: Nanomedicine 7.6 (Dec. 2011), pp. 780–788.

[118] Greg Wilson et al. “Best practices for scientific computing”. en. In: PLoS Biol. 12.1 (Jan.

2014), e1001745.

109

[119] Greg Wilson et al. “Good enough practices in scientific computing”. en. In: PLoS Comput.

Biol. 13.6 (June 2017), e1005510.

[120] Geir Kjetil Sandve et al. “Ten simple rules for reproducible computational research”. en.

In: PLoS Comput. Biol. 9.10 (Oct. 2013), e1003285.

[121] Hadley Wickham. “Tidy Data”. In: Journal of Statistical Software 59.10 (2017).

[122] Hadley Wickham and Lionel Henry. “Tidyr: Easily Tidy Data with “Spread ()” and “Gather
()” Functions. 2018”. In: URL https://CRAN. R-project. org/package= tidyr. R package
version 0. 8 3 (2019), p. 11.

[123] Hadley Wickham. ggplot2: Elegant Graphics for Data Analysis. en. Springer, June 2016.

[124] Team, R Core. R: A language and environment for statistical computing. Vienna, Austria,

2013.

[125] Hadley Wickham et al. “dplyr: A grammar of data manipulation”. In: R package version 0.

4 3 (2015).

[126] H Wickham. Purrr: Functional programming tools. 2016.

[127] Carlos Salomon et al. “A gestational profile of placental exosomes in maternal plasma and

their effects on endothelial cell migration”. en. In: PLoS One 9.6 (June 2014), e98667.

[128] Ramkumar Menon et al. “Quantitative Proteomics by SWATH-MS of Maternal Plasma
Exosomes Determine Pathways Associated With Term and Preterm Birth”. en. In: En-
docrinology 160.3 (Mar. 2019), pp. 639–650.

[129] Dionne Tannetta et al. “Update of syncytiotrophoblast derived extracellular vesicles in
normal pregnancy and preeclampsia”. en. In: J. Reprod. Immunol. 119 (Feb. 2017), pp. 98–
106.

[130] Barry R Zeeberg et al. “Mistaken identifiers: gene name errors can be introduced inad-
vertently when using Excel in bioinformatics”. en. In: BMC Bioinformatics 5 (June 2004),
p. 80.

[131] Mark Ziemann, Yotam Eren, and Assam El-Osta. “Gene name errors are widespread in the

scientific literature”. en. In: Genome Biol. 17.1 (Aug. 2016), p. 177.

[132] H D Attwood and W W Park. EMBOLISM TO THE LUNGS BY TROPHOBLAST. 1962.

[133] Georg Schmorl. Pathologisch-anatomische Untersuchungen über Puerperal-Eklampsie. de.

Vogel, 1893.

110

[134] Tina Cronqvist et al. “Syncytiotrophoblast vesicles show altered micro-RNA and haemoglobin

content after ex-vivo perfusion of placentas with haemoglobin to mimic preeclampsia”. en.
In: PLoS One 9.2 (Feb. 2014), e90020.

[135]

Jenny Z Kechagia, Johanna Ivaska, and Pere Roca-Cusachs. “Integrins as biomechanical
sensors of the microenvironment”. en. In: Nat. Rev. Mol. Cell Biol. 20.8 (Aug. 2019),
pp. 457–473.

[136] Els J van der Vlist et al. “Fluorescent labeling of nano-sized vesicles released by cells and
subsequent quantitative and qualitative analysis by high-resolution flow cytometry”. en. In:
Nat. Protoc. 7.7 (June 2012), pp. 1311–1326.

[137] Claire McQuin et al. “CellProfiler 3.0: Next-generation image processing for biology”. en.

In: PLoS Biol. 16.7 (July 2018), e2005970.

[138] Kazuki Tainaka et al. “Whole-body imaging with single-cell resolution by tissue decol-

orization”. en. In: Cell 159.4 (Nov. 2014), pp. 911–924.

[139] Etsuo A Susaki et al. “Advanced CUBIC protocols for whole-brain and whole-body clearing

and imaging”. en. In: Nat. Protoc. 10.11 (Nov. 2015), pp. 1709–1727.

[140] Hao Chen et al. “Cytofkit: A Bioconductor Package for an Integrated Mass Cytometry Data

Analysis Pipeline”. en. In: PLoS Comput. Biol. 12.9 (Sept. 2016), e1005112.

[141]

Johannes Schindelin et al. “Fiji: an open-source platform for biological-image analysis”.
en. In: Nat. Methods 9.7 (June 2012), pp. 676–682.

[142] Ramkumar Menon et al. “Protein Profile Changes in Circulating Placental Extracellular
Vesicles in Term and Preterm Births: A Longitudinal Study”. en. In: Endocrinology 161.4
(Apr. 2020).

[143] Yen-Rei A Yu et al. “A Protocol for the Comprehensive Flow Cytometric Analysis of
Immune Cells in Normal and Inflamed Murine Non-Lymphoid Tissues”. en. In: PLoS One
11.3 (Mar. 2016), e0150606.

[144] Denis Bedoret et al. “Lung interstitial macrophages alter dendritic cell functions to prevent

airway allergy in mice”. en. In: J. Clin. Invest. 119.12 (Dec. 2009), pp. 3723–3738.

[145] Svetoslav Chakarov et al. “Two distinct interstitial macrophage populations coexist across

tissues in specific subtissular niches”. en. In: Science 363.6432 (Mar. 2019).

[146]

Joey Schyns et al. “Non-classical tissue monocytes and two functionally distinct populations
of interstitial macrophages populate the mouse lung”. en. In: Nat. Commun. 10.1 (Sept.
2019), p. 3964.

111

[147] E Ruoslahti. “RGD and other recognition sequences for integrins”. en. In: Annu. Rev. Cell

Dev. Biol. 12 (1996), pp. 697–715.

[148] Thomas C Sroka et al. “The minimum element of a synthetic peptide required to block
prostate tumor cell migration”. en. In: Cancer Biol. Ther. 5.11 (Nov. 2006), pp. 1556–1562.

[149] P J Newman et al. “PECAM-1 (CD31) cloning and relation to adhesion molecules of the
immunoglobulin gene superfamily”. en. In: Science 247.4947 (Mar. 1990), pp. 1219–1222.

[150]

Jia Wei et al. “Placental trophoblast debris mediated feto-maternal signalling via small
RNA delivery: implications for preeclampsia”. en. In: Sci. Rep. 7.1 (Oct. 2017), p. 14681.

[151] Mancy Tong et al. “Micro- and Nano-vesicles from First Trimester Human Placentae Carry
Flt-1 and Levels Are Increased in Severe Preeclampsia”. en. In: Front. Endocrinol. 8 (July
2017), p. 174.

[152] Carolina Motta-Mejia et al. “Placental Vesicles Carry Active Endothelial Nitric Oxide
Synthase and Their Activity is Reduced in Preeclampsia”. en. In: Hypertension 70.2 (Aug.
2017), pp. 372–381.

[153] Safinur Atay et al. “Trophoblast-derived exosomes mediate monocyte recruitment and

differentiation”. en. In: Am. J. Reprod. Immunol. 65.1 (Jan. 2011), pp. 65–77.

[154] Adrian Erlebacher et al. “Constraints in antigen presentation severely restrict T cell recog-

nition of the allogeneic fetus”. en. In: J. Clin. Invest. 117.5 (May 2007), pp. 1399–1411.

[155] Malin Hedlund et al. “Human placenta expresses and secretes NKG2D ligands via exosomes
that down-modulate the cognate receptor expression: evidence for immunosuppressive
function”. en. In: J. Immunol. 183.1 (July 2009), pp. 340–351.

[156] Lucia Mincheva-Nilsson and Vladimir Baranov. Placenta-Derived Exosomes and Syncy-
tiotrophoblast Microparticles and their Role in Human Reproduction: Immune Modulation
for Pregnancy Success. 2014.

[157] Chin-Siean Tay et al. “Cis-acting pathways selectively enforce the non-immunogenicity
of shed placental antigen for maternal CD8 T cells”. en. In: PLoS One 8.12 (Dec. 2013),
e84064.

[158] Olivia J Holland et al. “Minor histocompatibility antigens are expressed in syncytiotro-
phoblast and trophoblast debris: implications for maternal alloreactivity to the fetus”. en.
In: Am. J. Pathol. 180.1 (Jan. 2012), pp. 256–266.

[159]

Jin Jin and Ramkumar Menon. “Placental exosomes: A proxy to understand pregnancy
complications”. en. In: Am. J. Reprod. Immunol. 79.5 (May 2018), e12788.

112

[160] Soumyalekshmi Nair et al. “Human placental exosomes in gestational diabetes mellitus
carry a specific set of miRNAs associated with skeletal muscle insulin sensitivity”. en. In:
Clin. Sci. 132.22 (Nov. 2018), pp. 2451–2467.

[161] Peng Zhang et al. “Ultrasensitive detection of circulating exosomes with a 3D-nanopatterned

microfluidic chip”. en. In: Nat Biomed Eng 3.6 (June 2019), pp. 438–451.

[162] Dionne Tannetta et al. “Syncytiotrophoblast extracellular vesicles - Circulating biopsies

reflecting placental health”. en. In: Placenta 52 (Apr. 2017), pp. 134–138.

[163] Carlos Salomon et al. “Placental Exosomes During Gestation: Liquid Biopsies Carrying
Signals for the Regulation of Human Parturition”. en. In: Curr. Pharm. Des. 24.9 (2018),
pp. 974–982.

[164] Helena C Christianson et al. “Cancer cell exosomes depend on cell-surface heparan sulfate
proteoglycans for their internalization and functional activity”. en. In: Proc. Natl. Acad.
Sci. U. S. A. 110.43 (Oct. 2013), pp. 17380–17385.

[165] Dimitris Missirlis et al. “Substrate engagement of integrins 
5
1 and 
v
3 is necessary,
but not sufficient, for high directional persistence in migration on fibronectin”. en. In: Sci.
Rep. 6 (Mar. 2016), p. 23258.

[166] E Ruoslahti and M D Pierschbacher. “New perspectives in cell adhesion: RGD and inte-

grins”. en. In: Science 238.4826 (Oct. 1987), pp. 491–497.

[167] Thomas C Sroka, Michael E Pennington, and Anne E Cress. “Synthetic D-amino acid
peptide inhibits tumor cell motility on laminin-5”. en. In: Carcinogenesis 27.9 (Sept.
2006), pp. 1748–1757.

[168] Hui Wang et al. “Tumor cell alpha3beta1 integrin and vascular laminin-5 mediate pulmonary

arrest and metastasis”. en. In: J. Cell Biol. 164.6 (Mar. 2004), pp. 935–941.

[169] Christopher S Stipp. “Laminin-binding integrins and their tetraspanin partners as potential

antimetastatic targets”. en. In: Expert Rev. Mol. Med. 12 (Jan. 2010), e3.

[170] Lema F Yousif, Jacopo Di Russo, and Lydia Sorokin. “Laminin isoforms in endothelial and
perivascular basement membranes”. en. In: Cell Adh. Migr. 7.1 (Jan. 2013), pp. 101–110.

[171] Oscar P B Wiklander et al. “Extracellular vesicle in vivo biodistribution is determined by
cell source, route of administration and targeting”. en. In: J Extracell Vesicles 4 (Apr. 2015),
p. 26316.

[172] F Hytten. “Blood volume changes in normal pregnancy”. en. In: Clin. Haematol. 14.3 (Oct.

1985), pp. 601–612.

113

[173] M Knight et al. “Shedding of syncytiotrophoblast microvilli into the maternal circulation
in pre-eclamptic pregnancies”. en. In: Br. J. Obstet. Gynaecol. 105.6 (June 1998), pp. 632–
640.

[174] Carlos Salomon et al. “Exosomal signaling during hypoxia mediates microvascular en-

dothelial cell migration and vasculogenesis”. en. In: PLoS One 8.7 (July 2013), e68451.

[175] Tina Cronqvist et al. “Syncytiotrophoblast derived extracellular vesicles transfer functional
placental miRNAs to primary human endothelial cells”. en. In: Sci. Rep. 7.1 (July 2017),
p. 4558.

[176] Kirsten Ridder et al. “Extracellular vesicle-mediated transfer of genetic information between
the hematopoietic system and the brain in response to inflammation”. en. In: PLoS Biol.
12.6 (June 2014), e1001874.

[177] B Sauer. “Manipulation of transgenes by site-specific recombination: use of Cre recombi-

nase”. en. In: Methods Enzymol. 225 (1993), pp. 890–900.

[178] B Sauer. “Inducible gene targeting in mice using the Cre/lox system”. en. In: Methods 14.4

(Apr. 1998), pp. 381–392.

[179] Etsuo A Susaki et al. “Whole-brain imaging with single-cell resolution using chemical

cocktails and computational analysis”. en. In: Cell 157.3 (Apr. 2014), pp. 726–739.

[180] Kyosuke Kagami et al. “Three-dimensional visualization of intrauterine conceptus through

the uterine wall by tissue clearing method”. en. In: Sci. Rep. 7.1 (July 2017), p. 5964.

[181] Weizhe Li, Ronald N Germain, and Michael Y Gerner. “Multiplex, quantitative cellular
analysis in large tissue volumes with clearing-enhanced 3D microscopy (Ce3D)”. en. In:
Proc. Natl. Acad. Sci. U. S. A. 114.35 (Aug. 2017), E7321–E7330.

[182] Weizhe Li, Ronald N Germain, and Michael Y Gerner. “High-dimensional cell-level analy-
sis of tissues with Ce3D multiplex volume imaging”. en. In: Nat. Protoc. 14.6 (June 2019),
pp. 1708–1733.

[183]

Julie Helft et al. “GM-CSF Mouse Bone Marrow Cultures Comprise a Heterogeneous
Population of CD11c(+)MHCII(+) Macrophages and Dendritic Cells”. en. In: Immunity
42.6 (June 2015), pp. 1197–1211.

[184] Steven Hann et al. “A Wisp3 Cre-knockin allele produces efficient recombination in sper-
matocytes during early prophase of meiosis I”. en. In: PLoS One 8.9 (Sept. 2013), e75116.

[185] Mandar Deepak Muzumdar et al. “A global double-fluorescent Cre reporter mouse”. en.

In: Genesis 45.9 (Sept. 2007), pp. 593–605.

114

[186] D W Bianchi et al. “Male fetal progenitor cells persist in maternal blood for as long as 27

years postpartum”. en. In: Proc. Natl. Acad. Sci. U. S. A. 93.2 (Jan. 1996), pp. 705–708.

[187] Stephanie Pritchard et al. “Comprehensive analysis of genes expressed by rare microchimeric

fetal cells in the maternal mouse lung”. en. In: Biol. Reprod. 87.2 (Aug. 2012), p. 42.

[188] Anoek Zomer et al. “Studying extracellular vesicle transfer by a Cre-loxP method”. en. In:

Nat. Protoc. 11.1 (Jan. 2016), pp. 87–101.

[189] Dionne S Tannetta et al. “Characterisation of syncytiotrophoblast vesicles in normal preg-
nancy and pre-eclampsia: expression of Flt-1 and endoglin”. en. In: PLoS One 8.2 (Feb.
2013), e56754.

[190] Carlos Salomon et al. “Extravillous trophoblast cells-derived exosomes promote vascular

smooth muscle cell migration”. en. In: Front. Pharmacol. 5 (Aug. 2014), p. 175.

[191] Yu Wang et al. “Fetal cells in mother rats contribute to the remodeling of liver and kidney

after injury”. en. In: Biochem. Biophys. Res. Commun. 325.3 (Dec. 2004), pp. 961–967.

[192] D W Bianchi. “Fetal cells in the maternal circulation: feasibility for prenatal diagnosis”.

en. In: Br. J. Haematol. 105.3 (June 1999), pp. 574–583.

[193] E A Bonney and P Matzinger. “The maternal immune system’s interaction with circulating

fetal cells”. en. In: J. Immunol. 158.1 (Jan. 1997), pp. 40–47.

[194] Yutaka Fujiki et al. “Fetomaternal trafficking in the mouse increases as delivery approaches
and is highest in the maternal lung”. en. In: Biol. Reprod. 79.5 (Nov. 2008), pp. 841–848.

[195] Tobias G Kapp et al. “A Comprehensive Evaluation of the Activity and Selectivity Profile

of Ligands for RGD-binding Integrins”. en. In: Sci. Rep. 7 (Jan. 2017), p. 39805.

[196] Samantha Sheller-Miller et al. “Environmental pollutant induced cellular injury is reflected

in exosomes from placental explants”. en. In: Placenta 89 (Jan. 2020), pp. 42–49.

[197] Charles P Lai et al. “Dynamic biodistribution of extracellular vesicles in vivo using a

multimodal imaging reporter”. en. In: ACS Nano 8.1 (Jan. 2014), pp. 483–494.

[198] Maarten P Bebelman et al. “Real-time imaging of multivesicular body-plasma membrane
fusion to quantify exosome release from single cells”. en. In: Nat. Protoc. 15.1 (Jan. 2020),
pp. 102–121.

[199] Bong Hwan Sung et al. “A live cell reporter of exosome secretion and uptake reveals
pathfinding behavior of migrating cells”. en. In: Nat. Commun. 11.1 (Apr. 2020), p. 2092.

115

[200] Elizabeth Q Littauer et al. “H1N1 influenza virus infection results in adverse pregnancy
outcomes by disrupting tissue-specific hormonal regulation”. en. In: PLoS Pathog. 13.11
(Nov. 2017), e1006757.

[201] Géraldine Engels et al. “Pregnancy-Related Immune Adaptation Promotes the Emergence
of Highly Virulent H1N1 Influenza Virus Strains in Allogenically Pregnant Mice”. en. In:
Cell Host Microbe 21.3 (Mar. 2017), pp. 321–333.

116